Thursday
July 18, 1991
Part il



Environmental

Protection  Agency

40 CFR Parts 141 and 142
National Primary Drinking Water
Regulations; Radionuclides; Proposed
Rule

-------

-------
33050
Federal Register / Vol.  56,  No. 138  / Thursday, July 18, 1991  / Proposed Rules
ENVIRONMENTAL PROTECTION
AGENCY

40 CFR Parts 141,142

[WH-FRL 3956-4]

RIN 2040-AA94

National Primary Drinking Water
Regulations; Radionuclides

AGENCY: Environmental Protection
Agency.
ACTION: Notice of proposed rulemaking.

SUMMARY: In this action under the Safe
Drinking Water Act (as amended in
1986), the Environmental Protection
Agency (EPA) is proposing Maximum
Contaminant Level Goals (MCLGs) and
National Primary Drinking Water
Regulations for the following
radionuclides: radon-222, radium-226,
radium-228, uranium, alpha emitters,
and beta particle and photon emitters.
These radionuclides are classified as
group A human carcinogens according
to EPA's classification scheme; also,
uranium is toxic to the kidneys. This
notice proposes MCLGs, Maximum
Contaminant Levels (MCLs), monitoring,
reporting, and public notification
requirements for these radionuclides.
DATES: Written comments should be
submitted by October 16,1991. A public
hearing will be held on September 6,
1991 in Washington, DC beginning at 9
a.m. A second public meeting will be
held on September 12,1991 in Chicago,
Illinois at 9 a.m. Washington hearing
speakers should register by August 23.
Chicago hearing speakers should
register by August 30.
ADDRESSES: Send written comments to
Comments Clerk—Radionuclides,
Drinking Water Standards Division,
Office of Ground Water and Drinking
Water (WH-550D), Environmental
Protection Agency, 401 M Street, SW.,
Washington, DC 20460. A copy of all
public comments and supporting
documents for this proposed regulation
will be available for review at EPA,
Ground Water and Drinking Water
Docket, 401 M Street, SW., Washington,
DC 20460. For access to the docket
materials, call 202-382-3027 between 9
a.m. and 3:30 p.m. Commenters are
requested to submit one original and
three copies of their written comments.
Commenters who wish to receive
acknowledgement of receipt of their
comments should include a self
addressed stamped envelope. All
comments must be post marked or
delivered by hand by October 16,1991.
No facsimiles (faxes) will be accepted,.
as EPA is  not equipped to receive the
                        large volume of comments expected to
                        arrive near the close of the comment
                        period, and cannot assure that faxes will
                        be delivered to the docket. Major
                        supporting documents cited in the
                        reference section of the proposed rule
                        will be available for inspection at the
                        Drinking Water Supply Branches in
                        EPA's Regional Offices listed below:
                        I. JFK Federal Bldg., (One Congress Street,
                         llth floor), Boston, MA 02203, Phone: (617)
                         565-3610, Jerome Healey
                        n. 26 Federal Plaza, Room 824, New York, NY
                         10278, Phone: (212) 264-1800, Walter
                         Andrews
                        III. 841 Chestnut Street, Philadelphia, PA
                         19107, Phone: (215) 597-9873, Dale Long
                        IV. 345 Courtland Street, Atlanta, GA 30365,
                         Phone: (404) 347-3633, Wayne Aeronson
                        V. 230 S. Dearborn Street, Chicago, IL 60604,
                         Phone: (312) 353-2650, Ed Walters
                        VI. 1445 Ross Avenue, Dallas, TX 75202,
                         Phone: (214) 655-7155, Thomas Love
                        VII. 726 Minnesota Avenue, Kansas City, KS
                         66101, Phone: (913) 236-2815, Ralph
                         Langemeir
                        VIII. One Denver Place, 999918th Street,
                         Suite 1300 Denver, CO 80202-2413, Phone:
                         (303) 293-1424, Patrick Grotty
                        IX. 75 Hawthorne Street, San Francisco, CA
                         94105, Phone: (415) 974-8073, Bruce Macler
                        X. 1200 Sixth Avenue, Seattle, WA 98101,
                         Phone: (206) 442-1225, Jan Hastings
                         Public hearings will be held in the
                        following locations;
                        Washington DC—Crystal City Marriott
                         Hotel, 1111 Jefferson Davis Highway,
                         Arlington, VA
                        Chicago, Illinois—J.C. Kluczynski
                         Federal Building, 230 Dearborn Street,
                         16th Floor, Chicago, IL
                         Members of the public who plan to
                        make a statement at either public
                        hearing should contact Danesha Reid to
                        register, EPA (WH-550D), 401 M Street,
                        SW., Washington, DC 20460, telephone
                        (202) 382-7575. Unregistered speakers
                        will be heard after all registered
                        speakers have made their statements.
                        FOR FURTHER INFORMATION CONTACT:
                        The Safe Drinking Water Hotline,
                        telephone (800) 426-4791, or Gregory
                        Helms, Drinking Water Standards
                        Division, Office of Ground Water and
                        Drinking Water (WH-550D),
                        Environmental Protection Agency, 401 M
                        Street, SW., Washington, DC 20460,
                        telephone (202) 382-7575.

                        Abbreviations Used in This Notice

                        BAT: Best Available Technology
                        BEIR: Committee on the Biological
                            Effects of Ionizing Radiation
                        CWS: Community Water System
                        EMSL: EPA Environmental Monitoring
                            and Support Laboratory (Cincinnati
                            or Las .Vegas)
                        ede: effective dose equivalent
                        GAG: Granular Activated Carbon
ICRP: International Commission on
    Radiation Protection
MCL: Maximum Contaminant Level
MCLG: Maximum Contaminant Level
    Goal
MDL: Method Detection Limit
Mr/hr: milliroentgen per hour
mgd: Million Gallons/Day
mrem/yr: millirem/year
NIPDWR: National Interim Primary
    Drinking Water Regulation
NPDWR: National Primary Drinking
    Water Regulation
NTNC: Non-transient, non-community
    water system
pCi/1: picocurie/liter
POE: Point-of-Entry Technologies
POU: Point-of-Use Technologies
PQL: Practical Quantitation Level
PTA: Packed Tower Aeration
PWS: Public Water System
Ra-226: Radium-226
Ra-228: Radium-228
RIA: Regulatory Impact Analysis
Rn-222: Radon-222, or radon
SDWA: Safe Drinking Water Act, or the
    "Act", as amended in 1986
SMR: Standard Mortality Ratio
WLM: Working Level Month

Table of Contents
I. Summary of Today's NPRM
II. Background
  A. Statutory Authority and Requirements
  1. MCLGs, McLs and BAT
  2. Variances and Exemptions
  3. Primacy
  4. Monitoring, Quality Control, and
    Recordkeeping
  5. Public Notification
  B. Applicability
  C. Regulatory Background
  D. Comments by the Science Advisory
    Board and the Public on the ANPRM
  1. SAB Comment
  a. General comments and generic issues
  a.l. General comments and generic issues.
  b. Responses to the five specific questions
  c. Comments on important issues in the
    criteria documents
  2. Public Comment on the ANPRM
  E. Other EPA Radon and Radiation
   Programs
  F. Basics of Radiation
III. Occurrence and Exposure
  A. Radium-226
  B. Radium-228
  C. Radon
  1. Occurrence
  2. Assessing individual radon exposure
   from inhalation and ingestion
  D. Uranium
  E. Beta and Photon-Emitting Radionuclides
  F. Alpha-Emitting Radionuclides
IV. Proposed MCLGs for Radionuclides
  A. Setting MCLGs
  B. Estimating Health Risks for
   Radionuclides
  C. Adverse Health Effects of the
   Radionuclides
  1. Radium-226 and Radium-228
  2. Radon

-------
                Federal Register / Vol. 56, No.  138 / Thursday,  July 18,  1991 / Proposed Rules
                                                                       33051
  a. Radon Risks from Inhalation
  b. Radon Risk via Ingestion
  3. Uranium
  a. Carcinogeniclty
  b. Non-cancer Effects
  4. Beta Particle and Photon Emitters
  5. Alpha Emitters
  D. MCLG Determinations
V. Proposed Maximum Contaminant Levels
Summary of the Proposal
  A. BATs and Associated Costs
  B. Best Available Technologies (BATs)
  1. Radium-226 and Radium-228
  a. Lime Softening
  b. Ion Exchange
  c. Reverse Osmosis
  2. Radon
  a. Aeration
  b. Secondary Effects of Aeration: Estimate
    of Risks from PTA Emissions of Radon
  c. Granular Activated Carbon
  3. Uranium
  a. Coagulation/Filtration
  b. Ion Exchange
  c. Lime Softening
  d. Reverse Osmosis
  4. Beta and Photon Emitters
  5. Alpha Emitting Radionuclides
  C. Waste Treatment and Disposal
  D. Analytic Methods
  1. Description of Analytic Methods
  2. Cost of Performing Analyses
  3. Method Detection Limits and Practical
    Quantitation Levels
  E. Laboratory Approval and Certification
  1. Background
  2. Acceptance Limits for Radionuclide
    Contaminants
  F. Proposed MCLs and Alternatives
    Considered
  G. Proposed Monitoring and Reporting
    Requirements
  H. State Implementation
  I. Variances and Exemptions
  1. Variances
  2. Exemptions
  3. Unreasonable Risks to Health (URTH)
VI. Public Notice Requirements
VII. Economic Impacts and Benefits
  A. Regulatory Flexibility Analysis
  B. Paperwork Reduction Act
VIII. References
  Appendix A—Fundamentals of
    Radioactivity in Drinking Water
  Appendix B—Beta Particle and Photon
    Emitters
  Appendix C—Alpha Emitters

I. Summary of Today's NPRM

Applicability
  The regulations proposed in this
notice would apply to all community
and all non-transient, non-community
public water systems. The proposed
regulations would not apply to private
water supplies (i.e., systems serving
fewer than 25 persons).

Proposed MCLGs and MCLs

3. Radon-222 	
4. Uranium 	

5. Beta and photon
emitters (excluding
Ra-228).
6. Adjusted gross
alpha emitters
(excluding Ra-226,
U, and Rn-222).
MCLG
zero 	
zero 	

zero 	


zero 	



MCL
300 pCi/1.
20 jig/1 (30 pCi/
1).
4 mrem ede/yr.


15pCi/1.




1. Radium-226 	 	
2, Radkjm-228 	
MCLG
zero 	
zero 	
MCL
20 pa/1.
20 pa/1.
  Note: EPA recognizes that most radionuclides
emit more than one kind of radiation as they decay.
The lists of compounds labeled "alpha" or "beta"
emitters identifies the predominant mode of decay.
  Note: In this document the unit mrem ede/yr
refers to the dose committed over a period of 50
years to reference man (ICRP 1975) from an annual
intake at the rate of 2 liters  of drinking water per
day.
Proposed BATs Under Section 1412 of
theSDWA
Radium 226/228: Ion exchange, lime
    softening, reverse osmosis
Radon: Aeration
Uranium: Coagulation/filtration, ion
    exchange, lime softening, reverse
    osmosis
Beta and photon emitters: Ion exchange,
    reverse osmosis
Alpha emitters: Reverse osmosis
Proposed BAT Under Section 1415 of the
SDWA
  The same as BAT under Section 1412.
Coagulation and filtration and lime
softening are not BAT for small systems
(those with S500 connections) for the
purpose of granting variances because
they are  not technologically feasible for
small systems.

Proposed Compliance Monitoring
  (a) The proposed initial monitoring
requirements for radon are:
  (1) For ground water systems and
mixed ground and surface water
systems, four consecutive quarterly
samples  for one year, and then annual
samples  for the remainder of the first
three year compliance period.  States
could grant monitoring waivers to
systems  that demonstrate compliance
with the MCL reliably and consistently
in the initial compliance period,
allowing systems to collect only one
sample per three year compliance period
for the remainder of the nine year
compliance cycle. Systems relying solely
on surface water are not required to
monitor for radon, because radon is a
highly volatile gas and is not expected
to be found in surface water.
Laboratories would be expected to
accurately measure radon down to
levels of 300 pCi/1 at the time of
sampling.
  (2) Systems that violate the MCL
would be required to monitor quarterly
until the average of four consecutive
quarterly samples is below the MCL.
  (b) The proposed monitoring
requirements for gross alpha, radium-226
and uranium are:
  (1) Three annual gross alpha screens,
to be initiated in the compliance period
starting January 1996; if gross alpha is
less than the MCLs for radium-226,
uranium, and adjusted gross alpha,
screening would be reduced to
monitoring once per three year
compliance period. Laboratories would
be expected to measure radium 226 and
uranium down to 5 pCi/1 and gross
alpha down to 15 pCi/1.
  (2) If gross alpha exceeds the radium-
226, uranium, or adjusted gross alpha
MCLs, specific analysis for uranium
and/or radium-226 must be conducted. If
the contaminant-specific analyses show
that the radium-226 or uranium MCL
was exceeded, quarterly monitoring for
that contaminant is required. If neither
MCL is exceeded, monitoring for
radium-226 and uranium (or gross alpha
screen in lieu of radium or uranium) may
be reduced to one sample every 3-year
compliance period after 3 annual
samples. Sampling may be reduced to
one sample every 9-year compliance
cycle if the state finds,  through a
monitoring waiver, that the system
meets the MCL reliably and
consistently.
  (3) Systems that violate the MCL
would be required to monitor quarterly
until four consecutive quarterly samples
is below the MCL.
  (c) The proposed monitoring
requirements for radium-228 are as
follows: Three annual radium-228
analyses would be required; if the
radium-228 MCL is exceeded, quarterly
monitoring would be required. If the
system is consistently below the MCL,
then the annual period may be reduced
to one sample per three year compliance
period. Monitoring may be further
reduced to once every 9-year
compliance cycle by the issuance of a
monitoring waiver if the state finds that
the system meets the MCL reliably and
consistently. A gross beta test may be
used as a screen for radium 228.
Systems that violate the MCL would be
required to monitor quarterly until four
consecutive quarterly samples is below
the MCL.
  (d) Gross beta monitoring. Only
supplies deemed vulnerable to
contamination would be required to
monitor for beta and photon emitters.
Vulnerable systems would be required
to measure gross beta quarterly and
tritium and strontium annually. The
presumptive screen for compliance with
the MCL would be 50 pCi/1. Because

-------
33052
Federal  Register / Vol. 56, No. 138  / Thursday, July 18,  1991 / Proposed Rules
 only vulnerable systems would be
 required to monitor, no reduction in
 monitoring would be allowed. Systems
 that violate the MCL would be required
 to monitor monthly until three
 consecutive samples is below the MCL.
   (e) Systems having historical data that
 has been collected in accord with the
 analytic chemistry requirements may
 use the data to determine compliance.

 Point-of-use (POU) devices, point-of-
 entry (POE) devices and bottled water
  POE would be allowed to be used to
 achieve compliance with MCLs;
 however, POE would not be BAT.
  POU and bottled water would not be
 allowed to be used to achieve
 compliance with the MCLs; however,
 either could be, at State discretion, a
 condition of granting a variance or
 exemption, except for radon (POU may
 not be used for radon because POU fails
 to address radon risks).

 Variances and Exemptions
  Primacy States may require public
 water systems to implement additional
 interim control measures suclvas
 installation of additional centralized
 treatment or POU devices or distribution
 of bottled water to each customer as
 measures to reduce the health risk
 before granting a variance or exemption.
 The State may not issue a variance or
 exemption if an unreasonable risk to
 health exists, as determined by the State
 using EPA guidance. States must require
 public water systems to provide POE/
 POU devices, bottled water or other
 means, as appropriate to the risks
 present (i.e., no POU or bottled water for
 volatile contaminants, such as radon),  to
 reduce exposure below unreasonable
 risk to health values before granting a
 variance or exemption. EPA is presently
 developing guidance for determining
 affordability to systems serving fewer
 than 3300 people of different water
 treatments, for purposes of granting
 variances. This guidance is expected to
 be proposed later this year.
 II. Background

A. Statutory Authority and
Requirements
  Section 1412 of the Safe Drinking
 Water Act, as amended in 1986
 ("SDWA" or "the Act"), requires the
Environmental Protection Agency (EPA)
 to publish Maximum Contaminant Level
 Goals (MCLGs) and promulgate
National Primary Drinking Water
Regulations (NPDWRs) for
 contaminants in drinking water which
may cause any adverse effect on the
health of persons and which are known
or anticipated to occur in public water
                       systems. Under section 1401, the
                       NPDWRs are to include Maximum
                       Contaminant Levels (MCLs) and
                       "criteria and procedures to assure a
                       supply of drinking water which
                       dependably complies" with such MCLs.
                       Under section 1412(b)(7)(A), if it is not
                       economically or technically feasible to
                       ascertain the level of a contaminant in
                       drinking water, the NPDWR may require
                       the use of a treatment technique instead
                       of an MCL.
                         Under section 1412(b), EPA is to
                       establish MCLGs and promulgate
                       national primary drinking water
                       regulations for 83 contaminants in public
                       water systems. The radionuclides
                       included in today's proposal are among
                       these 83 contaminants.
                       1. MCLGs, MCLs and BAT
                         Under section 1412(b)(4) of the Act,
                       EPA is to establish MCLGs at the level
                       at which no known or anticipated
                       adverse effects on the health of persons
                       occur and which allow an adequate
                       margin of safety. MCLGs are non-
                       enforceable health goals based only on  .
                       health effects and exposure information.
                         MCLs are enforceable standards
                       which the Act directs EPA to set-as^
                       close to the MCLGs as is feasible.   .' . '
                       "Feasible" means, feasible with the use
                       of the best technology, treatment
                       techniques, and other means which the
                       Administrator finds available (taking
                       cost into consideration) after
                       examination for efficacy under field
                       conditions and not solely under
                       laboratory conditions (SDWA section
                       1412(b)(5)). Also, the SDWA requires the
                       Agency to identify the best  available
                       technology (BAT) for meeting the MGL
                       for each contaminant.
                       2. Variances and Exemptions
                         Section 1415 authorizes the State to
                       issue variances from NPDWRs (the term
                       "State" is used in this preamble to mean
                       the State agency with primary
                       enforcement responsibility,  or
                       "primacy," for the public water supply
                       system program or EPA if the State does
                       not have primacy). The State may issue
                       a variance if it determines that a system
                       cannot comply with an MCL despite
                       application of the best available
                       technology (BAT). Under Section 1415,
                       EPA must propose and promulgate its
                       finding identifying the best available
                       technology, treatment techniques, or
                       other means available for each
                       contaminant, for purposes of section
                       1415 variances, at the same  time that it
                       proposes and promulgates a maximum
                       contaminant level for such contaminant.
                       EPA's finding of BAT, treatment
                       techniques, or other means for purposes
                       of issuing variances may vary,
 depending upon the number of persons
 served by the system or for other
 physical conditions related to
 engineering feasibility and costs of
 complying with MCLs, as considered
 appropriate by the EPA. The State may
 not issue a variance to a system until it
 determines that an unreasonable risk to
 health (URTH) does not exist. EPA has
 developed draft guidance, "Guidance in
 Developing Health Criteria for
 Determining Unreasonable Risks to
 Health" (EPA 1990k) to assist States in
 determining when an unreasonable risk
 to health exists. EPA expects to issue
 final guidance for determining when
 URTH levels exist later this year. When
 a State grants a variance, it must at the
 same time prescribe a schedule for (1)
 compliance with the NPDWR and (2)
 implementation of such additional
 control measures as the State may
 require.
  Under section 1416(a), the State may
 exempt a public water system from any
 MCL and/or treatment technique
 requirement if it finds that (1) due to
 compelling factors (which may include
 economic factors), the system is unable
 to comply,  (2) the system was in
 operation on the effective date of the
 MCL or treatment technique          !
 requirement, or, for. a newer system, that
 no reasonable alternative source of
 drinking water is available to that
 system, and (3) the exemption will not
 result in an unreasonable risk to health.
 Under section 1416(b), at the same time
 it grants an exemption the State is  to  .
 prescribe a compliance schedule and a
 schedule for implementation of any
 required interim control measures. The
 final date for compliance may not
 exceed three years after the initial d^te
 of issuance of the exemption unless the
 public water system establishes that: (1)
 The system cannot meet the standard   -
 without capital improvements which
 cannot be completed within the period
 of such exemption; (2) the system has
 entered into an agreement to obtain
 financial assistance for necessary
 improvements; or (3) the system has  .
 entered into an enforceable agreement
 to become part of a regional public
 water system. For systems that serve
 500 or fewer service connections and
 which need financial assistance to  come
 into compliance, the State may renew
 the exemption for additional two-year
 periods if the system is taking all
 practicable steps to meet the above
 requirements.
 3. Primacy

  As indicated above, States may
 assume primary enforcement
responsibility (primacy) for public water

-------
                Federal Register  /  Vol. 56,  No. 138 /  Thursday,  July 18, 1991  /  Proposed Rules
                                                                     33053
systems under section 1413 of the
SDWA. To assume or retain primacy,
States need not adopt the MCLGs but
must adopt, among other things,
NPDWRs that are no less stringent than
those EPA promulgates. States may also,
at their discretion, adopt standards more
stringent than the NPDWRs.
4. Monitoring, Quality Control, and
Recordkeeping
  Under section 1401(1)(D) of the Act,
NPDWRs are to contain "criteria and
procedures to assure a supply of
drinking water which dependably
complies with such maximum
contaminant levels; including quality
control and testing procedures to insure
compliance with such levels * *  *." In
addition, section 1445 (a)(l) states that
"every person who is a supplier of water
* * * shall establish and maintain such
records, make such reports, conduct
such monitoring and provide such
information as the Administrator may
reasonably require by regulation to
assist him hi establishing regulations
* * * in evaluating the health risks of .
unregulated contaminants, or in advising
the public of such risks." Section 1445
also requires  EPA to promulgate
regulations requiring every public water
system to conduct a monitoring program
for unregulated contaminants, and EPA
has established a number of specific
requirements.
5. Public Notification
  Section 1414(c) of the Act requires the
owner or operator of a public water
system which fails to comply with an
applicable maximum contaminant level
or treatment technique requirement,
testing procedure, or section 1445(a)
monitoring requirement to give notice to
the persons served by the water system.
Owners and operators of public water
systems for which variances or
exemptions are in effect, or which fail to
comply with the  requirement of any
schedule assembled pursuant to a
variance or exemption, must also give
notice. Section 1445(a)(5) also requires
public water  systems to notify the
persons served by the water system and
the Administrator of the EPA of the
availability of the results of monitoring
for unregulated contaminants. Public
notification regulations are codified at
40 CFR 141.32.
B. Applicability
  These proposed regulations would
apply to all community water systems
(CWSs) and all non-transient, non-
community public (NTNC) water
systems.
  Public water systems are defined hi
the Act as those systems which provide
piped water for human consumption and
have at least 15 connections or regularly
serve at least 25 people. Section
1401(1)(D)(4). The category "public
water system" is composed of
community and non-community water
systems. A community water system is
one which serves at least 15 connections
used by year-round residents or
regularly serves at least 25 year-round
residents (40 CFR 141.2). Non-
community systems, by definition, are
all other public water systems. Non-
community systems include transient
systems (e.g., restaurants and service
stations having independent water
sources) and non-transient systems
which EPA has defined as facilities that
have their own water supply and
regularly serve at least 25 of the same
persons for at least six months a year
(see 52 FR 25712, July 8,1987).
  Transient non-community water
systems would not be covered by these
proposed regulations. Environmental
levels of these contaminants pose public
health hazards over a long period of
exposure. Occasional and infrequent
exposure to environmental levels of
these contaminants pose minimal risks
to the public and do not warrant
regulation under the SDWA.
  EPA solicits public comment on the
application of these regulations to
community and non-community
nontransient public water supplies.
C. Regulatory Background
  In 1976, EPA promulgated the National
Interim Primary Drinking Water
Regulations (NIPDWRs) for combined
radium-226 and radium-228 at 5 pCi/1,
gross alpha particle emitters  at 15 pCi/1,
and beta particle and photon emitters
(also referred to  as the "man-made"
radionuclides) at a total dose equivalent
of 4 mrem/year to any organ or whole
body (40 CFR 141.15). These levels are
currently in effect and enforceable.
When these NIPDWRs were developed,
the Agency did not have sufficient
health and occurrence data on uranium
and radon to develop standards.
Therefore, there  are no existing primary
drinking water regulations for these two
radionuclides. As part of an effort to
develop better information for these
regulations, EPA sponsored a workshop
on radioactivity  in drinking water
(Health Physics,  1985).
  On September 30,1986, EPA published
an advance notice of proposed
rulemaking (ANPRM), (51 FR 34836,
Sept. 30,1986) concerning the
radionuclides  contained in today's
proposed action. The ANPRM discussed
EPA's understanding of the occurrence,
health effects, and risks from these
radionuclides, as well as the available
analytical methods and treatment
technologies and sought additional data
and public comment on EPA's planned
regulation. This notice builds on and up-
dates the information assembled for the
1986 ANPRM.
  The information in the ANPRM on
occurrence was estimated from the
nationwide compliance data for the
standards in place, several nationwide
and regional studies, and State data
bases. Although the  occurrence data for
uranium and radon were not as
complete as for the other regulated
radionuclides, the available  data
showed that uranium, radium, and radon
are seldom found together in high
concentrations. Relatively higher levels
of radium were found in the  midwest
and Appalachian region, natural
uranium in the Rocky Mountains, and
radon in the northeast. When the
ANPRM was published the available
data indicated that radon and uranium
generally were distributed at low levels
in water supplies throughout the United
States. In some areas, however, ground
water supplies had much higher levels of
radon. Compliance monitoring data on
radium indicated moderate occurrence,
primarily in the midwestern  states. Beta
particles and photon emitters were not
detected above the 50 pCi/1  screening
levels.
  The ANPRM summarized  the types of
cancer associated with each
radionuclide, the toxic effects of
uranium on the kidney, and the
estimated annual national risks posed
by each radionuclide in drinking water.
Several analytical methods were
mentioned and were presented along
with treatment technologies  and
estimated costs.

D. Comments by the Science Advisory
Board and the Public on the ANPRM

1. SAB Comment

  The EPA's Science Advisory Board's
(SAB) Radiation Advisory Committee
(RAG) reviewed the ANPRM and the
four draft criteria documents which
supported it prior to publication of the
ANPRM in the Federal Register (51 FR
34836; September 30,1986). EPA
subsequently revised the criteria
documents and resubmitted  them to the
SAB/RAC for review during the summer
of 1990. EPA has now revised the
criteria documents based on this latest
review (SAB/RAC, 1990) and presents a
summary of the SAB/RAC comments
and EPA's replies to them here. More
detail on these issues may be found in
the latest revised criteria documents
themselves.

-------
                Federal Register /  Vol.  56, No. 138 / Thursday,  July  18, 1991 / Proposed Rules
                        '\
   a. General comments and generic
 issues. In requesting review of the
 health criteria documents in 1990, the
 EPA requested the SAB/RAC to focus
 on five questions in their review, in
 addition to providing any additional
 comments the reviewers believed to be
 relevant. The five questions asked were:
   1. Are the estimates of the absorption,
 distribution and excretion of uranium,
 when ingested, appropriate and
 supported by the data?
   2. Do the estimates in the documents
 form an appropriate basis for  assessing
 the risks of directly ingesting water
 containing radon?
   3. What is an appropriate basis for
 estimating the risks from radon in
 water?
   4. What relative emphasis should be
 placed on the epidemiology data and
 modeled risk estimates for evaluating
 radium risks?
   5. Is the methodology for assessing
 risks from man-made radionuclides
 (both individually and collectively)
 appropriate?
   The  SAB/RAC reviewers also
 commented  on the overall quality of the
 draft documents and commented on
 several additional subjects.
   The  SAB/RAC comments were
 organized as follows: General
 Comments and Generic Issues;
 Responses to the Five Specific
 Questions; and Comments on Important
 Issues  in the Criteria Documents and
 Related Reports. EPA's replies to these
 comments follow the SAB/RAC
 organization, and are as follows:
   a=l. General comments and generic
 jfssues.The SAB/RAC made the
 following general comments:
   1. The general  quality of the
 documents was not good.
  EPA Reply: Full criteria documents,
 rather than only Quantification of
 Toxicological Effects sections, have
 been prepared, with careful review by
 ORP and ODW. Irrelevant information
 and incorrect definitions have been
 deleted, and definitive descriptions of
 the dosimetric models have been
 included in each Criteria Document.
 Except where noted in the Criteria
 Documents, the bases for selecting
 models is the same as those given by the
 ICRP in their publication ICRP 30 (ICRP
 1979). Material on chemical and physical
properties has been included, consistent
with OW format  for preparation of
 Criteria Documents. The five documents
have been made consistent in their
 approaches to risk assessment.
 Comments by the SAB/RAC made
 during the 1987 review have been
considered and addressed in the revised
Criteria Documents. EPA believes 0 the
overall quality of the revised documents
 is substantially improved, and will
 continue to udate the documents, as
 needed, between proposal and
 promulgation of this regulation.
   2. Technical decisions contrary to
 SAB and NAS recommendations were
 presented without discussion of
 alternatives or justification for the
 Agency's choices.
   EPA Reply: Detailed discussions are
 provided in the criteria documents of
 issues raised by the SAB, as indicated
 for document-specific comments below.
 The basis for adoption of SAB and NAS
 recommendations is presented in
 individual criteria documents and
 described briefly below. EPA's adoption
 of advice and guidance has attempted
 most appropriately resolve potentially
 conflicting recommendations, and
 strives to be consistent both internally
 and with other Federal Agencies in its
 assessments of radiation risks. EPA's
 modification of the ICRP dosimetric
 models  is used for assessing doses and
 risks from radium, uranium and gross
 alpha emitters, and for estimating doses
 used in  calculating the effective dose
 equivalent, which serves as the basis of
 the standard for beta and photon
 emitters.
   3. Uncertainties were not adequately
 addressed.
  EPA Reply: A new chapter has been
 added to each criteria document
 addressing uncertainties associated
 with the range of assumptions and
 models  considered and those arising
 from parameter variability (chapter IX
 in each  document).
  b. Responses to the five specific
 questions.
  1. Are the estimates of the absorption,
 distribution and excretion of uranium,
 when ingested, appropriate and
 supported by the data?
  The SAB/RAC believed the
 absorption, distribution and excretion
 estimates presented in the draft uranium
 criteria  document needed to be
 discussed in more detail and better
 supported by the criteria document. In
 particular, the SAB/RAC disagreed with
 use of 0.20 as the fi (gastrointestinal
 absorption factor) and cited a 1985
 review sponsored by the EPA as
 recommending an fi value of 0.014. SAB/
 RAG also urged that the value chosen be
 identified as representing the average
population or any special sensitive
groups.
  EPA Reply: EPA has extensively
reviewed the literature available on this
issue and believes that a value of 0.05 is
 appropriate. However, published studies
present  a wide range of possible values
for the uranium uptake factor. While
EPA believes a value of 0.05 is
supportable based on the literature,  the
 uncertainty associated with this value  ,
 may be great, perhaps a factor of 4
 greater or less than the value chosen.
 The basis for this uncertainty
 assessment is presented in the revised
 uranium health criteria document. EPA
 believes 0.05 to be a best estimate for
 the general population, and not a highly
 conservative value for the fi factor.
   2. Do the estimates in the documents
 form an appropriate basis for assessing
 the risks of directly ingesting water
 containing radon?
   The SAB/RAC urged EPA to better
 justify use of a fresh tap water
 consumption value of 0.66 liters/day, a
 value different than the 2 liters daily
 consumption usually used in assessing
 exposure to drinking water
 contaminants, and other assumptions
 about radon loss from water during
 consumption. The SAB/RAC also noted
 that the approach used for assessing the
 risks of radon in drinking water differs
 from that used for assessing risks from
 other volatile contaminants in drinking
 water.
   EPA Reply: A separate document was
 prepared by EPA to describe the data
 available for both the selected rate of
 ingestion and for radon loss during
 water consumption, and the rationale
 for the selected values. Consistency
 with previous regulations of volatile
 contaminants in drinking water is also
 addressed. These points are summarized
 in the Radon Criteria Document
 (sections IV.C.l and VIII.B.2) and the
 uncertainties are discussed in chapter IX
 of the Radon Criteria Document (section
 IX.B.1).
  The available data on tap water
 consumption is presented in "Radon in
 Drinking Water: Assessment of
 Exposure Pathways" (EPA, 1991h). EPA
 continues to believe a value other than 2
 liters per day is appropriate for
 assessing risks from ingested radon, and
 has used a value of 1 liter daily intake of
 fresh tap water, as a reasonable
 maximum, in the revised documents.
 EPA believes this is appropriate because
 radon is a volatile gas and will not be
 present in water used for cooking or
 making tea or coffee, or water that has
 been standing for some  time. EPA has
 therefore estimated consumption of
 water that is promptly consumed, i.e.
 water drawn from the tap and consumed
 immediately, for assessing radon
 exposure via the  ingestion route. EPA
 has previously used the 2 liter daily
 water consumption estimate in
 assessing risk for other volatiles in
 drinking water because  no separate
inhalation exposure and risk assessment
was performed. The exposure to other
volatile contaminants via ingestion

-------
                Federal Register  /  Vol.  56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                                     33055
predicted by 2 liters daily consumption
served as a surrogate to compensate for
the lack of a separate inhalation
exposure and risk assessment. Because
there are data on the transfer of radon
from water to the indoor air of homes,
an exposure assessment by the
inhalation route for radon derived from
water can be made. EPA has estimated
the inhalation exposure and risk and the
ingestion exposure and risk resulting
from radon in water separately, and
added the two assessments together in
estimating overall risks from radon in
water. The radon exposure pathways
document also describes the basis for
estimating 20% loss of radon from water
before it is consumed.
  3. What is an appropriate basis for
estimating the risks from radon in
water?
  The SAB/RAG asserted that use of a
generic tap water to air transfer factor
overlooks potential high radon
concentrations at the point of release,
such as during showering, and urged
inclusion of such an exposure
assessment in the revised documents.
SAB/RAG stated that all contributions
to total exposure should be considered,
and that uncertainties in all the
estimates must be addressed. SAB/RAG
also stated that there were differences
in the draft criteria document from a
draft radon risk assessment previously
submitted to SAB/RAG by EPA's ORP
for review.
  EPA Reply. An analysis of exposure
to radon during showering  and other
household uses of water was performed
by EPA and is presented in "Radon in
Drinking Water: Assessment of
Exposure Pathways" [EPA, 1991h). The
analysis was summarized in the Radon
Criteria Document (sections IV.C.2 and
VIII.B.2) and the uncertainties are
discussed in chapter DC of the Radon
Criteria Document (sections IX.A.3 and
IX.B.2).
  This document reviews the available
data and methods for evaluating
inhalation exposure to radon released
from water. These include  several
empirical studies as well as several
modeled approaches to exposure
assessment. EPA concluded in this
analysis that although mass balance
modeling can be performed for radon
from showering and other water use,
assessing risk based on this information
is difficult. Human activity patterns are
highly variable with regard to factors
that have a large influence on exposure,
such as temperature and length of
shower, shower flow rate,  timing of
multiple showers within a household,
and location and use of clothes washing
machines. Also, significant unanswered
questions remain about the equilibrium
of radon with its progeny in the shower
and bathroom, the unattached fraction,
and aerosol particle size in a shower
and behavior of water aerosols in the
respiratory tract. Modeling does allow
for risks from showers to be broadly
bounded, and EPA has done so in its
review. EPA concluded that integrated
exposure and risk estimates developed
from modeled water use through out the
house (including showering) differ only
slightly from the results obtained from
use of an average water to air transfer
factor such as 10,000:1 (i.e., 10,000 pCi/1
radon in water increases indoor air
levels by about 1 pCi/1), based on the
empirical data.
  In response to the final point of the
SAB/RAG, there is no overall
quantitative difference in the risk
assessment presented in the draft Radon
Health Criteria Document and the draft
submitted to the SAB for review in
February of 1990. Both present the
average unit risk value of 360 deaths per
106 working level months of exposure to
radon and its progeny, in air, as a
central estimate, consistent with EPA's
letter of November 23,1988 (EPA, 1988f]
to the SAB/RAG, which first used this
as the unit risk for radon. The source of
disagreement was apparently the lack of
separate presentation of risks to
smokers and nonsmokers. Risks to
smokers and non-smokers were not
presented separately and in detail as in
the February 1990 paper. EPA has added
this discussion to the revised Radon
Criteria Document (sections VI.C and
VIILB.2, Table VI-1). The preparation of
the Radon Criteria Document was
coordinated with the evolving ORP
position on indoor radon risks to the
extent the regulatory and review
schedules allowed, and EPA will
continue to update the document, as
needed, between proposal and        A
promulgation of the final rules.
  4. What relative emphasis should be
placed on the epidemiology data and
modeled risk estimates for evaluating
radium risks
  The SAB/RAG urged EPA to base its
risk assessment for radium on human
epidemiology data on radium watch dial
painters, rather than on modeled
estimates, and urged EPA to present its
rationale for adopting the modeling
approach for radium risk assessment.
The SAB/RAG also requested that EPA
better describe its dosimetric model in
the revised criteria document, including
calculated doses and risks to organs,
and that if EPA continued to use the
modeling approach, uncertainties in the
modeling be addressed.
  EPA Reply: The Agency carefully
reconsidered this issue. First it should be
pointed out that all risk estimates are
based on both epidemiologic data and
require mathematical modelling. The
EPA uses the wealth of epidemiologic
data on human exposure and risk of
radiogenic cancers, including radium
dial painters and epidemiologic data on
bone sarcomas resulting from injected
Ra-224.
  The watch dial painter data indicate
that the incidence of bone sarcomas
may follow a dose-squared response,
especially at higher exposures. EPA
policy, supported by recommendations
of SAB/RAG, is to assess cancer risks
from ionizing radiation as a linear
response. Therefore, use of the dial
painter data requires either deriving a
linear risk coefficient from significantly
non-linear exposure-response data, or
abandoning EPA policy and SAB/RAG
advice in this case. Two analyses were
recommended as alternatives by the
SAB/RAC,  those of Mays  et al. (1985)
and  of Schlenker (1982). Both analyses
used the same cohorts, calculated doses
and  definitions of incidence, and
differed primarily in the statistical
approach to deriving a linear slope that
would not be rejected by the
epidemiology data. The two resulting
values differ by about 60%. EPA was not
able to determine whether this degree of
agreement resulted from the use of
identical data, but took into account the
caution of the BEIRIV Committee (NAS,
1988) that there was no unique way to
derive a linear risk coefficient for bone
sarcomas from the dial painter data.
  There are, however, serious problems
in applying the watch dial painter
epidemiologic data. These include
uncertainties in intake, due to variability
in retention of radium and to lack of
measurement of Ra-228. There may also
be uncertainty in these data due to
possible bias in identification and
measurement of workers,  and the lack of
a unique way to specify the appropriate
extrapolation of the observed quadratic
response among workers at high intakes
with known abnormal bone physiology
to a linearized response consistent with
the lack of observed sarcomas among
lower-intake cohorts. There may also be
problems in extrapolating to continuous
intakes across years from a single
intake, and in assessing latency and
duration of plateau based on the
radium-224 data. The dial painter data
and the issues involved in extrapolation
are extensively discussed in the Radium
Criteria Document (sections III.B, VI.B.l,
VIII.B.2, IX.A.1 and K.A.2, Tables VI-1
to VI-3, VIII-1 and VIII-2), and a
thorough discussion of the RADRISK
model has been incorporated (sections
III.D, VILB, VIII.B.2, IX.A.2 and DC.B.2,
Tables III-l and VIII-3 to VIII-5).

-------
33056
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991 /  Proposed Rules
   An alternative to the dial painter data
 for deriving a linear coefficient is the
 experience with patients injected with a
 short lived isotope of radium. The BEIR
 III committee (NAS, 1980) found that
 these epidemiology data were consistent
 with a linear relationship between dose
 and bone sarcoma incidence, and
 derived a linear risk coefficient. Because
 of the difference in the toxicokinetics
 between the short-lived and the long-
 lived isotopes of radium, modelling is
 required to use  the BEIR III risk
 coefficient. The use of models
 introduces some uncertainty into the
 assessment of risk but has the
 advantage that  differing patterns of
 exposure can be evaluated (e.g. constant
 lifetime exposure).
   The RADRISK model (Sullivan et al,
 1981; Dunning et al., 1980; EPA, 1989a)
 used by EPA to assess risk from
 radionuclides also allows calculation of
 radiologic doses to and cancer risks in
 organs other than bone, based on
 epidemiology data on cancer risks from
 several studies of effects of ionizing
 radiation.
   One concern the SAB/RAC had with
 this model was that the predicted
 incidence of leukemias was higher than
 observed in the dial painter cohorts.
 EPA reexamined this prediction and
 revised the calculation of the high-LET
 radiation risk to bone marrow to be
 more consistent with the predictions of
 the watch dial painter study and the
 observations in the spondylitic  and
 Thorotrast studies. The predictions of
 the RADRISK model were adjusted to
 give a relative incidence of leukemias
 and bone sarcomas more consistent
 with observed data, which is also more
 consistent with  the watch dial painter
 data (as described in section VIII.B.2,
 Table VIII-5 of the revised criteria
 document). Data on the leukemia
 incidence reported in patients injected
 with Thorotrast, a thorium-based
 radiologic contrast agent were also
 examined (NAS 1988). EPA has also
 added head carcinoma risk to the model
 (for radium-226), consistent with the
 watch dial painter studies.
  As a result of this reconsideration
 EPA continues to incorporate the
 estimate  of the bone sarcoma risk
 coefficient derived from epidemiology
 data that show a linear dose-response
 curve (the data for radium-224), a
revised bone marrow risk coefficient
 and hence leukemia risk, and has added
 a risk coefficient for radium-226 induced
head carcinomas. These issues and
EPA's conclusions are discussed in the
revised radium health Criteria
Document, and as  requested by SAB/
RAG, an expanded description of the
                       RADRISK model and assessment of
                       uncertainties have been added.
                         5. Is the methodology for assessing
                       risks from man-made radionuclides
                       (both individually and collectively)
                       appropriate?
                         The SAB/RAC urged EPA to include
                       risks from man-made alpha emitters as
                       well as beta emitters, urged use of EPA
                       "official" risk estimates, and urged that
                       the results be presented without
                       reference to likely regulatory levels.
                         EPA Reply: EPA has revised its risk
                       assessment numbers to correspond to
                       previous estimates generated by the
                       RADRISK model, and will incorporate
                       any revisions based on the
                       recommendations of the BEIR V report
                       only after SAB/RAC has had an
                       opportunity to review and comment on
                       them as a separate issue and not in the
                       context of this proposed regulation.
                       Similarly, only unit risk and dose
                       assessments are presented in the
                       revised criteria document, without
                       reference to possible regulatory levels.
                         c. Comments on important issues in
                       the criteria documents. SAB/RAC also
                       made the following comments on the
                       draft Criteria Documents:
                         i. Uranium criteria document. 1. The
                       document fails to explain  selective
                       adoption of the recommendations of the
                       BEIR IV report, in particular the BEIR IV
                       use of analogy with radium as the basis
                       for risk evaluation of uranium, and the
                       BEIR IV conclusion that any cancer risk
                       from uranium is from bone sarcoma, not
                       other organs as predicted  by the EPA
                       model.
                         EPA Reply: For a number of reasons
                       discussed above, EPA has continued to
                       rely on its risk model for assessing
                       radium cancer risks, and uses this
                       approach for assessing uranium cancer
                       risks as well. EPA, like the ICRP,
                       evaluates dose and risk for a  number of
                       organs and tissues and combines them
                       as appropriate to obtain the risk
                       estimate. EPA believes that all emitters
                       of ionizing radiation are carcinogenic.
                       EPA has reviewed and revised a key
                       parameter value used in this model, the
                       fi value, according to SAB/RAC
                       recommendations, and has also revised
                       the predicted risks of leukemia, as
                       described above for radium, and the
                       risks to kidney. These revisions are
                       discussed in greater detail in sections IV
                       and VIII of the revised uranium Criteria
                       Document.
                         2. The uncertainty  in the risk
                       assessment for uranium must be
                       discussed.
                         EPA Reply: An analysis of the
                       uncertainties in the uranium cancer risk
                       estimate has been prepared and is
 presented in section IX of the revised
 uranium Criteria Document.
   3. If a modeled approach is chosen,
 EPA must justify selection of the models
 and parameter values, in particular the
 f1 value of 0.20 used in the draft criteria
 document, and the work of Wrenn et al.
 (1985) and Spencer et al. (1990; as cited
 in EPA, 1991e) must be addressed.
 Quality of the data and the possible
 effect of diet and eating habits on the
 uptake of uranium must be considered.
   EPA Reply: As discussed above, EPA
 has reviewed and revised the fi value
 used in estimating uranium risks. The
 work of Wrenn et al., and Spencer et al.,
 were considered in this review.
 Evaluation of the data quality and the
 possible effect of diet and habits, i.e.,
 iron deficiency and the "no-breakfast
 syndrome," are presented in the
 uncertainty discussion of the revised
 Uranium Criteria Document.
   4. Comments and recommendations of
 the 1987 Drinking Water Subcommittee
 review have not been incorporated into
 the document, and the document
 includes irrelevant information (on
 inhalation studies) and some incorrect
 definitions.
   EPA Reply: EPA has reviewed the
 comments made by the 1987 Committee
 review, and addressed those that remain
 pertinent to the revised documents.
 Studies by exposure routes other than
 ingestion have been included in the
 Criteria document, where those studies
 indicate systemic effects and especially
 where data by the ingestion route are
 sparse. Terms and definitions have been
 reviewed and corrected where found to
 be incorrect.
   ii. Radium criteria document. 1.
 Extrapolation of risk from dial painter
 data.
  EPA Reply: This issue is addressed hi
 question number 4 above.
  2. Uncertainties are not adequately
 addressed.
  EPA Reply: EPA has added an
 assessment of the uncertainties in the
 risk evaluations to  all of the revised
 Criteria Documents.
  3. The estimate of radium absorption
 from Maletskos et al. (1966) should be
 discussed further and uncertainties
 addressed.
  Response: The discussion has been
 expanded (section III.A) and
uncertainties are addressed (section
IX.B.1).
  4. The issue of sensitivity of children
to  non-cancer effects of radium should
be revised.
 Response: This recommendation was
followed (sections III.B, VI.C, VIII.A and
IX.A.1).

-------
                Federal Register / Vol.  56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                                     33057
  5. The RADRISK model should be
described in more detail, and the over-
prediction of leukemias, lack of
prediction of head carcinomas, and
relative risks of Ra-226 and Ra-228
should be addressed.
  EPA Reply: This recommendation was
followed (sections IH.D, VH.B, VHI.B.2,
IX.A.2 and IX.B.2, Tables ffl-1 and Vni-3
to VIII-5 of the radium Criteria
Document). As described above, the
estimates of leukemia risk for radium
220 and 228 have been revised and the
head carcinoma risk for radium-226
added, consistent with the watch dial
painter data. Organ doses and risks to
bone and other organs are also
presented. On review, EPA discovered
an error in the estimated radhim-228
dose to bone marrow and bone surface,
and has revised the dose, and hence risk
estimate for radium-228 to be consistent
with the dose estimated by the IGRP 30
model (EPA, 1991b). It should be noted
that the 2.5 fold higher potency of
radium-228 in inducing bone sarcomas
among the watch dial painters relates to
an instantaneous intake of radium, and
less of a difference would be expected
for continuous lifetime exposure
(Rowland et. al, 1978).
  iii. Documents related to radon. 1.
Inconsistencies in the relative
conservativeness of the assumptions
across the criteria documents for
radionuclides and with regulation of
other volatile chemicals in water should
be addressed.
  EPA Reply: Parameter values used in
the criteria documents have been
reviewed and are now more consistent
with regard to their degree of
conservativeness.
  As discussed above, radon is the first
drinking water contaminant for which
the inhalation pathway is specifically
addressed as a separate exposure
pathway. This involves adjusting the
ingestion risk downwards to account for
loss of radon from tap water used for
cooking and in other ways that would
cause radon loss (making coffee, tea,
etc.) and also separately quantifying
inhalation exposure from all household
uses of water. The Agency made an
extensive analysis of the exposure to
radon by ingestion and by inhalation of
radon released from household uses of
water, including short-term exposure
during showering. This analysis is
presented in a  separate document (EPA
1991h) and summarized in the Radon
Criteria Document (sections IV.C.l,
IV.C.2, Vin.B.2, IX.A.3, IX.B.1 and
IX.B.2).
  2. Uncertainties should be discussed,
particularly of variability of important
parameters in the risk assessment. -
  Response: Chapter IX of the Radon
Criteria Document addresses
uncertainties both from the range of
assumptions and models and from
parameter variability.
  3. The discussion of radon health risks
should be updated and made consistent
with the ORP approach, and the
appendix discussions of non-cancer
health effects of radiation exposure
should be omitted.
  EPA Reply: The discussion of miner
data, including Lubin et al.  (1990, as
cited in EPA, 1991c), has been updated
(Radon Criteria Document section
VI.B.2) and risks of inhaled radon decay
products have been listed separately for
smokers and nonsmokers (sections VI.C
and VHI.B.2, Table VI-1). Genetic effects
are discussed in the Radon Criteria
Document (sections VI.B.l,  VI.B.2,
VHI.C, IX.A.4 and IX.B.3, Tables VIII-7
to VIII-9) because these may be relevant
in the context of radon in drinking
water.
  4. The basis for the rate of
consumption of tap water and the loss of
radon should be presented  and
defended.
  EPA Reply: This has been done in a
separate document (EPA 1991h) and
summarized in the Radon Criteria
Document (sections IV.C.l, VIII.B.2 and
IX.B.1).
  5. The basis for the selection of the
transfer factor for waterborne radon
contribution to indoor air radon levels
should be presented and defended.
  Response: This has been done in a
separate document (EPA 1991h) and
summarized in the Radon Criteria
Document (sections IV.C.2, VIII.B.2 and
IX.A.3 and IX.B.2).
  6. The daily acute exposure from
showering should be considered,
including the degree of radon
equilibrium.
  EPA Reply: This has been done in a
separate document (EPA 1991h) and
summarized in the Radon Criteria
Document (sections IV.C.2, VIII.B.2 and
IX.A.3 and EX.B.2).
  7. Additional analysis of the ingestion
model by Crawford-Brown (1990) would
be useful, including extending the
analysis  of uncertainty.
  EPA Reply: The analysis of
uncertainty in radon ingestion risks is
extended in the Radon Criteria
Document (sections IX.A.2  and IX.B.1).
The model of Crawford-Brown (1990),
which has been published in peer-
reviewed journals (Risk Anal, 11:135-
143,1991), was considered  to be the best
analysis  available for assessing risks of
ingested radon.
  8. The  document should not contain
incorrect definitions of fundamental
technical terms or basic fallacies.
  EPA Reply: The Radon Criteria
Document has undergone extensive
internal Agency review to correct
inaccurate terminology.
  iv. Manmade Radionuclides
Document. 1. The document on
manmade radionuclides used risk
factors inconsistent with the other
radionuclides discussed here and used
an ad hoc extrapolation of risk factors
based on an assessment of the BEIR V
report that has not been submitted for
review by the SAB, in spite of a previous
agreement to do so.
  EPA Reply: As described above,
EPA's established risk factors have been
used in the revised Criteria Document.
Use of risk factors based on the BEIR V
report will be delayed until EPA has
reviewed these with the SAB/RAC in a
separate evaluation.
  2. The evaluation of risks should be
based on the ICRP effective dose
equivalent concept.
  EPA Reply: EPA has used its own
dosimetric model (the RADRISK model),
based to a large degree on ICRP models
and parameters, in the revised criteria
document on beta and photon emitters.
  3. The document should define the
potential risks of exposure, rather than
define the regulatory value of 4 mrem
ede/yr.
  EPA Reply: The revised beta and
photon emitter Criteria Document
assesses risks and does not present a
regulatory value. Regulatory values for
the beta and photon emitters, based on
the unit risks in the Criteria Document,
are presented in appendix B of this
notice.
  4. The document fails to adequately
discuss uncertainties associated with
the values of parameters selected and
overall uncertainty of the evaluation.
  EPA Reply: An assessment of the
uncertainties in the risk estimates has
been added to each of the Criteria
Documents.
  5. Tables A-l and V-l are misleading
or difficult to understand as presented.
  EPA Reply: These tables have been
revised to clarify the information
presented in them.

2. Public Comment on the ANPRM
  EPA requested comments on all
aspects of the September 30,1986
ANPRM. A summary of the major
comments, and the Agency's response to
the issues raised, are presented below.
A detailed enumeration of the comments
received and the Agency's responses is
presented in the document "Response to
Comments Received on the NPDWRs:
Radionuclides in Drinking Water—
Advanced Notice for Proposed
Rulemaking of September 30,1986,"

-------
33058
Federal  Register / Vol. 56, No.  138 / Thursday, July 18, 1991  /  Proposed Rules
 (EPA, 1991J) which is available in the
 public docket for this rulemaking.
   EPA received 44 written comments on
 the ANPRM. Of the comments received,
 2 were from individuals, 2 were from
 Federal agencies, 11 were from States, 3
 from local governments, 15 were from
 companies, 4 were from public water
 supplies and 8 were from public or
 professional organizations.
   EPA held a public hearing on
 November 13,1986. Representatives of a
 professional organization and of a
 company each made a statement and
 two local government representatives
 reported on levels of radionuclides in
 their water.
   Because some of EPA's approaches to
 risk evaluation and regulation have
 been revised since 1986, some of the
 comments and issues are addressed
 only in the comment response
 document. Those still considered
 significant are discussed here.
   a. EPA's proposal to set a MCLG and
 MCL for natural uranium.  A total of 16
 commenters addressed EPA's advanced
 notice for regulating uranium.
 Commenters raised four major issues
 regarding the uranium regulation:
   (1) Toxicity Versus Carcinogenicity
   (2) No-Observed-Adverse-Effect-Level
   (3) Risk Estimates
   (4) Economic Impact
   (1) Uranium toxicity versus
 carcinogenicity
   Comments: Ten commenters
 questioned EPA's proposal on the basis
 of insufficient scientific evidence to
 show that natural uranium is a
 carcinogen. One commenter disagreed
 with EPA's rationale to regulate uranium
 based on similarities to radium and
 another maintained that EPA's
 comparison of radium and uranium was
 flawed because EPA ignored the fact
 that uranium will expose tissues at a
 much lower dose and dose rate. In '
 support of EPA's proposal, one
 commenter urged EPA to set MCLGs at
 zero because of the lack of available
 data on the radiotoxic effects of uranium
 and because of similarities between
 radium and uranium.
   EPA Response: Uranium, like radium,
 is a source of ionizing radiation which
 decays and emits alpha particles
 internally, thereby irradiating internal
 tissues. Uranium also concentrates in
 bone as does radium, and kidney.
 Ionizing radiation has been shown in
 many studies to be carcinogenic in
 humans and EPA has classified it as a
 group A carcinogen. Uranium has
 caused cancers at multiple sites in
 laboratory animals, as would be
 expected from a source of ionizing
 radiation. Furthermore, the human
 carcinogenic risk from  ingested radium
                       is well-established (EPA, 1991b). For
                       these reasons, the Agency is proposing
                       to establish an MCLG for natural
                       uranium based on it being a carcinogen.
                       Uranium also is believed to be toxic to
                       the kidneys, and below, EPA discusses
                       exposure levels that would be
                       considered safe for this adverse effect.
                       In setting a standard, EPA will ensure
                       that the eventual MCL is protective for
                       both the carcinogenic potential of
                       uranium and for kidney toxicity. For the
                       purposes of this rule the MCL is based
                       on uranium's potential for kidney
                       damage.
                          Comment: One commenter stated that
                       information presented at the National
                       Workshop for Radioactivity in Drinking
                       Water held in May 1983, indicated that
                       carcinogenic risks were negligible from
                       uranium, as well as from radium and
                       radon.
                         EPA Response: The risk level
                       estimated in the 1983 Workshop on
                       uranium was 6 x 10 ~7 per pCi/1 lifetime
                       cancer risk (Mays et al., 1985). Since the
                       1983 Workshop, EPA has continued to
                       assess the hazards of all the
                       contaminants in this proposed rule. The
                       Agency still believes the risk from
                       uranium to be approximately 6 x 10~7
                       per pCi/1 (EPA, 1991e). EPA does not
                       regard this risk as negligible.
                       Longstanding and carefully considered
                       EPA policy for regulating carcinogens in
                       drinking water is that the lifetime
                       individual risk target is one in 10,000
                       (10~*) to one in 1,000,000 (10"6) risk. As
                       uranium occurs in water used as a
                       source of drinking water at levels posing
                       risks within this target range, the
                       Agency believes regulation is
                       warranted. Uranium is also toxic to
                       kidney at concentrations that may be
                       found in drinking water, and protection
                       against this potential hazard is also
                       warranted. In addition, regulation of
                       uranium in drinking water is  required by
                       the 1986 amendment to the SDWA.
                       (2) No-Observed-Adverse-Effect-Level
                         Comments: Two commenters cited
                       data showing that the lowest
                       concentration of uranium shown to
                       cause kidney damage is 3 jxg per gram of
                       kidney with 1 fig/gram kidney being a
                       kidney concentration well below the
                       level causing kidney damage. The
                       commenter stated that this
                       concentration in water is approximately
                       equivalent to an exposure of 1,00 pCi/1.
                       Another commenter believed there is no
                       reason to develop a regulatory limit for
                       uranium of less than 5 mg/1, asserting
                       that 5 mg/1 is the accepted, nontoxic
                       level for natural uranium from heavy
                       metal toxicity.
                         EPA Response: The study that the first
                       commenter is referring to (Wrenn et al.,
1985) goes on to derive an intake limit
for uranium in drinking water based on
the 1 fig per gram of kidney as a no-
toxic-effects concentration level. Using a
GI absorption estimate of 1.4% for
humans at environmental levels of
uranium .intake, a safety factor of 50,
and  a 1.711/day water intake, the study
recommends a 100 jug/1 limit for uranium
in drinking water.
   Based on evidence from a number of
chronic and subchronic toxicity studies
with several species of animals, EPA
has identified a lowest-observed-
adverse-effect-level (LOAEL) of 2.8 mg
uranium/kg/day based on moderately
severe renal damage following 30 days
of dietary administration of uranyl
nitrate to rabbits (EPA, 1991e). From this
LOAEL, the Agency calculated a
reference dose (RfD), or daily exposure
for humans likely to be without
appreciable risk of adverse health
effects during a lifetime.  The RfD is 3 x
10~3mg/kg/day (EPA, EPA, 1991s).
   When estimating drinking water
contaminant levels for contaminants or
effects associated with identified
thresholds, EPA calculates a Drinking
Water Equivalent Level (DWEL), a
drinking-water specific lifetime
exposure for the contaminant at which
adverse non-carcinogenic health effects
are not anticipated to occur. This DWEL
for uranium was calculated to be 0.10
mg/1 (or 100 ftg/1) using kidney toxicity
to adults as an endpoint. When setting
an MCLG based on an identified
threshold, the DWEL is multiplied by the
relative source contribution (RSC) for
water (the fraction of total  exposure that
derives from drinking water) to form the
basis for the MCLG. EPA examines the
available data on other exposure
sources to identify the RSC, and uses a
value of 20% as a default value if data
are not available or are of poor quality;
that  is the case with uranium. This
would give an MCLG of 20  ju.g/1, or
approximately 30 pCi/1. This level is
well below the level cited by the second
commenter as an accepted, nontoxic
level for natural uranium. These issues
are discussed in greater detail in
Sections III. and IV. below.

(3) Risk Estimates

   Comment: One commenter stated that
EPA's risk estimates for uranium are
flawed because they were developed
using a linear dose-response curve that
overestimated risk from lifetime
exposure to water supplies having up to
100 pCi/1 of uranium. This commenter
urged EPA to  consider the BEIRIV
report which contains information
concerning the extrapolation of the

-------
                Federal Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed Rules
                                                                     33059
biological effects of all alpha emitters,
including uranium.
  EPA Response: The BEIR HI report
(NAS, 1980) recommends linear dose
response curves for use in assessing
risks from all alpha emitters, and as
appropriate for uranium, since it is an
alpha emitter. The BEIR IV (NAS, 1988)
report makes no clear recommendation,
but rather discusses the implications of
making different choices among the
possible alternative approaches.
(4) Economic Impact
  Comments: Two commenters argued
that the cost of treatment for uranium is
too high, especially for small water
systems, considering the lack of data
showing that uranium is carcinogenic.
  EPA Response: As stated above, EPA
believes that there is adequate scientific
evidence to show that uranium is
carcinogenic to humans.
  Costs for uranium removal are
dependent on water system size,
concentration of uranium in source
water, and the type of removal
treatment used. EPA has determined
that proposed BATs for uranium
removal are affordable by regional and
large public water systems (EPA, 1991i).
EPA also evaluates total, or national
compliance costs as well as household
costs and cost-effectiveness in assessing
feasibility of treatment. EPA considers
the cost of the health protection
afforded by the proposed MCLs to be
reasonable (EPA, 1991i). While
affordability assessments are based on
cost to regional and large water
systems, variances and exceptions may
be available for some small systems if
required conditions are met (i.e., see
section V.IJ. Variances or exemptions
may not be granted if doing so would
result hi an unreasonable risk to health.
The Agency has specified proposed
BATs for variance purposes for small
water systems (see section V.B) and is
continuing to evaluate what costs are
reasonable for public water systems.
  b. EPA's proposal to set separate
MCLGs and MCLs for radium-226 and
radiam-228. A total of 11 commenters
submitted comments regarding the
appropriateness of establishing
combined or separate MCLGs and MCLs
for radium-226 and radium-228.
  Comments: Most commenters on this
Issue supported the establishment of
separate MCLGs and MCLs for the two
contaminants, citing several reasons:
each appears to be different
toxicologically, each has different
degrees of biological effectiveness, and
each has different risk levels associated
with identical concentrations.
  In opposition to separate MCLGs and
MCLs for radium-226 and radium-228,
one commenter maintained that the
database for radium-228 is insufficient
to warrant separate regulations for the
two isotopes.
  EPA Response: The Agency does not
agree that the database for radium-228
is insufficient to warrant separate
regulation. There is sufficient scientific
evidence that carcinogenic risks from
radium-228 are not qualitatively
different from radium-226 risks (EPA,
1991b).
  As discussed above, and in detail in
the revised health criteria document for
radium, EPA has classified radium-228
as a group A human carcinogen.
Radium-228 is a beta emitter that
irradiates the bone and other organs
where it is deposited; EPA has classified
all ionizing radiation as a group A
carcinogen. Use of human epidemiology
data in conjunction with the RADRISK
model estimate the lifetime cancer risk
from radium-228 at approximately
3X10~6per pCi/1. The epidemiology
studies addressing radium-226 and -228
directly also indicate that two types of
cancer, bone sarcomas and head
carcinomas, are elevated in persons who
have been exposed to ingested radium.
Rowland et al. (1978) compared the
relative effectiveness of radium-226 and
radium-228 in inducing bone sarcomas
and concluded that radium-228 was
more effective in inducing bone
sarcomas than radium-226. In addition,
they demonstrated that incidence of
head carcinomas were associated only
with exposure to radium-226, not
radium-228. This would be expected if
the accumulation of radon gas in the
mastoid air cells and paranasal sinuses
is important in the etiology of these
tumors.
  EPA also included radium-228 in its
MRS  survey of ground water systems
nationwide (EPA, 1988b). EPA therefore
has extensive data on the occurrence of
radium-228 in public water supply
ground water, as described in section III.
below. EPA also has data supporting the
analytic chemistry methods to
determine compliance with the radium-
228 MCL, and treatment information
showing the levels to which it can be
removed from drinking water, as
described in section V. below.
  Finally, analytical methods for
radium-226 and radium-228 differ; and,
analysis of NIRS co-occurrence data
suggests that in coupling regulation for
the two isotopes and using the interim
monitoring scheme, about half of actual
violations were not detected since in
most cases only a gross alpha test or
radium-226 test were done (the interim
monitoring requirements only required
radium-228 monitoring when the gross
alpha measurement exceeded 5 pCi/1, 40
 CFR 141.26(a)(l)(i); EPA, 1988d). The
 proposed revision to the monitoring
 requirements described in section V.
 below would rectify this problem.
   c. Disposal of radioactive waste
 generated from treatment of water for
 radionuclides. A total of 15 commenters
 discussed the need for EPA to address
 technical, regulatory, and economic
 aspects of treatment and disposal of
 radioactive waste resulting from water
 treatment to remove radionuclides.
   Comments: Commenters urged  EPA to
 address the issue of disposing radium-
 contaminated sludge from lime softening
 treatment, uranium-containing spent
 alumina, and uranium-contaminated
 sludge from coagulation treatment using
 alum or iron salts.
   Commenters pointed out that the
 waste streams generated by reverse
 osmosis and electrodialysis treatments
 for uranium could contain triple the
 uranium concentration of the raw
 material, and that a large problem
 associated with reverse osmosis
 treatment for uranium would be disposal
 of large volumes of brine generated by
 the process and disposal of the uranium-
 contaminated salts remaining after brine
 water evaporation.
   EPA Response: At the present time
 there are no federal regulations
 specifically addressing the disposal of
 wastes generated by water treatment
 processes on the basis of their
 radionuclide  content. There are
 regulations that apply to disposal of
 radioactive wastes in general, and these
 would apply to drinking water treatment
 wastes that are radioactive.
   In order to guide water treatment
 facilities and State and local regulators
, toward safe waste management
 practices for water treatment plant
 wastes containing radionuclides above
 background levels, EPA has reviewed
 regulations and guidelines which
 address the handling and disposal of
 wastes containing naturally occurring
 radionuclides originating from industries
 other than water treatment.
   Based on these regulations and
 guidelines, EPA has developed
 suggested guidelines for disposal
 options and institutional controls which
 would be pertinent for drinking water
 treatment wastes containing naturally-
 occurring radioactive contaminants at
 various ranges of concentration. These
 guidelines are presented in "Suggested
 Guidelines for the Disposal of Drinking
 Water Treatment Wastes Containing
 Naturally-Occurring Radionuclides"
 (EPA, 1990a).
   For disposal of liquid wastes, or
 brines, EPA suggests discharge to
 surface water, discharge to sanitary

-------
33060
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991  / Proposed Rules
sewer, deep well injection, or
evaporation or chemical precipitation
followed by land disposal, as permitted
by State and local regulations. For
disposal of solid wastes, or sludges, EPA
suggests disposal in a municipal landfill,
a stabilized or institutionally controlled
landfill, a hazardous waste disposal site,
a permitted or licensed naturally-
occurring or accelerator-produced
radioactive material (NARM) facility, or
a licensed low-level radioactive waste
disposal facility (should the waste
become low-level radioactive waste).
The selection of a waste disposal option
may be influenced by a variety of
federal, state  or local regulatory
constraints and water treatment facility
site specific conditions. Waste disposal
is discussed in greater detail  in Section
V.C below.
  Comments: Eight commenters were
concerned that disposal costs for water
treatment waste would significantly
raise the  treatment costs presented in
theANPRM.
  EPA Response: The treatment and
disposal of wastes generated by the
treatment processes could increase
overall treatment costs and may be
beyond affordability for some small
systems. However, hi establishing
proposed BAT, EPA identified the
treatment and disposal technologies that
are reasonably available for large
metropolitan regional drinking water
systems (systems which serve 50,000 to
75,000 persons). In this determination,
EPA evaluates total, or national
compliance costs as well as household
costs. EPA has determined that disposal
of waste from treatment for
radionuclides does not significantly
increase the total water treatment costs
for large systems and that the proposed
regulations are, overall, affordable. EPA
has also included the estimated cost of
waste disposal in its overall evaluation
of cost of the proposed regulations (EPA,
19911). Estimates of waste generation
and cost of disposal are described in
Tables 12-14, in section V.C below.
  As previously mentioned, under
certain conditions, variances and
exceptions from any MCL requirement
or NPDWR treatment technique
requirement may be available for some
small systems (see section V.I).
Variances and exemptions may be
granted by the States to systems if
installed BAT does not achieve
compliance, or for compelling economic
reasons, as long as granting such a
variance would not result in an
unreasonable risk to the health of the
water supply  customers.
  d. EPA's proposal to set a MCLG and
MCL for gross alpha radiation.  A total
of 19 individuals or organizations
                        submitted comments on EPA's proposal
                        for regulating gross alpha particle
                        activity.
                          Comments: A majority of the
                        commenters responding to this issue
                        disagreed with EPA's proposal to
                        regulate gross alpha radiation with an
                        MCLG and MCL; favoring the idea that
                        gross alpha be used as a screening
                        device only.
                          In support of a MCL, one commenter
                        asserted than a total alpha activity MCL
                        must be promulgated because Congress
                        included "gross alpha particle activity"
                        as one of the 83 contaminants specified
                        for MCL development under the SDWA.
                          EPA Response: Compliance
                        monitoring has only occasionally
                        detected naturally-occurring
                        radionuclides in drinking water other
                        than radium-226, radium-228, uranium,
                        or radon-222. Nevertheless, EPA
                        believes that this  does not preclude the
                        possible presence of other alpha
                        emitters, including transuranic man-
                        made alpha emitters, and believes that a
                        MCLG and MCL for gross alpha particle
                        activity will provide  adequate protection
                        from alpha emitters that could
                        potentially occur in drinking water. EPA
                        believes an MCL for gross alpha particle
                        activity will also provide a ceiling on the
                        aggregate exposure and aggregate risk
                        from all alpha emitting radionuclides.
                        EPA is also obligated to develop an
                        MCL for gross alpha  emitters by the 1986
                        amendments to the SDWA, which listed
                        gross alpha as one of the  83
                        contaminants to be regulated.
                          Gross alpha measurements  will also
                        be used as a screen for radium-226  and
                        uranium compliance  and may reduce
                        monitoring costs.
                          e. EPA's proposed amendment to the
                        definition of gross beta and photon
                        emitters. Seven commenters provided
                        comments on EPA's proposed definition
                        of gross beta and photon emitters.
                          Comments: Three commenters stated
                        that the definition is misleading because
                        some naturally occurring radionuclides
                        (e.g., potassium-40 and carbon-14) decay
                        by beta emission.
                          Another commenter pointed out that
                        some radionuclides which decay by
                        processes other than alpha or beta
                        decay, such as electron capture or alpha
                        emission accompanied by photon  ,
                        emission, would be excluded by the
                        proposed definition.
                          EPA Response: EPA is proposing to
                        regulate approximately 200 beta and
                        photon emitting radionuclides of which
                        most, but not all are man-made. EPA
                        considers an overall MCL for beta and
                        photon emitters to be more appropriate
                        than specific MCLs because of the low
                        possibility of occurrence.
  Radionuclides which decay by
processes such as electron capture or
alpha emission accompanied by photon
emission would not be excluded from
the definition.
  f. Comments on risk models used to
determine estimated risk values. A total
of 14 commenters addressed the
appropriateness of using an absolute
risk model versus a relative risk model,
and the appropriate application of risk
values generated by the two models.

(1) Risk Model Selection
  Comments: One commenter believed
either a relative or absolute risk model
was appropriate, but that the selection
of a model should depend on the
biological endpoint to be evaluated. This
commenter added that the relative risk  •
model could overestimate risk. Another
commenter urged that EPA consider a
quadratic dose-response risk model for
radium-228 and for its risk in causing  ,'•
bone sarcomas. Two commenters stated
that there are data to show sensitivity to
radionuclide induced cancer decreases
with age, and suggested that the relative
risk model would be appropriate for
younger age groups. One commenter
stated that EPA should select the risk
model with the most supporting data
and address the upper range of risk
estimates as generated by that model.
One commenter believed that either
relative or absolute risk models are
acceptable because both methods yield
negligible risk.
  EPA Response: EPA recognizes that
there has been no model developed to
date which perfectly and consistently
describes the carcinogenic risks
associated with exposure to radiation
and that all existing risk models can
potentially over- or under-estimate
actual risks. However, radiation risks
are among the most studied and best
understood, and there is a general
consensus among the scientific
community that for solid rumors other
than bone, the relative risk model
appears to most appropriately describe
how carcinogenic risk develops over age
and time. Leukemia and bone cancer
appear to better fit a model in which
risk peaks a few years after exposure
and then decreases subsequently. This
view is supported by a variety of
sources (UNSCEAR, 1988), (NRPB, 1988),
(RERF, 1987), (NAS, 1980,1988). The risk
models described in these sources use
age- and organ-specific risk coefficients
so that any age sensitivity to radiation
induced cancer is incorporated in the
models.
  Comment: One commenter
encouraged EPA to assume input
parameters for risk assessment models

-------
                Federal Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed Rules
                                                                     33061
that are mean values as opposed to
using conservative values.
  EPA Response: In its risk assessments
for radionuclide risks the Agency
generally does use best estimates rather
than cmservative values.
  g. CL.  merits on the appropriateness
of setting one dose equivalent MCL
standard for all radionuclides found in
drinking water.
  A total of 11 commenters addressed
the appropriateness of a combined MCL
standard for all radionuclides found in
drinking water.
  Comments: Most commenters opposed
establishing a combined MCL for alpha-
emitting radionuclides for the following
reasons: biological endpoints vary
among isotopes, radionuclides differ
with respect to occurrence and
toxicology; one standard would mislead
the public; and a combined standard
would require an extensive effort to
perform a dose assessment for each
radionuclide.
  One commenter noted that although it
is conceptually valid to establish a
combined MCL, its implementation
would be more difficult due to higher
analytical costs.
  EPA Response: The Agency agrees
that a combined MCL for all alpha-
emitting radionuclides would not be an
appropriate regulatory approach for two
reasons. First, the effective dose
equivalent (EDE) estimates for alpha
particle emitters would be too uncertain
to be the basis for risk assessment
intended to support standards, because
the range of alpha emissions is  so short
and the pharmacokinetics of alpha
emitters are so complex (although the
Agency believes they are reliable
enough to be the basis for comparisons
among the radionuclides). Second, it is
known that some alpha-emitting
radionuclides (i.e., uranium, radium and
radon) are more widespread than others
(EPA, 1985a; 1988b) and have more well-
established carcinogenicity. Proposed
monitoring requirements (i.e. gross-
alpha screening) would serve to identify
other, lesser occurring alpha-emitting
radioactive contaminants in an effective
and cost efficient manner.
  The Agency agrees with the statement
made by one commenter that
implementation of a combined MCL
would have higher costs due to the
extent of unnecessary monitoring that
might occur.
  h. Comments on regulation of man-
made radionuclides as a class. A total
of 15 individuals or organizations
commented on the appropriateness of
establishing a MCLG and MCL as
opposed to a health advisory for the
entire class of man-made radionuclides.
  Comments: A total of eight
commenters felt that EPA should not
establish MCLGs or MCLs for man-
made radionuclides. Seven of these
commenters expressed the view that
EPA should not establish MCLGs or
MCLs for man-made radionuclides
because the presence of these
contaminants in drinking water is
generally the result of accidental
discharges already addressed by other
federal regulations. Five commenters
stated their support for the
establishment of non-regulatory Health
Advisories for man-made radionuclides
rather than MCLGs or MCLs.
  In support of establishing both an
MCLG and MCL for man-made
radionuclides, two commenters
proposed that EPA require monitoring of
gross beta activity for a water system
only long enough to establish that
noncompliance with the MCL was
unlikely. However, one of these
commenters added that gross beta
monitoring should be conducted if an
event occurred that was expected to
result in radionuclide contamination of
the water supply.
  Another commenter suggested that if
an MCLG and an MCL are set, a cost-
effective alternative to the requirements
of the NIPDWR for gross beta
monitoring would be to drop strontium
and tritium from the required analyses,
except in the case of an accident
causing greater than 50 pCi/1 of gross
beta emissions.
  EPA Response: The Agency agrees
that the presence of man-made
radionuclide contamination in drinking
water generally results from accidental
discharges. EPA believes that because
these contaminants are known
carcinogens and one potential exposure
pathway is through drinking water,
setting an MCLG and MCL and requiring
periodic monitoring for this class of
radionuclides is appropriate, especially
when a potential source of chronic
contamination exists. In addition, EPA is
obligated under the 1986 amendments to
the SDWA to set an MCL for beta and
photon emitters.

E. Other EPA Radon and Radiation
Programs
  EPA has  developed the Radon Action
Program, a primarily non-regulatory
program, to reduce the health threat of
indoor radon hi air. Radon from soil gas
is the principal source of radon in the air
of homes, and EPA recommends that all
homes be tested for radon. The relative
risks of radon in air and water are
discussed in more detail in section V.F.
below.
  EPA's Radon Action activities are
conducted under the authority of the
Indoor Radon Abatement Act (IRAA).
They include: National and state radon
surveys to measure radon levels in
homes and schools; the Radon
Measurement Proficiency (RMP)
program, which evaluates radon testing
companies; the Radon Contractor
Proficiency (RCP) program, which trains
and evaluates radon mitigation
contractors; the establishment of four
regional training centers across the
country; and the development of model
standards for construction of new
housing to prevent elevated radon in
new homes.
  EPA has also prepared a variety of
public information materials to educate
the public about radon and to encourage
people to test their homes and reduce
elevated radon levels. EPA's "Citizen's
Guide to Radon," (EPA, 1986f) which
recommends that indoor air radon levels
above 4 pCi/1 in homes be mitigated, is
currently being updated to incorporate
the latest health risk information on
radon from both soil and water, as well
as mitigation technology. EPA also
works with the Advertising Council on a
national media campaign to motivate
the public to test homes and fix elevated
levels. EPA also conducts public
outreach activities with the American
Lung Association on a variety of
outreach activities in States across the
country, including media events and
workshops held during Radon Action
Week last October.
  Public information materials on radon
testing and mitigation in the home can
be obtained from the national radon
hotline at 1-800-SOS-RADON.
  There are also regulatory programs
that restrict radon and other
radionuclide exposures. In November of
1989 EPA issued final regulations
restricting radon emissions to the air
from several categories of point sources,
under section 112 of the Clean Air Act.
EPA also has standards for both existing
and new uranium mining and mill-
tailings piles.

F. Basics of Radiation

  The study of radiation is a specialized
scientific field and much of the public
water supply industry and public
affected by this regulation may have
only a limited understanding of it. To
help provide a better understanding of
radiation and these proposed
regulations, appendix A presents a
discussion of the fundamental concepts
of radiation, its nomenclature, and its
measurement.

III. Occurrence and Exposure

  There are approximately 2,000 known
radioisotopes, or radionuclides. These

-------
33062
Federal  Register / Vol. 56, No.  138 / Thursday, July 18, 1991  / Proposed Rules
isotopes emit radiation as they undergo
radioactive decay (alpha particles, beta
particles and gamma rays or photon
radiation). They can be classified
generally into two categories: natural •
and man-made, and are also frequently
categorized by their primary mode of
radioactive decay, i.e, by alpha or beta
or gamma emission. Most radionuclides
are mixed emitters to some degree, and
each has a primary mode of
disintegration with some smaller
percentage of the atoms present
decaying by others. The natural
radionuclides are largely alpha particle
emitters with some beta particle activity
from the progeny. The most significant
natural radionuclides (as determined by
their levels of occurrence in drinking
water and their potential to cause
adverse health effects by this exposure
route) are radon-222, radium-226,
radium-228, and uranium. Some other
alpha emitting radionuclides have
                        occasionally been found in drinking
                        water.
                         In setting drinking water MCLs, the
                        agency generally sets individual
                        contaminant standards. In this notice,
                        EPA is proposing to set MCLs for the
                        most prevalent radionuclide
                        contaminants, and standards for broad
                        categories of other much less prevalent
                        radionuclide contaminants. Because in
                        this notice EPA is proposing to set MCLs
                        near the 10"4 estimated lifetime risk
                        level for the contaminants regulated,
                        concern about co-occurrence of these
                        contaminants at the MCL levels arose
                        (EPA, 1988a). Water supply systems
                        having two  or more of these
                        contaminants at the MCLs could be
                        placing then1 customers at total risk
                        higher than EPA's target of 10~4 lifetime
                        risk. In addition, co-occurrence of
                        several that can be removed using the
                        same treatment could make removals
                        more cost-effective. Because the data
                        examined to date are limited, EPA
solicits additional data on co-occurrence
to enable a more complete assessment
of the potential for co-occurrence of
these contaminants near the proposed
MCLs.
  The natural radionuclides involve
three decay series which start with
uranium-238, thorium-232 or uranium-
235. These three series are shown in
Figure 1. These are called the uranium,
thorium, and actinium series,
respectively. Each series decays through
stages of various nuclides which emit
either an alpha or beta particle as they
decay and ends with a stable isotope, of
lead. A number of radionuclides also
emit gamma rays, which accompany the
alpha or beta decay. The uranium-238
series contains both radium-226 and
radon-222 in the decay series and ends
with the stable lead-206. The thorium-
232 series contains radium-228 and ends
with the stable lead-208.
BILLING CODE 6560-50-M

-------
        Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
33063
                            86
Figure 1.  Uranium and thorium isotope decay series
               THE ACTINIUM SERIES

-------
33084
Federal Register / Vol. 56, No. 138 / Thursday, July 18,1991 / Proposed Rules
                                87
                   THE THORIUM SERIES

-------
                  Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
33065
88
THE URANIUM SERIES
23SJJ
92
4.5xl09 yr

234Pa
91
6.8 hr
T A
234Th
90 6
24 da


««U
92
2.5X105 vr

P
a
"°Th
90
7.5xl04 yr

a
226Ra
88
1,600 yr

a
222Rn
86
3.8 da

a
218Po
84
3 min
i
214p
82
27 n
a
b
tin

Z14Po 210pn
844 84° ,
1.6xlOHsec 138 da
i A
214Bi 210Bi
83 P S3 p a
20 min 5 da

210pb 206pb
P 82 ft 82
22.3 yr Stable
BILLING CODE 656O-50-C

-------
33066
Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
  The man-made radionuclides fall into
two subcategories. For those
radionuclides of an atomic weight higher
than uranium in the Periodic Table (the
transuranics), generally both alpha and
beta particle decay modes occur. By
contrast, almost no radionuclides below
thallium (A=81) exhibit alpha particle
decay properties. They undergo decay
by beta and/or gamma ray emission.
  Of the radionuclides that comprise the
natural decay series, radium, uranium
and radon are most commonly found at
detectable levels in drinking water.
Many of the man-made radionuclides
have half-lives too short to allow them
to be transported completely through a
drinking water system. (The half-life of
an isotope is the time required for one-
half of the atoms present to decay.)
However, approximately 200 man-made
radionuclides do have half-lives long
enough to be considered potential
contaminants in drinking water, and
there are a few reported cases of high
levels of naturally occurring beta
emitters (e.g., lead-210) in private wells.
Thus, the 200 man-made and naturally-
occurring radionuclides are included  as
a class of beta and photon emitters in
this discussion.
  The estimates of radon, radium, and
uranium levels in drinking water of
public water systems presented in this
section are based on EPA's National
Inorganics and Radionuclides Survey
(MRS) (EPA, 1988b). Also presented for
each radionuclide is a summary of the
findings of the survey on the Nationwide
Occurrence of Radon and Other Natural
Radioactivity in Public Water Supplies
(EPA, 1985a). The title is shortened to
"The Nationwide Radon Survey" in the
following discussions. NIRS was
initiated in the early 1980s to
characterize the occurrence of a variety
of substances, including the naturally
occurring radionuclides covered in this
proposal. Estimations of radionuclide
levels derived from other available
nationwide monitoring data were
presented in the Advance Notice of
Proposed Rulemaking (ANPRM) (51 FR
34836, Sept. 30,1986). These have been
revised based on the NIRS data, and the
impacts and benefits of this proposal are
estimated based on an analysis of the
NIRS data.
  The NIRS (EPA, 1988b) survey was
designed as a stratified sample based on
the population served. The universe of
public groundwater supplies was
stratified into four size categories (by
population): very small (serving 25-500),
small (serving 501-3,300), medium
(serving 3,301-10,000) and large/very
large (serving more than 10,000). There
are approximately 60,000 community
                       water systems nationwide. Of these,
                       approximately 48,000 are served
                       primarily by groundwater, 33,000 of
                       which serve 500 or fewer people, about
                       10,000 serve people in communities of
                       500 to 3,300, 2,400 serve communities of
                       3,300 to 10,000, and about 1,200 serve
                       10,000 or more people. A total of 1,000
                       sites were selected randomly in
                       proportion to the number of public
                       groundwater supplies in each category.
                       Approximately 2.1 percent of the
                       drinking water supplies in each size
                       category were selected. (Note: Sample
                       results for the various constituents were
                       reported for 990 of the 1,000 sites
                       selected.)
                         The national occurrence estimates for
                       radon, radium-226, radium-228 and
                       uranium were obtained through
                       statistical modeling of occurrence
                       distributions derived from the results of
                       NIRS. Lognormal distributions were
                       computed for each radionuclide for each
                       of the four size strata noted above.
                       These distributions were computed
                       using statistical techniques that allowed
                       for the "non-detects" (referred to as
                       censored data) to be taken into account
                       in calculating the parameters of these
                       distributions. The details of the
                       methodology are provided in the
                       occurrence documents prepared by EPA
                       for these contaminants.
                       A. Radium-226
                         According to the NIRS (EPA, 1988b)
                       data approximately 40% of the systems
                       sampled in NIRS had radium-226 above
                       0.18 pCi/1, the Minimum Reporting Level
                       (MRL). However, less than 9% of the
                       systems exceeded 1 pCi/1 and only
                       about 1% exceeded 5 pCi/1. The
                       maximum level reported was 15.1 pCi/1.
                       The mean and median of the positive
                       values (those above the MRL) were 0.87
                       and 0.4 pCi/1, respectively. The overall
                       mean value was 0.4 pCi/1, assuming a
                       value of 0.9 pCi/1 (i.e., one-half the
                       MRL) for those systems with results
                       below the MRL.
                         NIRS also computed population-
                       weighted averages for the states having
                       supplies sampled in NIRS, and reported
                       that the highest values were found in
                       Illinois, Wisconsin, Minnesota,  and
                       Missouri, a region recognized by others
                       (e.g., Hess et al., 1985) for having high
                       radium-226 levels.
                         The national occurrence estimates
                       derived from NIRS indicate that
                       approximately 25,000 community and
                       non-transient non-community ground
                       water supplies in the U.S. have  radium-
                       226 level above 0.18 pCi/1.
                       Approximately 600 of these supplies are
                       expected to have radium-226 above 5
                       pCi/1 (half of which serve 500 or fewer
                       people), and approximately 70 are
expected to have levels exceeding 20
pCi/1 (20 of which serve 3,300 or fewer
people, and 40 of which are estimated to
serve 3,300 to 25,000 people) (EPA,
  Based on those occurrence estimates,
it is also estimated that 3.4 million
people using ground water systems are
exposed to radium-226 levels exceeding
5 pCi/1, and 890,000 are exposed to
levels above 20 pCi/1 (EPA, 1991i).
  Quantitative estimates of radium-226
occurrence and exposure in public water
supplies using surface water sources
could not be generated due to the lack of
comprehensive national survey data.
However, based on the  information
discussed in Hess et al. (1985), it
appears reasonable to conclude that the
overwhelming majority of surface water
supplies have levels between 0.1 and 0.5
pCi/1.

B. Radium-228

  NIRS (EPA, 1988b) reported that
radium-228 was found to exceed the
MRL of 1 pCi/1 in approximately 12% of
the systems sampled in NIRS. Less than  "
4% had levels above 5 pCi/1, and the
maximum value reported was 12.1 pCi/
1. The mean and median of the positive
values were 2.0 and 1.5  pCi/1,
respectively. The overall mean, using 0.5
pCi/1 for those systems below the MRL,
was 0.7 pCi/1 (EPA, 1990n).
  The national occurrence estimates for
radium-228 indicate that approximately
500 ground water supplies have levels
exceeding 5 pCi/1 (400 of which serve
3300 or fewer people), approximately 40
systems exceed 20 pCi/1 (most serving
3300 or fewer people) and 15-20 exceed
30 pCi/1. The corresponding exposure
estimates are that 1.3 million people
using ground water supplies receive
water with radium-228 levels above  5
pCi/1, and 164,000 are exposed to water
exceeding 20 pCi/1, and about 82,000 are
exposed to water exceeding 30 pCi/1
(EPA, 1990n; 1991i).
  Similar to radium-226, there are
inadequate survey data to estimate
national occurrence of radium-228 in
water-supplies using surface water
sources. However, Hess et al. (1985) also
reported that surface water levels for
radium-228 are low in comparison to
ground water levels.
C. Radon
1. Occurence.

  NIRS (EPA, 1988b) reported that radon
was found to exceed the MRL of 100
pCi/1 in approximately  72% of the
supplies sampled in NIRS. About 11% of
the NIRS systems were found to have
levels above 1,000 pCi/1, and 1%

-------
                Federal Register / Vol. 56,  No. 138  /  Thursday,  July 18, 1991 / Proposed Rules
                                                                     33067
reported radon levels above 10,000 pCi/
1. The maximum value reported was
25,700 pCi/1. The mean and median
values for the positive sites were 881
and 289 pCi/1, respectively. The overall
mean, using a value of 50 pCi/1 for those
sites below the MRL, was reported to be
648 pCi/1.
  Based on the NIRS data, it is
estimated that approximately 45,000
community and non-transient
noncommunity ground water supplies in
the U.S. have radon levels above 100
pCi/1. About 25,900 are estimated to
have levels exceeding 300 pCi/1, with
9,400 exceeding 1,000 pCi/1.
Approximately 80-85% of all systems
exceeding any of these values serve 500
or fewer people. It is also estimated that
47 million people are served by those
systems having radon levels above 100
pCi/1,17 million by those having radon
levels above 300 pCi/1,  and 2.7 million
by those with levels above 1,000 pCi/1
(EPA, 1991i).
  Quantitative estimates of the
occurrence  of radon in public water
supplies using surface water sources
could not be developed due to the lack
of data. However, based on the limited
information provided in the Nationwide
Radon Survey it appears that levels in
such supplies are very low compared to
levels observed in ground water
supplies. Of 25 surface water systems in
the Nationwide Radon Survey for which
data were available, 23 (92%) had levels
below 100 pCi/1. The mean level was 34
pCi/1, with a maximum level reported at
240 pCJ/1. The Agency requests that
data on radon levels in water supplies
using surface water sources in this
notice be submitted, if such data are
available.
  Radon levels in ground water can also
vary on a diurnal or longer term basis.
Data on radon variability were
developed by Kinner et. al (Kinner, 1990)
during their study of radon treatments.
A review of the monitoring data taken
from several of the wells used in the
treatment studies showed up to 2 fold
variations in radon levels at various
wells over periods of one year or less.
Variability  over the course of a single
day was generally less than over the
longer periods. EPA has also funded an
ongoing study by the State of
Connecticut to investigate the variability
in radon levels in water. EPA will
incorporate these results when they are
available.
  Because of this variability in radon
levels in water, EPA is proposing more
frequent monitoring for radon than the
other contaminants proposed for
regulation here, but will also allow
averaging of results for  determining
compliance, as described in section V.G
below. EPA solicits additional data on
the variability of radon levels in water,
and on use of these data in establishing
compliance monitoring requirements.

2. Assessing individual radon exposure
from inhalation and ingestion
  Because it is a volatile contaminant,
radon poses exposure issues not
encountered in estimating exposures
(and risks) for other drinking water
contaminants. In assessing exposure
and risk from radon, EPA has generated
two separate exposure (EPA, 1991h) and
risk assessments (EPA, 1991c), by the
inhalation and ingestion exposure
routes.
  For other volatile contaminants
regulated under the SDWA, EPA has
continued to use its estimate of 2 liters
of daily water consumption to assess
overall exposure and risk. EPA
estimated that while a volatile
compound may be lost from water used
for cooking or to make tea or coffee  (and
therefore the ingestion exposure would
be lost), there would be an inhalation
exposure to the contaminant
approximately equivalent to the amount
lost in cooking, etc., via contaminant
release to the air (from all water uses in
the house). Because adverse health
effects for the VOCs were systemic
rather than route specific, exposure
route was not critical if overall exposure
was adequately estimated. In addition,
there were few data on inhalation
exposures to volatile drinking water
contaminants. Therefore, continued  use
of the 2 liters daily water consumption
served as an adequate surrogate for
total exposure by both routes.
  In considering exposure and risk
estimation for radon there were two
critical differences that led EPA to its
present approach of generating two
route specific exposure and risk
assessments. First, it was possible to
generate a reliable average estimate of
inhalation exposure, although there can
be substantial individual variability.
Empirical studies have been conducted
on the transfer of radon from water to
the air of a house (Hess et. al., 1991),
and several published modeling
approaches to assessing exposure are
available. EPA's assessment of these is
described in detail in the background
document "Radon in Drinking Water:
Assessment of Exposure Pathways"
(EPA 1991h). Second, there are
important route-specific considerations
in assessing radon risks. While radon is
considered a known human carcinogen
by both ingestion and inhalation, the
type and quality of information on
which to base a risk assessment is
different for the two routes. Risk of lung
cancer by inhalation from radon and its
progeny is based on a series of human
epidemiology studies, as described
below, and has many elements specific
to radon with its progeny in the air. The
target organ for these studies was the
lung only. Risk by ingestion is based on
modeled estimates of radiation dose and
risk to all body organs as a result of
consuming water containing radon.
  In assessing indoor air exposure to
radon resulting from its presence in
drinking water, EPA has used an overall
average estimated factor for transfer of
radon from water to air of 10,000 to 1
(i.e., 10,000 pCi/1 radon in water
contributes 1 pCi/1 to air). EPA
extensively reviewed both the empirical
data and the various modelling
approaches that are available, including
exposure to radon during showering.
EPA's review is presented in "Radon in
Drinking Water: Assessment of
Exposure Pathways" (EPA 1991h). As
described above in EPA's reply to
comments from the SAB/RAG, EPA
concluded that although mass balance
modeling can be performed for radon
from showering and other water use,
assessing risk based on this information
is difficult. Human activity patterns are
highly variable with regard to factors
that have a large influence on exposure,
such as type and length of shower, flow
rate, timing of multiple showers within a
household, and location and use of
clothes washing machines. Also,
significant unanswered questions
remain about the equilibrium of radon
with its progeny in the shower and
bathroom, the unattached fraction, and
aerosol particle size in a shower and
behavior of water aerosols in the
respiratory tract. Modeling does allow
for risks from showers to be broadly
bounded, and EPA has done so in its
review. EPA concluded that exposure
and risk estimates developed from
modeled water use throughout the house
(including showering) differ only slightly
from the results obtained from use of an
average water to air transfer factor such
as 10,000:1, based on the empirical data.
EPA is therefore using the 10,000:1
transfer factor as an average for
purposes of assessing national risks to
radon in drinking water.
  In assessing exposure and risk due to
ingestion of radon in water EPA used a
value less than its standard assumption
of 2 liters daily water consumption.
Because radon is a volatile gas, only
water freshly drawn from the tap and
directly consumed will have appreciable
amounts of radon. Even water directly
consumed after being drawn will have
less radon than would be measured by
carefully drawing a sample from the tap
for monitoring purposes, because of

-------
33068
Federal Register / Vol. 56, No. 138 / Thursday,  July 18,  1991 / Proposed Rules
aeration and agitation of the water in
the process of drawing the water and
consuming it. EPA therefore applied a
correction factor of 0.20 (i.e., reduced by
20%) to fresh, directly consumed tap
water to account for radon loss resulting
from the act of drawing and drinking the
water (EPA, 1991h). EPA also reviewed
the available data on water ingestion,
rates, and presents its analysis in the
background document (EPA, 1991h).
This analysis separately estimates fresh
tap water intake, total tap water intake,
and total fluid intake. Because only
freshly drawn and directly consumed
tapwater is expected to contain radon,
the direct tap water intake values were
considered for assessing exposure to
radon via ingestion. Based on this
analysis, EPA estimated an average
direct tapwater intake of 0.65 liters
daily, rounded to 0.7 liters. However,
EPA has considered its 2 liter daily
intake to be a "reasonable maximum"
estimate, and believes water intake for
assessing radon exposure via ingestion
should be consistent with this. As noted
in the analysis, Ershow and Cantor
(1989) found that fresh tapwater intake
was 55% of total tapwater. Using  this
percentage with the 2 liter assumption
results in a reasonable maximum fresh
tap water exposure of 1.1 liters daily.
EPA has rounded this value to 1 liter of
daily directly consumed tap water for
assessing radon exposure via ingestion
of drinking water (in addition to
airborne exposures).
  EPA solicits public comment on the
radon exposure issues discussed here.
Specifically, EPA solicits public
comment on use of an average water to
air transfer factor of 10,000 to 1 for
inhalation exposure to radon and its
progeny, and on possible alternative use
of models to assess exposures,
especially during possible high exposure
activities such as showering, and
especially focusing on dosimetry issues
in this exposure scenario. EPA also
specifically solicits comment on its use
of 1 liter daily consumption of freshly
drawn, directly consumed tap water as
a reasonable maximum estimate for
assessing exposure to radon via
ingestion, and possible alternative use
of the average value of 0.7 liters daily
water intake. Finally, EPA solicits
comment on the estimated 20% loss of
radon from water before consumption.
D, Uranium
  Natural uranium contains three
isotopes: uranium-234, uranium-235 and
uranium-238. The corresponding
percentages of occurrence in rock for
these isotopes are 0.006, 0.72 and 99.27
percent by weight, respectively.
However, the percent occurrence of
                        these isotopes relative to each other is
                        not constant in drinking water.
                        Uranium-238 and uranium-234 are
                        responsible for most of the uranium
                        radioactivity in natural waters. The
                        overall activity-to-mass of uranium ratio
                        for the three natural isotopes of uranium
                        in rock is approximately 0.68 pCi/jug
                        and is frequently used to estimate the
                        activity of total uranium measured as
                        mass (EPA, 1988b; EPA/ORNL, 1981).
                        The 0.68 pCi/jug value is based on the
                        natural crustal abundance of isotopes.
                        The uranium-234/uranium-238 activities
                        ratio of one, that is inherent in this
                        assumption, may not be appropriate for
                        samples taken from water. The
                        Nationwide Radon Survey (EPA, 1985a),
                        which measured uranium as well as
                        radon, reported a range of uranium-234
                        to uranium-238 activity ratios in water
                        of 0.7 to 32 with an arithmetic mean of
                        4.4 and a geometric mean of 2.7. Using
                        the uranium-234 to uranium-238 activity
                        ratio of 2.7, an overall activity to mass
                        ratio of 1.3 pCi/jag was calculated for
                        uranium as it occurs in drinking water
                        (EPA, 1990h; 1991o). The 1.3  factor was
                        applied to the NIRS results to convert
                        those data from mass (/xg/1)  to activity
                        (pCi/1) for total uranium.
                         Approximately 72% of the sites in
                        NIRS had uranium levels above.0.1 pCi/1
                        (0.08 ftg/1). Most of these (70%) had
                        levels between 0.1 and 20 pCi/1
                        (approximately 0.08 and 15 ju.g/1).
                        Uranium was found to exceed 30 pCi/1
                        (20 ju.g/1) in only about 1% of the systems
                        in NIRS. The maximum value found was
                        115 pCi/1 (88.2 jxg/1) (EPA, 1991o).
                         Based on an analysis of the NIRS
                        data, national occurrence estimates for
                        community and non-transient non-
                        community water supplies (both ground
                        and surface water) indicate that
                        approximately 1500 will have levels
                        exceeding 20 jtg/1, serving
                        approximately 875,000 people (EPA,
                        1991i). Of the 1500 systems exceeding 20
                        (Ltg/1,1460 are estimated to serve 3300 or
                        fewer people. The  available  data on
                        uranium in surface water supplies was
                        limited. Although levels are expected to
                        be lower than for ground water systems,
                        unlike radium and radon they may not
                        be insignificant. As a conservative
                        estimate of occurrence, the ground
                        water occurrence distributions were
                        applied to surface  water systems to
                        derive the above estimate (EPA, 1991i;
                        1991o).
                         Uranium is a kidney toxin (as well as
                        a carcinogen) and EPA is proposing to
                        base the MCL on kidney toxicity, as
                        discussed in sections IV.C.3 and V.F
                        below, because kidney toxicity may
                        occur at levels below the 10~4 cancer
                        risk level. The MCLG is being proposed
as zero, and the relative contribution of
exposure from other sources is not
usually considered. However, because
kidney toxicity is the limiting toxic
endpoint of concern for regulation,
uranium exposure from sources other
than drinking water was reviewed, to
derive a relative source contribution
(RSG) factor, to ensure that the MCL is
set at a safe level.
  In determining how to consider
exposures by routes other than drinking
water in establishing standards, EPA
first reviews all relevant exposure data
on the contaminant. This typically
involves reviewing dietary intake data,
and assessing the relative contributions
of diet and drinking water to total
intake. The fraction of total intake
accounted for by drinking water as a -
source is the relative source contribution
factor for drinking water. When data are
inadequate to confidently estimate this
value, a default value of 20% is used. A
ceiling of 80% for the relative source
contribution is also used. EPA's
approach to determining relative source
contributions is decribed in more detail
in the Federal Register published May
22,1989, on pages 22069-20070.
  The data available on uranium intake
from various food sources are described
in the occurrence document for uranium
(EPA 1990h; 1991o). Those data indicate
that median dietary uranium intake from
food is generally low, approximately 1.3
pCi/day as an average, with a 90th
percentile of approximately 5 pCi/day.
However, these data represent residents
of only three cities, on the east coast  -
and west coast, with no assessment of
dietary intake for residents of the
midwest or west, where uranium in soil
and water may be higher.            .
  EPA is proposing to use the 20%
default value as the RSC for use in
calculating a uranium MCL because of
the poor data base for estimating dietary
exposures. EPA recognizes this may be
a conservative assessment, but believes
it is warranted because the available
data on uranium intake via food do not
include areas of the country expected to
have uranium in the soil and water.
Those areas may need lower water
contributions to total uranium intake in
order to maintain total uranium intakes
low enough to ensure safety from kidney
toxicity. EPA solicits public comment on
use of the default value of 20% RSC for
uranium. EPA is especially interested in
additional data on uranium intake from
food to better estimate an alternative
RSC value between 20% and 80%. EPA
also solicits public comment on its
general approach to determining the
relative source contribution factor,

-------
                Federal Register / Vol.  56, No. 138 / Thursday, July 18, 1991  /  Proposed Rules
                                                                      33069
including its method of calculation and
20% and 80% boundaries.
E. Beta and Photon-Emitting
Radfonuclides
  The availability of data on the
occurrence of man-made radionuclides
in public water supplies  is very limited.
The major source of relevant
information is the ERAMS
(Environmental Radiation Ambient
Monitoring System), the  data for which
are published in the quarterly ERD
(Environmental Radiation Data; as
reported in EPA, 1989c) reports. The
ERD reports provide data on gross beta,
tritium, strontium-90, and iodine-l3l for
78 sites (all surface water sources) that
are either major population centers or
selected nuclear facility  environs.
  The data presented in the ERD reports
for 1985 through 1987 indicate that gross
beta levels ranged from 0.3 to-17.8 pCi/1,
with an average across all three years of
less than 3 pCi/1 (EPA, 1989c). There
were no instances where gross beta
exceeded 50 pCi/1. Tritium levels in this
period were reported to range between 0
and 2,500 pCi/1, with average values
across all three years generally falling
between 100 and 300 pCi/1. Strontium-90
values did not exceed 0.9 pCi/1, with
typical values falling below 0.2 pCi/1.
Iodine-131 levels were all below 0.4 pCi/
1, with average values below 0.1 pCi/1
EPA, 1989c).
  As is apparent from these data,
nuclear facilities routinely release very
small amounts of these materials to the
environment during their normal
operations. These releases are of
concern only to a few drinking water
supplies, i.e., those supplies downstream
from nuclear facilities or using a water
source that may be affected by nuclear
facility releases. While normal releases
pose very low risks, accidental or
unscheduled releases could be of
concern.
  One naturally occurring beta and
photon emitter potentially of concern is
lead-2lO, the first long-lived progeny of
radon-222. Lead-210 was not monitored
in the NIRS survey, and data on its
occurrence in drinking water supplies
are limited (EPA, 1991g). However, the
drinking water concentration estimated
to correspond to 4 mrem ede/yr
(assuming 2 liters daily intake) is \ pCi/
1, a level low enough to potentially
warrant health concern,  and below the
PQL for the gross beta screen. As
discussed in section V.G below, EPA is
proposing unregulated contaminant
monitoring of lead-210 in public water
supplies, to better assess any risk posed
and to evaluate the possible need to
develop an MCL for lead 210.
F. Alpha-Emitting Radionuclides

  Gross alpha is a measure of the alpha
particle emissions from total non-
volatile alpha emitting radionuclides. .
Since radium 226 and uranium are alpha
emitters that are proposed to be
regulated separately, the gross alpha
occurrence assessment is adjusted to
eliminate these radionuclides. The term
"adjusted gross alpha" represents total
gross alpha measurements less radium
226 and uranium contributions. EPA is
proposing an "adjusted gross alpha"
MCL as a means of limiting exposures to
a number of other radionuclides that do
not occur frequently enough to warrant
a national regulation but may be present
in some water supplies. These include
several of the progeny of the
radionuclides for which contaminant-
specific standards are being proposed
today. The  adjusted gross alpha MCL is
distinguished from the gross alpha
laboratory measurement to avoid
confusion.
  The evaluation of the NIRS (EPA,
1988b) database for adjusted gross alpha
entails the manipulation of three sets of
data (i.e., gross alpha, radium 226, and
uranium). Each data set has its own
detection limit and inherent uncertainty,
and the analysis of all three data sets
together to  estimate occurrence
increases the overall uncertainty of the
results. To create the most meaningful
data set of adjusted gross alpha, the
NIRS data were evaluated in terms of a
reasonable worst case approximation,
which represents the highest reasonable
estimate of gross alpha concentrations
(EPA, 1991f). An attempt to estimate the
lower bound was unproductive, because
when lower bound assumptions  were
made in evaluating the three data sets
together, there were too few positive
data points to model national
occurrence.
  Due to the lack of national data,
quantitative estimates of the occurrence
of adjusted gross alpha in surface water
supplies could not be generated
independently. As a conservative
estimate, the ground water occurrence
distributions were applied to surface
water systems (EPA, 1991f).
  Based on the upper bound
approximation, 17% of the systems
sampled in  NIRS reported adjusted
gross alpha above 2.6 pCi/1, the
minimum reporting level for gross alpha.
The maximum level was 94 pCi/1. The
overall mean and median levels were 2.7
and 1.8 pCi/1, respectively. Fewer than
7% reported levels above 5 pCi/1, 3%
reported levels above 10 pCi/1, 2%
reported levels over 15 pCi/1 and only
1% had levels over 20 pCi/1 (EPA, 1991f).
  National occurrence estimates based
on the upper bound approximation for
adjusted gross alpha indicate about 1200
water supplies (serving 5 million people)
exceeding 5 pCi/1, 300 systems (serving
1.8 million people) exceeding 10 pCi/1,
130 systems (serving 900,000 people)
exceeding 15 pCi/1 and 65 systems
(serving 500,000 people) exceeded 20
pCi/1 (EPA, 1991f, EPA, 1991i).
Approximately 90% of the systems
affected at any of these levels serve
3300 or fewer persons.
  EPA notes however, that this analysis
has a high degree of uncertainty, due to
the simultaneous assessment of the
three data bases together. Also, analytic
problems with the gross alpha
measurements in the NIRS survey
preclude a more refined analysis. EPA
considers the uncertainty in this
estimate to be large, and that it likely
over predicts occurrence.
  EPA also conducted a search of the
published literature to identify reports of
alpha emitting radionuclides in water
(EPA, 1991f). While these data are not
nationally representative, and not all
were measurements made in potable
water, they do provide some indication
of the alpha emitters that may be found
in public water supplies in some
instances. The most frequently occurring
alpha emitter was polonium 210, which
was identified in ground water at levels
up to 2500 pCi/1 in one sample in
Florida, and at 3100 pCi/1 in one sample
in a uranium rich area of New Mexico.
Most measurements were below these
levels, in the 1 to 10 pCi/1 range. Various
radioisotopes of thorium were also
found in ground water, although most
were at or below 1 pCi/1. The same
uranium rich area of New Mexico
showed some higher thorium
measurements. Finally,  various
plutonium isotopes were found in
surface waters around the country,
mostly at levels below O.Ol pCi/1. These
levels are most likely present as nuclear
fallout from above-ground nuclear
explosions.
  Another source of relevant
information is the ERAMS
(Environmental Radiation Ambient
Monitoring System), the data for which
are published in the quarterly ERD
(Environmental Radiation Data; as
reported in EPA, 1991f) reports. The ERD
reports provide data on a number of
beta emitters as well as plutonium-238,
-239 and -240 for 78 sites (all surface
water sources) that are either major-
population centers or selected nuclear
facility environs. Average plutonium
levels were generally below 0.01 pCi/1,
although values as high as 0.8 pCi/1

-------
 33O7O
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991  / Proposed Rules
 were reported for piutonium-238 at two
 sites.

 IV. Proposed MCLGs for Radionuclides

' A. Setting MCLGs

   MCLGs are set at concentration levels
 at which no known or anticipated
 adverse health effects would occur,
 allowing for an adequate margin of
 safety. Establishment of a specific
 MCLG depends on the evidence of
 carcinogenicity from drinking water
 exposure or the Agency's reference dose
 (RfD), which is calculated for each
 specific contaminant.
   Establishing the MCLG for a chemical
 is generally accomplished in one of
 three ways depending upon its
 categorization (Table 1). The starting
 point in EPA's analysis is the Agency's
 cancer classification (i.e., A, B, C, D, or
 E). Each chemical is analyzed for
 evidence of carcinogenicity via
 ingestion. In most cases, the Agency
 places Group A, Bl, and B2
 contaminants into Category I, Group C
 into Category II, and Group D and E into
 Category III. However, where there is
 additional information on cancer risks
 from drinking water ingestion (taking
 into consideration weight of evidence,
 pharmacokinetics and exposure)
 additional scrutiny is conducted which
 may result in placing the contaminant
 into a different category. Asbestos and
 cadmium are examples where the
 categorization was adjusted based on
 the evidence of carcinogenicity via
 ingestion.
   EPA's policy is to set MCLGs for
 Category I chemicals at zero. The MCLG
 for Category II contaminants is
 generally based on the RfD/DWEL .
 (drinking water equivalent level, as
 described below) with an added margin
 of safety to account for cancer effects or
 is based on a cancer risk range of 10~5 to
 10~6 when non-cancer data are
 inadequate for deriving an RfD.
 Category III contaminants are based on
 the RfD/DWEL approach.

   TABLE 1.—EPA's THREE-CATEGORY
  APPROACH FOR ESTABLISHING MCLGs
                       TABLE 1.—EPA's THREE-CATEGORY AP-
                         PROACH  FOR ESTABLISHING MCLGs—
                         Continued
Category
1 	

Evidence of
carcinogenicity via
ingestion

considering
weight of
evidence,
pharmacokine-
tics, and
: exposure..
MCLG setting
approach


Category
II 	
Ill 	

Evidence of
carcinogenicity via
ingestion

considering
weight of
evidence,
pharmacokine-
tics, and
exposure.
animal evidence.
MCLG setting
approach

with added
safety
margin or
1
-------
                Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991 / Proposed Rules
                                                                                                   33071
consumption of 2 L per day. The DWEL
assumes the total daily exposure to a
substance is from drinking water
exposure. The MCLG is determined by
multiplying the DWEL by the percentage
of the total daily exposure contributed
by drinking water, called the relative
source contribution. Generally, EPA
assumes that the relative source
contribution from drinking water is 20
percent of the total exposure, unless
other exposure data for the chemical are
available. The calculation below
expresses the derivation of the MCLG:

        NOAEL or LOAEL
        uncertainty factor
   mg/kg/body weight/
         day
                        (1)
 DWEL*
   mg/L
    RfDXbody weight

dally water consumption in
         L/day
                        [2]

                        (3)
MCLG=DWELX drinking water
         contribution

  For chemicals suspected as
carcinogens, the assessment for non-
threshold toxicants consists of the
weight of evidence of carcinogenicity in
humans, using bioassays in animals and
human epidemiological studies as well
as information that provides indirect
evidence (i.e., mutagenicity and other
short-term test results]. The objectives
of the assessment are (1) to determine
the level or strength of evidence that the
substance is a human or animal
carcinogen and (2) to provide an upper
bound estimate of the possible risk of
human exposure to the substance in
drinking water. A summary of EPA's
carcinogen classification scheme (51FR
33992, September 24,1986) is:
   Group A—Human Carcinogen based
on sufficient evidence from
epidemiological studies.
   Group Bl—Probable human
carcinogen based on at least limited
evidence of carcinogenicity to humans.
   Group B2—Probable human
carcinogen based on sufficient evidence
in animals and inadequate  or no  data in
humans.
   Group C—Possible human carcinogen
based on limited evidence of
carcinogenicity in animals in the
absence of human data.
  Group D—Not classifiable based on
lack of data or inadequate evidence of
carcinogenicity from animal data.
  Group E—No evidence of
carcinogenicity for humans  (no evidence
for carcinogenicity in at least two
adequate animal tests in different
species or in both epidemiological and
animal studies).

B. Estimating Health Risks of
Radionuclides

  During the years since the publication
of the National Interim Primary Drinking
Water Regulations (41 FR 28404, July 9,
1976), which established MCLs for
radium, gross alpha, and gross beta, a
great deal of additional data and better
understanding of the risks posed to
human health by the radionuclides
discussed in this notice have been
obtained. Many of these new data are
presented and discussed in the ANPRM
(51 FR 34836, Sept. 30,1986) and the
health criteria documents supporting
this proposal.
   Several different approaches have
been used in assessing the risks posed
by exposure to radionuclides. These fall
into two broad categories: Risk
assessment based directly on the  results
of individual scientific studies of
specific compounds (either human
epidemiology studies  or experimental
studies on animals) for developing a risk
assessment for that radionuclide,  or risk
assessment based on dosimetric models
which integrate the results of a large
number of studies on  a variety of
radioactive compounds and radiation
exposure situations into an overall
model which is then used to estimate
risks for many different radionuclides.
Studies used to create such models
include both human epidemiology
studies and animal studies, and include
the results of research on subjects such
 as the metabolic fate  of different
radioisotopes, risks posed by different
kinds of radiation,  effects of dose rate,
 sensitivity of internal organs to
 radiation, identification of sensitive sub-
 populations, and other relevant subjects.
 The Criteria Documents developed in
 support of this proposed regulation
 present both studies which could
 individually be used as the basis  for
 estimating risks, and also dosimetric
 models (EPA, 1991a; 1991b; 1991c; 1991d;
 1991e). As described  below, and in the
 Criteria Documents, EPA has generally
 used the dosimetric model approach to
 estimating risks to the radionuclides
 (except for radon lung cancer risk), and
has used specific studies to make
several adjustments to the modeled
estimates.
  There are several examples of using
individual scientific studies of specific
radionuclides as the basis for risk
estimation for those radionuclides.
These include the radium watch dial
painters studies of Rowland et al. (1978)
and the risk assessment developed by
Mays et. al. (1985), and studies of radon
exposure to uranium mine workers.
They also include a series of studies of
patients injected with Thorotrast, a
thorium-based contrasting agent used in
medical radiology, which were reviewed
by the BEIRIV committee (NAS, 1988).
Another approach is combined analysis
of several studies or cohorts of miners
exposed to radon gas, as was done by
the BEIR IV committee in assessing
radon lung cancer risks (NAS, 1988).
  In addition, there are several
community ecologic studies of
exposures to radionuclides in drinking
water supplies and the disease rates in
these communities. However, these
studies do not show consistent increases
in specific tumor types across studies of
the same radionuclide as do the watch
dial painter studies and the underground
miner studies of radon. There is
considerable difficulty in controlling for
confounding factors in such studies and
they generally do not have the
specificity or statistical power to serve
as the basis for a quantitative
estimation of cancer risk, although some
of them do give indications of possible
effects and may point to future research
needs. Therefore, although reviewed in
the criteria documents, these are not
used to estimate risks for radionuclides
in drinking water.
  Because all radiation has identical
health effects, dosimetric models which
integrate a large body of information on
radiation in general as well as
individual radionuclides  can apply to a
large number of radionuclides. This is
an advantage because information on
one radionuclide can be extrapolated to
estimate risks from other radionuclides
for which there may be fewer data.
Models can also be used to estimate
radiation dose, and risk,  to tissues that
are atlower risk and therefore not
identified as target organs in
epidemiology studies. Several such
models have been developed. The
International Council for Radiation
Protection (ICRP) is one group that has
developed and made several revisions
to a model for predicting and controlling

-------
 33072
       Federal  Register / Vol. 56, No.  138 /  Thursday, July 18, 1991 / Proposed Rules
 radiation doses to workers exposed to
 radionuclides and to provide for worker
 safety. EPA uses a dosimetric model
 that is very similar to the ICRP model in
 a computer program called "RADRISK"
 which uses the ICRP type models to
 estimate risk to the qeneral population
 due to environmental exposures. EPA
 views use of dosimetric models as a
 means of integrating all information on
 the risks posed by radionuclides into a
 more complete evaluation of the risks,
 and tries to appropriately use all
 information in establishing the model
 parameters.

 C. Adverse Health Effects of the
 Radionuclides

   The radionuclides for which NPDWRs
 are proposed in today's Notice are all
 classified in Group A, known human
 carcinogens. For radium and radon this
 classification is based on direct human
 epidemiological evidence. In the case of
 uranium, the classification is based on
 the knowledge that uranium is deposited
 in the body, delivering calculable doses
 of ionizing radiation to the tissues. This
 is also true of beta, gamma, and photon
 emitters. Despite differences in radiation
 type, energy or half-life, the health
 effects of radiation are identical.

    TABLE 2.—CLASSIFICATION OF THE
   CARCINOGENICITY OF RADIONUCLIDES
 Isotope
Rn-222....
Ra-226	
Ra-228	
Cancer
group
         A	
                      Summary of basis
        Lung cancer caused by inha-
         lation of radon  and  its
         short-lived,    radioactive
         decay products. Increased
         lung cancer  mortality in
         numerous epidemiological
         studies  of   underground
         miners exposed to elevat-
         ed levels of  Rn-222 and
         its radioactive decay prod-
         ucts. Animal studies show
         similar results (NAS, 1988;
         UNSCEAR,  1988;  EPA,
         1991b; 19910).
        Bone sarcomas and head
         carcinomas in workers oc-
         cupationally  exposed  to
         radium-containing   paints
         via ingestion.  Supporting
         human  data from  studies
         of increased  cancer inci-
         dence in patients  treated
         with Ra-224 via injection
         and supporting animal evi-
         dence  from  studies  of
         mice injected with Ra-226
         and beagle dogs  Injected
         with  Ra-226  and 228
         (NAS, 1988;  UNSCEAR,
         1988; EPA, 1991b; 1991c):
       Same  as   Ra-226   except
         head carcinomas are not
         believed  to be associated
         with ingestion of  Ra-228
         (NAS,  1988;  UNSCEAR,
         1988; EPA, 1991b; 1991c).
                                TABLE 2.—CLASSIFICATION OF THE CAR-
                                  CINOGENICITY   OF   RADIONUCLIDES—
                                  Continued
                                 Isotope
                                Uranium...
                                Beta/
                                 gamma.
           Cancer
           group
Summary of basis
                  Emission of ionizing radiation
                    (alpha,   beta    and/or
                    gamma  radiation)  by U
                    and its decay products. Al-
                    though there is little direct
                    evidence of U carcinogen-
                    icity, U is found in soft tis-
                    sues and concentrates in
                    kidney and  bone.  These
                    body burdens deposit cal-
                    culable amounts of ioniz-
                    ing radiations  in  tissue.
                    These tissues are expect-
                    ed to respond as they
                    would to  any other ionizing
                    radiation  and  be  at in-
                    creased  risk from cancer.
                    These   conclusions  are
                    supported by the results of
                    animal   studies (Hodge,
                    1973;  Maynard  et  al.,
                    1953;  NAS, 1988; EPA,
                    1991e).
                  Extensive human epidemio-
                    logical data in a number of
                    irradiated     populations
                    show  increasing risks of
                    various types of cancers
                    with increasing doses of
                    ionizing radiation; most no-
                    tably, the Japanese atomic
                    bomb survivors. Also sup-
                    ported by animal study re-
                    sults (NAS, 1988).
 1. Radium-226 and Radium-228

   The Agency has placed radium-226 in
 Group A based upon clear evidence of
 carcinogenicity to humans and animals
 (EPA, 1991b; 1991p). Most information
 on human health effects of radium
 comes from epidemiologic studies of two
 groups: (1) Radium-dial painters in the
 early part of this century who ingested a
 considerable amount of radium paint
 [containing various proportions of
 radium-226 and radium-228) by
 sharpening the point of the paint brush
 with the lips and (2) patients in Europe
 injected with a short-lived isotope of
 radium, radium-224, for treatment of
 spinal arthritis and tuberculosis
 infection of the bone (NAS, 1988; EPA,
 1991b). Radium-226 and radium-228 are
 category I contaminants.
  Harmful effects of radium result from
 tissue damage caused by the
 radioactivity of radium and its
 daughters (ATSDR, 1990). The dosimetry
 of radium is controlled by its chemical
 and radiological properties. Because
radium is chemically similar to calcium,
it is sequestered in bone, so ingestion or
 inhalation over a short period results in
 long-term accumulation. The two main
isotopes of radium are: radium-226, with
 a half-life of 1,600 years, and radium-
 228, with a half-life of 5.75 years
 (ATSDR, 1990). The alpha, beta, and
 gamma radiation released by the decay
 of radium and their progeny cause
 ionization of cellular components and
 the subsequent death or mutation of
 affected  cells (EPA, 1989a).
   For about half of known radium dial
 workers, radium exposure has been
 calculated from measured body burdens
 (Rundo et al., 1986). In most cases, only
 radium-226 was detected, so that
 exposure to radium-228 is estimated
 from reports of the ratio of radium-228 to
 radium-226 in the place of employment.
 This ratio varied both over time and
 among companies (Sharpe, 1974;
 Stebbings et al., 1984). Total radium
 intake was back extrapolated using the
 Norris retention function (Norris et  al.,
 1955) and based on the gastrointestinal
 absorption factor of 20 percent found by
 Maletskos et al. (1966,1969), ingestion
 was assumed to be five times the intake
 to the blood (Mays et al., 1985).
   At higher levels of exposure to
 radium, several non-cancer health
 effects occur: benign bone growths,
 osteoporosis, severe growth retardation,
 tooth breakage, kidney disease, liver
 disease, tissue necrosis, cataracts,
 anemia, immunological suppression and
 death (ATSDR, 1990). The most sensitive
 indicator of non-cancer effects is bone
 necrosis scored by X-ray (Keane et  al.,
 1983). Thirty or more years after
 exposure, the incidence of bone necrosis
 in female radium dial painters with  total
 ingestion of radium-226 or radium-228
 above 50 /tCi was significantly higher
 than in unexposed controls (Keane et
 al., 1983). However,  levels of exposure
 from naturally-occurring radium are
 much lower than this threshold, and so
 bone necrosis and other non-cancer
 health effects are usually not of concern
 for radium in drinking water (EPA,
 1991b; EPA, 1990g; EPA, 1990n).
  Scientists have long recognized that
 exposed radium dial painters have
 elevated rates of two rare types of
 cancer, bone sarcomas (osteosarcomas,
 fibrosarcomas and chondrosarcomas)
 and carcinomas of head sinuses and
 mastoids  (Evans et al., 1944; Sharpe,
 1974). A recent quantitative analysis of
 the epidemiologic data (Rowland et  al.,
 1978) found a highly significant excess
 of bone sarcomas and head carcinomas
in a cohort of measured women first
 employed before 1930. The relative
effectiveness of radium-226 and radium-
228 in inducing bone sarcomas was
estimated to be 1:2.5. The incidence of
head carcinomas was associated 'vith
exposure to radium-226, but not radium-
228 (Rowland et al., 1978). This is

-------
                Federal Register  /  Vol. 56,  No. 138  /  Thursday,  July  18, 1991 / Proposed Rules
                                                                     33073
expected if these cancers are due to
accumulation of radon gas in the
maatoid air cells and paranasal sinuses,
because the radon daughter of radium-
228, radon 220, decays to Ra-224 too
quickly for substantial diffusion to air
cells (NAS, 1988). In this cohort, a dose-
squared relationship was the best fit of
the data for radium-226 and radium-228
induction of bone sarcomas, while a
linear relationship was the best fit for
radium-226 induction of head
carcinomas (Rowland et al, 1978).
However, the shape of the dose-
response curves are uncertain because
radium intake is not known for about
one third of the cases of bone sarcomas
and head carcinomas.
  Patients medically treated  with
radium-224, a daughter of radium-228,
also show an increase in bone
sarcomas, but not head carcinomas
(Mays and Speiss, 1984). These data are
consistent with a linear dose-response
relationship (NAS, 1988). The risk
coefficient for bone  cancer which is
used in the RADRISK model  is derived
from data on exposure to radium-224
(NAS, 1980; EPA, 1991b) because actual
exposures to radium-226 and radium-228
to the watch dial painters is not well
known, and because of the uncertainty
that would be introduced in deriving a
linear risk coefficient from significantly
non-linear data.
  No statistically significant increase in
cancers other than bone sarcomas and
head carcinomas have been found in
cohorts of radium dial painters
(Stebbings et al., 1984). Increases in
breast cancer and multiple myeloma are
better correlated with duration of
employment, a surrogate for external
dose of gamma irradiation, than with
radium intake (Stebbings et al., 1984).
The lack of an increase in leukemias is
unexpected, because the accumulation
of radium in bone would be expected to
provide substantial irradiation of
potentially leukemogenic cells (Mays et
al., 1985), and external irradiation has
clearly been established as a cause of
leukemia in humans (NAS, 1980).
Possible explanations for the lack of
observable increase in leukemias
include alterations in bone architecture,
non-uniformity of irradiation, lethality of
irradiation to marrow cells, low
frequency of leukemogenic cells in
irradiated regions, misdiagnosis of bone
marrow diseases, incomplete
 ascertainment of the cohort, and
 overestimation either of the  risk
 coefficient for beta and gamma
 irradiation or of the relative
 effectiveness of alpha irradiation (EPA,
 1991b).
  Possible correlations between cancer
rates and radium in drinking water have
been examined in three studies in the
United States. Petersen et al. (1966)
found an elevated rate of fatalities from
bone malignancies among residents of
Iowa and Illinois with elevated radium-
226 in drinking water, but the statistical
significance was marginal and
confounding factors could not be ruled
out (NAS, 1988). Bean et al. (1982) found
an increased incidence  of 4 out of the 10
cancers investigated among Iowa
residents of small communities with
elevated radium-226 content of the
water supply. However, confounding by
radon exposure could not be ruled out
and cancer sites were different from
those observed in dial painters: bladder
and lung cancer for males and breast
and lung cancer for females. Lyman et
al. (1985) found a small but consistent
excess of leukemias in Florida counties
with elevated radium-226 or radium-228
in private wells, but there was no
evidence of a dose-response trend.
Rates of colon, lung and breast cancer
and lymphoma showed no consistent
excess (Lyman and Lyman, 1986).
  Animal studies have shown that
exposure to radium causes bone
sarcoma in mice, rats and dogs and
leukemia in mice (ATSDR, 1990). Evans
et al. (1944) produced bone sarcomas in
rats by both oral exposure for 20 days
and intradermal exposure for 2 days to
radium-226. Experiments at Argonne
National Laboratory using large
numbers of CFl female mice injected
once with radium-226 demonstrated a
clear increase in bone sarcomas (Finkel
et al., 1969). Studies at  the University of
California at Davis using beagle dogs
injected with radium-226 eight times at
two-week intervals demonstrated a
clear dose-response trend in premature
deaths and incidence of bone sarcomas
(Raabe et al., 1981). In addition to bone
sarcomas, other malignancies
associated with radium exposure in
animals are eye melanomas in beagle
dogs injected with radium-226 or
radium-228 (Taylor et al., 1972) and
leukemias in mice injected with radium-
224 (Humphreys et al.,  1985; Muller et
al., 1988).
   Quantitative estimates of the risks of
low level exposure to radium in drinking
 water were generated by the RADRISK
 model and adjusted for over-prediction
 of leukemias lack of separate prediction
 of head carcinomas by radium-226, and
 for under-prediction of bone dose and
 sarcoma risk by radium-228. The
 resulting risks corresponding to lifetime
 intake of water containing 1 pCi/1 are
 4.4 x 10~6 for radium 226 and 3.8 x 10"6
 for radium 228 (EPA, 1991b). An
alternative approach to evaluating the
risks of radium in drinking water was
presented by Mays et al., (1985). These
investigators derived linear risk
coefficients from the dial painter
epidemiologic data, which, as noted
above showed a significantly non-linear
response for bone sarcoma incidence.
Mays et al. (1985) calculated the risks
corresponding to lifetime intake of water
containing 1 pCi/1 radium to be 8.4 x
10~6 for radium 226 and 8.8 x 10"6 for
radium 228. The adjusted risk
coefficients used by the Agency in
evaluating the risks of radium in
drinking water are about half those
calculated by Mays et al. (1985), but are
considered to be better estimates
because of the quantitative uncertainties
in the dial painter data concerning
ingested dose, cancer incidence, and
particularly  low dose extrapolation.
  There may be several sources of
uncertainty in the risk estimates. These
are discussed in detail in the Criteria
Document (EPA, 1991b), and are briefly
summarized here. They include the use
of non-linear data for bone sarcomas as
one part of a linear low dose
extrapolation, lack of statistically
significant increases in cancers other
than head carcinomas and bone
sarcomas in the watch dial painters,
even though predicted by the model.
While there may be uncertainties in the
modeled risk estimates, EPA has
evaluated all the available data and
believes the approach selected is likely
to have fewer uncertainties than  other
approaches to assessing radium risks at
environmental intake levels.
   EPA solicits public comment on its
estimation of risks from radium in
drinking water. In particular, EPA
solicits public comment on use of the
RADRISK model to assess risks, use of a
linear risk model to extrapolate to low
doses, and the adjustment of estimated
leukemia risks  and addition of the head
carcinoma risks to the risk estimate, and
 adjustment of the radium-228 bone
 sarcoma risks.
   In summary,  the Agency's assessment
 of risk of drinking-water exposure to
radium is based on the following:

Radium-226
   • Excess incidence of bone sarcomas
 and head carcinomas among humans
 occupationally exposed to radium-226.
   • Excess incidence of bone sarcomas
 among laboratory animals injected with
 radium-226.
   • A calculated mortality risk from
 lifetime ingestion of radium-226 in
 drinking water of 4.4 x 10"6/pCi/l,
 assuming 2 liters consumption per day.
 A lifetime mortality risk of 10"4 would

-------
33074
Federal  Register / Vol. 56, No.  138 / Thursday, July 18, 1991 /  Proposed Rules
exist at approximately 22 pCi/1 radium
226 in water.

Radium-228
  • Excess incidence of bone sarcomas
among humans occupationally exposed
to radium-228.
  • Excess incidence of bone sarcomas
among laboratory animals injected with
radium-228.
  • A calculated mortality risk from
lifetime ingestion of radium-228 in
drinking water of 3.8 x 10~6/pCi/l,
assuming 2 liters consumption per day.
A lifetime mortality risk of 10" 4 would
exist at approximately 26 pCi/1 radium
228 in water.

2. Radon
  EPA's primary concern in regulating
radon in drinking water is risk from
radon released from water to the air in
residences. Inhalation is the primary
exposure route of concern,  lung is the
target organ, and lung cancer is the
endpoint of primary concern. EPA also
believes that some cancer risk to
internal organs is posed by ingesting
water containing radon, and breathing
radon gas, and has developed
dosimetric models for estimating risks to
internal organs from these exposures
(EPA, 1991c).
  The Agency has classified radon-222
as a Group A carcinogen based on
sufficient evidence for a causal
association between exposure to radon
and lung cancer in humans (EPA, 1991c;
NAS, 1988). In addition, data from
studies with experimental animals also
provide sufficient evidence for the
carcinogenicity of radon. The fact that
ionizing radiation is classified as a
group A carcinogen provides the basis
for considering radon to pose cancer
risk when ingested and for  radon gas
that is inhaled, absorbed and distributed
(EPA, 1991p].
  a. Radon risks from inhalation.
Human epidemiologic data have been
obtained from groups of underground
metal-ore miners mainly in the United
States (Colorado Plateau), Canada
(Ontario, and Eldorado)
Czechoslovakia, Sweden (Malmberget),
Newfoundland and Great Britain. These
studies have been reviewed by NCRP
(1984a,b), NIOSH (1987), ICRP (1987),
NAS (1988), DOE (1988), and EPA
(1989a).
  The Colorado Plateau study
represents a large, clearly defined, well-
traced population having individual
smoking histories and exposure records
and a follow-up period exceeding 20
years (as reported in EPA, 19901). As of
1982, the lung cancer deaths had
increased to 255 compared with about 50
expected (Standard Mortality Ratio,
                       SMR=510) in a cohort of 3,366 white
                       and 780 nonwhite male miners. The
                       major weaknesses of this study are the
                       great number of mines (2,500) involved
                       (some with few radon exposure
                       measurements), self-reported work
                       histories, and high exposure levels.
                         The cohort in the Ontario study
                       consisted of 15,094 persons who worked
                       for 1 or more months in uranium mines
                       during the 1954-74 period (as reported in
                       EPA, 19901). Of those with a cumulative
                       Working Level Month (WLM) exposure
                       of 340 WLM or greater by 1986,14 cases
                       of lung cancer were observed compared
                       with 3-4 expected (SMR=412). (One
                       WLM of exposure is approximately
                       equal to being exposed to radon and its
                       progeny at 200 pCi/1 in air for 170 hours,
                       or 8 hours daily for 20 days.) This study
                       involved low mean cumulative
                       exposures with reasonably good
                       working histories but limited smoking
                       histories.
                         The Czechoslovakian cohort consisted
                       of 2,433 miners who began mining
                       uranium ore in 1948-52 and had worked
                       at least 4 years underground (as
                       reported in EPA, 19901). For exposures
                       of 12 years or longer, the dose-related
                       increase hi lung cancer had been
                       established. For exposures of less than
                       12 years, a nonlinear relationship
                       existed, so that increasing  dose (WLM)
                       did not result in increased risk if
                       exposure was less than 5.6 to 9.5 years.
                       In the 23.5 year group exposed to the
                       highest level of radon (716 WLM), 82
                       lung cancers were observed compared
                       with 10 expected (SMR=820). Recently,
                       a significant excess of lung cancer was
                       observed in exposure categories below
                       50 WLM (Sevc et al., 1988). The mean
                       attributable annual cancer risk after
                       about 30 years of observation in the
                       whole study was approximately 20
                       cases per year per WLM/106 persons,
                       and in persons starting exposure after 30
                       years of age the risk was approximately
                       30 cases per year per WLM/106 persons.
                         The Malmberget retrospective
                       mortality study involved a  cohort of
                       1,415 miners who had worked
                       underground for more than one calendar
                       year from 1897 to 1976 (as reported in
                       EPA, 19901). Mean exposure of these
                       miners to radon was estimated to be
                       93.7 WLMs. The major source of
                       airborne radon and radon progeny was
                       radon dissolved in groundwater. Excess
                       lung (50 observed vs 12.8 expected,
                       SMR=390) and stomach (28 observed vs
                       15.1 expected, SMR=185) cancers were
                       reported. The excess risk for lung cancer
                       first become evident 20 years after the
                       beginning of underground mining. The
                       low exposure levels, long follow-up
                       period, and stability of the  work force
                       are the strengths of this study.
  The Eldorado Beaverlodge
retrospective cohort study involved
8,487 male miners exposed during 1948
to 1980 (as reported in EPA, 19901). A
dose-related increase in lung cancer was
seen, although no increased risk was
evident at 5 WLM or less. For lung
cancer deaths occurring during the 1950-
80 period, 54 were observed in the
mining group versus 28.27 expected
(SMR=191). For those exposed to 150
WLM or greater, 10 cases were observed
versus 1.04 expected (SMR=961).
  In general, the response in animals to
inhaled radon daughters is qualitatively
similar to that in humans. However,
species response has varied with
respect to tumor type and latency
period. The  animal studies have
demonstrated that radon and radon
progeny can induce lung cancer in rats
and dogs (EPA, 1991c).
  Several risk assessments have been
conducted to quantify the risk to miners
exposed to radon and radon progeny.
Recent concern with exposure of the
general public to radon in the home
environment has prompted the NAS
(1988) and the ICRP (1987) to conduct
risk assessments.
  The NAS  (1988) assessment is
commonly referred to as the BEIRIV
report. The Colorado Plateau, Ontario,
Malmberget and the Eldorado
Beaverlodge miner cohort data  set were
analyzed by NAS. It was concluded that
the appropriate model would involve the
computation of relative risk with •
consideration of the change in risk with
time since exposure (TSE Model). The
age-specific lung cancer mortality was
calculated for cumulative radiation
exposure, in WLM, incurred between 5
and 15 (Wi) or > 15 years (W2) before
age using the equation:
r(age, period, dose history) =r0 (age)
  [1+0.025y (age) (Wi +0.5 W2)]
where y has a value of 1.2 for persons
    younger than 55 years, a value  of 1 for
    persons 55 to 65 years old, and 0.4 for
    persons older than 65. Based on this
    equation, the excess lifetime lung cancer
    mortality for males was 5.06X10"4
    cases/WLM of lifetime exposure, and the
    risk for females was 1.86X10"4 cases/
    WLM of lifetime exposure. Assuming
    equal numbers of males and females in
    the U.S. population, 253 and 93 lung
    cancer cases in 500,000 exposed males
    and 500,000 exposed females would
    result each year (i.e. 350 lung cancer
    deaths/106 person-WLM of lifetime
    exposure).
  The ICRP (1987) employed a
somewhat different approach. Only
three epidemiological sets were
considered (Colorado Plateau,
Czechoslovakia and Ontario). These
were analyzed by both absolute and

-------
                Federal Register / Vol. 56,  No. 138  /  Thursday, July 18, 1991 / Proposed Rules         33075
relative-risk projection models.
However, the proportional hazard model
(constant relative risk] was selected for
analysis of radon risk in the indoor
environmental. It was assumed that the
lung cancer rate is proportional to radon
exposure and is proportional to the
normal lung cancer rate without radon
exposure.
  The equation for the constant relative-
risk, proportional hazard model is:
X(t) = Xo(t) [1 + /„'-' r(tj R[tJ dtj = the
    mortality rate at age, t
where:
  X<,(t)=age-specific lung cancer rate at age, t
  r(te)=risk coefficient at age of
    exposure,!,
  R(t«)=age-dependent exposure rate
  retime lag (minimal latency)
A correction of 0.8 was used to account
for the other carcinogens present in
mines but not present in indoor
buildings. Another adjustment of 0.8
was made to account for differences in
dose to the bronchial epithelium for
indoor as compared with miner
exposure. This resulted in a risk
reduction factor of 0.64. The ICRP also
considered the potential for increased
sensitivity of young people and assigned
an increased risk factor of 3 for
exposure to persons age 20 or less. Thus,
the final relative-risk coefficients were
0.64%/WLM for those >2Q years of age
and 1.9S/WLM (3X0.64) for those <20
years of age.
  Employing a 10-year lagtime and the
1980 U.S. lifetable and vital  statistics at
an exposure level of 0.001 WLM/year,
ICRP calculated 610 lung cancer deaths/
10 8 WLM for males and 204 for females
(i.e. a combined risk of 420 lung cancer
deaths/10 6 WLM).
  The current EPA estimates for lung
cancer risk from radon exposure are
based on an averaging of the results of
the BEIRIV and ICRP 50 analyses with
slight modifications (EPA, 1989a; EPA,
1991c). The EPA has accepted the BEIR
IV conclusions that the dose and risk
per WLM exposure in residences and in
mines are basically identical, and thus
no compensation is made for age- and
sex-specific tracheobronchial
deposition. The ICRP 50 (1987) results
have been slightly modified by deleting
the risk reduction factor of 0.8 used by
ICRP to compensate for differences in
dose to bronchial epithelium between
household residents and miners.
Therefore calculations in the ICRP 50
model were made using risk coefficients
of 0.8%/WLM for those >20 years and
2.496/WLM for those <20 years of age
(EPA, 1989a).
  The EPA's risk estimate was adjusted
for an assumed background exposure of
0.25 WLM/year; the average radon
exposure rate was based on 1980 U.S.
vital statistics and Nero's radon in
residence distribution estimate (Nero et
al., 1986).
  EPA estimated the excess lifetime risk
in the general population due to
constant low-level lifetime exposure,
based on an average of the BEIR IV and
ICRP 50 estimates and the modifications
discussed above, at 550 and 190/10 6
WLM for males and females,
respectively, or a combined  risk of 360
lung cancer deaths/10 6 WLM, with an
estimated range of 140 to 720 lung
cancer deaths per 10 6 WLM (EPA,
1989a).
  The occupancy factor of 0.75*is based
on studies by Moeller and Underhill
(1976) and Oakley (1972), which
estimated radiation exposure and
population dose in the United States and
is supported by more recent reports. An
equilibrium factor of radon with its
progeny of 0.50 was estimated (EPA,
1991i), and EPA estimates that 10,000
pCi/1 radon in water will contribute
about 1 pCi/1 to the air of a  house, on
average (EPA, 1991h).
  The risk estimates for excess lung
cancer deaths due to  inhalation of radon
can be used in estimating the risk of
radon in water (EPA, 1991c). Using the
above assumptions, the risk estimate of
360 deaths/10 6 WLM is converted to
units of deaths/pCi/1 water as follows:
Risk(pCi/lwater)=
  (360 deaths /10 6 WLM) X (51.6 WLM/WL-
    yr) X (70 yr) X (0.5 WL/100 pCi/1^) X (10~4
    (pCiAUj/CpCi/WI) X (0.75)
  =4.9X10"'
Lifetime individual risk for lung cancer
of 5X10~7 deaths per pCi/1 water was
estimated for inhaled radon daughters
(EPA, 1991c).
  However, EPA is in the process of
reviewing and revising its estimate of
radon risk. This review is based on the
conclusions of the recent report by the
National Academy of Science entitled
"Comparative Dosimetry of Radon in
Mines and Homes" (NAS, 1991), on
results of the National Residential
Radon Survey and also on comments
received by EPA on the background
document supporting revisions  to the
Citizen's Guide to Radon. The study by
NAS was funded by EPA to help reduce
the uncertainties of using miner data to
estimate radon risks in the home. EPA
has submitted a revised risk assessment
to the SAB/RAG for then- review, and
will revise the risks estimated here, if
appropriate, when the SAB/RAC
completes its review and provides EPA
comments. This revised risk evaluation
was discussed by the SAB/RAC at a
meeting held May 20 and 21, 1991. EPA
anticipates that the lung cancer risk
estimate for radon by inhalation (based
on the epidemiology studies) may be
reduced by as much as 30% in the final
revised estimate (EPA, 19911).
  As a volatile gas, radon may also be
absorbed via inhalation and distributed
throughout the body, posing some risk to
internal organs. The human
epidemiology studies do not account for
this risk. EPA estimated the risk to
internal organs from inhaled radon gas,
using the RADRISK model, the 0.75
occupancy factor, an estimated
breathing rate of 22,000 liters daily
(EPA, 1989a) and the 10,000:1 water to
air transfer factor (EPA, 1991h), as
2X10~8 deaths per pCi/lwater. Details of
this calculation are provided in the
Health Criteria Document for radon
(EPA, 1991c).
  EPA has also reviewed information on
the interaction of smoking and lung
cancer risk from radon. The BEIR IV
committee (NAS 1988) concluded that
the data show a multiplicative
interaction between smoking and radon
exposure in causing lung cancer, not an
additive interaction. In reviewing the
relative risks from radon to smokers
EPA (EPA 1990i; EPA 1991c) estimated
risk multipliers applicable to the
population average risks for different
categories of smokers. The categories
include non-smokers, former smokers,
and current smokers of different
numbers of cigarettes. For non-smokers,
estimated risks from radon are about
20% of the overall average population
risk; for former smokers, radon risks are
about 80% of the average risk. For
current smokers, estimated risks range
up to about 450% of the average
population risk (40+ cigarettes per day),
with a smoker average of 180% of
overall average population risk. Heavy
smokers are therefore at considerably
greater risk from radon exposure than is
the general population.
  b. Radon risk via ingestion. EPA's
assessment of the risk associated with
radon when ingested is less certain than
the estimate of risk by the inhalation
exposure route. No experimental or
epidemiologic data link exposure via
ingestion to increased cancer rates.
  In the present assessment, EPA has
estimated the risk from ingestion of
radon-222 in drinking water using data
on organ doses recently developed for
the Agency by Crawford-Brown (1990).
In developing these dose estimates,
Crawford-Brown used the results of
biokinetics studies carried out by
Correia et al., (1987; 1988) using xenon-
133, a gas  that behaves similarly to
radon-222. Hess and Brown (1991) have
also studied retention and clearance
rates of radon gas when ingested in
water.

-------
33076
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991  / Proposed Rules
   Crawford-Brown developed
 mathematical models of the movement
 and accumulation of radon-222 within
 the various organs of the body following
 ingestion. Rate constants for movement
 of radon-222 within the various body
 organs were also developed. Using these
 models, the concentration of radon-222
 in body organs was calculated under
 steady-state conditions.
   EPA used these dose factors, an
 estimated 1 liter daily intake of freshly
 drawn directly consumed tap water and
 a 20% correction for radon loss from
 water during the process of drawing and
 consuming a glass of .water (discussed in
 Section III.C above, and EPA, 1991h), in
 estimating the risk from ingested radon.
 EPA calculated the lifetime risk from
 ingestion of radon-222 in drinking water
 to be 1.5 X1CT7 per pCi/1 (EPA, 1990c;
 1991c). This is about 20% of the risk
 estimated from inhalation of radon-222
 progeny from domestic use of water.
   The total estimated risk for radon in
 water is 6.6 X10" 7 per pCi/1. This gives
 an estimated IX 10~4 individual lifetime
 risk at approximately 150 pCi/1 in water
 for all water related exposure to radon
 (EPA, 1991c).
   EPA estimates that approximately five
 percent of total indoor air radon is
 attributable to radon from drinking
 water on average, for homes  served by
 groundwater. The MRS occurrence
 survey showed average radon levels in
 public water ground water supplies to
 be 650 pCi/1, with a maximum reported
 level of 26,000 (although many private
 wells are known to have higher levels;
 EPA, 1990f). EPA estimates that
 approximately 200 (75 to 400) cancer
'fatalities per year are attributable to
 radon in drinking water, 80%, or 160 of
 which are estimated to  be due to lung
 cancer (EPA, 19911). Of these,
 approximately 85% may involve
 synergism with smoking. Overall, radon
 in homes is estimated to account for
 approximately 8,000 to 40,000 lung
 cancer deaths annually (EPA, 1989g;
 1990m). Individual risks at the 4 pCi/1
 indoor air action level are
 approximately 1-5 in 100 (EPA, 1986f).
   There may be several sources of
 uncertainty in the radon risk estimates.
 These are discussed in detail in the
 Criteria Document (EPA, 1991c), and are
 briefly summarized here. They include
 variability in the contribution of radon
 in water to indoor air radon levels,
 differences in homes  and the mine
 environment, and estimates in
 distribution and effective dose to tissue
 of ingested radon. While there may be
 uncertainties in the risk estimates, EPA
 has evaluated all the  available data and
 believes the approach selected is likely
                        to have fewer uncertainties than other
                        approaches to assessing radon risks.
                         EPA solicits public comment on its
                        assessment of risks from radon in
                        drinking water. In particular, EPA
                        requests comment on its estimate of
                        water contributions to indoor air levels
                        of radon and exposure during
                        showering, and its estimate of risks due
                        to directly ingesting radon hi water.
                        3. Uranium
                         Exposure to uranium (U) is of concern
                        because of the radioactive nature of
                        uranium and its ubiquitous occurrence
                        in the environment, including Water
                        supplies. Kidney toxicity and
                        carcinogenicity are the primary adverse
                        effects of concern associated with
                        exposure to uranium (EPA, 1991e). EPA
                        proposes to regulate uranium at the level
                        that will be protective of both its kidney
                        toxicity, and its carcinogenic potential
                        as well. Studies in both humans and
                        animals show uranium toxicity to the
                        kidneys. The EPA has also classified
                        uranium in Group A as a human
                        carcinogen (sufficient evidence of
                        carcinogenicity in humans) based on the
                        fact that uranium emits alpha radiation,
                        a well-established carcinogen (which is
                        also classified in Group A; EPA, 1991p),
                        and uranium is an analogue of radium-
                        226, a well-known human carcinogen  in
                        bone  (EPA, 1991e).
                         a. Carcinogenicity. The carcinogenic
                        effects of uranium have been
                        characterized based on effects of
                        ionizing radiation generally, the
                        similarity of uranium to isotopes of
                        radium and on the effects of high
                        activity uranium. Ionizing radiation has
                        been  classified by EPA as a Group A
                        carcinogen, and EPA considers all
                        emitters of ionizing radiation to be
                        carcinogenic  (EPA, 1991p). Studies have
                        also shown that uranium, like radium,
                        accumulates primarily in bone, and that
                        bone  sarcomas may result from radium
                        ingestion (EPA, 1991b; 1991e). The
                        induction of bone sarcomas is regarded
                        as a common property of both radium
                        and uranium, which is believed to result
                        from the alpha emissions of these nuclei
                        as they decay. Finally, studies of
                        enriched and high activity isotopes of
                        uranium have shown them to be
                        carcinogenic in animal studies.
                         Studies using natural uranium do not
                        provide direct evidence of carcinogenic
                        potential (EPA, 1991e). Malignant
                        tumors were observed in mice following
                        injection of uramum-232 or uranium-233
                        (at levels greater than 0.1 jnCi/kg), but
                        not following injection of natural
                        uranium (Finkel, 1953), probably
                        because radiation dose levels were
                        about 100-fold lower than the dose at
                        which the tumors were observed for
uranium-232 and -233 by injection.
Highly enriched uranium (i.e., uranium
enriched with the more radioactive
isotopes) has been shown to induce
bone sarcomas in rats (NAS, 1988).
  Existing human epidemiology data are
inadequate to assess the carcinogenicity
of uranium ingested in drinking water
(EPA, 1991e). However, some
epidemiological data do suggest that
inhalation exposure to uranium or direct
exposure to uranium deposits may be
carcinogenic in humans. Polednak and
Wilson (as cited in Dupree et al., 1987)
found nonstatistically significant
increases in cancers of the digestive
organs in workers exposed to airborne
uranium, although confounding
variables were present (EPA, 1991e).
Wilkinson (1985) reported higher
mortality rates from gastric cancer hi
New Mexico counties located over
uranium deposits. However, other
etiological factors (such as radon
progeny and trace elements) may be
involved (EPA, 1991e).
  EPA estimated the carcinogenic risk
associated with uranium exposure using
the RADRISK dosimetric model, as
described in the revised Drinking Water
Criteria Document for uranium (EPA,
1991e). EPA's earlier draft of this
document (EPA, 1989f) and earlier risk
assessment used a gastrointestinal
uptake (fi) factor of 0.20, which is
revised in the updated Criteria
Document (EPA, 1990e; 1991e) to 0.05 in
response to comments by the SAB/RAG.
While EPA believes the 0.05 value
represents a best estimate, the wide
range of values reported in the literature
for the uranium ft (from less than 0.01 to
0.30) indicate that there may be
substantial uncertainty associated with
the 0.05 value. The individual studies
bearing on this issue are described in
the updated Criteria Document (EPA,
1991e). EPA solicits public comment on
the issue of the uranium 6. value.
  Using a gastrointestinal uptake (fi)
factor of 0.05, risks of fatal cancer
estimated using the RADRISK model
indicated that uranium in water poses
cancer risk of approximately 5.9X10"7
per pCi/1, assuming 2 liters daily intake.
Concentrations in water of 1.7 pCi/1,17
pCi/1 and 170 pCi/1 correspond to
lifetime mortality risks of approximately
1X10~6, !X10"5andlXlO~4,
respectively.
  b. Non-cancer effects. The major
target organ of uranium's chemical
toxicity is the kidney (Hodge, 1973;
Leggett, 1989; EPA, 1991s). Based on
available toxicity data, rabbits have
been identified as the most sensitive
species (data summarized in Table 3). In
humans, symptoms of transient

-------
                Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                                      33077
albuminuria and edema of the skeletal
muscle developed in several laboratory
workers exposed to combined vapors of
uranium hexafluroride, uranium
oxyfluoride, and hydrofluoric acid
(Howland, 1949). However, vapor  ,
concentration was not measured. Some
of the effects may have been
attributable to the direct action of
fluoride, since the workers were
exposed to a mixture of chemicals; the
transient renal effects, however, may be
related to the toxic action of absorbed
uranium (Haven and Hodge, 1949).
             TABLE 3.—A COMPARISON OF 30-DAY, I-YEAR, AND 2-YEAR NOAELs/LOAELs FOR URANIUM TOXICITY
Species/compound
Rat:
UOjFa 	
UOi(NOi)] 	 	 	 	 	
UF4 	
Dog:
UOjFi 	 	 	
Rabbit:
UOjtNOjJj

NOAEL (mg U/kg/day)
30-day
19
24
760
4
ND
1-yr
19
24
760
8
NT
2-yr
19
24
760
NT
NT
LOAEL (mg U/kg/day)
30-day
39
120
7,600
8
2.8
1-yr
39
120
7,600
19
NT
2-yr
39
120
7,600
NT
NT
   NT » Not tasted.
   ND - Not determined.
   NOAEL -= no observed adverse effect level
   LOAEL » lowest observed adverse effect level
   Source: Maynard and Hodge (1949, as cited in U.S. EPA, 1991 a).
  Nephrotoxicity has been reported in
rats, rabbits, and/or dogs fed various
soluble uranium compounds for periods
of 30 days, 1 year, or 2 years (Maynard
and Hodge, 1949; Maynard et al, 1953).
Treatment-related histopathological
changes were observed in the kidneys of
rats fed UO^, UO2(NO3)2. BHzO, and
UGli. No histopathologic changes were
found in the kidneys of rats fed
insoluble uranium compounds. Acute (30
day) exposure of rabbits to uranyl
nitrate down to 2.8 mg/kg/day in the
diet resulted in renal damage at all dose
levels (Maynard and Hodge, 1949; EPA,
1991s).
  Renal toxicity has also been
demonstrated in rats and dogs following
administration of various uranium
compounds in the diet for 1 or 2 years. A
summary of NOAEL and LOAEL values
derived from these studies is presented
in Table 4.
                                  TABLE 4.—SUMMARY OF NOAEL AND LOAEL VALUES
Compound
Uranyl nitrate.... 	
Uranyl ftuorid@..»» 	

Uranyl tetrafluorfde, 	
Uranium dioxide effects 	

NOAEL
percent
0.1
0.05
0 1
2


LOAEL
percent
0.5
0.1
05
20
20

Effect
Body weight depression, mild tubular necrosis of kidneys.
Body weight depression.
Body weight depression, kidney changes.
Body weight depression, kidney changes.
No toxic

   Source: Maynard and Hodge (1949, as cited in U.S. EPA, 1991e);  Maynard et al. (1953, as cited in U.S. EPA, 1991e).
  The mechanism of action of uranium
in renal toxicity is not fully understood
(Leggett, 1989). Nephritis and changes in
urine composition are the primary
symptoms (EPA, 1991e).
Morphologically, the most evident
changes occur in the proximal
convoluted tubule of the nephrons.
Necrosis of the tubular lining occurs
first, followed by a clogging of the  ,
tubules with cellular debris and
appearance of the debris (casts) in the
urine. Regeneration of tubular lining
cells within 2 to 3 weeks can occur in
nonfatal cases, but the  cells are not
normal in appearance. The mechanism
of action may involve interference with
sodium transport across membranes,
damage to lysosomes, or destruction of
functional properties in mitochondria
(EPA, 1991e).
  In addition to renal effects, animal
studies also indicate that exposure to
uranium may be associated with dermal,
ocular, teratogenic/reproductive, and
hepatic effects as well as lethality, at
higher exposures (EPA, 1991e).
Histopathological changes (distortion of
centrilobular and perilobular zones)
were observed in the livers of rats fed 20
mg (9.5 mg U/kg) uranyl nitrate.
  Oral administration of uranium to rats
and mice has resulted in embryo
lethality, adverse fetal and neonatal
development, increased fatal resorption,
reduced fetal body weight and length,
adverse functioning of the reproductive
system, and increased number of dead
young/litter at birth and at lactation
(Paternian et al., 1989; Domingo et al.,
1989a; 1989b; Maynard et al., 1953).
Brandom et al. (1978) found a significant
increase in the prevalence of
chromosomal aberrations in uranium
miners as compared with controls.
  EPA identified the LOAEL as 0.02 ppm
uranyl nitrate hexahydrate in food,
converted to 2.8 mg uranuim/kg/day,
based on the kidney toxicity in rabbits
(Maynard and Hodge, 1949; See Table
3). EPA applied a 1000 fold uncertainty
factor to derive an RfD of 3X10~3mg/
kg/day (EPA, 1991s; 1991e). EPA
multiplied the RfD by 70 kg and divided
by 2 liters daily water intake, to derive
the DWEL of 100 jug/1. If EPA were
basing the MCLG on kidney toxicity, the
20% relative source contribution would
be applied as discussed above. This
would result in a MCLG based on
kidney toxicity 20 ju.g/1.
  EPA is proposing to set the MCLG at
zero because of uranium's
carcinogenicity. However, EPA is
proposing to limit the MCL because  of

-------
                 Federal  Register / Vol. 56, No. 138 /  Thursday, July 18,  1991 / Proposed Rules
 33078 		

 kidney tbxicity, because of the low
 carcinogenic potency of uranium. EPA
 believes drinking water MCLs must be
 protective of the public against all
 adverse health effects.
   There may be several sources  of
 uncertainty in the uranium risk
 evaluation. These are discussed  in detail
 in the Criteria Document (EPA, 1991e),
 and are briefly summarized here. They
 include in particular for the uranium
 cancer risk estimate the fi factor, as well
 as the lack of confirmation of uranium's
 carcinogenicity in the available
 epidemiology studies. For kidney
 toxicity, uncertainties in uranium
 exposures from other sources may lead
 to uncertainty. While there may be
 uncertainties in the assessment of
 uranium's adverse affects, EPA has
 evaluated all the available data and
 believes the approach selected is likely
 to have fewer uncertainties than other
 approaches to assessing uranium risks.
   EPA solicits public comment on the
 proposed MCLG of zero for uranium,
 including the fi factor, uranium's
 carcinogenicity, and kidney toxicity as
 the limiting adverse health effect.

 4. Beta Particle and Photon Emitters
   "Beta and photon emitters" are a
 broad group of mostly man-made
 radionuclides which characteristically
 decay by beta and photon emissions,
 which are ionizing radiation. EPA has
 classified ionizing radiation as a  group
 A carcinogen (EPA, 1991p). Accordingly,
 the Agency considers beta and photon
 emitters Group A human carcinogens.
   Beta and photon emitters are
 radionuclides that decay primarily by
 electron and/or photon emissions and
 are usually man-made. These low
 energy radiation emitters (low-LET)
 include beta emitters (electrons or
 positrons), gamma emitters, and x-ray
 emitters.. There are a large number of
 radionuclides of concern, and each
 radioriuclide/element has different
 absorption and retention properties and
 decay schemes. Differences in energy of
 irradiation, type, and geometry of
 irradiation also exist.
  Despite differences in radiation type,
 energy, or half-life, the health effects
 from radiation are identical (EPA,
 1991d), but may occur in different target
 organs arid at different activity levels.
Nonstochastic effects occur at relatively
high doses of radiation but not at doses
of typical environmental exposure and
regulatory interest. Radionuclides ,
having a half-life of 1 hour or less are
not considered in the group proposed for
regulation, since they will decay prior to
consumption of drinking water. For a
stochastic effect such as cancer, the
probability of the effect increases with
 increasing dose, and it is assumed that a
 threshold does not exist. The cancers
 produced by radiation cover the full
 range of carcinomas and sarcomas.
 Many forms of cancer have been shown
 to be induced by radiation (ICRP, 1977;
 NAS, 1990). The epidemiological basis
 for risk estimates specific to irradiation
 have been reviewed in detail in BEIR III
 (NAS, 1980) and by U.S. EPA (1989a).
 Since the available data suggest that
 lowered dose rates of low-LET radiation
 yield a lowered cancer risk, the use of
 risk coefficients from A-bomb survivors
 (which are the result of very high dose
 rates) will probably not underestimate
 risk from low-LET radiation (EPA,
 1991d).
   The methodology used in risk
 calculations is formalized in the
 RADRISK computer code. The
 calculations assume an average lifetime
 of 70.7 years and a cohort of 100,000
 persons (Dunning et al., 1980; Sullivan et
 al, 1981; EPA, 1989a; 1991d). Equivalent
 organ doses consider the concentration
 of the radionuclide, the intake of water,
 the absorption of the radionuclide from
 the gastrointestinal tract tato the
 bloodstream, the distribution to various
 organs or compartments, the retention,
 and the radiologic decay in each organ.
 The absorption (characterized by fi) and
 fraction deposited in the organ or
 compartment (f2') are functions of the
 chemical form and of age. The values of
 fi and f2! and the retention functions for
 each radionuclide and chemical form
 are taken mostly from the tabulations in
 ICRP Publication 30 (ICRP, 1979; 1980;
 1981); Sullivan et al., (1981) and Dunning
 et al. (1984). Organ masses are values
 from ICRP Publication 23 for adults
 (ICRP, 1975). The model integrates the
 organ burden for each year of life to
 obtain an annual burden, which is
 corrected for age with a nuclide-specific
 S-factor. The S-factors (units of dose
 equivalent per Ci-day) are derived by
 calculating the number of decays in the
 organ during residence and the energy
 absorbed  as the  result of the decays.
 Using these parameters, the dose
 delivered  to each organ as the result of a
 unit intake of each radionuclide is
 calculated to. obtain the annual dOse
 rate. The target organs for dose
 estimation specified by the RADRISK
 code are ovaries, testes, breast, red
 marrow (for leukemia), lungs, thyroid,
 endosteal cells, stomach, lower and
 upper large intestine, small intestine,
 kidneys, bladder, spleen, uterus, thymus,
 thyroid, liver, and pancreas (EPA, 1989a;
 1991d).
  The risk factor associated with
 exposure to 1 Sv (Sievert; 1 Sv=100
rems) that is adopted is 39,000/106
persons (or for 1  rem, 4xlO~4persons;
 EPA, 1989a;). This risk factor is an age-
 adjusted estimate for cancer resulting
 from low-level, whole-body, low-LET
 radiation. At an exposure rate of 1
 mrem/year, based on the above risk
 factor, and a lifetime of 70.7 years, the
 lifetime probability (P) of a radiation-
 induced fatal cancer is 2.8xlO~5 per
 mrem ede per year. For the purpose of
 setting standards, the EPA generally
 considers allowable values for lifetime
 risk to lie between 10~6 and 10~4. A
 lifetime cancer risk of approximately
 10~4corresponds to 4 mrem ede/yr.
   Appendix B lists the concentrations, in
 pCi/1 that correspond to 4 mrem ede/   •
 year for each beta emitter, assuming
 lifetime intake of 2 liters of water daily.
   There may be several sources of
 uncertainty hi the beta and photon
 emitters risk evaluation. These are
 discussed in detail in the Criteria
 Document (EPA, 1991d), and are briefly
 summarized here. They include
 uncertainty in the metabolic model,
 including absorption, distribution and
 dosimetry, and the risk coefficients used
 for calculating risk. While there may be
 uncertainties in the assessment of beta
 and photon emitter risks, EPA has
 evaluated all the available data and
 believes the approach selected is likely
 to have fewer uncertainties than other
 approaches to assessing risks from beta
 and photon emitters.
   EPA solicits public comment on its
 approach to estimating risks from beta
 and photon emitters in drinking water.

 5. Alpha Emitters

  EPA considers all ionizing radiation to
 be carcinogenic, and has classified the
 ionizing radiation released during alpha
 decay as a Group A carcinogen (EPA,
 1991p). Therefore, as a class, alpha
 emitting radionuclides are considered
 group A carcinogens. There are also
 adequate data on some individual alpha
 emitters to conclude that they are
 carcinogenic. Accordingly, the Agency
 has placed alpha emitting radionuclides
 as a class in Group A as known human
 carcinogens (EPA, 1991a).
  Alpha emitters are primarily naturally
 occurring, deriving from the uranium
 and thorium decay series. There are a
 more limited number of alpha emitting
 radionuclides (than beta emitters) that
 are of potential concern in public water
 supplies, as only a few alpha emitters
 have ever been reported in the
 published literature to occur in water. In
 addition to  the naturally occurring
radionuclides, plutonium and
 americium, man-made alpha emitters,
may also be of concern, although these
have only been found at very low (less

-------
                Federal Register  /  Vol. 56,  No. 138  /  Thursday,  July 18, 1991 / Proposed Rules
                                                                     33079
than 0.1 pCi/1) concentrations in
drinking water (See section III.F).
  As for the beta and photon emitters,
risks from ingestion of alpha emitters
can be evaluated using a modelling
approach, combined with radionuclide-
speclfic epidemiology or animal studies
where available. Despite differences hi
radiation type, energy, or half-life, the
health effects from radiation are
identical, although they may occur in
different target organs and at different
activity levels. Nonstochastic effects
occur at relatively high doses of
radiation but not at doses of typical
environmental exposure and regulatory
interest. For a stochastic effect such as
cancer, the probability of the effect,
increases with increasing dose, and it is
assumed that a threshold does not exist.
The cancers produced by radiation
cover the full range of carcinomas and
sarcomas. Essentially every form of
cancer has been shown to be induced by
radiation (ICRP, 1977; NAS, 1988; 1990).
The type of cancer caused depends
largely on where the radionuclides
localize in the body as a result of
metabolism. Radionuclides that are
deposited hi bone frequently cause bone
sarcomas. Widely distributed
radionuclides may increase cancer risk
for many organs (EPA, 1991a). The
epldemiologic basis for risk estimates
specific to irradiation have been
reviewed in detail in BEER IEC (NAS,
1980) and by EPA (1989a). For internally
deposited alpha emitters, the BEIRIV
report (NAS, 1988) reviewed available
information.
  Risk assessment for the alpha emitters
was performed using the RADRISK
model (EPA, 1991a). The criteria
document for alpha emitting
radionuclides describes the metabolic
model as do Dunning et al. (1980):
Sullivan et al. (1981); and EPA (1989a),
and as described above for the beta
emitters. The model estimates radiation
dose to organs, the dose is used to
calculate risk to the organs, and the
risks to organs are summed to estimate
overall risk. Levels of the alpha emitters
representing 10"* lifetime risk in
drinking water (assuming ingestion of 2
liters daily) are presented in Appendix
C.
  Specific alpha emitters of interest
include polonium, thorium, plutonium
and possibly americium as these have
been found in water. There are some
human epidemiology and animal data
available to help in assessing the risks
posed by the individual contaminants.
However, there are not complete enough
data on any of them to form the basis of
a risk assessment, and EPA has
determined that the RADRISK modelling
approach will provide the best estimates
of the hazards posed by these
contaminants (EPA, 1991a).
  Polonium-210 is in the uranium-238
decay series, and is the daughter of
lead-210, the first long-lived daughter of
radon 222. The BEIR IV (NAS 1988)
report reviewed the available literature
on polonium. Polonium was reported to
cause lymphomas in mice, and various
soft tissue tumors in rats given
polonium. In addition, a number of non-
neoplastic adverse effects were reported
in test animals, including sclerotic
changes in the blood vessels, atrophy of
the seminiferous epithelium and
hyperplasia of the interstitial (Leydig)
cells in the testes, and other effects, but
all at relatively high doses. Effects in
exposed humans including hematologic
changes, impairment of the liver, kidney
and reproductive organs, were reported
by the BEIR IV committee. The BEIR IV
committee concluded that there is no
direct measure of risk for most polonium
isotopes based on the human data, and
suggested several possible means of
estimating risk. EPA, as discussed, has
relied on its RADRISK model in
assessing polonium risk. The model
estimates that polonium at 14 pCi/1 in
water (assuming 2 liters  daily intake)
would pose an approximate lifetime
cancer risk of 1X10"4 (EPA, 1991a).
Several public water supplies and
private wells have exceeded this value,
although most reported polonium levels
were in the range of 1 to 10 pCi/1 (EPA,
1991f).
  The BEIR IV committee also reviewed
available information on the adverse
effects of thorium. Substantially better
information (than for polonium) is
available for human exposure because a
colloidal form of thorium dioxide (ThO2;
Thorotrast) was used in medical
radiology as a contrast agent from the
1920's until about 1955. Patients were
injected with the Thorotrast. The
colloidal particles posed a radiation risk
to the reticuloendothelial system in
which they were ultimately sequestered
after injection. Various studies of the
Thorotrast patients showed clear
increases in liver cancers, as well as
possible increases in leukemia.
However, the BEIR committee discussed
the limitations of these data for
assessing the risk due to other forms of
thorium. Forms of thorium other than
ThO2 would have a different metabolic
fate than the Thorotrast, and would
affect different organs. Therefore, EPA
believes a dosimetric approach, as
contained in the RADRISK model,
provides the best available basis for
assessing risk from the various thorium
isotopes. Based on the model results,
EPA estimates that the various thorium
isotopes pose lifetime cancer risks of
1X10~4 at drinking water concentrations
ranging from 50 pCi/1 to approximately
125 pCi/1 (EPA, 1991a). Most reported
thorium occurrence in drinking water
was at levels near 1 pCi/1 (EPA, 1991f).
  Plutonium is widely present at very
low levels in the environment, largely as
a result of atmospheric nuclear weapons
testing from 1945 to 1963. It is also found
in nuclear power reactors and could be
released in the event of an accident. The
BEIR IV committee reviewed available
data on plutonium and other
transuranics. They concluded that
studies in animals clearly indicate bone,
liver, and lung (by inhalation) cancers
caused by plutonium exposure.
However, available (and limited) human
epidemiology studies have not yet
shown unequivocal association between
plutonium exposure and cancer at any
particular anatomical location. The
Committee recommended risk
assessment based on analogy with other
radionuclides and high LET radiation
exposure risks. EPA has used its
RADRISK model to assess plutonium
risks. The RADRISK model estimates
that lifetime cancer risks of
approximately 1X10"4 are posed by
drinking water plutonium concentrations
of about 7 pCi/1 for the different
plutonium isotopes (EPA, 1991a).
Reported plutonium levels in drinking
water were less than 0.1 pCi/1 (EPA,
1991f).
  Estimated risks for these and other
alpha emitting compounds can be found
in appendix C.
  There may be several sources of
uncertainty in the alphas risk
evaluation. These are discussed in detail
in the Criteria Document (EPA, 1991a),
and are briefly summarized here. They
include uncertainty in the metabolic
model, including absorption, distribution
and dosimetry, and the risk coefficients
used for calculating risk. While there
may be uncertainties in the assessment
of risks, EPA has evaluated all the
available data and believes the
approach selected is likely to have
fewer uncertainties than other
approaches to assessing risks from
alpha emitters.
  EPA solicits public comment on its
assessment of risks from alpha emitting
radionuclides in drinking water.

D. MCLG Determinations
  For the reasons stated in the
preceding sections on health effects and
risks (e.g.,  the fact that all of these
radionuclides are Group A, known
human carcinogens) and based on the
Agency's policy of setting MCLGs for

-------
                  Federal Register  / Vol.  56,  No.  138 /  Thursday, July 18, 1991  / Proposed Rules
 33080  —	

 known or probable human carcinogens
 at zero, the Agency is proposing to set
 MCLGs of zero for radon, radium-226,
 radium-228, uranium, and alpha and
 beta particle and photon emitters.

 V. Proposed Maximum Contaminant
 Levels

 Summary of the Proposal
   The SDWA directs the Agency to set
 an enforceable standard for a
 contaminant (MCL) as close to the
 health goal for the contaminant (MCLG)
 as is "feasible". Feasible is defined as
 the use of the "best technology,
 treatment techniques and other means
 which the Administrator finds * * * are
 available (taking cost into
 consideration)". (Section 1412(b)(5).) In
 determining MCLs, the Agency
 considers a number of factors. The
 Agency evaluates the availability and
 performance of the various technologies
 capable of removing the contaminants,
 identifying those that have the highest
 removal efficiencies, that are compatible
 with other types of water treatment
 processes, and that are not limited to
 application in a particular geographic
 region. As the MCL levels must be
 generally enforceable, EPA also
 considers the ability of laboratories to
 measure reliably for the contaminants in
 water. EPA derives practical
 quantitation levels (PQLs) which reflect
 the contaminant concentration that can
 be measured by good laboratories under
 normal operating conditions within
 specified limits of precision and
 accuracy.
   The Agency also considers the health
 risk associated with the contaminant.
 The Agency estimates both the
 incidence of disease and the risk to
 individuals. EPA has historically set a
 reference risk range for carcinogens at
 10~4to 10"6 lifetime individual risk; risks
 within this range have been considered
 acceptable.

 The Role of Costs in Setting MCLs
   In setting MCLs, the Agency also
 considers a number of cost elements. In
 the past, EPA has generally limited
 consideration of economic costs under
 the SDWA to whether a technology is
 affordable for large municipal water
 systems. (52 FR 42225 Nov. 3,1987; "the
 legislative history indicates that EPA is
 to base MCLs on treatment technology
 affordable by the largest public water
 systems"). However, EPA has
 determined that nothing in the statutory
language, the legislative history or EPA's
prior constructions of the statute
precludes consideration of cost-
 effectiveness in setting MCLs under the
 SDWA (EPA, 1990o).
   EPA's focus on affordability for large
 systems in the past is consistent with
 statements in the 1974 House Committee
 Report:
   In determining what methods are generally
 available, the Administrator is directed to
 take costs into account. * * * It is evident
 that what is a reasonable cost for a large
 metropolitan (or regional) public water
 system may not be reasonable for a small
 system which serves relatively few users.
 The Committee believes, however, that the
 quality of the Nation's drinking relatively few
 users. The Committee believes, however, that
 the quality of the Nation's drinking water can
 only be upgraded if the systems which
 provide water to the public are organized as
 to be most cost-effective.  In general, this
 means larger systems are to be encouraged
 and smaller systems discouraged. For this
 reason, the Committee intends that the
 Administrator's determination of what
 methods are generally available  (taking cost
 into account) is to be afforded by large
 metropolitan or regional public water
 systems.

 H.R. Rep. No. 93-1185, A Legislative
 History of the Safe Drinking  Water Act,
 97th Cong, second session, pp. 549-550
 (1982) (emphases supplied). Far from
 prohibiting cost-effective solutions to
 the Nation's drinking water problems,
 the legislative history indicates that
 Congress wanted to encourage cost-
 effectiveness, but thought that
 promoting consolidation of small
 drinking water systems into larger ones
 would promote cost-effective solutions
 in the circumstances that prevailed in
 1974. EPA has concluded that in light of
 changing circumstances since 1974,
 including the large number of MCLs that
 have been established  in the meantime
 under the SDWA, it is no longer
 appropriate to focus exclusively on large
 system costs in order to promote cost-
 effective solutions that protect human
 health from contaminants in the nation's
 drinking water.
  In addition to the statements in the
 1974 House Committee Report, a 1986
 floor statement by Senator Durenberger
 might be read to  suggest that
 consideration of large system
 affordability is the only permissible role
 for considering costs under the SDWA.1
  1 In the floor debates on passage of the
conference report for the 1986 amendments to the
SDWA, Senator Durenberger stated that the
amendments were "not an instruction for the
administrator to conduct a cost-benefit analysis to
determine the MCL. The law emphatically does not
provide that the administrator will set the MCL at a
level where benefits outweigh costs, nor does it
require EPA to balance costs and benefits in any
other way. Cost only enters into the judgment of the
administrator in defining which treatment
technologies are to be considered best available
technologies. And availability in this instance is
considered only in the context of the largest supply
systems." 132 Cong. Rec. S6287 (May 21,1986).
 However, EPA believes that, in context,
 the 1986 Durenberger floor statement
 was not in fact intended to preclude
 consideration of cost-effectiveness, as
 opposed to cost-benefit analysis.2
 Nowhere in his floor statement does
 Senator Durenberger reject
 considerations of cost-effectiveness (as
 opposed to cost-benefit) in setting
 MCLs. On the contrary, later in the same
 floor statement, Senator Durenberger
 refers to considering cost-effectiveness
 with approval in the context of using
 granular activated carbon (GAG)
 technology to establish MCLs. 132 Cong.
 Rec. S6294. EPA believes that it would
 be anomalous and contrary  to
 Congressional intent to sanction using
 cost-effectiveness considerations hi
 setting MCLs using one technology
 (GAG), which Congress clearly
 intended, but to prohibit consideration
 of cost-effectiveness in setting MCLs
 using other technologies which raise
 very similar issues.
    Similarly, neither the statutory
 language nor EPA's prior constructions
 preclude considering the cost-
 effectiveness of requiring additional
 increments of technology or
 contaminant control in establishing
 MCLs. The  statute requires EPA to
 establish the MCL as close to the
 maximum contaminant level goal
 ("MCLG") "as is  feasible." SDWA Sec.
 1412(a)(4), 42 U.S.C. 300g-l(a){4). The
 term "feasible" is in turn defined as:
   * * *  feasible with the use of the best
 technology, treatment techniques and other
 means which the Administrator finds, after
 examination for efficacy under field
 conditions and not solely under laboratory
 conditions, are available [taking cost into
 consideration}.
   SDWA sec. 1412(a)(5), 42 U.S.C.  3009-
 l(a)(5) (emphasis supplied). The
 dominant emphasis in the statutory
 language is on achieving practical
 results.3 Furthermore, far from
   EPA's approach to considering cost-effectiveness
 in this rule is consistent with the literal language of
 Senator Durenberger's floor statement, in that EPA
 is considering cost-effectiveness in the context of
 determining which technology should be deemed
 "best available technology," and these cost-
. effectiveness considerations apply to large as well
 as small systems.
   2 In any event, even if the Durenberger floor
 statement had been intended to restrict EPA's
 discretion to consider costs in any way other than
 large system affordability (which EPA does not
 believe that it was), legally it could not have that
 effect Floor statements by individual legislators,
 while entitled to some weight, do not effectively
 restrict agency discretion to adopt statutory
 interpretations which are otherwise reasonable and
 consistent with the statute, as recent Supreme Court
 cases have made clear. See, e.g. Brock v. Pierce
 County, 476 U.S. 253, 263 (1986).
   3 One factor indicating that Congress intended
 the standard-setting exercise to focus on obtaining
                                 Continued

-------
                Federal Register / Vol.  56,  No. 138  /  Thursday, July 18, 1991 /  Proposed Rules
                                                                      33081
precluding consideration of economics
in setting MCLs, EPA is specifically
instructed to "tak[e] cost into
consideration" in determining whether a
technology is available.
  EPA does not believe that Congress's
instruction to take economic costs into
account in determining whether
technologies are available was intended
to preclude consideration of economic
cost-effectiveness in determining which
technology, or level of technology, is
"best." As a matter of the ordinary
meaning of language, if two technologies
or levels of control achieve comparable
or nearly comparable results, but one of
them is much more efficient or cost-
effective than the other, the more
wasteful and expensive one could
hardly be  said to be the "best."
  This plain language interpretation of
"best" available technology as
permitting some weighing of economic
costs is reinforced by the fact that EPA
has construed "best available
technology" requirements in other
environmental statutes to encompass
cost-effectiveness, but generally not
cost-benefit considerations. For
example, EPA has long interpreted
"best" available control technology for
purposes of the PSD program under the
Clean Air Act as incorporating cost-
effectiveness considerations [EPA,
1979a; unreasonable adverse economic
effects of an available control
technology, as demonstrated by an
incremental analysis of that option
relative to others, are an adequate basis
to reject an alternative). EPA's
interpretation of the "best" available
technology requirement under the Clean
Air Act as allowing consideration of
cost-effectiveness has been upheld by
the courts. See, e.g., Northern Plain
Resource Council v. EPA,  645 F.2d 1349
(9th Cir. 1981) (affirming EPA's rejection
of an available control option on cost-
effective grounds under the "best
available control technology"
requirements of the PSD program under
the Clean Air Act).
  From the standpoint of facilitating
sound environmental policy, it makes
little sense to set multiple MCLs based
solely on considerations of the ability of
large systems to afford each individual
MCL. Rather, EPA believes that the
overall purposes of the Safe Drinking
Water Act, assuring the overall safety of
the Nation's drinking water supply, will
be best effectuated by a consideration of
cost-effectiveness. In addition, EPA has
historically set its MCLs based on 10~4
to 10~6 lifetime risk to an exposed
 practical results is that EPA is adjured to set
 standards based on practical results obtainable in
 the field, rather than solely in the laboratory.
individual. See for example 40 CFR part
300 [National Contingency Plan), 40 CFR
part 61 [Benzene NESHAPs, 54 FR
Contingency Plan), 40 CFR part 61
(Benzene NESHAPs, 54 FR 38044,
September 14,1989), and 52 FR 25700-
25701, July 8,1987 (Final VOC MCLs). If
cost-effectiveness could not be
considered, EPA would be required to
set each individual MCL at the limits of
technology that could be afforded. This
does not necessarily maximize the
overall health benefits to the drinking
water supply as a whole. The limited
resources which are "affordable" would
achieve greater health benefits for
people served by the drinking water
supply as a whole if these resources are
deployed where they can achieve the
greatest health benefits in the aggregate.
Indeed, in prior SDWA rules, EPA has in
fact taken cost-effectiveness into
account in setting MCLs for certain
volatile organic chemicals, albeit
without extensive discussion, 52 FR
25699 (July 8,1987), since failure to do so
would lead to absurd results.
  In sum, EPA has concluded that
limited consideration of cost-
effectiveness in setting MCLs will
further the overall goal that Congress set
in the SDWA, which was to maximize
the health and safety of the country's
drinking water supply as a whole.
Setting each individual MCL at  the
limits of economic affordability would,
in EPA's judgment,  actually impede  that
goal by misallocating limited resources
to achieving comparatively small or
nonexistent health benefits based on the
order in which EPA sets MCLs,  rather
than where the greatest health benefits
can be achieved. Therefore, EPA
believes it has an obligation to
maximize the overall health benefits
that accrue from all its actions affecting
the nation's drinking water supply. For
all of these reasons, EPA believes that it
is consistent with the language,
legislative history, and Congress's
overall purposes to interpret the SDWA
to allow EPA to consider cost-
effectiveness in setting the level of
control which is considered feasible
using the best technology.

Radionuclides MCLs
   In selecting MCLs for the four
radionuclides that have the greatest
frequencies of occurrence in public
water supplies (radon, radium 226 and
228, and uranium), the Agency
considered the factors described above.
The Agency was able to collectively
analyze these contaminants because
they have the unique characteristic  that
all radionuclides cause  cancer by the
same mechanism, i.e., delivering ionizing
radiation to tissue (by either external
exposure, or internally when ingested or
inhaled). These individual contaminants
may be viewed as vehicles for internal
delivery of that ionizing radiation to
different parts of the body. Indeed, the
Agency has classified ionizing radiation
(as well as the individual contaminants
proposed for regulation here), as a class
A carcinogen. This classification applies
to alpha, beta and photon, and gamma
ray emitters. It is therefore possible to
make comparisons of either the
radioactivity in water or removed from
water (in pCi or uCi), or the radiation
dose delivered by each of the
radionuclides in terms of "reins ede".
Rems ede are a way of normalizing for
different radionuclides the radiation
dose to the body taking into account the
effect of different types of ionizing
radiation on tissue as well as the
distribution of dose (largely determined
by metabolic destination of the
radionuclides) in the body of the
ingested or inhaled radionuclide. These
measures permit comparison of the
overall reduction in either total
radioactivity or the effective dose of
ionizing radiation that can be achieved
with different control level options.
   These common characteristics  of the
radionuclides allow comparisons among
radon, radium 226 and 228, and uranium
in terms of overall reduction in ionizing
radiation in drinking water and
radiation dose delivered via drinking
water by implementing different control
levels, and the relative cost of such
reductions. These comparisons were
considered in evaluating alternative
MCLs.
   EPA is interested in soliciting public
comment on the applicability of cost-
effectiveness to other drinking water
contaminants.

Radon
   Radon is estimated to cause about
8,000 to 40,000 (EPA, 1989g) lung cancer
deaths annually. Typical indoor radon
levels (1-2 pCi/1) pose estimated lifetime
lung cancer risks near 1 in 100. The most
significant contributor to indoor radon is
soil gas. However, volatilization  of
radon from drinking water during
household use also increases indoor
radon levels thereby contributing to
increased risk of lung cancer. Direct
ingestion of radon may also pose some
risk of stomach and other cancers (EPA,
1991c).
   EPA estimates that more than 26,000
public water systems have radon in
water at levels exceeding an
approximate 10" 4 individual lifetime risk
level. Because radon is significantly
more prevalent in drinking water than
radium 226 and radium 228 or uranium,

-------
 33082
Federal  Register / Vol. 56, No.  138 / Thursday, July 18, 1991  / Proposed Rules
 radon poses the greatest risk on a nation
 wide basis of any of the radionuclides
 found to occur in drinking water (EPA,
 1991i). Accordingly, the Agency first
 determined the appropriate MCL for
 radon.
   In determining what radon MCL to
 propose, the Agency evaluated the
 availability and performance of the
 various technologies capable of
 removing radon. Based on this
 evaluation, the Agency determined that
 only aeration fulfills the requirements of
 the SDWA as best available technology
 for radon removal as discussed in
 Section V.B.2. Based on treatability,
 radon could theoretically be reduced to
 100 pCi/1 or lower in most water supply
 systems.
   The Agency also considered whether
 the ability of laboratories to measure
 reliably for radon in water imposes any
 limits on where the Agency can set the
 MCL. The Agency determined that the
 radon PQL could be established at 300
 pCi/1 (although other researchers
 variously believe the number could be
 either lower or higher; see Sections V.C
 and V.D below).
   The Agency then analyzed the costs
 of implementing a 300 pCi/1 standard.
The Agency estimated that costs for
 large systems to achieve 300 pCi/1
 would be very low ($4 per household per
 year). The costs for small systems, while
 greater ($170 per household per year in
 systems serving 25 to 100 persons), were
 found to be affordable by the Agency.
 Because of the large number of water
 systems that would need to install
 treatment to reach the 300 pCi/1
 standard, the annual nationwide costs
would be  approximately $180 million.
While this is a significant cost,  the
Agency concluded that these costs are
reasonable in view of the substantial
reduction in exposure to ionizing
radiation and the resulting risk
reduction that would be achieved. At
this level, EPA estimated that up to 8300
uCi, representing approximately 200,000
person-rems ede would be removed
from drinking water each year (EPA,
     .
  Finally, the Agency estimated the
health risks at the 300 pCi/1 level to be a
lifetime risk of approximately 10~4 (i.e.,
2X10-4). The Agency concluded that
this level would be adequately
protective of public health since it is
within the target risk range of
approximately 10~4to 10~6.
  Taking these factors into account, the
Agency is proposing to set the MCL for
radon at 300 pCi/1.

Radium and Uranium
  Radium 226 and 228 and uranium are
also naturally occurring contaminants.
                       Although they are less prevalent than
                       radon, they are present in a significant
                       number of water systems. The total
                       person-rems ede and associated
                       population risk attributable to these
                       contaminants collectively are much
                       lower than for radon alone, although in
                       some communities individual risks from
                       these contaminants exceed the target
                       risk range. The Agency identified
                       several technologies that are highly
                       efficient in removing radium 226 and 228
                       and uranium from water. Based on this
                       evaluation, radium 226 and 228 could
                       each be theoretically treated to a level
                       lower than 2 pCi/1; uranium could be
                       theoretically treated to a level lower
                       than 5 pCi/1 (see section V.B below).
                       The Agency also established PQLs for
                       these three radionuclides at 5 pCi/1 for
                       each (see sections V.C and V.D below).
                       EPA's analysis indicated that it is
                       technologically feasible to achieve
                       control levels of 5 pCi/1 for radium 226
                       and 228, and uranium.
                         The Agency 'then analyzed a number
                       of cost factors. The cost of reducing
                       radioactivity and rems ede of delivered
                       dose by removing radium and uranium
                       to the technically feasible level is much
                       greater than the cost of reducing
                       radioactivity and rems ede by removing
                       radon (EPA, 19911). First, the cost of the
                       treatments for radium and uranium on a
                       household basis, would be
                       approximately $20 to $60/year for large
                       systems and $700 to $800 per year for
                       the  smallest systems. These costs are far
                       greater than for treatment of radon
                       which would be approximately $4 per
                       house per year for large systems and up
                       to $170 per house per year for the
                       smallest systems. The total number of
                       both uCi and rems ede that would be
                       removed by controlling radon at 300
                       pCi/1 is much greater than the number
                       that would be removed by controlling
                       radium and uranium at the technically
                       feasible levels. At the 300 pCi/1
                       proposed standard for radon, nearly
                       8300 uCi annually, representing
                       approximately 200,000 person-rems ede
                       per year would be removed from
                       drinking water. The total annual costs
                       for removing this radiation by treating
                       radon is about $180 million. In contrast,
                       at the technically feasible levels, 150
                       uCi, representing 86,000 rems ede of
                       radium and uranium would be removed
                       annually, at a cost of nearly $400
                       million. The cost of removing radiation
                       by controlling radium and uranium is
                       approximately 200 fold greater per uCi
                       removed and 5 fold per rem removed
                       greater than that for radon treatment.
                         The Agency concludes that the cost of
                       reducing radioactivity and rems ede of
                       delivered dose by removing these three
                       contaminants to the technically feasible
 level is disproportionate to the cost of
 reducing radioactivity and rems ede by
 removing radon. The Agency does not
 believe it would be reasonable to select
 MCLs that would impose such
 disproportionate costs.
   Since it is not cost-effective to set the
 MCLs for radium and uranium at the
 technically feasible levels, EPA
 examined alternatives at the 10~4
 lifetime individual risk level, which are
 approximately 20 pCi/1 for radium 226,
 20 pCi/1 for radium 228 and 20 jxg/1 for
 uranium. These levels are less costly but
 still assure that persons served by
 public water systems will not be
 exposed to a greater than approximately
 !XlO~4risk. In addition, the uranium
 value is protective against kidney
 toxicity, which may occur at levels far
 below the 10"4 lifetime risk level for
 uranium. For drinking water
 contaminants, EPA has set a reference
 risk range for carcinogens (after
 regulation) at 10~4to 10~6 excess
 individual risk from lifetime exposure
 and therefore considers an
 approximately 10~4risk protective of
 public health. Based on these
 considerations, EPA proposes to set the
 MCL for radium 226 at 20 pCi/1, for
 radium 228 at 20 pCi/1, and uranium at
 20 jig/1.
   Following is a detailed discussion of
 the factors considered in developing this
 proposal.                        ~ ,

 A. BATs and Associated Costs

   Section 1412(b)(6) of the Act states
 that each national primary drinking
 water regulation which establishes an
 MCL shall list the technology, treatment
 techniques, and other means which the
 Administrator finds to be feasible for
 purposes of meeting the MCL. In order
 to fulfill the requirements of section
 1412(b)(6), the EPA has identified best
 available technologies (BAT) for each
 radionuclide covered in this proposal.
  Technologies are judged to be BAT
 based upon the following factors: High
 removal efficiency, general geographic
 applicability, cost, reasonable service
 life, compatibility with other water ,
 treatment processes, and the ability to
 bring all of the water in a system into
 compliance.
  Table 5 summarizes the BATs
 identified by EPA for the removal of the
 subject drinking water contaminants,
 and their respective removal
 capabilities.
  Table 6 shows theoretical technology
limits of BATs. The achievable effluent
concentrations are based upon
maximum removal of influent levels

-------
                Federal Register /  Vol. 56,  No. 138 / Thursday, July 18, 1991  / Proposed Rules
                                                                      33083
from the NIRS survey data and            rates. PWSs applying these BATs will     achieve these estimated low effluent
maximum demonstrated BAT removal     not need to design treatment systems to    concentrations.

                                      TABLE 5. BAT CONTAMINANT REMOVAL RATES "•
Contaminant
Radon 	
Radium (226 and 228) . 	

Beta Emitters
j_131 	



Ion exchange

80-97%
65-99%
95-99%

95-99%
2 90-99%

Lime softening

75-95%
85-99%





Aeration
Up to 99.9%







Reverse
osmosis

87-98%
98-99.4%
90-99%
90-99%
90-99%
96-99%

Coagulation/
filtration


80-95%





    'Information regarding removal efficiencies, test conditions and other factors are contained in the EP/j'Technology and Cost documents (T&C}i and cost
supplamants to each T&C, I.e., for uranium, radium, radon and manmade radionuclides  (EPA, 1984b; 1985b; 1986b;  1986c; 1987b, 1987c, 1987d, 19886).
    * Mixed bed or two bed (antonic/cationic) exchange resins. Removal rate does not include I •""
                            -131.
   TABLE 6.—TECHNOLOGY LIMITS FOR
        RADIONUCLIDE REMOVAL
Contami-
nant/
Technolo-
gy-(BAT)
Ration
Aeration..
Racfium-
226
IE.,,.....,....
LS 	
RO 	 .....
ftadlum-
223
IE...™......
LS 	 	
RO... 	
Uranium
IE 	
LS 	 ..
RO 	 .....
CF~ 	
Bota
Emitters
Two bed
km
ox-
changa
RO.™, 	
Greatest
percent
removal

99.9


97
95
98


97
95
98

99
99
99
95





99
99
Maximum *
Influent
(pCi/l)

26,000


15
15
15


12
12
12

88
88
88
88





*
*
Achievable
effluent
(PCi/l)

26


0.45
0.75
0.30


0.36
0.60
0.24

0.9
0.9
0.9
4.4





*
*
   1 Maximum levels In groundwater sources of drink-
 Ing water as reported In NRS.
   Note: IE (Ion exchange); LS (lime softening); RO
 (rovorse osmosis); CF (coagulation/filtration).
   Source:  (EPA,  1984b; 1985b; 1986b;  1986c;
 1987b 1987c; 1987d; 1988e).
  The total costs for the removal of
specific radionuclide contaminants,
using the proposed BATs, are
summarized in Table 7. Tables 8 and 9
display the total capital cost and annual
operation and maintenance costs,
respectively. Costs cited in Tables 7, 8
and 9 are based on treatment conditions
that would require removal of fairly high
levels of contamination. The assumed
removal rates are as follows: 50 percent
for radium; 80 percent for radon; and 60
percent for uranium. The general
assumptions used to develop the
treatment costs include: chemical costs,
capital costs amortized over 20 years at
a 10 percent interest rate, current
engineering fees, contractor overhead
and profit,  late 1986 power and fuel
costs and labor rates (EPA, 1984b; 1985b;
1986b; 1986c; 1987b; 1987c; 1987d; 1988e).
Costs as evaluated  here assume the
existence of no residential POE water
treatment such as water softening for
aesthetic reasons which might
incidentally reduce some pollutant
levels. The prevalence of such home
treatments is extremely difficult to
estimate and incorporate into a national
level analysis.
   EPA is presently  conducting a study of
treatment for very small water systems.
All of the small system treatments for
radionuclides, and also other
contaminants, are included, and
verifying treatment costs is one element
of this study. EPA will make this study
available to the public when it is
completed. EPA solicits public comment
and data on treatments that may be
especially well suited to small systems
and any treatment systems designed for
small systems, including data on
treatment efficiencies, adaptability of
designs to different size systems, and
cost to install and operate treatment
systems designed for small public water
suppliers.
  Costs may vary significantly from
those shown, depending on local
circumstances. Costs of treatment will
be less  than shown on Table 8 if
contaminant concentration levels
encountered in the raw water are lower
than those used for the calculations.
However, costs of treatment will be
higher if additional treatment or storage
requirements need to be satisfied. The
costs in Tables 7, 8 and 9 do not include
those attributable to the treatment and
disposal of wastes generated by water
treatment plants. Waste disposal
techniques and associated costs are
discussed in section C, following a
discussion of BATs.
   TABLE 7.—TOTAL PRODUCTION COST OF CONTAMINANT REMOVAL BY BAT * NOT INCLUDING WASTE BY-PRODUCT DISPOSAL COST
                                      (DOLLARS/1,000 GALLONS, LATE 1986 DOLLARS)

Radium (50% removal):



Radon (80% removal):
Uranium (60 removal):
Ion exchange 	
Population served :
25-100
2.60
6.40
3.50
5.10
0.94
4.40
4.10
100-500
1.50
3.00
1.70
4.00
0.50
2.10
2.70
500-1,000
0.90
1.30
0.78
2.70
0.26
0.83
2.00
1,000-3,300
0.58
0.67
0.39
2.30
0.15
0.38
1.70
3,300-
10,000
0.33
0.54
0.11
1.30
0.07
0.10
1.10
> 1,000,000
0.17
0.16
0.01
0.72
0.05
0.02
1.00

-------
33084
Federal Register /  Vol. 56, No. 138 / Thursday,  July 18,  1991 / Proposed Rules
  TABLE 7.—TOTAL PRODUCTION COST OF CONTAMINANT REMOVAL BY BAT1 NOT INCLUDING WASTE BY-PRODUCT DISPOSAL COST
                                (DOLLARS/1,000 GALLONS, LATE 1986 DOLLARS)—Continued

Lime softening, modified 	
Reverse osmosis 	

Population served
25-100
4.20
6.20
100-500
2.10
4.70
500-1,000
0.93
3.50
1,000-3,300
0.47
2.70
3,300-
10,000
0.20
1.50
> 1,000,000
0.03
0.89
Notes:
    ^echnologies and cost documents, and cost supplements for radium, radon, and uranium (EPA, 1984b; 1985b; 1986b; 1986c; 1987b; 1987c- 1987d- 1988e)
 form the basis for costs. Costs  were revised  in May,  1990 to  account for new  system level treatment design flows  adopted  by EPA (EPA, 1990d).

                               TABLE 8.—CAPITAL COST OF CONTAMINANT REMOVAL BY BAT (1>
                                                 (Kilo Dollars, Late 1986 Dollars)

Radium (50% removal):
Ion exchange 	
Lime softening, new 	
Lime softening, modified 	
Reverse osmosis 	
Radon (80% removal):
Packed tower aeration 	
Uranium (60% removal):
Coagulation/filtration, modified 	
Ion exchange 	
Lime softening, modified 	
Reverse osmosis 	 	 	

Population served
25-100
36
79
33
51
15
27
41
43
64
100-500
91
130
74
160
33
55
100
91
200
500-1,000
180
180
140
340
58
96
200
160
500
Notes:
1,000-3,300
280
240
200
820
78
130
310
220
960

3,300-
10,000
350
540
150
1,000
100
100
330
300
1,400

> 1,000,000
31,000
55,000
400
177,000
13,000
480
31,000
480
249,000

    1 Technologies and cost documents, and cost supplements for radium, radon, and uranium (EPA, 1984b; 1985b: 1986b; 1986c: 1987b: 1987c- 1987d- 1988e)
form the basis for costs. Costs were revised in  May, 1990  to account for new system  level treatment design  flows adopted by EPA (EPA/ 1990d)!

       TABLE 9.—OPERATION AND MAINTENANCE COST OF CONTAMINANT REMOVAL BY BAT (K$/YEAR, LATE 1986 DOLLARS)

Radium (50% removal):
Ion exchange 	
Lime softening, new 	
Lime softening, modified 	
Reverse osmosis 	 ; 	
Radon (80% removal):
Packed tower aeration 	
Uranium (60% removal):
Coagulation/filtration, modified 	
Ion exchange 	
Lime softening, modified 	
Reverse osmosis 	 	 	

Population served
25-100
1.1
3.8
3.2
4.5
0.2
5.7
3.4
3.5
5.1
100-500
2.8
11
6.4
16
0.6
12
12
7.4
18
500-1,000
7.5
20
8.2
45
1.4
15
39
10
52
Notes:
1,000-3,300
17
28
9.5
100
3.1
16
110
13
120

3,300-
10,000
43
73
9.1
200
7.6
14
250
16
230

> 1,000,000
13,000
9,700
1,100
50,000
3,400
1,400
95,000
3,200
59,000

t~ 1 Technologies and cost documents, and cost supplements for radium, radon, and uranium (EPA, 1984b; 1985b; 1986b; 1986c; 1987b: 1987c: 1987d- 1988e)
form the basis for costs. Costs were revised in May, 1990  to account for new system level  treatment  design flows adopted by  EPA (EPA, 1990d)!
B. Best Available Technologies (BATs)

  1. Radium-226 and radium-228. The
Agency proposes that of the
technologies capable of removing
radium from source water, lime
softening, ion exchange and reverse
osmosis fulfill the SDWA requirements
as BAT for radium removal. While
radium-226 and radium-228 are
radiologically different, they are
chemically the same. Therefore, the
same BATs, with the same removal
efficiencies, apply to both. All of these
                       technologies have demonstrated high
                       radium removal efficiencies and have
                       been determined to be of low cost for
                       large public water systems. All of these
                       technologies are currently available and
                       have been installed in public water
                       supplies and are compatible with other
                       water treatment processes currently in
                       use. The full range of technical
                       capability and unit costs for each of the
                       proposed BATs for radium removal is
                       summarized in the EPA publication,
                       "Technologies and Costs for the
                       Removal of Radium from Potable Water
Supplies" (EPA, 1984b), and the
supplementary cost document for
radium (EPA, 1987d). Treatments
applicable to smaller systems have also
been identified (EPA, 1988g; 1988h).
  a; Lime softening. Lime softening is
capable of achieving removal
efficiencies for radium of 75 to 95
percent. At optimum pH levels (between
10 and 10.6) removal efficiencies of 94 to
95 percent can be achieved. Lime
softening can also be used to reduce
TDS, turbidity and heavy metals as well
as radium and total hardness. The

-------
                Federal  Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed  Rules
                                                                      33085
estimated cost for an existing lime
softening system to be modified to
remove radium ranges from $3.50/1,000
gallons treated for systems serving from
25-100 persons to $.01/1,000 gallons
treated for systems serving more than
1,000,000 persons. However, if a new
lime softening plant was built to remove
radium its cost would range from $6.40/
1,000 gallons to $0.16/1,000 gallons for
the same system sizes.
  For a utility planning to use or
currently using lime softening
technology to remove radium, waste
disposal concerns deserve ample
consideration. Radium-226 and radium-
228 concentrations in lime softening
sludge have been reported by Snoeyink
et al. (EPA, 1985d) to range from about 1
to 22 pCi Ra-226/g and from 2 to 12 pCi
Ra-228/g dry solids. Extended sludge
drying in an impoundment may increase
the dry solids content to 70 percent or
greater, with a corresponding increase
in sludge contaminant concentration.
Backwash waters may contain radium
concentrations of 6 to 50 pCi/1. (EPA,
1985d).
  b. Ion exchange. Cation exchange
systems are capable of removing from 80
to 97 percent of radium from drinking
water. Estimated costs range from $2.60/
1,000 gallons treated for systems serving
25-100 persons, to $0.17/1,000 gallons
treated for systems serving over
1,000,000 persons for removal of radium
from ground water. Ion exchange
softening systems are adaptable for both
large and small systems, and are
acceptable as either a new installation
or an add on to an existing facility.
Sodium cation exchange resins and ion
exchange equipment are readily
available commercially. Finished
("softened") waters may be corrosive to
distribution system materials. However,
a bypass of some unsoftened water,
blended with the finished water, may
provide adequately protective levels of
calcium carbonate, reducing the finished
water corrosivity. Disposal of
concentrated waste brines containing
relatively high TDS and radioactivity
should be given full consideration before
implementing this technology.
   c. Reverse osmosis. Reverse osmosis
 (RO) membranes are capable of
removing between 87 to 98 percent of
 the radium present in source water. RO
 has been primarily used for removing
 total dissolved solids (TDS) from water
 in treatment of brackish and sea waters
 for desalination purposes. At pressures
 of 200 and 425 psi, RO is capable of 95
 and 98 percent radium removal,
 respectively. At reduced pressures this
 process is adaptable to fresh water
 sources. The RO method can be used by
both large and small systems. If
operated to remove 50 percent of the
influent radium, costs would range from
approximately $5.10/1,000 gallons
treated for systems serving 25-100
persons to $0.72/1,000 gallons treated for
systems serving over 1,000,000 persons.
If removal of TDS is also a goal, then
using reverse osmosis is a very cost
effective solution in the removal of
radium from ground waters.
  RO performance is adversely effected
by the  presence of turbidity, iron,
manganese, silica or scale producing
constituents in the source water. If
pretreatment is not already in place to
remove these constituents, the cost to
install the pretreatment facilities may be
an important factor. Disposal of waste
brine, the reject flow representing 20 to
50 percent of the feed (source) water,
and the quantity of available feed water
to accommodate this technology, would
require consideration by a water system
in its initial evaluation of alternative
technologies for radium removal.
  2. Radon. The Agency proposes that,
of the technologies capable of removing
radon from source water, only aeration
fulfills the requirements of the SDWA as
BAT for radon removal. Aeration has
demonstrated radon removal
efficiencies in excess of 99.9 percent.
This technology is currently available,
has been installed in public water
supplies, and is compatible with other
water  treatment processes in different
regions. The full range of technical
capabilities for this proposed BAT is
discussed in the EPA technologies and
costs document for radon (EPA, 1987b),
and summarized below.
  Granular activated carbon (GAG) can
also remove radon from water, and was
evaluated as a potential BAT for radon.
However, the long empty bed contact
time required for radon removal renders
it infeasible for large municipal
treatment systems, and it is therefore
not considered a BAT for radon.
   a. Aeration. Aeration techniques for
removal of radon from drinking water
include active processes such as diffuse
 aeration, packed tower aeration (PTA),
 slat tray aeration and free fall, with or
without spray aerators, and passive
processes such as free-standing, open-
 air storage of water for reduction of
 radon. Radon reduction by decay (into
 the daughter products of radon) may
 also occur in storage tanks and in
 pipelines which distribute drinking
 water, reducing radon by approximately
 10 to 30 percent, with 8 to 30 hour
 detention periods. Aeration is
 considered BAT for meeting the
 proposed radon MCL due to high
 removal efficiencies, its relative
simplicity as a technology, relatively
low cost and ease of operation,
availability, and compatibility with
other treatment processes. The aeration
technique that a system chooses for
radon reduction will depend upon
source water quality (including radon
and other contaminants removed or
otherwise affected by aeration),
institutional or manpower constraints,
site-specific design factors, and local
preferences.
  The costs associated with the various
technological options for radon
reduction, such as packed tower
aeration (PTA) and diffused aeration
installations, have been examined (EPA,
1987b). Ninety-nine percent reduction of
radon by PTA is estimated to cost from
$1.20/1,000 gallons treated for very
small systems which serve 25 to 100
persons, to $0.07/1,000 gallons treated
for systems serving 1,000,000 persons.
Eighty percent reduction of radon by
PTA is estimated to range from $0.94 to
$0.05 per 1,000 gallons for the same
system sizes.
  The following factors influence the
aeration processes and therefore affect
radon removal rates:
—temperature of water and ambient air
—air to water ratio
—contact time between air and water
—available area for transfer of radon
  from water.
  PTA provides the most efficient
transfer of radon from water to air,  with
the ability to remove greater than 99
percent of radon from water. A supply
which requires a smaller reduction  of
radon, for example 50 percent, could opt
to install PTA and treat 50 percent of its
source water and subsequently blend
the treated with raw water, or it may
design a shorter packed tower to
achieve compliance with the MCL.
Other advantages of PTA include:
removal of hydrogen sulfide, carbon
dioxide, and VOCs, and oxidation of
iron and manganese. No pilot or full-
scale packed columns have yet been
constructed for removal or radon at
large municipal supplies. However, field
tests have been performed by EPA,
documented by Khmer et al. (1989;  1990),
which verify the efficacy of aeration for
radon removal. Since radon acts
 similarly to some highly volatile organic
 compounds, and packed columns have
been shown to be the most efficient form
 of aeration for VOC removal, PTA is
 appropriate as BAT for radon.
   Diffused aeration, which may remove
 up to 97 percent of radon in water
 possesses the following advantages: the
 potential for modifying existing basins
 or storage vessels for diffused aerator

-------
33086
Federal  Register / Vol. 56, No.  138 / Thursday, July 18, 1991  /  Proposed Rules
 installation; no packing media costs; and
 reduced pumping requirements. The
 Radon Technology and Cost document
 (EPA, 1987b) summarizes the case study
 of a full-scale diffused aeration plant in
 Belstone, England which was built to
 remove influent radon, and provided
 long-term removal efficiency of 97
 percent. The disadvantages of diffused
 aeration include the requirement for
 increased contact time, the
 impracticality of large air-to-water
 ratios because of air pressure drops, and
 overall less efficient mass transfer of
 radon from water. The level of contact
 between air and water achievable in a
 packed tower aerator is difficult to
 obtain in a diffused air system. The
 above-referenced Belstone, England
 plant treated 2.5 mgd water, with 2,800
 air diffusers, each designed to supply a
 maximum of 0.8 cubic feet per minute,
 and with a  24-minute retention time,
 achieved an air-to-water ratio of 8 to 1
 for 97 percent radon reduction. In a field
 test of a diffused bubble aeration
 system, Kinner et al. (1990) report that
 removals of 90 to 99 percent were
 achieved at air-to-water ratios of 5 to 1
 and 15 to 1.
  Disadvantages which have been
 identified by Kinner et al. (1989; 1990)
 are the potential for bacteria fouling and
 iron and manganese precipitation,  which
 may clog or otherwise impede
 operations  at an aeration facility (PTA
 or diffused  bubble type). These
 secondary effects may occur, however
 they would be highly dependent on
 source water quality conditions.
  Spray aerators direct water upward,
 vertically, or at an angle, in such a
 manner that the water is broken into
 small droplets (by fixed nozzles on a
 pipe grid) which provide large surface
 areas for radon migration out of the
 water to the air. Most of the advantages
 cited above for diffused aeration also
 apply to spray aeration. Disadvantages
 include the need for a large operating
 area and operating problems during cold
 weather months when the temperature
 is below the freezing point. Costs
 associated  with this option (for all sizes
 of water treatment plants) have not
 been developed by EPA.
  EPA has  evaluated other, less
 technology-intensive ("low-tech"),
                        options which may be suitable for small
                        water systems, and which may cost less
                        than the above options to install and
                        operate (Kinner et al., 1989; 1990). These
                        options include: Open air storage, free
                        fall with nozzle-type aerator, bubble
                        aerators, and slat tray aerators. With 24
                        to 48 hours detention, open air storage
                        may reduce radon levels by 30 to 50
                        percent; a free fall of 2 feet with simple
                        nozzle attachment was found to reduce
                        radon by 65 to 75 percent with 8 hrs
                        detention time; and a two foot free fall
                        into a tank equipped with garden hose
                        (punctured) bubble aerators, supplied by
                        a laboratory air pump, yielded 85 to 90
                        percent radon reduction with 8 to 12
                        hour detention time. The above-
                        referenced study concluded that very
                        effective radon reduction can be
                        achieved by simple aeration
                        technologies that may be easily applied
                        in small communities.
                         EPA has developed cost estimates for
                        the above mentioned alternative low-
                        tech water treatment methods, suitable
                        for small systems that may need partial
                        radon removal to meet the drinking
                        water MCL. Cost estimates for small
                        systems installing 9-hour storage/
                        detention, diffused aeration, spray
                        aeration, slat tray aeration, and PTA are
                        presented in an addendum to the EPA
                        Radon Technology and Cost Document
                        (EPA, 1988e).
                         b. Secondary effects of aeration:
                        Estimate of risks from PTA emissions of
                        radon. Since this notice contains a
                        proposal to reduce radon concentrations
                        in drinking water by setting an MCL,
                        and the EPA is proposing aeration as
                        BAT for meeting the MCL, the Agency
                        undertook an evaluation of risks
                        associated with potential air emissions
                        of radon from water treatment facilities
                        due to aeration of drinking water. It is
                        logical to assume that radon, removed
                        from drinking water and released to the
                        atmosphere, could result in some
                        degradation of air quality and possibly
                        pose some incremental health risk  to the
                        general population. However, the risks
                        due to potential human exposure to PTA
                        emissions appear very small in
                        comparison to the risks associated with
                        radon in drinking water (EPA, 1988c;
                        1989b).
  In one evaluation of risks associated
with potential radon emissions from
aeration of drinking water (EPA, 1988c),
EPA used radon data from 20 drinking
water systems in the U.S. which,
according to the Nationwide Radon
Survey (EPA, 1985a), contained the
highest levels of radon in drinking water
and affected the largest populations
and/or drinking water communities.
EPA estimated the potential annual
emissions (in pCi radon/yr), from PTA
treatment facilities, assuming 100
percent radon removal, and these were
applied to appropriate dispersion
models. Estimates were made for the
following parameters: Air dispersion of
radioactive emissions, including radon
and progeny isotopes of radon decay;
concentrations in the air and on the
ground; amounts of radionuclides taken
into the body via inhalation of air and
ingestion of meat, milk, and fresh
vegetables, dose rates to organs and
estimates of fatal cancers to exposed
persons within a 50 kilometer radius of
the water treatment facilities. Estimates
of individual risk and numbers of annual
cancer cases were completed for each of
the 20 water systems, as well as a crude
estimate of U.S. risks (total national
risks) based on a projection of results
obtained for the 20 water systems.
These estimates were based on
exposure analyses on a limited number
of model plants, located in urban,
suburban and rural settings, which were
scaled to evaluate a number of facilities.
(A similar approach has been used by
the Agency in assessing risks associated
with dispersion of coal and oil
combustion products.)
  The risk assessment results for the 20
systems indicate the following: A
highest maximum lifetime risk to
individuals at one system of 4 X 10~s,
with a maximum incidence at the same
location of 0.0060 cancer cases per year;
an estimate of annual cancer cases for
all 20 systems of 0.016/yr; and a crude
U.S. estimate  of 0.4 cancer cases/year
due to air emissions at all drinking
water supplies to meet a hypothetical
MCL of 200 pCi/1. The results of the risk
assessment for potential radon
emissions from drinking water facilities
are given in Table 10.
  TABLE 10.—ESTIMATES OF RISKS AT 20 SITES DUE TO POTENTIAL RADON EMISSIONS FROM PTA UNITS AND CRUDE ESTIMATE OF
                                                     U.S. RISK1
Scenario
20 Facilities:
1 	
Concentra-
tion in water
(pCi Rn/1)
1.839
Emission
from PTA
(Ci Rn/yr)
2.79
Maximum
life,
individual
risk
6 X 10-'
Cancer
cases/year
.0003

-------
               Federal Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed Rules          33087
 TABLE 10.—ESTIMATES OF RISKS AT 20 SITES DUE TO POTENTIAL RADON EMISSIONS FROM PTA UNITS AND CRUDE ESTIMATE OF
                                               U.S. RISK1—Continued
Scenario
2 	
o 	


g 	
•j

Q 	
•JQ 	
^•j 	 : 	



•j5 	
•jg 	
17
18 	
•jg 	

AH 20 facilities 	
All U S drinking water plants 	

Concentra-
tion in water
(pCi Rn/1)
5,003
2,175
1,890
1,310
1,329
4,085
10,640
3,083
3,270
2,565
4,092
16,135
3,882
1,244
2,437
996
7,890
9,195
7,500

<200
<500
<1,000
Emission
from PTA
(Ci Rn/yr)
6.22
2.85
20.89
1.81
91.80
2.26
1.18
0.55
9.04
3.54
1.75
2.23
0.27
1.03
1.35
8.94
0.87
1.02
1.04
161
4,200
2,000
900
Maximum
life,
individual
risk
1 x 10-6
6 x 10-
1 X 10-
9 X 10-
2 X 10-
5 X 10-
2 X 10-
1 X 10-
4 X 10-
1 X 10-
3 X 10-
4 X 10-
1 X 10-
5X 10-
7 X 10-
2 X 10-
4 X 10-"1
5 X 10-
5 X 10-
4 X 10-

—
Cancer
cases/year
.0008
.0004
.0004
.0000
.0040
.0001
.0000
.0000
.0060
.0023
.0001
.0001
.0000
.0001
.0000
.0008
.0000
.0000
.0000
.016
.4
.2
.09
   1 Estimates of risk assessed using AIRDOSE-EPA, RADRISK and DARTAB air dispersion and lifetime risk computer codes (EPA, I988c).
  Numerous assumptions were applied
in conducting the above analysis,
including the following:
  • PTA treatment applied, removing
100 percent of radon;
  • typical (not site-specific)
meteorology is used at the model plants,
and flat terrain is assumed;
  • 1980 census data were used, with
people located in "population centroids"
representative of census districts;
  • 70-year residency at same location,
and exposure to air and radon emissions
persists throughout 70 yrs.;
  • additive impact of exposure to
emissions from more than one plant
emitting radon was not accounted for.
  To further investigate potential health
risks due to PTA radon emissions, EPA
used the MINEDOSE model developed
to determine compliance of radon point
sources regulated under EPA's
NESHAPS standards [EPA, 1989b). In
that study, worst case scenarios
representing systems with radon levels
ranging from 1,330 to 110,000 pCi/1 were
identified and their potential emissions
modeled. These systems represent what
may be the greatest potential among
PWSs to increase risks via air
emissions. Only systems with very high
flow rates posed any potential for
increasing ambient air radon exposure
appreciably. The one modeling run'that
did indicate a potential problem
assumed that all radon emissions came
from a single point source (i.e, the entire
production flow was treated through a
single aeration tower). However, the
community modelled relies on numerous
widely dispersed wells for its total
water supply, and aeration treatment
could be installed at individual wells,
thereby dispersing the emissions to the
ambient outdoor air. This modeling also
found that systems having very high
radon levels, (100,000 pCi/1) but lower
flow rates, did not appreciably increase
ambient air radon levels and risks.
  Given the uncertainties in calculating
such risk estimates, EPA views the
above estimates as "order of magnitude
estimates." Nevertheless, it is apparent
that the risks to the U.S. population, and
to the individual drinking water
communities, due to potentially aerated
radon from source water are much
smaller (in most cases 2 to 4 orders of
magnitude smaller) than the risks due to
radon in water if no treatment were
applied.
  EPA is aware that some states allow
no emissions from PTA regardless of
downwind risks. EPA has reviewed the
few available data on removal of radon
from air by carbon. Based on these data,
EPA believes air phase  removal of
radon by GAG may not be feasible.
Systems trying to meet local air
emissions requirements may need to
rely on GAG in the water phase.
  c. Granular activated carbon. Pilot
plant studies have shown that granular
activated carbon (GAG) is capable of
removing radon in drinking water at
efficiencies of 90 to 99 percent (Khmer et
al., 1989). The efficiency of removal is
dependent upon radon concentration,
the mass of carbon in the GAG column,
empty bed contact time (EBCT) and
contactor configuration (i.e., upflow or
downflow). The pilot studies have
shown radon to require a longer EBCT
than other adsorbable (e.g., organic)
materials. Thus, to achieve a 90 percent
removal efficiency with a radon influent
concentration of 10,000 pCi/1, an EBCT
of approximately 70 minutes may be
required. The need for such a lengthy
EBCT means that GAG may not be
practical for large municipal treatment
systems (EPA, 1987b) and it is therefore
not considered BAT.
  Another disadvantage associated with
the use of carbon for radon removal is
the buildup of radiation inside  and
surrounding the GAG contactor. The
radionuclides that may build up on the
GAG media are the progeny of radon,
specifically the radioactive isotopes of
lead, polonium and bismuth. The short-
lived radon progeny include Pb-214 and
Bi-214. Long-lived radon progeny include
Pb-210, Bi-210, and Po-210. The level of
gamma radiation surrounding the GAG
vessel depends on the amount  of radon
removed; gamma intensity drops sharply
with increased distance from the GAG
vessel. Due to the buildup of radon
daughter products, such as lead-210, a
beta particle emitter, the GAG  unit can
become a source of low-level radiation,
and may present a disposal problem as
well. Studies have shown that the
radiation level is usually less than 1.0
mR/hr. at a distance of three (3) feet
from the GAG tank surface (Kinner et

-------
33088
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991 / Proposed Rules
al., 1989). EPA's guidelines for
radioactive waste disposal (EPA, 1990a]
provide guidance on the disposal of
GAG waste containing naturally
occurring radionuclides, and appropriate
occupational guidance.
  The estimated cost for small GAG
water treatment systems for 80 percent
removal of radon ranges from $6.60/
1,000 gallons of water serving 25 to 100
people to $1.40/1,000 gallons of water
serving 3,000 to 10,000 persons,
exclusive of the cost to dispose of spent
carbon. Due to the problems identified
above, i.e., of radiation build-up, waste
disposal, and contact time, the Agency
has judged that GAG cannot be
designated as BAT for radon removal
(EPA, 1987b; 1988e; 1991i).
  3. Uranium. The Agency proposes
that, of the technologies capable of
removing uranium from source water,
coagulation/filtration, ion exchange,
lime softening and reverse osmosis
fulfill the requirements of the SDWA as
BAT for uranium removal. These
technologies have demonstrated
effective uranium removal, are currently
available, have been installed in public
water supplies, and are compatible with
other water treatment processes in
different regions. The full range of
technical capabilities for each of these
proposed BATs is discussed in the EPA
technologies and costs document for
uranium (EPA, 1985b), and summarized
below.
  a. Coagulation/filtration.  Laboratory
and pilot plant studies have shown that
pH and coagulant dosage significantly
impact uranium removal efficiency in
water treatment. Iron and aluminum
based coagulants are generally more
effective in aiding the removal process
at pH values near 6 and 1. Removal by
coagulation appears to be low at pH 8
due to stability and charge
characteristics of uranyl species in
solution. In one study, removal
efficiencies of greater than 80 percent
were reported (Sorg, 1988) in tests using
20 mg/1 doses of ferric sulfate, ferrous
sulfate, or aluminum sulfate coagulants.
Influent uranium levels were about 83
ju.g/1 in that study. Coagulation/filtration
has been demonstrated to achieve
removal efficiencies as high as 95
percent when using aluminum sulfate
dosed at 10 mg/1 or more, at pH 10 (Sorg,
1988).
  Coagulation/filtration as a new
process designed specifically to remove
uranium may not be cost effective,
particularly for smaller utilities.
However, where the reduction of
turbidity in the source water is also a
concern, this method can be very
effective.
                         Estimated costs for an existing
                        coagulation/filtration facility to modify
                        treatment for 60% removal of uranium
                        from ground water sources range from
                        $4.40/1,000 gallons of water for systems
                        serving a population of 25-100 persons,
                        to $0.02/1,000 gallons of water for
                        systems serving over 1,000,000 persons.
                         b. Ion exchange. Anion exchange
                        systems for the removal of uranium and
                        other soluble ions have demonstrated .
                        uranium removal efficiencies of between
                        65 and 99 percent. Ion exchange devices
                        are available for most applications. The
                        estimated costs for removal of uranium
                        from ground water by ion exchange
                        range from $4.10/1,000 gallons of water
                        treated for systems serving 25-100
                        persons, to $1.00/1,000 gallons of water
                        treated for systems serving more than
                        1,000,000 persons. Disposal of
                        concentrated waste brine must also be
                        considered, as discussed above.
                         c. Lime softening. Lime softening is
                        capable of achieving removal
                        efficiencies for uranium of up to 99
                        percent. At optimum pH levels of 10.6 to
                        11.5 removal efficiencies of 85 to 99
                        percent can be achieved. Best results
                        can be achieved by increasing the
                        dosage  of lime to approximately 250 mg/
                        L and maintaining the pH above 11.
                        Lower dosages of Ca(OH)2, 50 to 100
                        mg/1, have achieved 85 percent uranium
                        removal. This treatment should be given
                        serious  consideration if removal of
                        hardness from source water is also a
                        desired objective. It may not be cost
                        effective for a system to build a new
                        lime softening treatment facility
                        specifically to remove uranium. The
                        estimated cost to modify an existing
                        lime softening treatment facility to
                        remove uranium from ground water
                        ranges from $4.20/1,000 gallons of water
                        serving 25-100 persons to $0.03/1,000
                        gallons  of water serving more than
                        1,000,000 persons.
                         d. Reverse osmosis. Reverse osmosis
                        (RO) membranes are capable of
                        removing uranium and many other
                        contaminants in source water, at high
                        efficiencies. RO has been used primarily
                        for removing total dissolved solids
                        (TDS) from water in the treatment of
                        brackish and sea waters for
                        desalinization purposes. At reduced
                        pressures RO is adaptable to fresh
                        water sources. Using cellulose acetate
                        membranes, at 250 psi pressure, RO has
                        successfully achieved 98 to 99.4 percent
                        removal efficiencies. However, RO
                        performance is adversely affected by the
                        presence of turbidity, iron, manganese,
                        silica or scale producing constituents in
                        source water. If pretreatment is not
                        already in place to remove these
                        constituents, the cost to install the
pretreatment facility would be an
important factor.
  The RO system is adaptable to all size
systems with costs ranging from $6.20/
1,000 gallons for systems serving 25-100
persons to $0.89/1,000 gallons for
systems serving over 1,000,000 persons.
If reducing TDS is also a goal of the
treatment process then reverse osmosis
is a very cost effective solution for the
removal of uranium from source waters.
Disposal of waste brine, the reject flow
representing 20 to 50 percent of the feed
water, and the quantity of available feed
(source) water to accommodate this
technology, would require consideration
by a water system in its initial
evaluation of alternative technologies
for radium removal.
  4. Beta and photon emitters. The
Agency proposes that of the
technologies considered to remove beta
particle emitters from drinking water,
ion exchange and reverse osmosis
would fulfill the requirements of the
SDWA as BAT for gross beta particle
removal. The subject radionuclides
originate from the nuclear fuel cycle,
defense related industrial activities,
institutions such as hospitals, research
foundations and universities,
commercial/industrial users of
radioisotopes,  and atmospheric or
surface detonation of nuclear devices.
Some beta-emitting radionuclides
originating from such sources have
occurred in drinking water sources and
have been partially removed by drinking
water treatment processes.
  Levels of gross beta above the
maximum contaminant  level are likely
to occur only in transient situations
following a contaminating event. The
following technologies may be effective
in lowering the contaminant level below
the MCL value. The full range of
technical capability of the proposed
BATs is summarized in the EPA
document "Technologies and Costs for
the Removal of Man-Made
Radionuclides from Potable Water
Supplies" (EPA, 1986b). The
technologies listed are available and
compatible with other water treatments
in all regions of the United States.
  a. Ion exchange. Ion exchange has
been successfully employed by the
nuclear power industry in treating liquid
radioactive wastes as well as chemical,
laboratory, and laundry wastes
containing various ionic species. Cation
exchange resins have exhibited a 95 to
99 percent removal efficiency for low
level and trace amounts of the following
contaminants: barium-137, barium-140,
cadmium-115, cesium-137, lanthanum-
140, scandium-46, and strontium-89.
Anion exchange resins have exhibited a

-------
                Federal Register  /  Vol. 56,  No. 138 / Thursday,  July 18, 1991  /  Proposed Rules
                                                                                                           ,33089
94 to 99 percent removal efficiency for
the following contaminants: niobium-95,
tungsten-185, zirconium-95, scandium-46,
and yttrium-91. Mixed bed ion exchange
may effectively remove between 90 and
99.9 percent of all contaminants listed
above. Therefore ion exchange
technology is proposed as BAT for beta
and photon emitters. Disposal of waste
brine may pose difficulty due to the high
concentration of radionuclides in the
brine, the availability of disposal
options for the liquid wastes, and State
or Federal limitations which may
prevail.
  The cost for removal of beta-emitting
radionuclides utilizing ion exchange
would be highly dependent upon type
and amount of contamination. The cost
supplement (EPA, 1987c) to the above
cited Technologies and Cost document
contains estimated cost for removal of
beta emitters from public water systems
using two-bed ion exchange system  (i.e.,
cationic and anionic).
  b. Reverse osmosis. Reverse osmosis
(RO) membranes can effectively remove
more than 99 percent of radioactive
contaminants such as strontium, cesium,
and iodine from water. Pilot studies
have demonstrated removal efficiencies
of 90 to above 99 percent of dissolved
iodine-131, strontium-89, and cesium-
134. The cost of removing man-made
radionuclides from source water
utilizing RO may be  similar to the costs
cited in Tables 7, 8 and 9 for removal of
uranium from drinking water. However,
cost would be highly dependent on type
and degree of contamination.
  RO performance is adversely affected
by the presence of turbidity, iron,
manganese, silica, or scale-producing
constituents in the source water. If the
pretreatment is not already in place to
remove these constituents, the cost to
install the pretreatment facilities may be
an important factor. Disposal of waste
brine may be problematic due to the
high concentration of radionuclides  in
the brine, or due to local requirements or
regulations affecting discharge.
  The cost supplement (EPA, 1987c) for
the Technology and Cost document  cited
above contains estimated cost for
removal of beta emitting radionuclides
from public water systems using reverse
osmosis technology.
   c. Coagulation/filtration. Some beta-
emitting radionuclides which exist as
suspended material  in water may be
removed by coagulation/filtration. In
laboratory studies involving many
soluble radionuclides, it was reported
that coagulation employing aluminum
sulfate, ferric chloride and/or ferrous
sulfate was more effective for removal
of soluble cations of valences 3,4 or 5
which include: niobium-95, cerium-141,
phosphorus-32, zirconium-95, cobalt-58
and -60, ruthenium-103, and sulfur-35.
  Full-scale studies in municipal
filtration plants downstream from
nuclear reactor sites have indicated
removal of chromium-51, scandium-46,
arsenic-76 and seven other nuclides at
efficiencies of 28 to 87 percent, using
alum as the coagulant. Activated silica
or clay can be added when needed to
enhance flocculation, coagulation and
precipitation. Ninety percent removal of
strontium requires iron coagulant
dosages greater than 500 mg/1 at a pH of
11. Efficiencies of removal of specific
radionuclides by the coagulation
process can range from 0 to 99 percent.
  Due to the variability cited above in
the removal efficiencies,  and because of
the lack of information on removal of
many beta emitting radionuclides, EPA
proposes that coagulation/filtration
does not meet the requirements to be
proposed as a BAT for beta emitters.
  d. Lime softening. Lime softening with
soda-ash addition can remove
approximately 90 percent of strontium
and other radiological contaminants
present in source water. To achieve this
percent removal the sodium carbonate
concentration should be  three times the
equivalent permanent calcium hardness.
Using 68 to 205 mg/1 of lime and 68 to
154 mg/1 of soda ash, 90 percent removal
of the following radionuclides may be
achievable: barium-140, cadmium-115,
zirconium-95, lanthanum-140, scandium-
46, niobium-95, strontium-89, and
yttrium-91.
  Due to the lack of information on
removal of many of the beta emitting
radionuclides addressed by this
proposed regulation, EPA proposes that
lime softening not be designated as a
BAT for beta emitters.
  5. Alpha emitting radionuclides.  In
order to determine BAT for the removal
of alpha-emitting radionuclides, the
Agency required information regarding
the identity and treatability of those
radionuclides which occur or may occur
in potable water supplies (other than
radium, radon and uranium). Alpha
emitters identified above that may occur
in water systems include polonium-210
(Po-210), thorium 228,230, and 232 (Th-
228, 230, 232)  and at very low levels,
plutonium 238, 239 and 240 (Pu-238, 239,
240).
  EPA summarized available treatment
data from field studies and from public
water systems in the document
"Technologies and Costs for the
Removal of Alpha Emitters from Potable
Water Supplies (EPA, 1991k). EPA has
found no treatability information on the
radionuclide thorium, a fact likely due to
the insolubility of and the difficulties
associated with measuring this
contaminant. Relatively little
information was available on
treatability of plutonium in water
supplies. However, plutonium appears
to be removed by coagulation and
filtration technology, particularly where
the contaminant is associated with
turbidity in surface waters or with
colloidal hydroxide particulates. Surface
water contaminated with trace amounts
plutonium 239 and 240, such as Lake
Michigan (fallout derived plutonium)
and the Savannah River (downstream
from a nuclear power plant), have been
treated for industrial and municipal use
with coagulation/filtration technology.
Raw influent waters contained 1 to 2
femtocuries of plutonium per liter of
water. Removals of plutonium at these
facilities have been recorded in the
range of 25 to 96 percent. The addition
of carbonates through lime and soda ash
appears to contribute to the  coagulation
and removal of colloidal plutonium from
natural surface waters. Plutonium
removal efficiency was found to
increase with higher plutonium
concentrations. Nonetheless, in regard
to the application of coagulation/
filtration for removal  of plutonium from
water, EPA finds that the wide range  of
efficiencies that have been documented
preclude its  designation as a BAT for
alpha emitters.
  EPA has undertaken to identify BATS
that effectively remove polonium-210
from drinking water to achieve
compliance with the gross alpha
standard. The results of treatability
studies conducted in Maine  on well
water containing high levels of
polonium-210 are discussed  in detail in
the Cost and Technologies Document
cited above. In the Maine field study
conducted over 2 months during 1990-
1991, anion exchange,  reverse osmosis,
and granular activated carbon (GAG)
were tested. These tests showed (after
correction of some clogging  and fouling
of the ion exchange and carbon units)
reverse osmosis with the highest
removal rates (98-99%), and GAG (69-
93%) and ion exchange (52-83%)
showing somewhat lower removal rates.
Water pH may affect polonium removal
rates for GAG and ion exchange, but
this has not  been documented.
  The Maine treatability studies and  the
Technologies and Cost document form
the basis for a decision by EPA to
propose a BAT for removal of alpha
emitters. RO has provided the highest
removal efficiencies and is proposed as
BAT for alpha emitter removal.

C. Waste Treatment and Disposal

  The  treatment and disposal of waste
by-products generated by the treatment

-------
33090
Federal Register / Vol.  56,  No. 138  / Thursday, July 18, 1991 / Proposed Rules
processes increases overall water
treatment costs, especially for small
systems. However, in establishing BAT,
EPA identifies the treatment and
disposal technologies that are
reasonably available for large
metropolitan regional drinking water
systems (i.e., systems which service
                        50,000 to 75,000 persons]. Disposal of
                        wastes from treatment for radionuclides
                        does not significantly increase the total
                        treatment costs for large systems.
                        Several waste disposal techniques and
                        estimates of associated costs are
                        identified in Table 11. Technologies and
                        costs related to the disposal of the
granular activated carbon that may in
some cases be used for radon removal
have not been determined by EPA. GAG
is not a BAT for radon removal for
reasons outlined in section B, part 2(c),
above.
           TABLE 11.—RANGE OF BRINE AND SLUDGE DISPOSAL COSTS IN REMOVAL OF RADIONUCLIDE CONTAMINANTS 1
                                             [Cents/1,000 Gallons of Water Treated]
Treatment process
Brine Disposal:
Ion Exchange 	
Reverse osmosis 	

Treatment process
Sludge Disposal:
Coagulation/Filtration 	
Lime Softening 	

Direct
discharge
l2)
2-95

Discharge
to sewer
1-190
(v\

Discharge to
sewer
4-110
1 0-230

Non-mechanical
dewatering and
land disposal (3)
65-360
30-600

Evaporation
pond/land
on-9^n
(Z\

Dewatering and
land disposal (5)
75-2 800
/4\

Chemical
precipitation

/4\





   Notes:
   1 From "Technologies and Costs for the Treatment and Disposal of Waste Byproducts from Water Treatments for the Removal of Inorganic and Radioactive
Continants (EPA,1986d). Cost ranges represent disposal costs for very large to very small water systems.
   2 Data not available.
   3 Non-mechanical dewatering alternatives for sludges include sand drying beds and dewatering lagoons.
   * Disposal option too expensive.
   6 Mechanical dewatering may include utilization of pressure filtration.
  Liquid wastes, or brines, are
generated by ion exchange, reverse
osmosis, and activated alumina. The
most economical disposal method for
concentrated brines is discharge to a
sanitary sewer, and for reverse osmosis,
direct discharge of the concentrated
waste stream to a receiving body of
water, if these methods are acceptable
to applicable regulatory agencies and
meet Clean Water Act requirements for
direct and indirect discharges to surface
water. Underground injection may be an
option, subject to the requirements of
the Underground Injection Control
Program. Other possible though more
expensive alternatives include
evaporation pond dewatering followed
by land disposal, and chemical
precipitation followed by non-
mechanical drying and land disposal.
Sludges are generated by coagulation/
filtration, greensand filtration, and lime
softening. The most economical disposal
method for sludges is discharge to a
sanitary sewer. Again, this method may
be restricted by state or local
requirements and pre-treatment
requirements under the Clean Water Act
(see generally 40 CFR part 403]. An
alternative option may be non-
mechanical drying (lagoons or drying
beds] followed by land disposal.
Mechanical methods tend to be higher in
cost, though technically feasible, for all
sludges.
                          At the present time there are no
                        federal regulations which specifically
                        address the disposal of water treatment
                        wastes containing radionuclides.
                        However, the selection of waste by-
                        product disposal alternatives may be
                        determined by federal, state, and local
                        regulatory constraints and site specific
                        conditions. Regulatory constraints may
                        include industrial pretreatment
                        requirements for sanitary sewer
                        discharges (including requirements
                        applicable to sewage sludge use and
                        disposal under section 405 of the Clean
                        Water Act], requirements under the
                        Underground Injection Control (UIC)
                        program, RCRA requirements for
                        hazardous waste disposal and
                        protection of groundwater, and effluent
                        limitations and water-quality based
                        limits for the discharge of some
                        contaminants into local receiving waters
                        (groundwaters and surface waters]
                        under the NPDES program. Site-specific
                        conditions which influence waste
                        management include the  availability of
                        sewage disposal, location of disposal
                        sites, climatic factors, cost of land, and
                        other local or regional factors including
                        available manpower and infrastructure
                        characteristics.
                          EPA's report entitled "Suggested
                        Guidelines for the Disposal of Naturally
                        Occurring Radionuclides Generated by
                        Drinking Water Treatment Plants,"
                        (EPA, 1990a] outlines the Agency's
understanding of the technical issues
and the existing regulatory framework
that may be relevant to systems which
remove naturally-occurring radioactive
substances from drinking water
supplies. In this report, EPA
recommends types of treatment and
disposal options and institutional
controls which would be pertinent for
solid and liquid wastes containing
radioactive contaminants, at various
ranges of concentration. The report also
makes recommendations regarding
radiation in the water treatment plant
and protection of workers at the plant
and during waste disposal operations.
EPA solicits public comment on its
waste disposal guidance, and waste
disposal issues in general.
  EPA and others have studied the
treatment technologies available for the
removal of radionuclides from drinking
water and characterized some of the
waste residuals of treatment. These
studies were  conducted on source
waters naturally high in radioactivity
and produced data which may be useful
for the purpose of characterizing  solid
and liquid wastes from the treatment of
drinking water and for comparison with
the EPA Suggested Guidelines cited
above. Table 12 summarizes  some data
that EPA has gathered on water
treatment wastes containing radium and
uranium.

-------
                 Federal Register / Vol.  56, No.  138 /  Thursday, July  18,  1991  /  Proposed Rules
                                                                                               33091
  Table 13 outlines the options for
sludge disposal suggested in the EPA
guidelines. Notwithstanding these
suggested guidelines, solid wastes and
liquid wastes generated by drinking
water treatment plants should be
disposed of in compliance with Federal,
State and local requirements, State-
                     adopted criteria of 40 CFR part 257,
                     which contains RCRA groundwater
                     protection criteria, and municipal solid
                     waste landfill regulations under 40 CFR
                     part 258.
                       Similarly, from the same EPA report
                     cited  above, EPA guidelines were
developed for disposal of liquid wastes,
or brines, which result from the
treatment of drinking water containing
radionuclides. These are outlined in
Table 14. EPA solicits public comment
on the waste disposal guidance and
estimated disposal costs.
              TABLE 12.—SUMMARY OF WATER TREATMENT DATA ON WASTES CONTAINING NATURAL RADIONUCLIDES
Treatment wastes







Ion exchange brino/rogon

	

Rn-226 	




References
(EPA, 1985d)


(EPA, 1987f)

(EPA, 1985d)

(Schliekelman, 1976) 	

(Schliekelman, 1976 and EPA, 1985d) 	
(Sorg et al., 1 980) 	

(EPA, 1985d and EPA , 1987a)



Concentration range

1-22pCi/g(dry).
2-12 pCi/g (dry).

57-171 pCi/g (dry).

6-50pCi/l.
11 0-530 pCi/l.
3,500 pCi/l.
upto610 pCi/l.

7-38 pCi/l.

21-106pCi/l.
5.7-83 pCi/l.

TABLE  13.—DISPOSAL  GUIDELINES  FOR
   RADIOACTIVE SOLID  WASTES  RESULT-
   ING FROM  DRINKING  WATER  TREAT-
   MENT PROCESSES l
  Waato characteristics
 I. Solids/sludges
  containing less than 3
  pd/g of radium or
  laad-210, or less than
  30 pd/g uranium.
 II. Solids/sludges
  containing 3 to so
  pd/g ol radium or
  load-210, or 30 to 500
  pd/g uranium.
 III. Solids/sludges
   containing 50 to 2,000
   pd/g of radium or
   load-210, or 500 to
   2.000 pd/g of
   uranium.
    Disposal option
Sludge should be
  dewatered, and mixed
  in landfill.
Sludge should be
  dewatered, and
  disposed of within a
  stabilized landfill to
  Isolate and to avoid
  Inappropriate usage of
  the site.
Case-by-case
  determination, to
  Include consideration
  of standards for
  uranium mill tailings
  (40 CFR 192), NARM
  disposal, and long-
  term institutional
  control of disposal
  sites. RCRA
  hazardous waste units
  should also be
  considered. NRC
  provisions may apply.
TABLE 13.— DISPOSAL GUIDELINES FOR
RADIOACTIVE SOLID WASTES RESULT-
ING FROM DRINKING WATER TREAT-
MENT PROCESSES J— Continued
Waste characteristics
IV. Solids/sludges
containing more than
2,000 pCi/g of natural
radioactivity.







Disposal option
Should be disposed of in
a low-level radioactive
waste disposal facility
operated under the
provisions of the
Atomic Energy Act, as
amended, or at a
State or EPA-permitted
facility for NARM
disposal. Uranium
recovery may be
possible. NRC
provisions may apply.
Dept. of
Transportation
regulations would
apply.

Note: Water treatment, facilities should keep
records of the amount and composition of radioac-
tive wastes they generate, and the manner and
location of disposal.
1 From EPA Suggested Guidelines (EPA, 1990a).

TABLE 14.— DISPOSAL GUIDELINES FOR
RADIOACTIVE LIQUID WASTES GENER-
ATED BY WATER TREATMENT PLANTS *
Disposal option
A. Disposal into surface
water.

Requirements (Federal
and other)
(1) Federal, State and
local discharge limits
and NPDES permit
requirements apply.


TABLE 14.— DISPOSAL GUIDELINES FOR
RADIOACTIVE LIQUID WASTES GENER-
ATED BY WATER TREATMENT PLANTS x—
Continued
Disposal option

B. Discharge into
sanitary sewers (if Ra-
226 is less than 400
pC\/\, Ra-228 less
than 800 pCi/l, total
uranium less than 1
(j.Ci/1, and yearly total
discharge less than 1
curie).



C. Disposal of
radioactive wastes
through injection wells
(under conditions
consistent with 40
CFR 144
classifications of
wells). Shallow
injection banned.
D. Evaporation,
precipitation, drying, or
other treatment.



Requirements (Federal
and other)

(1) State limits on
discharge of
hazardous or
radioactive wastes.
(2) Limits on discharge
of radium and uranium
into sanitary sewers —
per NRC standards for
discharge by licensees
Ou CFR 20, part 303).
(3) Federal, State, and
local pretreatment
requirements.
(1) Authorization of any
injection of liquid
wastes under the
Underground Injection
Control (UIC) program
regulations in 40 CFR
144.6(a)(2), and
144.12(c).

(1) Residual solids
should be disposed
per solid waste
regulations and per
EPA guidelines for
water treatment solid
was es
'From EPA Suggested Guidelines (EPA, 1990a).
D. Analytic Methods
The SDWA directs EPA to" set an MCL
for contaminants for which there are
MCLGs, "if, in the judgement of the
Administrator, it is economically and

-------
33092
Federal Register /  Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
technologically feasible to ascertain the
level of such contaminants in water in
public water systems." (SDWA section
1401[l][C][iij). NPDWRs are also to
"contain[s] criteria and procedures to
assure a supply of drinking water which
dependably complies with such
maximum contaminant levels; including
quality control and testing procedures to
insure compliance with such levels."
(SDWA section 1401[1][D]). The analytic
methods described and evaluated here
are the testing procedures EPA
identified to insure compliance with the
MCLs. EPA evaluated the availability,
cost, and the performance of these
analytical techniques, as well as the
ability of laboratories to use  these
methods to measure radionuclide
contaminants consistently and
accurately in a compliance monitoring
setting.
  The reliability of analytic methods at
the maximum contaminant level is
critical to implementing and enforcing
                         the MCLs. Therefore, each analytical
                         method was evaluated for accuracy or
                         recovery (lack of bias) and precision
                         (good reproducibility over the range of
                         MCLs considered). The primary purpose
                         of this evaluation is to determine:
                           •  Whether analytical methods are
                         available to measure the regulated
                         radionuclide contaminants in drinking
                         water;
                           •  The ability of recently developed
                         analytical method(s) to measure
                         radionuclide contaminants in drinking
                         water;
                           •  Reasonable expectations of
                         technical performance by analytical
                         laboratories conducting routine analysis
                         at or near the MCL levels; and
                           •  Analytical costs.
                           The selection of analytical methods
                         for compliance with these regulations
                         includes consideration of the following
                         factors:
  (a) Reliability (i.e., precision/accuracy
of the analytical results over a range of
concentrations, including the MCL);
  (b) Specificity in the presence of
interferences;
  (c) Availability of adequate equipment
and trained personnel to implement a
national compliance monitoring program-
(i.e., laboratory availability);
  (d) Rapidity of analysis to permit
routine use; and
  (e) Cost of analysis to water supply"
systems.
  1. Description of analytic methods.
Analytical methods exist to measure
each radionuclide contaminant covered
by today's proposed regulations. Table
15 lists these analytical methods. EPA
believes these methods are technically
sound, economical,  and generally
available for radionuclide monitoring,
and is proposing their use for monitoring
to determine compliance with the MCLs.
                           TABLE 15.—PROPOSED METHODOLOGY FOR RADIONUCLIDE CONTAMINANTS
Contaminant
Naturally Occurring:
Gross alpha and beta...
Gross alpha 	
Radium 226 	
Radium 228 	 	 	
Radon 222 	
Uranium 	
Man-Made:
Radioactive Cesium 	
Radioactive Iodine 	
Radioactive Strontium
89, 90.
Tritium 	
Gamma and photon
emitters.
Methodology
Evaportation 	
Co-precipitation
Radon Emanation
Radiochemical

Liquid Scintillation
Lucas Cell 	
Radiochemical
Fluorometric 	
Alpha Spectrometry 	
Precipitation
Precipitation 	
Precipitation . ...
Radiochemical 	
Liquid Scintillation
Gamma Ray
Spectrometry.
References (Method or Page Number) - .
EPA1
900.0
903.1
903.0
904.0

908.0
908.1
901.0
902.0
905.0
906.0
901.1
EPA2
pp. 1-3
pp. 16-23
pp. 24-28
pp. 4-5
pp. 29-33
pp. 108-114
pp. 34-40
EPA3
00-01
00-02
Ra-03
Ra-05
Ra-05
00-07
Sr-04
H-02
EPA4
Pi
p. 19
p. 19
p. 33
1-01
p. 65
p. 87
SM5
7110 B
750-Ra B
7500-Ra
D*
7500-U B
7500-U C
7500-Cs B
7500-1 B
7500-Sr B
7500-3H B
ASTM6
D 1943-81
D 3454-86
D 3972-82
D 2907-83 .
D 2334-88
D 2476-81
(87)
D-3649-85
USGS*
R-1 120-76
R-1141-76
R-1 142-76
R-1 180-76
R-1 182-76
R-1 11 0-76
R-1 1 60-76
R-1171-76
DOE8
E-U-03
E-U-04
E-Cs-01
E-Sr-01
4.5.2.3
Other
9 N.Y.
8 N.Y.,
10 N.J.
"913,
12 LS
12 LC
   10'P!jeTribed Prooedures f°r Measurement of Radioactivity in Drinking Water," EPA Environmental Monitoring and Support Laboratory, Cincinnati, OH (EPA-600/
   2 "Interim Radiochemical Methodology for Drinking Water," EPA-600/4-75-008, March 1976. (EPA 1976)
   'Eastern Environmental Radiation Facility,  Montgomery, AL 36109, "Radiochemical  Procedures Manual," EPA 520/5-84-006, August 1984. (EPA,  1984a)*
   «. Radiochemical Analytical Procedures for Analysis of Environmental Samples," EMSL-LV-0539-17, March 1979. (EPA, 1979b)
r, „ !• sta.ndard. Methods for the Examination of Water and Wastewater," 17th edition, American Public Health Association, American Water Works Association, Water
Pollution Control Federation, 1989. (APHA, 1989)
   *,}?*£. Annual Book of ASTM Standards, Vol. 11.02, American Society for Testing and Materials, 1916 Race Street, Philadelphia, Pa. 19103. (ASTM, 1989)
., ./MM™*i for Determination of Radioactive Substances in Water and Fluvial Sediments," Book 5, 1989, Techniques of Water-Resources Investigations of the
United States Geological Survey, Chapter AS. (USGS, 1989)
   8 Environmental Measurements Laboratory, U.S. Department of Energy, "EML PROCEDURES MANUAL, 27th edition." (DOE, 1990)
     d^jerm'i98l? fNY2?Ra a"d 228Ra *Ra~02)l fadio'oQica1  Sciences Institute Center for Research—New York State Department of Health, January 1980
 ,  "."Determination of Radium 228 in Drinking Water," State of New Jersey—Department of Environmental Protection—Division of  Environmental Quality—Bureau
of Radiation and Inorganic Analytical Services, August 1990. (NJ DEO, 1990)
   " Method 913—Radon in drinking water by liquid scintillation, "Environmental Monitorinn and Support Laboratory, Las Vegas, NV. (EPA, 1991q)
i . . . App,end~ P'uAnal.vticSl Jest procedure,  "The Determination  of Radon in Drinking Water," p. 22, Two Test Procedures  for  Radon in Drinking  Water,
Interlaboratory Collaborative Study, EPA/600/2-87/082, March 1987. (EPA, 1987e)       or-,                                    a      .
  EPA believes that the analytical
methods listed in Table 15 are
technically and economically available
                         for radionuclide monitoring. Many of the
                         listed analytical methods have been
                         used for a number of years in water
analyses under the Interim Drinking
Water Regulations (see 40 CFR part 141,
subpart C) and in determining

-------
                Federal Register / Vol. 56. No. 138 / Thursday,  July 18, 1991 / Proposed Rules          33093
compliance with the current MCLs (see
40 CFR part 141, subpart B). EPA has
updated the original references to the
most recent editions of the manuals and
references, when applicable, i.e., EPA,
Standard Methods (SM), American
Society for Testing Materials (ASTM),
United States Geological Survey (USGS)
and Department of Energy (DOE).
Several more recently developed
methods are also listed. In addition, EPA
Method 909, "Determination of Lead-210
in Drinking Water" would be used for
the unregulated contaminant monitoring
for lead-210 (EPA, 1982).
  The reliability of these methods has
been demonstrated by a history of many
years' use by state, federal and private
laboratories. Most of the methods above
have undergone an interlaboratory
collaborative study (multilaboratory
tested), with the remainder being
subjected to single laboratory tests. The
majority of the validation studies were
EPA performed or sponsored. Those
validations performed by accredited
standard bodies, i.e., SM, ASTM, etc.
were reviewed by EPA personnel and
determined to be acceptable. The N.Y.
method for radium 226 and 228 had
"limited approval", previous to the
discontinuation of alternate test
procedures (ATPS) in the drinking water
program. The N.J. method for radium 228
is currently under review. EPA requests
comments on whether these techniques
should be considered available for
purposes of this proposed rule.
   Below is a brief description of the
proposed radionuclide techniques listed
in Table 15. Analysis generally requires
some sample preparation followed by
counting by one of several methods.
Radiation counting instruments include
various types of gas-flow proportional
counters, scintillation cells and
scintillation counters that are suitable
for measuring alpha- or beta-emitting
radionuclides, and sodium iodide or
germanium detectors coupled to
multichannel analyzers are available for
gamma spectrometry. General
description of the different basic
counting methods are presented,
followed by brief discussions of the
methods specific for each analyte.
Copies of the complete methods are
available in the Drinking Water Docket,
as well as hi several published reference
manuals. EPA refers readers to the
references for information on precision,
accuracy, counting efficiency,
background determination, sample and
source preparations, interferences and
calibration information on the proposed
analytical methods.
   a. Counting methods. L Alpha Emitting
Radionuclides (Gross alpha particle
activity, Radium-226, Radon-222 and
Uranium)—Alpha Counting Methods—
Alpha particles are characterized by an
intense loss of energy in passing through
matter. This intense loss of energy is
used hi differentiating alpha
radioactivity from other types by the
dense ionization or intense scintillation
it produces. Alpha counting methods,
which measure alpha radioactivity, are
applicable in the determination of gross
alpha particle activity, radium-226,
radon-222 and uranium. Alpha
radioactivity can be measured, after
various sample preparations, by one of
several types of detectors in
combination with appropriate electronic
components. The techniques for
measuring the alpha emitters use gas-
flow proportional counters, scintillation
cell systems and liquid scintillation
counters, in conjunction with electronic
components such as high voltage power
supplies, preamplifiers, amplifiers,
sealers and recording devices.
Additional techniques using fluorometry
and alphaspectrophotometric techniques
are being proposed for uranium
analysis.
  Proportional Counting. In proportional
counting, alpha particles are introduced
to the sensitive region of a proportional
counter and produce ionization of the
counting gas. The electrons are
accelerated towards the anode,
producing secondary ionization and
developing a large voltage pulse by gas
amplification. The total ionization is
proportional to the primary ionization
produced by the alpha particle.
Electronic voltage discrimination allows
for differentiation of alpha particles
from beta particles.
   Scintillation Counting. In scintillation
counting, the alpha particle transfers
energy to a scintillator disk, such as zinc
sulfide, which is enclosed within a light-
tight container. The transfer of energy to
the scintillator disk results in the
production of light at a wavelength
characteristic to the scintillator, and
with an intensity proportional to the
energy transmitted from the alpha
particle. The scintillator disk is placed
next to the sample and on the face of the
photomultiplier tube. The  light from the
scintillator strikes the photocathode
producing electrons, which are emitted
at levels proportional to the intensity of
the light. The photoelectrons are
amplified by the multiplier phototube
and a voltage pulse is produced at the
anode for measurement. An electronic
sealer (counter) records the individual
pulses which are proportional to the
number of alpha particles striking the
scintillation detector.
  A scintillation cell system for radon
gas counting performs alpha particle
counting using the principles of
scintillation counting as described
above. The exceptions are that a
scintillation flask ("Lucas Cell", a 100-
125 ml metal cup coated on the inside
with zinc sulfide and having a
transparent window) replaces the
scintillation disk in the apparatus. A
counting system compatible with the
scintillation flask is incorporated. The
scintillation cell system is used for the
specific measurement of radon. Radium-
226 can also be measured by Lucas Cell
counting of its radon-222 progeny.
  Direct, low volume liquid scintillation
(liquid scintillation) counting of alpha
emitters with a commercially available
instrument is also employed in the
proposed methods. A liquid scintillator
or organic phosphor is combined in an
appropriate mineral oil or other organic
base scintillator "cocktail" with the
water sample. Mixing achieves a
uniform dispersion before counting. This
replaces the planchet or disk
preparation that occurs before the
counting step in the scintillation
technique.
  Analyses performed using a
fluorometer require sample preparation
as mentioned above. Fluorometry is
used in one of the procedures for
uranium in this proposal. The
fluorometer measures the fluorescence
of the uranium from the sample that is
exposed to ultra violet light from the
instrument. The response to this
excitation is proportional to the
concentration of the analyte in the
sample.
  Alpha spectroscopy involves
identifying specific  alpha isotopes by
converting the kinetic energy of an
alpha particle to a charge pulse whose
magnitude is proportional to the alpha
particle energy absorbed by the
detector. The pulse is routed to a
multichannel analyzer where  energy
discrimination can be performed. This
alpha spectrometer is employed in some
of the techniques for the measurement of
uranium.
  ii. Beta Emitting Radionuclides (Gross
beta particle activity, Radium-228,
Cesium-134 and -137, lodine-131,
Strontium-89 and -90 and Tritium)-Beta
Counting Methods:  The large difference
in the specific ionization energy
produced by alpha and beta particles
permits pulse discrimination between
these radiations to allow for
identification. Beta  particles are
characterized as fast electrons emitted
by radioactive nuclei. The beta particles
from a particular radioactive element
are not all emitted with the same energy

-------
33094
Federal Register / Vol. 56, No. 138 / Thursday,  July 18,  1991 / Proposed Rules
but with energies ranging from zero up
to a maximum value which is
characteristic of the nuclide. This fact
makes it extremely difficult to
differentiate among beta emitters by
energy discrimination.
  Beta counting methods, which
measure beta radioactivity, use one of
several types of instruments (counters)
that consist of a  detector and an
amplifier, power supply, and sealer, etc.
As in alpha counting, there are various
sample preparations or chemical
separations necessary prior to counting.
The most widely used instruments are
proportional counters, but scintillation
systems are also used. These counting
techniques are applicable for the
measurement of beta radioactivity by
using beta emitting standards for
calibration and determination of
counting efficiency in the  analyses.
  iii. Gamma and Photon Emitting
Radionuclides-Gamma Counting
Method: Gamma rays are high energy
photons with discrete energies that are a
penetrating form of radiation. This
characteristic can be  used to measure
samples of any form,  as long as
calibration standards of the same form
are available and are counted using the
same geometry. Individual calibration
standards are used for identifying and
quantifying contributing gamma emitting
radionuclides using gamma counting or
gamma spectrometry. Gamma counting
is performed using solid detectors (Nal
or germanium), as opposed to gas-filled
detectors.
  In gamma-ray analysis or counting,
the detectors produce light photons
(scintillations) or electron-hole pairs
that are amplified into electrical pulses
within the counting system. These
output pulses, which are directly
proportional to the,amount of energy
produced,  are counted using a sealer or
analyzed by pulse height to produce a
gamma-ray spectrum, depending on the
detector employed. The use of a
multichannel analyzer allows for energy
discrimination and the identification
and quantification of the individual
nuclides.
  b. Specific analytic methods—i. Gross
alpha and gross beta  activity. The gross
alpha and gross beta activity methods
are the simplest of radio analytic
methods. A portion of the water sample
is simply evaporated to dryness on a
planchet, which is then counted for
alpha and beta activities. The different
types of alpha and beta counting
equipment used was described above.
The co-precipitation method, usually
applicable for gross alpha analysis, adds
one chemical separation step before
counting to reduce the total solids
present, thereby reducing  self
                        absorption and improving counting
                        efficiency. It also allows for the use of
                        larger samples for greater sensitivity.
                          In addition to being used to determine
                        compliance with the MCLs, these
                        methods would be used as screening
                        procedures  to determine if additional
                        analyses for the specific radionuclides
                        are necessary, if the appropriate
                        standard is  used for calibration. Gross
                        alpha measurement would be used as a
                        screen for radium 226 and uranium, and
                        gross beta would be used as a screen for
                        radium 228. If gross alpha methods are
                        to be used for screening for radium 226
                        and uranium compliance, the labs
                        however, would be required to calibrate
                        the counter  for uranium. Laboratories
                        would also be required to generate
                        standard curves for their counters
                        showing the change in counting
                        efficiency versus the total solids in the
                        water sample  (for both radium and
                        uranium), and use these curves to
                        correct for lower counting efficiencies
                        found with high solids samples. If these
                        corrections  are not made, gross alpha
                        measurements would not be considered
                        a valid screen for radium 226 and
                        uranium for determining compliance
                        with the MCLs. Valid gross beta
                        measurements can be made with waters
                        having a much larger dissolved solids
                        content than for alpha emitters. In beta
                        counting efficiency  does not change
                        appreciably with solids in water
                        samples but generation of self
                        absorption curves is still required. EPA
                        recommends use of strontium 90 for the
                        beta screen for radium 228. The gross
                        alpha screen would no longer be used to
                        screen for the  presence of radium 228 as
                        in the current interim monitoring
                        requirements,  as radium 228 is a beta
                        emitter and alpha screening could not be
                        expected to reliably serve as a screen.
                         The Agency believes that a pure alpha
                        particle emitter i.e., thorium 230 should
                        be used as a standard for calibration for
                        gross alpha  activity. Past use of
                        americium^241 tended to bias analytic
                        results low due to the  over estimate of
                        counting efficiency because of its higher
                        energy alpha particle.  Cesium 137 is
                        recommended for calibrating the gross
                        beta screen.
                         A co-precipitation method for gross
                        alpha activity  has also been included.
                        This method was reviewed and
                        evaluated in the report, "Test Procedure
                        for Gross Alpha Particle Activity in
                        Drinking Water" (EPA, 1985c). Water
                        samples that have high dissolved solids
                        (>500 mg/1), are likely to have high self
                        absorption of alpha particles which
                        reduces the  sensitivity of the
                        measurement.  When high solids are
                        present, the  Agency recommends use of
                        the coprecipitation method.
  ii. Radon. EPA is proposing two
methods for measurement of radon in   •
water. These are direct low volume
liquid scintillation counting, and by
radon de-emanation from the sample
into a Lucas Cell chamber for counting.
These two methods are described in the
report "Two Test Procedures for Radon
in Drinking Water, Interlaboratory
Collaborative Study" (EPA, 1987e). EPA
has slightly modified the liquid
scintillation procedure described in that
report and proposes to establish this
revised method as EPA Method 913.
  In direct, low volume liquid
scintillation measurement of radon, a
small volume of water (about 10 ml) is
placed in a vial with a scintillation
solution (mineral oil),  mixed, and the
vial placed in a liquid scintillation
counter. Counting time can range up to
100 minutes or more, depending on the
amount of radon in the sample and the
desired precision of analysis.
Companies using liquid scintillation
counting report that they can analyze
50-200 samples daily (EPA,  1989e; 1990J).
  In using the Lucas Cell method, radon-
free helium or aged air (to allow the
radon present to decay out) is bubbled
through a water sample in a bubbling
apparatus into an evacuated
scintillation chamber. After equilibrium
is reached (3 to 4 hours), this chamber is
placed in a counter and the  scintillations
are counted through its window. This
method generally allows measurement
of lower level of radon than does low
volume direct liquid scintillation.
However, this is a method that is
difficult to use, requiring specialized,
glassware and skilled technicians. Most
laboratories that currently measure
radon use liquid scintillation, and few
have the equipment to perform Lucas
Cell counting. Estimated start-up cost to
obtain Lucas Cell equipment would be
about $35,000 (to do 30-40 samples
daily), plus technician training (EPA,
1989d). Also, a variant of the Lucas Cell
method, requiring the same equipment
and skills, can be used to measure
radium 226 (because radon 222 is the
first daughter of radium 226). The
widespread use of Lucas Cells for radon
analysis would make the method less
available for radium 226 analyses. These
factors limit use of the Lucas Cell
method on a large scale for radon
measurement, and EPA believes it is not
appropriate as the sole basis for
compliance monitoring for radon in
water. EPA includes it here as an
adjunct to the liquid scintillation
method; the Lucas Cell method would be
allowed to be used for radon
measurement, but could not be relied on
to support a national sampling program

-------
                Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991  /  Proposed Rules          33095
for radon in water. EPA believes only
liquid scintillation would allow accurate
analysis of the large number of samples
required nation wide by these proposed
regulations.
  iii. Radium. Several methods are
available for the specific analysis of
radium 226 and 228 as listed in table 15.
Most of the methods in the interim
regulations for radium analyses are
technique dependent and time-intensive.
Some of the other methods listed appear
to be improvements over the existing
approved methods. For example, co-
precipitation steps are employed in
methods for both radium 226 and 228 to
purify the sample and reduce
interferences.
  Analysis of radium 226 by radon
emanation requires allowing the radium
228 to decay to radon (to equilibrium) in
the water sample, bubbling radon-free
helium gas through the water into an
evacuated Lucas Cell counting chamber,
and then counting the chamber. While
this method can produce good precision
and accuracy at relatively low radium
226 levels, it is as noted above, time
consuming and requires special
equipment and specially trained lab
technicians. These  factors may limit its
use on a large scale. EPA believes this
is, however, one of several appropriate
methods for radium 226. Appropriately
conducted gross alpha screens should
eliminate the need  for specific radium
228 analyses in many cases.
   iv. Uranium. Uranium can be
analyzed using fluorometric (mass) or
radiochemical methods, or using alpha
 spectrometry. The fluorometric method
 measures the mass of total uranium
 present in the sample. Because EPA is
 proposing an MCL expressed in mass
. units, this is the preferred method.
   However, should the final MCL be an
 activity standard, the results of
 fluorometric analysis may be converted
 to an activity level using the conversion
 factor 1.3 pCi/fig. This conversion factor
 is based on evaluation of the relative
 occurrence of the different radioisotopes
 of uranium in water samples. This value
 is somewhat different from uranium
 naturally occurring in soil, which has an
 estimated conversion factor of 0.68 pCi/
 ju,g. The need for conversion from mass
 to activity following analysis, and the
 potential for variability in the
 conversion factor would be a weakness
 of the fluorometric method in
 determining compliance with an activity
 MCL for uranium. EPA solicits public
 comment on the advisability of
 continuing to allow use of this method to
 measure  uranium activity levels.
   The radiochemical method for
 uranium involves chemical separation of
 uranium followed by counting in an
 alpha counter, as described below.
 Uranium is specifically precipitated
 from the  sample and the sample is then
 counted. In addition, uranium may be
 measured by alpha spectrometry which
 allows for the determination of
 individual isotopes of uranium and the
 calculation of the total mass of uranium
 present. These aforementioned methods
 may be found to be more expensive to
 perform than the fluorometric method,
however EPA believes that the results
will be more reliable.
  c. Sample Collection, handling and
preservation. In order to ensure that
samples arriving at laboratories for
analysis are in good condition, EPA is
proposing requirements for sample
collection, handling and preservation, as
described in table 16. For radium,
uranium and gross alpha and gross beta
analysis, sample collection should be
performed as for inorganic contaminant
monitoring as described in EPA's
"Manual for the Certification of
Laboratories Analyzing Drinking Water"
(EPA, 1990b).
  For radon, because it is a volatile gas,
special attention to sample collection is
required. Either the VOC sample
collection method, or one  of the methods
described in "Two Test Procedures for
Radon in Drinking Water,
Interlaboratory Collaborative Study"
(EPA, 1987e) should be used. In addition,
because plastics can absorb radon, glass
bottles with teflon lined caps must be
used. Finally, EPA's assessment of
laboratory performance is premised on
analysis of samples no longer than 4
days after collection. Laboratories
unable to comply with this holding time
maximum may have difficulty
performing within the estimated
precision and accuracy bounds. EPA
 solicits public comment on the proposed
 sample collection procedures for radon
 in  drinking water, including any
 available data on radon loss from water
 samples during collection by different
 methods.
                               TABLE 16.—SAMPLING HANDLING, PRESERVATION, HOLDING TIMES
Parameter

Gross &
-------
33096
Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
laboratories exist to help define costs;
(b) as the number of experienced
laboratories increases, the costs can be
expected to decrease; (c) analytical
costs are determined, to some extent, by
the quality control efforts and quality
assurance programs adhered to by the
analytical laboratory; (d) per-sample
costs are influenced by the number of
samples analyzed per unit time. EPA
solicits comments on its cost estimates
from laboratories experienced in
performing these analyses.

     TABLE 17.—ESTIMATED COST OF
     ANALYSES FOR RADIONUCLIDES
Radionuclides
Radium-226 	 	 	
Radium-228 	
Uranium (total) 	
U isotopic 	 	 	
Radon-222 	
Gross alpha emitters 	
Gross beta emitters 	
Radioactive Cesium 	
Radioactive Iodine 	
Radioactive Strontium 	
Total, 89 and 90 	
Tritium 	 	
Gamma emitters 	

Approxi-
mate cost
for
analysis
in
drinking
water
$85
100
45
125
50
35
35
100
100
105

50
110

  Source: (EPA, 1991m)
  Note: Estimated costs are on a per-sample basis;
analysis of multiple samples may have lower cost.

  3. Method detection limits and
practical quantitation levels. Method
detection limits (MDLs) and practical
quantitation levels (PQLs) are two
performance measures used by EPA to
estimate the limits of performance of
analytic chemistry methods for
measuring contaminants in drinking
water. An MDL is the lowest level of a
contaminant that can be measured by a
specific method under ideal research
conditions. A PQL is the level at which a
contaminant can be.ascertained with
specified methods on a routine basis,
(such as compliance monitoring) by well
managed laboratories, and within
specified precision and accuracy limits.
The proposed PQLs for the
radionuclides are listed in Table 18
below (EPA, 1991r).
  EPA considers PQLs in evaluating
alternatives for the MCL. Consideration
of the PQL is especially important for
those contaminants for which EPA is
proposing MCLGs at zero. The
feasibility of implementing an MCL at a
particular level is in part determined by
the ability of analytical methods to
ascertain contaminant levels with
sufficient precision and accuracy at or
near the MCL.
                         EPA usually defines the method
                       detection limit (MDL) as the minimum
                       concentration of a substance that can be
                       measured and reported with 99 percent
                       confidence that the true value is greater
                       than zero. The term MDL is used
                       interchangeably with minimum
                       detectable activity (MDA) in
                       radionuclide analysis, and is defined as
                       that amount of activity which in the
                       same counting time, gives a count which
                       is different from the background count
                       by three times the standard deviation of
                       the background count. Identifying an
                       MDL concentration is limited by the fact
                       that MDLs (MDAs) are specific to the
                       performance of a given measurement
                       system, and vary from system to system.
                         The concept of MDL is different for
                       radionuclide measurement than for non-
                       radioactive chemicals. Because counting
                       times can be expanded to days or even
                       weeks or longer in a research setting,
                       very small differences from background
                       can theoretically be detected depending
                       on research needs. These extremely long
                       counting times are unrealistic for
                       compliance monitoring for drinking
                       water. EPA has sometimes set
                       laboratory performance expectations at
                       a level 5 to 10 times the MDL. However,
                       MDLs (MDAs) are not necessarily
                       reproducible on a routine basis in a
                       given laboratory, even when the same
                       analytical procedures, instrumentation
                       and sample matrix are used. EPA has
                       therefore relied on actual performance
                       data generated in Performance
                       Evaluation and other studies in setting
                       standards for laboratory performance
                       for radionuclide monitoring.
                         The PQL is  determined through
                       evaluation of the results of
                       interlaboratory studies, such as
                       performance evaluation (PE) studies. In
                       these studies, prepared samples of
                       known concentration are distributed for
                       analysis to participating labs as
                       unknowns. The results of the analyses
                       by the participants are compared with
                       the known value and with each other to
                       estimate the precision and accuracy of
                       both the methods used and the lab's
                       proficiency in using the method. (See 54
                       FR 220624, May 22,1989; 52 FR 25699,
                       July 8,1987; and 50 FR 46906, November
                       13,1985 for further discussions on MDLs
                       and the concept of PQLs.) MDLs (MDA)
                       are lower than PQLs since the MDL
                       represents the lowest level at which
                       there is 99% confidence that the true
                       value is greater than zero, while the PQL
                       represents the level that can be
                       ascertained under practical and routine
                       laboratory conditions. The measurement
                       of radioactivity becomes limited at low
                       concentrations and small sample sizes
                       due to the random nature of radioactive
 decay and the resulting theoretical
 counting uncertainty. The counting
 uncertainty is the major contribution to
 the overall uncertainty. This uncertainty
 must be calculated and added to the
 result and other uncertainties to
 determine whether or not the analysis
 has demonstrated compliance (EPA,
 1991r; 1986a).
  The method for estimating the PQLs
 for radionuclides is based on the same
 criteria as that used for organic and
 inorganic compounds and incorporates,
 through the methodology, the counting
 time and background activity in each
 laboratory. The PQLs for radionuclides
 are estimated based on results from
 EPA's Water Supply Performance
 Evaluation and Intercomparison Cross
 Check Studies for radionuclides with the
 exception of radon, for which no PE or
 cross check data were available. These
 studies are conducted as a part of EPA's
 laboratory certification program by
 EPA's Environmental Measurement
 Systems Laboratory in Las Vegas. A
 number of laboratories, ranging from 60
 to 140 depending on the analyte, have
 participated annually and biannually,
 respectively, in the PE and cross check
 studies. There are approximately one
 hundred certified laboratories nationally
 that have the capability to conduct
 analyses for the radionuclides currently
 regulated (Ra-226 and 228, gross a and
 gross B, and also uranium). PE studies
 were used to estimate PQLs primarily
 because they are good indicators of
 laboratory performance. The fact that
 they are blind samples eliminates
 possible biases. The intercomparison
 studies cross  check study data served as
 an alternative source of data as well as
 a means of verifying laboratory
 performance.
  Because until recently there was not a
 standardized analytic method, nor a
 calibration standard for radon, no PE
 studies were done on radon. Both a
 standard method and calibration
 standard have now been developed and
 EPA plans to include radon in future PE
 studies. In the interim, EPA relied on
 two data sources for estimating
 performance of the available radon
 methods. One study was the report
 "Two Test Procedures for Radon in
 Drinking Water, Interlaboratory
 Collaborative Study" (EPA, 1987e),
 which evaluated performance of the
 radon methods down to 1600 pCi/1.
 Because EPA wanted to consider MCL
 alternatives lower than this, additional
 data on radon measurement was
generated by EPA. Radon samples,
 supported by radium 226 bound to a
resin, as low as 100 pCi/1 were tested by
12 labs using liquid scintillation and 4

-------
                Federal Register  /  Vol.  56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                                     33097
labs using Lucas Cells, and these data
were used to evaluate performance of
the methods and estimate the PQL [EPA,
1991n; 1991r). EPA considers the radon  ,
data to be a limited basis for deriving a
PQL, and solicits additional information
on radon analysis.
  The PQLs for the radionuclides were
derived applying a procedure described
in 50 FR 46908, Nov. 13,1985 and 54 FR
22100, May 22,1989. Data from all
reporting laboratories of Performance
Evaluations A and B, 1983-1990 [EPA,
1991r), which include EPA and State
laboratories, were used for radium,
uranium, gross alpha and gross beta. For
radon, data from the two studies
described above were used. The PQL
procedure generates acceptance limits
that are set around a "true" value. Using
the procedure described  in these
notices, the PQLs for all radionuclide
contaminants were set at a
concentration where  it was estimated
that at least 75 percent of all reporting
laboratories are within the specified
acceptance ranges.
  The radon PQL required some special
considerations. Because  of the practical
considerations involved  in analyzing a
radioisotope with a short half life (3.8
days), EPA has made allowance for
transport time from the water supply to
the laboratory in setting  the PQL. EPA
has premised its PQL on samples being
analyzed no longer than 4 days after
collection;  mail delays could reduce
accuracy for low level samples. The
sample collection date and time would
be required on all samples collected,
and will be used by the laboratory in
calculating radon levels  present at the
time of collection. This assumption,
along with the fact that the radon PQL
was based on more limited data than
the other radionuclides PQLs makes it
more uncertain than the  other PQL
values. If the PQL were premised on an
8 day time frame from collection to
analysis [to make greater allowance for
mail delays or back-ups  in laboratories),
the PQL could be 500 pCi/1. Similarly, if
the counting time were increased
(beyond the proposed 100 minutes), a
value somewhat lower than 300 pCi/1
might be achievable  as the PQL.
Similarly, should 100 minute counts
prove infeasible, a higher PQL may need
to be set. EPA solicits public comment
on these issues related to the radon
PQL.
   Different PQL values could also be
established using different acceptance
limits. At an acceptance limit of ±20%,
for example, the radon PQL would be
 about 500 pCi/1; at acceptance limits of
 ±40% the  PQL would be 200 pCi/1. In
 choosing an acceptance limit of ±30%
and PQL of 300 pCi/1, EPA considered
the likely reliability of the overall
compliance monitoring program, the
number of systems that would have
measurements within the error range,
and the risks of radon. With an error
band of ±40%, and a PQL of 200 pCi/1,
approximately 19,000 of the estimated
33,000 systems affected would fall
within the error band and would have
potentially unclear compliance  status,
potentially resulting in requests for re-
testing and additional burdens on states
to determine and achieve compliance.
When EPA chose a ±40% acceptance
limit for the vinyl chloride regulation,
only a  few hundred systems were
expected to exceed the MCL; care could
be taken to accurately determine
compliance status if it were  in doubt.
With a ±30% error band for radon at
300 pCi/1, only 5000 to 7000 systems
would have potentially unclear
compliance status because of data
uncertainty. While this number would
decrease with an even narrower error
band, the individual lifetime risks would
be higher. Therefore, on balance, EPA is
proposing to set the PQL at 300 pCi/1.
   EPA recognizes that some laboratories
may be able to achieve better
performance than ±30% at 300 pCi/1.
Lowry (1991) very recently published a
study indicating that radon  could be
measured using liquid scintillation
counting at 300 pCi/1 with an overall
error of less than ±10%, assuming 4
days from sample collection to analysis.
EPA is reviewing this study to identify
potential improvements in its own
procedures for measuring radon by LSC.
However, EPA does not now believe
most laboratories will be capable of the
levels  of precision and accuracy
achieved by Lowry. EPA will soon
conduct a series of performance
evaluation studies on radon analysis to
better gauge performance levels and to
develop a data set on which to base lab
certification determinations when the
regulations are final. In addition, Vitz
(1991) recently published a paper
evaluating the effect of several different
variables on error in measurements,
including the effect of the type of
scintillation cocktail used, the type of
vials and standardization procedure
used, and temperature control and
instrument settings. Vitz also
commented on sampling procedures.
Vitz (1991) overall reported that radon
levels of 200 pCi/1 may be measured
with 20% precision using a 20 minute
 count, if all parameters are  optimized.
EPA is reviewing this report to identify
improvements in its proposed radon
method, EPA Method 913.
  EPA solicits public comment on these
issues, and will continue to collect and
evaluate additional data to refine and
better substantiate the proposed PQL
and the constraints on regulation
imposed by limits on analytic methods.
EPA specifically requests comment on
information supporting PQLs higher than
the proposed PQL (such as 500 pCi/1),
and information supporting a lower PQL
than that proposed, such as 200 pCi/1.
  Public comments are requested on the
approach used to determine the PQLs
for radionuclide contaminants, on the
proposed PQLs for these contaminants,
and information is sought on any new
developments in methodology for the
radionuclide contaminants that may be
used to support development of these
regulations. EPA also solicits public
comment on the usefulness of PQLs in
setting standards, and the
appropriateness of alternative methods
for accounting for analytic methods
limitations in setting standards.

TABLE  18.—PRACTICAL  QUANTITATION
   LEVELS  (PQLs)  FOR   RADIONUCLIDE
   CONTAMINANTS
Contaminant
Radium-226 	
Radium-228 	

Radon-222 	
Gross alpha emitters 	


134 	
137 	


89 	
90 	
Tritium 	

PQL
(pCi/l)
5
5
5
300
15
30

10
10
20

5
5
1200

  (EPA, 1991r)

 E. Laboratory Approval and
 Certification
   1. Background. The ultimate
 effectiveness of the proposed
 regulations depends upon the ability of
 laboratories to reliably analyze
 contaminants at relatively low levels.
 The existing drinking water laboratory
 certification program (LCP) established
 by EPA requires that only certified
 laboratories may analyze compliance
 samples.
   External checks of performance to
 evaluate a laboratory's ability to
 analyze samples for regulated
 contaminants within specific limits is
 the primary means of judging lab
 performance and determining whether
 to grant certification. EPA provides
 performance evaluation samples to
 laboratories on a regular basis;

-------
 33098
Federal Register / Vol.  56, No. 138  /  Thursday, July 18, 1991 /  Proposed Rules
 participation in the PE program is
 prerequisite for a laboratory to achieve
 certification and to remain certified for
 analyzing drinking water compliance
 samples. Achieving acceptable
 performance in these studies of known
 test samples provides some indication
 that the laboratory is following proper
 practices. Unacceptable performance
 may be indicative of problems that
 could affect the reliability of the
 compliance monitoring data.
   Unacceptable performance on PE
 studies should trigger an investigation to
 establish the possible cause(s) and to
 take corrective action. EPA recognizes
 that even superior analytical
 laboratories occasionally produce  data
 which are outside the acceptance limits
 due to statistical reasons rather than
 from any actual analytical problems.
 EPA has incorporated the criteria of
 using fixed acceptance limits around the
 true value to overcome this
 misinterpretation of analytical results. A
 provision for rapid follow-up analysis is
 necessary if a laboratory fails the initial
 determination to decrease the likelihood
 of statistical error and to determine if a
 real problem exists.
   EPA's present PE sample program and
 the approaches to determine laboratory
 performance requirements were
 discussed in 50 FR 46907 (November 13,
 1985). In addition, guidance of minimum
 qualtor assurance requirements,
 conditions of laboratory inspections and
 other elements of laboratory
 certification requirements for
 laboratories conducting compliance
 monitoring measurements are detailed
 in the Manual for the Certification of
 Laboratories Analyzing Drinking Water,
 Criteria and Procedures Quality
 Assurance (EPA, 1990b). Participation
 by 300 or more laboratories in the
 interlaboratory studies required in the
 LCP demonstrates that laboratory
 capability and capacity for the
 radionuclide analyses necessary to
 support this  proposed regulation exists.
  Acceptable performance has
 historically been identified by EPA using
 one of two different approaches: (1)
 Regressions  from performance of
 preselected laboratories (using 95
 percent confidence limits), or (2)
 specified accuracy requirements.
 Acceptance limits based on specified
 accuracy requirements are developed
 from existing PE study data. EPA was
 able to use fixed acceptance limits for
 all of the contaminants proposed in
 today's rule because of the availability
 of PE data, with the exception of radon
in which an interlaboratory
collaborative study was used. EPA
would prefer to use the true value
                        approach because it is the better
                        indicator of performance and provides
                        laboratories with a fixed target. This
                        approach requires that each laboratory
                        demonstrate its ability to perform within
                        pre-defined limits. Laboratory.
                        performance is evaluated using a
                        constant yardstick independent of
                        performance achieved by other
                        laboratories participating in the same
                        study. A fixed criterion based on a
                        percent error around the "true" value
                        reflects the experience obtained from
                        numerous laboratories and includes
                        relationships of the accuracy and
                        precision of the measurement to the
                        concentration of the analyte. It also
                        assumes little or no bias in the
                        analytical methods that may result in
                        average reporting values different from
                        the reference "true" value. This concept
                        assures that reported results can be
                        related to the percentage of variance
                        from the PQL.
                         2. Acceptance limits for radionuclide
                        contaminants. EPA relied on the data
                        generated from the radon
                        interlaboratory collaborative study to
                        estimate acceptance limits (using the
                        approach described in 54 FR 22131-
                        22132, May 22,1989). The levels (100,
                        200, and 500 pCi/1 in lab samples,
                        corresponding to 200, 400 and 1000 pCi/1
                       in field samples analyzed 4 days after
                        collection) used in the study were  below
                        and above the PQL (300 pCi/1) proposed
                       in this regulation, demonstrating the
                       participating laboratory's ability to
                       measure at or around the proposed MCL
                       (EPA, 1991r).
                         Performance data are available for all
                       of the other radionuclide contaminants
                       at the levels proposed for regulation
                       (EPA, 1991r). The acceptance limits are
                       developed using the approach noted
                       above,  resulting in the specification of a
                       "plus or minus percent of true value" for
                       setting acceptance limits. The available
                       PE data indicate that both the precision
                       and accuracy attained for specific
                       radionuclide contaminants are
                       contaminant specific. The "plus or
                       minus percent of the true value"
                       acceptance limits have been derived for
                       each contaminant taking into
                       consideration past performance of the
                       laboratories and the expected precision
                       and accuracy (EPA, 1991r).
                         EPA believes  that the nature of
                       radionuclides analysis (i.e., background
                       counts, counting time, decay) requires
                       unique  analytical considerations. In
                       some cases this  may result in a greater
                       effort from laboratories to perform
                       analyses which meet the proposed
                       acceptance limits. The Agency believes
                       that these circumstances are to be
                       addressed by the individual
 laboratories, when executing the
 analyses using the proposed
 methodology.
   The proposed acceptance limits for
 the radionuclide contaminants are
 summarized in Table 19. The acceptance
 limits only apply to concentrations
 above the PQL.

   TABLE 19.—PROPOSED ACCEPTANCE
                 LIMITS
Contaminant
Radium-226 	 	
Radium-228 	 	
Uranium natural 	
Radon-222 	 ;..
Gross alpha emitters 	
Gross beta emitters 	
Radioactive Cesium:
134 	
137 	
Radioactive Iodine 	 	 	
Radioactive Strontium:
89 	
90 	
Tritium 	

Accept-
ance
limits at
the PQL
(percent)
+30
+50
+30
i +30
±50
+30
+20
+30
+20
+50
+30
+20

   1 Acceptance limits based on 100 minute count.
 (EPA, 1991r)
 F. Proposed MCLs and Alternatives
 Considered
   The sections below discuss derivation
 of each of the MCLs for the
 contaminants proposed for regulation.
 The first section presents an evaluation
 of radon in water and discusses special
 policy issues EPA considered in
 choosing the MCL to propose for radon.
 This is followed by the derivation of
 MCLs for radium, and uranium which
 are proposed today. This is followed by
 an alternative basis for regulation, the
 lowest technically feasible levels limited
 by affordability to large water suppliers,
 on which EPA requests public comment.
 Finally, proposed MCLs for alpha and
 beta emitters are discussed.
  1. Radon. Regulation of radon in water
 is a complex issue for several reasons.
 In evaluating the various alternatives for
 proposing a radon MCL, EPA considered
 the critical policy question of whether
 radon in water should be regulated like
 other drinking water contaminants, or
 whether it should be regulated more  in
 accord with its importance compared to
 overall radon exposures. In considering
 the radon MCL, EPA reviewed and
 evaluated alternatives over the range of
 200 to 2000 pCi/1.
  The primary health hazard posed by
radon in water is due to its volatilization
from water during household water use,
and enrichment of indoor air radon
levels, thereby contributing to increased

-------
                Federal Register / Vol. 56,  No. 138  /  Thursday,  July  18, 1991 / Proposed Rules
                                                                     33099
risk of lung cancer. Direct ingestion of
radon may also pose some risk of
stomach and other cancers. While on
average water makes a small
contribution to indoor air radon (about
5% for houses served by ground water),
it is prevalent in drinking water from
groundwater wells and does contribute
to the very substantial risks posed by
radon in the environment overall.
Because it is a volatile gas, very little
radon is expected to be found in surface
water, and no surface water systems are
anticipated to require treatment. EPA
estimates that 30,000 or more public
water systems serving 30 million or
more people may have radon in water at
levels exceeding an estimated 1X10"4
risk level [150 pCi/1 water).
   Outdoor background levels of radon
in air (about 0.1 to 0.5 pCi/1 air) present
estimated lifetime lung cancer risks of
about 1 in 1000, a risk level above those
generally accepted in EPA regulatory
programs. Typical indoor air radon
levels (1-2 pCi/1 air) pose estimated
lifetime lung cancer risks near 1 in 100.
Radon is estimated to cause 8000 to
40,000 (EPA, 1989g) lung cancer deaths
annually, of which about 75-400 may be
attributed to radon from drinking water.
As discussed in Section IV.C.2 above,
the SAB/RAC is presently reviewing a
proposed revision of the radon risk
estimate, which could result in an
approximate 3095 reduction in these
estimates. While the average water
contribution to indoor air radon is small
relative to the contribution of soil gas
(for most houses), it does represent a
substantial estimated number of annual
cancer cases and in many communities
poses individual lifetime risks above
EPA's lifetime cancer risk goal for
drinking water regulations of 10~4 to
ID'6 (52 FR 25698, July 8,1987). While
these risk estimates have inherent
uncertainties, they are no greater here
than for other contaminants regulated
by EPA using such a risk assessment
approach.
   A number of factors were considered
in deciding on the approach to
regulating radon. Radon in public water
systems can be treated centrally rather
than on a house-by-house basis as is the
case with radon from soil gas. Radon
can be  removed from drinking water
 efficiently and relatively inexpensively
 (compared with other drinking water
contaminants and treatments), although
 costs to small systems will be high.
Also, while EPA has no  authority to
regulate radon in private homes (or
 wells),  the Agency is required to
 regulate water delivered to customers
 by public water systems under the
 SDWA. Moreover, the 1986 amendments
to the SDWA require EPA to develop an
MCL for radon.
  Finally, while saving an estimated 57-
100 cancer cases annually (the
estimated benefit of regulating radon in
water in the range of 500 to 200 pCi/1,
respectively) is a small number
compared with the estimated 8,000-
40,000 annual cancer cases caused by
radon exposure (EPA, 1989g), it would
be a substantial public health benefit
compared with other drinking water
regulations and other environmental
regulation programs administered by
EPA. For example, regulation of vinyl
chloride in drinking water is estimated
to avoid 27 cancer cases annually; the
only other currently regulated individual
contaminant (out of some 50 standards)
with more estimated cancer risk avoided
is ethylene dibromide, with an estimated
72 cases avoided per year. EPA
concluded that regulation of radon in
water constitutes an opportunity to
achieve a substantial public health
benefit in an area of high environmental
risk, and to do so at relatively low cost.
  EPA also considered other factors in
developing its proposed radon MCL,
including the ability to accurately
measure radon in water and potential
implementation difficulties. As
discussed in Sections V.D and E, radon
poses some challenges in routine
measurement. Not only is it a volatile
gas, it also has a short radioactive half-
life (3.8 days). This means that samples
must be carefully collected and
promptly sent for analysis; analytic
sensitivity decreases by one half for
every 3.8 days after collection that the
sample is analyzed. While the count
time could in theory be extended to
compensate for this, the 300 pCi/1 PQL is
premised on a count of 100 minutes,
which EPA believes is at a reasonable
limit, and that overall, a PQL of 300 pCi/
1 is at the reasonable limit of the
analytic methods, based on available
data. Should additional data show that
it is difficult for labs to perform
consistent analysis at this level  with the
expected precision  (due perhaps to long
transport times), or if data uncertainty
near this value (i.e. the ±30% now
estimated and believed to be
acceptable) renders the MCL impossible
to implement, the PQL could possibly be
reviewed and revised upward. Similarly,
should new data show analysis easier at
low levels than now believed, the PQL
 could be revised downward. The recent
 study by Lowry (1991) indicates that
 some individual labs may achieve better
performance than the minimum
requirements proposed here.
   EPA also considered potential
 difficulties in implementing a radon
MCL at different levels in the range of
200 to 2000 pCi/1. Implementation was
considered to be a serious issue only in
the range of 200-500 pCi/1. A large
number of PWS would be affected at
any MCL in the range of 200 to 500 pCi/1,
but many more systems would be
affected at the 200 pCi/1 MCL option.
There are approximately 48,000
community and 20,000 non-community,
non-transient public water systems
served by ground water sources. At an
MCL of 200 pCi/1, EPA estimates that
33,000 PWSs would be required to take
action to meet the MCL; at 300 pCi/1,
26,000 systems would be affected; at 500
pCi/1, approximately 18,000 systems
would be affected. EPA is particularly
concerned about these impacts because
of the overall regulatory burden being
placed on water suppliers as the 83
mandated contaminants are regulated.
For example, 40,000 systems are
expected to need to treat to meet the
recently promulgated lead and copper
regulations. EPA solicits public comment
on consideration of implementation
issues in setting MCLs.
   Because radon is a problem only for
ground water dependent systems, a
large percentage of the  affected systems
are small (85% serve fewer than 500
people). While treatment for radon is
inexpensive for larger PWSs (on a per-
house basis), smaller systems will have
more difficulty installing treatment.
Also, exemptions are unlikely to be
available to these systems, as all of the
options considered are  in the 10""risk
range, which is the proposed limit for
identifying unreasonable risks to health
(URTH) posed by drinking water
contamination in the draft document:
"Guidance for Developing Health
Criteria for Determining Unreasonable
Risks to Health" (EPA,  1990k). EPA also
recognizes that there would be a
substantial State burden to implement
any radon MCL in the 200 to 500 pCi/1
range, but that it would be greater at the
lower MCL option. EPA solicits public
comment on how these considerations
should be factored into establishing the
radon MCL.
   EPA considered proposing radon
MCLs in the range of 200 to 2000 pCi/1.
However, 2000 pCi/1 represents an
estimated 10"3 risk, and this alternative
was rejected as inconsistent with the
 SDWA and Agency risk management
policy. EPA therefore concentrated
much of its effort on evaluating MCL
 alternatives in the range of 200 to 500
pCi/1. Based on considerations of
 available treatment technologies, cost,
risk, analytic capabilities and
 implementation concerns, EPA
 determined that 300 pCi/1 is the lowest

-------
 33100
Federal Register / Vol. 56, No.  138 / Thursday, July 18, 1991 / Proposed Rules
 feasible level at which radon can be
 regulated, and proposes to set the MCL
 at this level.
   EPA solicits public comment on this
 proposal, as well as all the alternatives
 considered, from 200 to 2000 pCi/1. In
 particular, comment is sought on 200
 pCi/1 as an  alternative, in light of new
 studies indicating that radon analysis
 may be improved in the'future and the
 greater health benefits at this level (an
 estimated 20 additional cancer cases
 avoided annually), and also on 500 pCi/1
 as an alternative, if analytic difficulties
 in a implementation setting become
 apparent (i.e., the PQL may be set higher
 if 4 day delivery to labs proves too
 short) and in light of the substantial
 implementation burden that would be
 imposed by lower values.
   Another issue of concern to EPA
 regarding radon regulation was
 application  of the MCL to private wells.
 The relative magnitude of risks from
 radon in water (vs soil gas) is important
 for home owners to bear in mind when
 applying any radon MCL to private
 wells. Because the soil gas contribution
 to indoor radon levels is in most cases
 much larger than the water contribution,
 testing and mitigation strategies for
 private homes should consider all
 sources of radon. The mitigation
 strategy which is most cost-effective
 overall for an individual home should be
 used. In a majority of cases, this will
 mean controlling the soil gas
 contribution to indoor radon before
 ensuring that the radon MCL is met. Soil
 gas contributes more radon to the indoor
 air than does water  in most houses.
 Economies of scale for treatment by
 public water systems make radon
 removal from water cost-effective for
 PWSs. Water treatment is unlikely to be
 the most cost-effective first step in
 mitigating radon in individual homes
 (relative to soil gas mitigation). EPA has
 prepared several publications for
 homeowners and private well owners to
 help them in addressing their radon
 problems effectively and for the lowest
 cost possible. These publications
 include, for general information on
radon risks,  testing in the home, and
mitigation of soil gas contributions to
indoor air, A Citizen's Guide to Radon
 and A Homeowners Guide to Radon;
and for radon in water, Radionuclides in
Drinking Water Fact Sheet. These
materials can be requested from either
the Safe Drinking Water Hotline, at 1-
800-426-4791, or from the radon
information hotline,  at 1-800-SOS-
RADON.
  EPA solicits public comment on these
issues regarding regulation of radon
under the SDWA.
                         2. Radium and Uranium MCLs. As
                        described above, all radionuclides cause
                        cancer by the same mechanism, i.e.,
                        delivery of ionizing radiation to tissues
                        (in the case of drinking water,
                        internally), and it is therefore possible to
                        make comparisons among them. Several
                        comparisons  may be made in the course
                        of developing regulatory standards
                        including the total radioactivity
                        removed from potable water in pCi/1 or
                        more conveniently, uCi/1 (one million
                        pCi equals one uCi), the pCi/1 (or uCi/1)
                        removed, or reins ede, the effective dose
                        to tissue. These comparisons allow
                        assessment of the relative cost-
                        effectiveness of controlling the different
                        radionuclides subject to today's rule.
                         The control options considered by
                        EPA for radium and uranium range from
                        the contaminant level that can be
                        reliably measured in routine laboratory
                        operations (PQL) to the level
                        representing an approximate 10~4
                        individual lifetime risk level, and for
                        uranium, the level at which kidney
                        toxicity concern arises. EPA also
                        considered the levels to which these
                        contaminants can be treated in drinking
                       water in assessing which control options
                        are technically feasible.
                         The Agency determined that it is
                       technically feasible to  achieve control
                       levels of 5 pCi/1 for radium 226, radium
                       228 and uranium. EPA  then considered a
                       number of cost factors related to the
                       removal of these contaminants. The high
                       cost of removing radium and uranium as
                       compared with radon was especially
                       apparent when the cost per uCi removed
                       from water was estimated. Radon
                       removal cost approximately $20,000 per
                       uCi removed,  where as radium and
                       uranium at the lowest technically
                       feasible levels cost from $2 million to $5
                       million per uCi removed. Even at radium
                       levels equal to the 10~4risk level, the
                       removal cost per uCi was $600,000 to $1
                       million per uCi (EPA, 1991i).  For
                       uranium at the kidney toxicity limit of 20
                       jxg/1 (representing a cancer risk of
                       approximately 10"5), the removal cost
                       was nearly $2 million per uCi. EPA also
                       reviewed the cost per rem removal for
                       these contaminants.  While the cost
                       differences are less dramatic, they are
                       still large, and in the same direction i.e.,
                       the cost per rem of removing radium and
                       uranium is far greater than the cost of
                       removing radon.
                         In assessing the MCL alternatives,
                       EPA also considered the chemical
                       toxicity of uranium to the kidneys.   .
                       While the 10~4 risk level is 170 pCi/1,
                       adverse effects on the kidneys may
                       occur at lower levels for naturally
                       occurring uranium in the environment.
                       EPA estimates that the DWEL for
 uranium is 100 jxg/1, and using a 20%
 RSC, as discussed in section IV above, a
 safe drinking water level would be 20
 fig/1, corresponding to approximately 26
 pCi/1 (using the conversion of 1.3 pCi/
 ftg; this value rounds to 30 pCi/1). This
 value is below the 10~4 lifetime
 individual cancer risk level and is
 protective' for kidney toxicity, the
 limiting adverse health effect level for
 naturally occurring uranium in drinking
 water.
   The SDWA directs EPA to consider
 cost in setting MCLs. The Agency does
 not believe  it would be reasonable to
 establish MCLs that would impose such
 disproportionate costs for removing
 what is effectively the same
 contaminant from drinking water.
 Therefore, EPA proposes to set MCLs
 for radium 226 and radium 228 and  .
 uranium at levels less stringent than
 may be technically feasible (if only
 affordability to large systems was taken
 into consideration). These levels are, for
 radium 226, 20  pCi/1, for radium 228, 20
 pCi/1, and for uranium, 20 jug/1. The    •
 proposed levels will assure that persons
 served by PWS will not be exposed to
 greater than 10"4 lifetime cancer risk,
 and will for uranium also protect against
 possible kidney toxicity.
   Table 21 compares some of the
 important considerations in establishing
 standards that  are cost-effective with
 the same considerations at the lowest
 technically feasible level.
   EPA recognizes that setting radium
 standards at levels less stringent than
 the interim standards may be disruptive
 to some state regulatory programs. The
 interim standard for radium is 5 pCi/1
 for radium 226 and 228 combined.
 Primacy states  have been implementing
 and enforcing this MCL since it was
 effective in 1976, with mixed results. A
 large percentage of water systems with
 radium problems have chronically
 exceeded the radium MCL, and continue
 to do so. States have  been working to
 bring these systems into compliance,
 and some may view a revision of the
 radium MCLs to 20 pCi/1 for radium 226
 and 20 pCi/1 for radium 228 as
 frustrating their program planning and
 expectations. EPA understands these
 concerns and has considered them in its
 deliberations. The Agency believes
 however, that it is appropriate to revise
 these MCLs  in light of the fact that the
 cost of removing radionuclides from
 drinking water by removing uranium
 and radium to the technologically
 feasible limit is disproportionate to the
 cost of removing radon.
  EPA solicits public  comment on this
 approach to  setting MCLs, and on the
MCL levels proposed. EPA also solicits

-------
                Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                                      33101
comments from systems that have
installed or need to install treatment to
meet the current interim standards.
  3. Alternative MCLs. EPA has
generally set MCLs at the lowest
technically achievable level, with cost
considered largely in terms of whether
the standards would be affordable to
large public water systems.
  Key technical information used in
assessing the lowest feasible levels has
been based on engineering and analytic
chemistry capabilities, with affordability
determinations based on the estimated
increase in residential water bills.
  Engineering feasibility is assessed
based on the treatments available as
BAT,  and the occurrence of the
regulated contaminants. The BAT
treatments for these contaminants are,
at maximum efficiency, capable of
achieving 9095  and greater removals for
all of the regulated contaminants. Radon
removal by aeration treatment can
exceed 99% removal. Occurrence of the
contaminants is reviewed in detail in
section III of this notice. The average
radon level in the MRS survey was
about 800 pCi/1, with a maximum of
26,000 pCi/1. Maximum radium 226 and
228 levels in the NIRS survey were both
below 20 pCi/1 (occurrence at higher
levels is based on a statistical projection
of the 1000 data points in NIRS to the
entire country). The maximum uranium
level in NIRS was 88 pCi/1. Based on
treatability and occurrence, radon  could
theoretically be treated to 100 pCi/1 or
lower in most water supply systems,
radium 226 and 228 could be treated to 2
pCi/1 or lower in most water supplies,
and uranium could be treated to 5 pCi/1
or lower as described in Table 20.
  In reviewing analytic capabilities,
EPA identifies the practical quantitation
level, or PQL. This is the level EPA
believes can be measured on a routine
basis in compliance monitoring,  within a
fixed error rate (often  ± 20%-40%), as
described in section V.E. In reviewing
the analytic capabilities, EPA
determined that the radon PQL could be
established at 300 pCi/1, and that
radium 226, radium 228, and uranium
PQLs can be set at 5 pCi/1.
  The cost of treatment for removal of
these contaminants ranges from about
$4 per household per year (for radon) to
$60 per household per year for radium.
These are  costs to large public water
systems serving 50,000 to 75,000,  and
cost to residents of small systems would
be higher.  All of these costs are within
the range that EPA considers to be
affordable for large public water supply
systems.
  Based on these considerations, EPA
would consider the lowest feasible
levels to which these contaminants
could be regulated are 300 pCi/1 for
radon, 5 pCi/1 for radium 226, 5 pCi/1 for
radium 228, and 5 pCi/1 for uranium
(kidney toxicity by uranium is not the
limiting factor here, as it is above) and
15 pCi/1 for adjusted gross alpha. EPA
solicits public comment on these levels
as possible alternative MCLs for the
radionuclides.
                                TABLE 20.—BACKGROUND INFORMATION ON RADIONUCLIDES


PQU(pCi/l) . 	 - 	 • 	 • 	

1x10"* 	 Lifetime risk level (pCi/I) 	 • 	



Rn-222
<100
300
$4
150
27M
195

Ra-226
<2
5
$60
22
890K
8

Ra-228
<2
5
$60
26
100K
2.1

U
<5
5
$20
170
50K
1.6

Alpha
<5
15
$130
n/a
n/a
n/a

    Source: EPA 19911
                        TABLE 21.—COMPARISON OF PROPOSED AND LOWEST FEASIBLE MCL OPTIONS

MCU Options (pCi/i):

Lifetime risk:
Alternate 	 	 • 	
Cases avotded/yr.:
Alternate 	 "••"•••
Fraction of total cases avolded/yr.:
Proposed 	 	 	
Alternate 	
No. Sya affected:

Total S/yn:

S/rom(K);

S/uCi:

Incr. S/caso;


Rn-222
300
300
2X10-"
2X10-4
80
80
6.41
0.41
26,000
26,000
S180M
S180M
$1K
$1K
$20K
$20K
$2.9M
$2.9M

Ra-226
20
5
1X10-1
2X10-5
3
5
0.38
0.63
70
590
$30M
$120M
$1.6K
$5K
$600K
$2M
$23M
$75M

Ra-228
20
5
8X10-=
2X10-5
0.2
0.6
0.03
0.19
40
500
$6M
$55M
$3.9K
$17K
S1.6M
$2M
$50M
$158M

U
20 (j.g/1
5
1X10-5
3X10-"
0.2 1
0.6
0.17
0.33
1500
7200
$55M
$225M
$380K
$700K
$2M
$4M
$57M


Alpha
15
15
n/a
n/a
n/a
n/a
n/a
n/a
130
130
$37M
$37M
n/a
n/a
n/a
n/a
n/a
n/a

    Source: EPA 19911
    1 Approximately 900,000 people also reduced to exposure level with increased probability of kidney toxicity.

-------
33202
Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 /  Proposed Rules
  4. Gross alpha and beta and photon
MCLs. Alpha and beta emitters are a
way of broadly grouping a large number
of radioactive contaminants based on
their radioactive characteristics.
Radioactive isotopes have characteristic
decay patterns which allow them to be
identified as being primarily alpha,  beta
or photon (gamma ray) emitters
(although many compounds decay by a
combination of these routes with one
being predominant). Alpha emitters are
primarily naturally occurring
compounds, although some are man-
made (such as plutonium). Beta emitters
are mostly man-made compounds, but
some are naturally occurring (such as
radium 228 and lead 210). The 1986
amendments to the SDWA direct EPA to
establish MCLs for these two categories
of radioactive contaminants (section
  Because they emit ionizing radiation
as they decay, they are all considered to
be group A human carcinogens, and the
proposed MCLG for both alpha and
beta/photon emitters is zero, as
described in Section IV-C above.
  The other radionuclides proposed for
regulation today all fall into one of these
categories (radium 226, radon and
uranium are alpha emitters, and radium
228 is a beta emitter). EPA has proposed
to set individual MCLs for radon, radium
and uranium because they occur in the
water of an important number of public
water supplies over substantial parts of
the country. This is not true for the
majority of radionuclides. Many of the
other alpha and beta emitters have
never been detected in drinking water,
and others only sporadically. Many of
the naturally occurring radionuclides
may be found in water because they are
radioactive progeny of the more
commonly occurring radionuclides for
which individual MCLs are being
proposed. The man-made radionuclides
may be found in water as a result of
their release from facilities where they
are produced, stored,  used or disposed
of. These could include nuclear power
plants, research or manufacturing
facilities, high or low level radioactive
waste disposal sites, and others.
  There are approximately 2000
nuclides that fall into these categories.
Many of these have very short half-lives,
and are not of concern in water; several
hundred have longer half lives and could
be important. EPA is proposing to
regulate these contaminants as classes
of compounds because they all cause
cancer by the same basic mechanism.
Also, EPA believes that none of them
individually occur with enough
frequency to warrant a national
regulation, but that as groups they are
                       found frequently enough to warrant
                       public health concern, and therefore
                       regulation. EPA further believes that
                       public water systems using water that is
                       known to have the potential to become
                       contaminated with nuclear reactor (or
                       other nuclear facility) releases, by either
                       scheduled or unscheduled release,
                       should monitor for these compounds and
                       that there should be standards in place
                       to protect the public should high levels
                       occur.
                          a. Gross alpha. There is currently an
                       interim MCL for alpha emitters which
                       was set as a screen for the occurrence of
                       both radium 226 and other alpha
                       emitting radionuclides that might be
                       present in drinking water. Few water
                       systems have ever exceeded the gross
                       alpha MCL (except when it is due to
                       high radium levels). The 15 pCi/1 MCL
                       was intended to limit overall exposure
                       to alpha radiation in drinking water, and
                       EPA continues to believe  that it is
                       important to limit overall alpha emitter
                       exposure. EPA is proposing to retain but
                       modify the gross alpha MCL. As
                       discussed in Section IV-G, alpha
                       emitters are carcinogenic, and EPA is
                       proposing to set the MCLG for gross
                       alpha at zero, in accord with EPA's
                       general policy for regulating carcinogens
                       occurring in drinking water.
                         Most alpha emitters in drinking water
                       occur naturally. Alpha emitters other
                       than radium and uranium that have
                       been found in drinking water include
                       polonium and thorium as  discussed in
                       section III-F above. In addition,
                       plutonium and americium may occur.
                       EPA believes the potential for
                       occurrence of these contaminants
                       indicates that a screening standard
                       would be appropriate to restrict the
                       limited exposure that may occur, while
                       not requiring that separate MCLs, with
                       required separate monitoring, be set.
                       The available data indicate that
                       occurrence of alpha emitters other than
                       those specifically regulated (i.e. radon,
                       radium and uranium) is infrequent. EPA
                       believes this limited occurrence means
                       that individual, nationally applicable
                       MCLs are not warranted, but that some
                       mechanism to detect potential
                       occurrence and reduce exposure when
                       alpha emitters do occur is warranted.
                       EPA believes that a gross  alpha MCL
                       would provide a mechanism to detect
                       and reduce exposure to alpha emitters,
                       while not overburdening water systems
                       with monitoring requirements.
                         EPA has reviewed the risks for these
                       contaminants, and discusses them in
                       Section IV-C above and in greater detail
                       in the alpha emitter criteria document.
                       As noted above, the MCLG for alpha
                       emitters is being proposed as zero,
 because all ionizing radiation is
 considered to be carcinogenic. Lifetime
 risks hi the 1X10"4 range for alpha
 emitters in drinking water are 14 pCi/1
 for polonium, 50-125 pCi/1 for various
 thorium isotopes, and 7 pCi/1 for
 plutonium (see appendix C).
   EPA has also reviewed the available
 treatment information to determine
 what levels of alpha emitters can be
 successfully removed. EPA has also
 conducted limited pilot scale studies to
 better determine the treatability of
 polonium (EPA, 1991k). BAT has been
 identified as reverse osmosis. Ion
 exchange, GAG, and coagulation and
 filtration have been shown to remove
 some of these contaminants, but data
 are inadequate to consider any of them
 BAT. RO can remove up to 99% of alpha
 emitters that may be present in drinking
 water.
   The analytic methods for measuring
 alpha emitters is the gross alpha test
 (EPA No.900.0) or gross alpha by
 coprecipitation, when high amounts of
 solids are present. As discussed in
 section V.D, the PQL for gross alpha is
 15 pCi/1, with ±40% error.
   While retaining the gross alpha MCL,
 EPA proposes to revise its approach to
 this standard. Because separate MCLs
 are being proposed for radium and
 uranium, the gross alpha MCL will not
 include them (the current gross alpha
 standard includes radium 226 but
 excludes uranium and radon). The alpha
 emitter MCL will be defined as gross
 alpha, less radium 226, and uranium
 (and not including radon). To avoid
 confusion of the regulatory use  of the
 term "gross alpha" and the laboratory
 measurement that is called gross alpha,
 EPA proposes to designate the MCL as
 "adjusted gross alpha", to indicate that
 compliance with the gross alpha MCL
 would be determined by first measuring
 gross alpha and if the value exceeds the
 MCL, measuring and subtracting out the
 radium 226 and uranium contributions
 (because of the way the test is
 conducted, any radon initially present hi
 a sample would be driven off by the
 sample preparation; therefore, while the
 adjusted gross alpha measure does not
 include radon, neither would radon be
 subtracted from the gross alpha
 measurement, as would radium 226 and
 uranium). EPA proposes that the
 "adjusted gross alpha" MCL would be
gross alpha minus radium 226 and minus
uranium, and proposes that the  adjusted
gross alpha MCL be set at 15 pCi/1. This
MCL would, overall, limit exposure to
other radionuclides and ensure that
risks from alpha emitting radionuclides
would not exceed the 10~4 to 10"6
lifetime risk range. EPA considers this to

-------
                Federal  Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed Rules
                                                                      33103
be the lowest level at which it is feasible
to set the adjusted gross alpha MCL,
bounded by 10"4 lifetime risk.
  EPA recognizes that there could be
situations in which several
radionuclides occur together hi drinking
water. Based on the data available
today, it appears unlikely that
radionuclides will co-occur at levels
near the proposed MGLs. Therefore, the
potential for overall risks to be greater
than 10~4 appears small. EPA solicits
public comment on its proposed MCLs
in regard to possible co-occurrence  of
radionuclides and possible approaches
to ensuring that overall risks do not rise
above the 10~4 level.
  Assessing the impacts of the proposed
adjusted gross alpha MCL is difficult
due to uncertainties in the available
data, and also because of its "screening"
nature. As a worst case, EPA estimates
that up to 130 systems could exceed an
adjusted gross alpha MCL of 15 pCi/1,
and believes the actual number of
systems would be far below that
number. No violators of the  current
gross alpha MCL have been identified in
a search of the EPA compliance data
base.
  b. Beta and photon emitters. There are
over 200 beta and photon emitters
covered by this regulation (see appendix
B). Most of these are man-made isotopes
and are the waste from nuclear power
plants, medical industry, nuclear
weapon development, and other
industries. The Agency regulated the
beta and photon emitters  as a class in
the NIPDWRs with an MCL of 4 mrem
per year effective dose equivalent
(whole body or any organ), and
proposes to retain the interim standard
as a final MCL.
  Strontium-90, strontium-89, cesium-
134, cesium-137, iodine-131,  and cobalt-
60 are the beta emitters with the highest
toxicity. These are also the most likely
to be found in reactor releases or
accidents.
  Ion exchange and reverse osmosis are
capable of removing up to 99% of these
isotopes, with several exceptions. Only
reverse osmosis is capable of removing
iodine. Also, while there is no treatment
for tritium other than use  of an alternate
water source, EPA considers an
alternate water source (including bottled
water) to be BAT for this limited
purpose. Both ion exchange and reverse
osmosis may be used to remove mixed
commercial radionuclides. The
treatment cost varies between $330 to
$540 per household per year for a small
system and between $84 to $230 per
household per year for a large system.
  Beta emitters are measured by the
gross beta method (EPA No. 900.0),
which has PQL of 30 pCi/1.
  At the time of the interim standards,
there was great concern about the
fallout of strontium 90 (and others) from
above-ground nuclear tests. Since the
ban on above-ground tests in 1963,
environmental levels have declined and
the concern now has shifted more
toward water which is vulnerable to
radionuclides released from industrial
and governmental (DOE) facilities and,
to a lesser degree, landfills. Controls are
in place for discharges from these
sources under the Clean Water Act,
RCRA, and NRG and DOE regulations.
These regulations are intended to be
protective of the environment and public
health. The drinking water standard
under these conditions becomes an
adjunct to these release restrictions, and
establishes values which would be used
in case of an accident or unscheduled
release, where these regulations are
violated. EPA nonetheless believes it is
necessary and appropriate to establish
the beta and photon emitter MCL to
ensure protection of public health hi
these circumstances, and is required to
set such a standard by the 1986
amendment to the SDWA, which listed
beta emitters as among the 83
contaminants for which MCLs must be
developed.
  The Agency is proposing to set the
beta MCL at 4 mrem ede per year. The
individual lifetime risk at 4 mrem ede/
year is estimated to be approximately
1X10~4.
  One naturally occurring beta emitter
of potential concern is lead-210. Lead-
210 is the first long lived progeny of
radon-222, and could be anticipated to
co-occur in ground water where radon
occurs. However,  there are few data on
lead-210 occurrence in water, and
modeling exercises of lead movement
through the environment indicate that
low levels (mass)  of lead may bind to
soils and be unavailable to water (EPA,
1986e). Because data on which to base
risk and regulatory impact estimates are
lacking, EPA is proposing to require
unregulated contaminant monitoring for
lead-210, as discussed below, and
consider it for possible regulation in the
future.
G. Proposed Monitoring and Reporting
Requirements
   Compliance monitoring requirements
are being proposed for determining
compliance with the MCLs. In
developing the proposed compliance
monitoring requirements for these
contaminants, EPA considered:
   (1) The likely source of contamination
of drinking water,
   (2) The differences between ground
water and surface water systems,
  (3) The collection of samples which
are representative of consumer
exposure,
  (4) The economic burden of sample
collection and analysis,
  (5) The use of historical monitoring
data to identify vulnerable  systems and
to specify monitoring requirements for
each of the individual systems,
  (6) The limited occurrence of some
contaminants, and
  (7) The need for States to tailor
monitoring requirements to site-specific
conditions.
  A major goal has been to make these
monitoring requirements consistent with
the monitoring requirements for other
regulated drinking water contaminants,
as described in the standardized
monitoring requirements. EPA wants to
develop monitoring requirements that
will meet the statutory goal of ensuring
compliance with the MCLs while
providing efficient utilization of State
and utility resources. The monitoring
program will focus on targeting the
monitoring efforts in individual water
supply systems to the contaminants that
are likely to be present. The general
approach taken by EPA includes:
  • Providing latitude to the States to
target monitoring efforts based on
vulnerability of the system to a
particular contaminant if its occurrence
is not widespread and thus avoiding
unnecessary monitoring efforts.
  • Allowing the use of recent
monitoring data in lieu of new data if
the system has conducted a monitoring
program using reliable analytical
methods.
  • Allowing the use of historical
monitoring data meeting specified
quality requirements and other available
records to make decisions regarding the
vulnerability of a system to
contamination.
  • Requiring all vulnerable systems to
conduct repeat monitoring unless the
system demonstrates that its
vulnerability status has changed.
  • Designating sampling locations  and
frequencies that permit simultaneous
monitoring for all regulated
contaminants, whenever possible and
advantageous.
  • Requiring that samples be taken
during high vulnerability times.
  EPA is proposing to require
monitoring to begin at the start of the
next 3 year period after the regulation is
effective, which is January 1,1996, in
accord with the standardized monitoring
requirements. However, under Section
1445, monitoring, reporting, and
recordkeeping regulations which may be
used to assist in determining compliance
may be made effective on the date that

-------
33104
Federal  Register / Vol. 56, No. 138 / Thursday, July 18,  1991 / Proposed  Rules
 the regulation is finalized. EPA solicits
 public comment on the effective date for
 the monitoring requirements,
 particularly whether monitoring should
 begin before January 1,1996.
  Surface water systems must sample at
 points in the distribution system which
 are representative of each source i.e., at
 each entry point to the distribution
 system which is located after any
 treatment and which is representative of
 each source. The number of samples will
 be determined by the number of sources
 or treatment plants. Sampling must be
 done at entry points to the distribution
 system for ground water systems and
 the number of samples will be
 determined by the number of entry
 points. This approach will make it easier
 to identify possible contaminated
 sources (wells) within a system. In both
 surface and ground water systems, the
 proposed sampling locations are such
 that the same sampling locations may be
 used for the collection of samples for
 other source-related contaminants such
 as the volatile organic chemicals and
 inorganic chemicals, which simplifies
 sample collection efforts.
  Because of the large number of
 regulations for drinking water
 contaminants that have been developed
 in recent years,  EPA recently sought to
 coordinate contaminant monitoring to
 simplify the requirements imposed on
 public water systems. This coordination
 is called the standardized monitoring
 framework. EPA announced this
 framework in January of 1991 (56 FR
 3526-3597, January 30,1991), and held a
 public meeting to discuss the concept
 and solicit public comment. Reaction of
 the water supply industry was generally
 favorable, and EPA has proceeded to
 implement the standardized monitoring
 framework in the context of individual
 rulemakings (56 FR 3526, January 30,
 1991}. The monitoring requirements for
 the radionuclides regulations will rely
 on the basic structure described in the
 documents on standardized monitoring.
 Initial monitoring will begin with the
 compliance period that begins January 1,
 1996, and would be required to be
 completed by January 1,1999. EPA
 solicits public comment on the use of the
 Standardized Monitoring scheme for the
radionuclides regulations.
  The monitoring requirements for the
 different radionuclides would vary
 depending on their likely occurrence.
For example for radon, all ground water
 systems would be required to collect
 one sample from each entry point to the
 distribution system quarterly at first and
annually after compliance is
 established, whereas surface water
                       systems are not required to test for
                       radon.
                         Only systems designated as
                       vulnerable would be required to monitor
                       gross beta for beta and photon emitters.
                       Vulnerability for beta and photon
                       emitters would be determined by states,
                       and would be based on the proximity of
                       the system to potential sources of man-
                       made radionuclides, such as nuclear
                       power facilities, universities or other
                       research facilities, or manufacturing
                       facilities that use radioactive material,
                       or radioactive waste disposal sites (for
                       either high or low level waste). EPA
                       suggests a 15 mile radius around such
                       facilities as the vulnerable area for
                       purposes of requiring gross beta
                       monitoring.
                         MCL exceedences would trigger
                       increased monitoring requirements,
                       which could be reduced to the base
                       monitoring requirements once
                       compliance with the MCL is re-
                       established.
                         Because these contaminants present
                       risks from long-term, chronic  exposure,
                       only community and non-community,
                       non-transient public water supplies
                       would be required to monitor for them.
                         1. Radon.—a. Radon monitoring for
                       surface water supplied systems.
                       Systems relying exclusively on surface
                       water as their water source would not
                       be required to sample for radon.
                       Systems that rely in part on ground
                       water would be considered groundwater
                       systems for purposes of radon
                       monitoring. Systems that use ground
                       water to supplement surface water
                       during low-flow periods would be
                       required to monitor finished water at
                       each entry point to the distribution
                       system for radon during periods of
                       ground water use, according to the
                       groundwater monitoring requirements.
                       Also, groundwater under the influence
                       of surface water would be considered
                       ground water for this regulation.
                         b. Radon monitoring for ground water
                       systems. Systems relying wholly or in
                       part on ground water would be required
                       to sample for radon quarterly for one
                       year at each well or entry point to the
                       distribution system. If the average of all
                       first year samples at each well is below
                       the MCL, monitoring would be reduced
                       to one sample annually per well or entry
                       point to the distribution system. All
                       samples would be required to be of
                       finished water, as it enters the
                       distribution system and after any
                       treatment.
                         c. Radon compliance and increased
                       and decreased monitoring requirements.
                       Compliance would be determined based
                       on an average of 4 quarterly samples in
                       the initial year of monitoring, and
 annual samples in the second and third
 years of the first compliance period. The
 reported values (rather than the bottom
 of the error band associated with the
 measurements) would be averaged
 together; systems with averages
 exceeding 300 pCi/1 at any well or
 sampling point would be deemed to be
 out of compliance. Systems exceeding
 the MCL would be required to monitor
 quarterly until the average of 4
 consecutive samples are less than the
 MCL. Systems would then be allowed to
 reduce monitoring to one sample
 annually per well or sampling point.
 States would be  allowed to reduce
 monitoring requirements to one sample
 per three-year compliance period per
 well or sampling point, if the state
 determines that the system is reliably
 and consistently below the MCL.
 Systems monitoring annually or once
 per three year compliance period that
 exceed the radon MCL in a single
 sample would be required to revert to
 quarterly monitoring until the average of
 4 consecutive samples is less than the
 MCL. Ground water systems with
 unconnected wells would be required to
 conduct increased monitoring only at
 those wells exceeding the MCL.
  EPA is proposing more frequent
 monitoring for radon than for the other
 radionuclides because levels are known
 to vary diurnally and over the course of
 a year. Variability may be 100% or more.
 EPA solicits public comment on the
 proposed radon monitoring
 requirements,  and on the advisability of
 allowing up to nine years between
 samples, and the criteria that might be
 used to identify systems very unlikely to
 exceed the MCL for which monitoring
 once every nine years may be adequate.
  2. Gross Alpha, Radium-226 and
 Uranium. All ground water and surface
 water systems would be required to
 monitor annually for gross alpha, and if
 the gross alpha measurement exceeds
 the MCL for radium 226 and/or uranium,
 specific analyses for the contaminant(s)
 exceeding the MCL would be required.
 Systems would be required to sample
 each well or entry point to the
 distribution system. Samples would be
 of finished water after any treatment.
 Systems exceeding the MCL would be
required to monitor quarterly until four
consecutive samples were less than the
MCL. For systems not exceeding the
MCL after three consecutive annual
samples are taken, sampling would be
reduced to one sample per three year
compliance period. States would be
allowed to reduce monitoring to once
per nine year compliance cycle if the
state determines that a system
consistently and reliably meets the

-------
                Federal Register / Vol. 56,  No. 138  / Thursday, July 18, 1991 / Proposed Rules
                                                                     33105
MCL. Systems with unconnected wells
would be required to conduct increased
monitoring only at those wells
exceeding the MCL.
  Gross alpha measurement would be
used both to determine compliance with
the adjusted gross alpha MCL and as a
screen for radium 226 and uranium,
provided the analytic requirements
described hi section V.D are met. These
requirements include appropriate
calibration of equipment to ensure that
neither radium 228 or uranium are
underestimated by the screen.
Compliance determinations for adjusted
gross alpha, radium 226 and uranium
based on gross alpha measurements are
listed hi Figure 2. Adjusted gross alpha
is defined as the gross alpha
measurement less radium 226 and less
uranium. Because the adjusted gross
alpha MCL is less than the radium 226
and uranium MCLs, one or both of these
may need to be specifically analyzed to
determine adjusted gross alpha
compliance even though the gross alpha
screen indicates that both the radium
226 and uranium MCLs have been met
(i.e., if the gross alpha is between 15 and
20 pCi/1).
  Systems with gross alpha less than
the radium 226 or uranium MCLs would
be considered to be in compliance with
those respective MCLs. Specific
analyses of either or both contaminants
would be required if the gross alpha
measurement exceeds the respective
MCL.
  For adjusted gross alpha, radium 226
and uranium, compliance would be
based on the average of an initial
sample exceeding the MCL and a
confirmation sample [as the reported
values, not the lower bound of the error
band associated with the measurement).
  EPA solicits public comment on the
proposed radium 226 and uranium
monitoring, and use of the gross alpha
screen for these contaminants,
especially hi light of the fact that the
uranium MCL is proposed to be set
based on mass rather than activity
measurements.
  3. Radium-228. All ground water and
surface water systems would be
required to monitor annually for radium
228.  Systems would be required to
sample each well or entry point to the
distribution system. Samples would be
of finished water after any treatment.
Systems exceeding the MCL would be
required to monitor quarterly until four
consecutive samples were less than the
MCL. For systems not exceeding the
MCL, sampling would be reduced to one
sample per three year compliance period
after three consecutive annual samples
are below the MCL. States would be
allowed to reduce monitoring to one
sample per nine year compliance cycle if
the state determines that a system
consistently and reliably meets the
MCL. Systems with unconnected wells
would be required to conduct increased
monitoring only at those wells
exceeding the MCL.
  Gross beta measurement would be
allowed to serve as a screen for radium
228 levels. Systems with gross beta
levels less than the radium 228 MCL
would be considered to be in
compliance with the radium-228 MCL.
Systems with gross beta levels
exceeding the radium-228 MCL would be
required to measure radium-228
specifically.
  For radium-228, compliance would be
based on the average of an initial
sample exceeding the MCL and a
confirmation sample (as the reported
values, not the lower bound of the error
band associated with the measurement).
  4. Beta and photon emitters. Because
of revisions in the estimated drinking
water concentrations of various beta
and photon emitters that correspond to
a yearly dose of 4 mrem ede, EPA is
proposing to revise and simplify the
monitoring requirements for beta and
photon emitters. The revised estimates
in general allow for less specific
monitoring and greater reliance on the
gross beta screen. In addition, because
of the special vulnerability
circumstances which could result in the
presence of man-made beta emitters in
drinking water, monitoring more
frequent than that required for other
contaminants under the standardized
monitoring program is being proposed.
  The current gross beta monitoring
program requires all vulnerable PWS
and all systems serving 100,000 or more
persons to perform a screen plus specific
analyses for several contaminants. EPA
proposes to revise these requirements so
that only vulnerable systems would be
required to perform gross beta
monitoring. States would make the
vulnerability determination for each
PWS, and it would be  based on the
proximity of the water source for the
system to facilities using or producing
radioactive materials.  EPA suggests that
all systems within a 15 mile radius of
these facilities be considered
vulnerable, as well as  systems using a
water source  clearly influenced by such
a facility. All systems  using water that
could be influenced by releases (either
scheduled or unscheduled) from
facilities such as nuclear power plants,
Department of Energy nuclear facilities,
Nuclear Regulatory Commission
licensees, low or high  level nuclear
waste storage or disposal facilities, or
other facilities using or making
radioactive material should be
considered vulnerable. Monitoring could
be required of either surface or ground
water dependent systems, depending on
their vulnerability.
  EPA considered two gross beta
monitoring programs. Under the first
alternative, the current 50 pCi/1 screen
for presumptive compliance, along with
additional specific monitoring for tritium
and strontium 90 would be required. If
the 50 pCi/1 screen were met, and
tritium and strontium were individually
and combined below the 4 mrem ede
value, the system would be considered
to be in compliance. The beta screen
would be required quarterly and the
tritium and strontium would be required
annually, as described in Figure 3.
Under the second alternative, the beta
screen would be set at the gross beta
PQL of 30 pCi/1, and only specific
analysis of tritium would be required.
The screen would be required quarterly
and the tritium analysis annually.
Because of the vulnerable status of
these systems, no reduced monitoring
would be allowed. Under either
alternative, water suppliers would be
required to identify the particular
contaminants present if the screen is
exceeded, and add the estimated doses
including tritium and strontium 90 under
the first alternative to ensure that the 4
mrem ede MCL is not exceeded. The
values in Appendix B would be used to
perform this calculation. EPA believes
that either of these monitoring plans
would ensure the safety the public
served by vulnerable water supplies.
   EPA proposes to establish the first
alternative, of retaining the 50 pCi/1
screen for presumptive compliance with
the gross beta MCL and specific
analyses for tritium  and strontium 90
(because 50 pCi/1 would not adequately
screen for tritium and Sr-90 at the 4
mrem ede level). EPA solicits public
comment on reducing the screen to 30
pCi/1 and eliminating the strontium 90
measurement.
   5. Monitoring schedule. In order to
moderate demand on analytic
laboratories,  the monitoring
requirements for determining
compliance with these regulations
would be phased-in over a 3 year period.
States would determine the  schedule for
phasing in monitoring, but all systems
would be required to have performed
their first year of sampling by the end of
the first 3 year compliance period (i.e,
December 31,1998).
   6.  Grandfathering data. Interim MCLs
have been in place and analytic
methods available for radium, gross
alpha and beta and photon emitters
since 1976. Validated analytic methods
for other radionuclides, including

-------
33106
Federal Register / Vol. 56,  No. 138  /  Thursday, July 18, 1991 / Proposed Rules
uranium, have also been available since
then. Most water supply systems that
would be covered by these proposed
regulations have been monitoring for the
regulated contaminants for several
years. Data collected in compliance with
the interim MCL requirements (i.e.,
analyses by certified laboratories)
would be allowed to be used to
determine compliance with the proposed
MCLs. While no EPA-approved radon
analytic method has been available,
EPA recognizes that many water
supplies have conducted some radon
monitoring in recent years. Data on
radon occurrence generated  using
                       methods and with laboratory
                       performance similar to those proposed
                       here would be allowed to be used to
                       determine compliance, at the discretion
                       of the State.
                         7. Monitoring for unregulated
                       contaminants. As discussed above,
                       available data are inadequate to
                       determine whether lead-210 occurs
                       frequently enough to warrant public
                       health concern. EPA is therefore
                       proposing to require all community and
                       non-community, non-transient public
                       water systems to collect one sample
                       from each well or entry point to the
                       distribution system, after any treatment,
and analyze the sample for lead-210.
States may require systems to collect
one confirmation sample. All regulated
systems would be required to collect
and analyze one sample for lead-210, so
that adequate data on which to assess
exposure may be obtained. EPA solicits
public comment on this proposed
monitoring for unregulated
contaminants.

  EPA solicits public comment on the
proposed monitoring requirements
described above.

BILLING CODE 6560-50-M

-------
             Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                     33107
                                       315

                    FIGURE 2.  GROSS ALPHA SCREENING
    COMPLIANCE
        WITH
         AGA
      Ra-226
     URANIUM
                NO
                          MEASURE GROSS ALPHA
NO
     GROSS  ALPHA
       >15  pCi/1
                                     YES
                              MEASURE Ra-226
                              &/OR URANIUM
                          ADJUSTED GROSS ALPHA
                                 <15 pCi/1
                       NO
                        YES
                NON-
             COMPLIANCE
                                       Ra-226
                                       <20 pCi/1
                                       URANIUM
                                       <30 pCi/1
                                                             UranS
                                                                    m
                          COMPLIANCE
AGA- Adjusted Gross Alpha

-------
33108
Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                316

          FIGURE 2.  GROSS ALPHA  SCREENING (Continued^
      NO
               MEASURE GROSS  ALPHA
                   GROSS ALPHA
                     >30 pCi/1
                    MEASURE Ra-226
                    &/OR URANIUM
                 ADJUSTED GROSS ALPHA
                           pCi/1
             NO
                          YES
                                                     Ra-226
                                                     <20  pCi/1
                                         URANIUM
                                         <30 pCi/1
                                                    COMPLIANCE
      NON-
   COMPLIANCE
                            COMPLIANCE

-------
            Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                                            33109
                                    317

                 FIGURE 3.  GROSS  BETA SCREENING OPTIONS
Option 1; Higher Screening Level
   YES
    YES
          MONITOR QUARTERLY
               MEASURE
             GROSS BETA
          Is  BETA <50 pCi/L
                    NO
          ANALYZE TO IDENTIFY
            INDIVIDUAL BETAs;
              SUM DOSES
Is ANNUAL DOSE
   FROM BETAS
   < 4mrem/yr
                    NO
            NON-COMPLIANCE
              COMPLIANCE
                                     MONITOR ANNUALLY
                                    MEASURE TRITIUM
                                       AND Sr~90
                                       IS TRITIUM
                                     <  60,000 pCi/1
                                              NO
                                              IS Sr-90 <42  pCi/1
                                                         NO
                                     NON-COMPLIANCE
                                       COMPLIANCE
YES
                                                                    YES

-------
331W
 Federal Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules
                                     318

           FIGURE  3.   GROSS BETA  SCREENING OPTIONS (Continued^

Option 2: Low Screening Level
    YES
    YES
           MONITOR QUARTERLY
               MEASURE
             GROSS  BETA
          IS BETA  <30 pCi/1
                     NO
          ANALYZE  TO IDENTIFY
           INDIVIDUAL BETAS;
             SUM DOSES
IS ANNUAL  DOSE
  FROM BETAS
 <4 mrem ede/yr
                     NO
                 NON-
              COMPLIANCE
              COMPLIANCE
                                      MOKITOk
                                     MEASURE TRITIUM
                                       IS TRITIUM
                                     < 60,000 pCi/1
                                                NO
                                             NON-
                                          COMPLIANCE
COMPLIANCE
                 YES
BILLING CODE 6560-50-C

-------
                Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 / Proposed Rules          33111
H. State Implementation.
  The Safe Drinking Water Act provides
that States may assume primary
implementation and enforcement
responsibilities. Fifty-four out of 57
jurisdictions have applied for and
received primary enforcement
responsibility (primacy) under the Act.
To implement the Federal regulations for
drinking water contaminants, States
must adopt their own regulations which
are at least as stringent as the Federal
regulations. This section of today's
proposal describes the regulations and
other procedures/policies that States
must adopt to implement today's
proposed rule. EPA has recently revised
its program implementation
requirements of 40 CFR part 142, on
December 20,1989 (54 FR 52126), and on
June 3,1991 (56 FR 25046).
  To implement today's proposed rule,
States will be required to adopt the
following regulatory requirements:
When they are promulgated:  § 141.25,
Radionuclide Sampling and Analytical
Requirements; § 141.32, General public
notice requirements; § 141.44, Special
monitoring for radionuclides; and
§ 141.64, MCLs for Radionuclides.
  In addition to adopting drinking water
regulations no less stringent than the
Federal regulations listed above, EPA is
proposing that States adopt certain
requirements related to this regulation hi
order to have their program revision
application approved by EPA. In various
respects the proposed NPDWRs provide
flexibility to the State with regard to
implementation of the monitoring
requirements by this rule.
  Today EPA is also proposing changes
to State recordkeeping and reporting
requirements. EPA's proposed changes
are discussed below. EPA requests
comments on these proposed
requirements.
  1. Special primacy requirements. To
ensure that the State program includes
all the elements necessary for an
effective and enforceable program, the
State's request for approval must
contain the following:
  (1) If the State issues waivers, the
procedures and/or policies the State will
use to conduct and/or evaluate
vulnerability assessments;
  (2) The procedures/policies the State
will use to allow a system to decrease
its monitoring frequency; and
  (3) A plan that ensures that each
system monitors by the end of each
compliance period.
  2. State recordkeeping. The current
regulations in § 142.14 require States
with primary enforcement responsibility
to keep records of analytical results to
determine compliance, system
inventories, sanitary surveys, State
approvals, enforcement actions, and the
issuance of variances and exemptions.
In this rule, States would be required to
keep additional records of the following:
(1) Any determination of a system's
vulnerability to contamination by beta
and photon emitters due to proximity of
an emitting source; and (2) any
determination that a system can reduce
monitoring for gross beta, uranium,
radium 226 or 228 or increase monitoring
frequency. The records must include the
basis for the decision, and the repeat
monitoring frequency.
  Systems that are located within a 15
mile radius of a nuclear facility, or
hospitals or other locations that use,
store or dispose of radioactive material
should be considered vulnerable to
contamination, and therefore, monitored
more closely.  Systems that are found not
to be vulnerable to contamination will
be listed as such. This information will
be available to EPA for review in a
similar manner to current records kept
by the State.
  3. State reporting. EPA currently
requires in § 141.15 that States report to
EPA information such as violations,
variances and exemption status,
enforcement actions,  etc. EPA proposes
in this notice that in addition to the
current reporting requirements, States
report to EPA:
  (1) A list of all systems on which the
State conducted a vulnerability
assessment, the dates of those
assessments, the results of that
assessment, and the basis for that
determination; and
  (2) A list of all systems on which the
State is requiring repeat monitoring for
Gross beta particle and photon emitters,
the results of that assessment, and the
basis for that determination.
  EPA believes that the State reporting
requirements  contained in this proposal
are necessary to ensure effective
oversight of State programs. Public
comments on these proposed State
reporting requirements are requested.
EPA particularly requests comments on
whether the proposed reporting
requirements  are appropriate.

I. Variances and Exemptions
  1. Variances. Under section
1415(a)(l)(A) of the SDWA, a State
which has primary enforcement
responsibility (i.e., primacy), or EPA as
the primacy agent,  may grant variances
from MGLs to those public water
systems that cannot comply with the
MCLs because of characteristics of the
water sources that are reasonably
available. At the time a variance is
granted, the State must prescribe a
compliance schedule and may require
the system to implement additional
control measures. The SDWA requires
that variances may only be granted to
those systems that have installed BAT
(as identified by EPA). However, in
limited situations a system may receive
a variance if it demonstrates that the
BAT would only achieve a de minimis
reduction in contamination (see
§ 142.62(c)). Furthermore, before EPA or
a State may grant a variance, it must
find that the variance will not result in
an unreasonable risk to health to the
public served by the public water
system. The levels representing an
unreasonable risk to health for each of
the contaminants in this proposal will
be addressed in subsequent guidance
(see discussion below). In general, the
unreasonable risk to health (URTH)
level would reflect acute and subchronic
toxicity for shorter-term exposures and
high carcinogenic risks for long-term
exposures (as calculated using the
linearized multistage model in
accordance with the Agency's risk
assessment guidelines; See URTH
Guidance, 55 FR 40205, October 2,1990).
   Under section 1413(a)(4), States that
choose to issue variances must do so
under conditions, and in a manner,
which are no less stringent than EPA
allows in Section 1415. Of course, a
State may adopt standards that are
more stringent than the EPA standards.
Before a State may issue a variance, it
must find that the system is unable to (1)
join another water system, or (2)
develop another source of water and
thus comply fully with all applicable
drinking water regulations.
   EPA specifies BATs for variance
purposes. EPA may identify as BAT
different treatments under section 1415
for variances than BAT under section
1412 for MCLs. EPA's section 1415 BAT
findings may vary depending on a
number of factors, including the number  •.
of persons served by the public water
systems, physical conditions related to
engineering feasibility, and the costs of
compliance with MCLs.
   Section 1415 Best Available
Technology for Radionuclides. Table 22
shows the BATs that EPA is proposing
for variance purposes under section 1415
for radionuclides. EPA has not proposed
coagulation/filtration or lime softening
as BAT for small systems (i.e., those
systems <500 connections) for the
purpose of granting variances because
they are not technologically feasible for
small systems, as discussed below.

-------
 33322         Federal Register / Vol.  56, No. 138 / Thursday,  July 18, 1991 / Proposed Rules
    TABLE 22.—PROPOSED BATs FOR
    VARIANCES UNDER SECTION 1415
Contaminant
Radon 222 	
Radium 226 	
Radium 228 	
Uranium (N) .„ 	
Alpha particle emitters 	
Beta particle and Photon emitters 	
BAT
1.
2, 3, 4
234
2345
3.
3, 6.
  Key to BATs:
  1= Aeration: Packed Tower, spray, slat tray and
other forms.
  2=lon exchange.
  3=Reverse osmosis.
  4=Lime softening; except for systems serving
^ 500 connections.
  5=Coagulation/filtration; except for systems serv-
ing ^500 connections.
  6=Mixed bed ion exchange.

  Coagulation/filtration and lime
softening for radionuclides (i.e.,
uranium, radium-226 and radium-228)
involve a greater degree of complexity
than is required for removing
conventional contaminants (i.e.,
turbidity removal). These differences
result in increased operating tune and
level of expertise needed to operate
coagulation/filtration and lime softening
systems. Specific differences include: (a)
Generally higher pH requirements for
lime softening removal of radium and
specific pH control for coagulation of
uranium; (b) higher doses of chemical
coagulants or lime for precipitation of
radionuclides than for conventional
turbidity removal or lime softening,
which can complicate treatment
operations with respect to chemical
supply, and waste by-product (sludge)
management; and (c) larger
sedimentation basins and possible two-
stage processes (one for turbidity
softening and one for radionuclides
precipitation). Consequently,
coagulation/filtration and lime softening
treatment  are considered too complex in
terms of operating time and levels of
technical and managerial expertise
usually available at  small systems.
  Costs of installing and operating some
of the BATs listed in Table 22 (reverse
osmosis and ion exchange) are high for
small systems relative to costs for large
systems, as shown by EPA estimates in
tables 7 through 9. EPA is requesting
comment on these technologies as BAT
for variance purposes for small systems.
EPA is continuing to evaluate what
costs are reasonable for public water
systems and hi this regard, commenters
are  encouraged to provide a basis for
their statements on what should
constitute  BAT for small systems.
  With regard to BAT established under
section 1415, EPA is  requesting comment
on:  (1) Whether other technologies
should be  considered BAT under section
1415 for radionuclides; (2) whether it is
appropriate to exclude coagulation/
filtration and lime softening for small
systems; and (3) the appropriateness of
reverse osmosis (RO) and ion exchange
as BAT under section 1415 for small
systems. EPA notes that RO offers the
benefit of multiple contaminant removal
and desalting, which makes RO
technology especially attractive for
some drinking water systems, including
small systems. EPA also notes that ion
exchange offers the benefit of water
softening (i.e., removal of hardness)
where hard water conditions prevail.
  Use ofPOU devices and bottled
water. Under section 1415(a)(l)(A)(ii),
the State is to prescribe a schedule for
implementation of any additional
control measures it may require. The
State may require the use of POU
devices, bottled water, or other
mitigation measures as an "additional
control measures" during the period of a
variance, as a condition to receiving the
variance, if an unreasonable risk to
health exists. The use of POU devices
and bottled water would not be allowed
for radon; only point of entry devices
would be allowed for radon.
  POU devices  fail to treat water for the
most significant risk from radon in
water, the inhalation risk. EPA also
recognizes that the use of POU devices
to reduce levels of radon in water could
present problems of disposal of the
devices when their useful life  is over. To
prevent potential disposal problems,
and to ensure that treatment required
under variance  provisions reduces risks,
EPA is proposing to disallow the use of
POU devices for radon for granting
variances. Public comment on this
proposed disallowance of POU devices
to remove radon is requested.
  2. Exemptions. Under Section 1416(a),
EPA or a State may exempt public water
systems from any requirements
respecting an MCL or treatment
technique requirements of an NPDWR, if
it finds that (1) due to compelling factors
(which may include economic factors),
the PWS is unable to comply with the
requirement; (2) the  exemption will not
result in an unreasonable risk to human
health; and (3) the PWS was in
operation on the effective date of the
NPDWR, or for  a system which was not
in operation by  that date, only if no
reasonable alternative source  of
drinking water is available to  the new
system.
  If EPA or a State grants an exemption
to a public water system, it must at the
same  time prescribe a schedule for
compliance (including increments of
progress) and implementation of
appropriate control measures that the
State requires the system to meet while
the exemption is in effect. Under section
1416(2)(A), the schedule must require
compliance within one year after the
date of issuance of the exemption.
However, section 1416(b)(2)(B) states
that EPA or the State may extend the
final date for compliance provided in
any schedule for a period not to exceed
a total of three years, if the public water
system is taking all practicable steps to
meet the standard and one of the
following conditions applies: (1) The
system cannot meet the standard
without capital improvements which
cannot be completed within the period
of the exemption; (2) in the case of a
system which needs financial assistance
for the necessary implementation, the
system has entered into an agreement to
obtain financial assistance; or (3) the
system has entered into an enforceable
agreement to become part of a regional
public water system. For public water   ,
systems which do not serve more than
500 service connections and which need
financial assistance for the necessary
improvements, EPA or the State may
renew an exemption for one or more
additional two-year periods if the
system establishes that it is taking all
practicable steps to meet the
requirements noted above. Section
1416(b)(2)(C).
  Under section 1416(d), EPA is required
to review State-issued exemptions at
least every three years and, if the
Administrator finds that a State has, in
a substantial number of instances,
abused its discretion in granting
exemptions or failed to prescribe
schedules in accordance with the statute
after following various procedures, the
Administrator may revoke or modify
those exemptions and schedules. EPA
will use these procedures to strictly
scrutinize exemptions from the MCLs
granted by States and, if appropriate,
will revoke or modify exemptions
granted.
  As a condition for receiving an
exemption, the State may require the
use of POU devices or bottled water for
the duration of the exemption. The
conditions are the same as those
referenced in the variance section.
  3. Unreasonable risks to health
(URTH). As a part of the variance and
exemption granting process, States must
determine whether granting such a
variance or exemption will pose an
unreasonable risk to the health of the
population served. While the granting of
variances and exemptions, and the
inherent URTH assessment, are State
determinations, they occur within the
overall context of State primacy and
EPA oversight of the State's
administration. EPA has therefore

-------
                Federal Register /  Vol. 56.  No. 138  / Thursday, July 18, 1991  /  Proposed Rules
                                                                       33113
developed guidance to assist States in
making URTH determinations (EPA,
1990k), and published a draft of the
guidance for public comment. For
carcinogens, the draft guidance
recommends that URTH be set at the
top of EPA's risk range that is generally
considered acceptable, 10~4 lifetime risk.
Because EPA is proposing to regulate
these contaminants at the most cost-
effective level, bounded by 10~4 risk, the
URTH values could be equal to the
proposed MCLs, except for adjusted
gross alpha and uranium. Adjusted gross
alpha is a screening MCL; an URTH
should not be considered to exist unless
the individual contaminants in the
adjusted gross alpha sample exceed a
10""* risk. Uranium is being regulated
based on its kidney toxicity; URTH
guidance would need to be developed
for uranium based on this toxic end
point.
  EPA solicits public comment on this
approach to establishing URTH
guidance for radionuclides.
VI. Public Notice Requirements
  Under section 1414{c)(l) of the  Act,
each owner or operator of a public
water system must give notice to
persons served by it of (1) any violation
of any MCL, treatment technique
requirement, or testing provision
prescribed by an NPDWR; (2) failure to
comply with any monitoring requirement
under section 1445(a] of the Act; (3)
existence of a variance or exemption;
and (4) failure to comply with the
requirements of a schedule prescribed
pursuant to a variance or exemption.
  The 1986 amendments required that
EPA amend its current public
notification regulations to provide for
different types and frequencies of notice
based on the differences between
violations which are intermittent or
infrequent and violations which are
continuous or frequent, taking into
account the seriousness of any potential
adverse health effects which may be
involved. EPA promulgated regulations
to revise the public notification
requirements on October 28,1987 (52 FR
41534). The revised regulations state
that violations of an MCL, treatment
technique or variance or exemption
schedule ("Tier 1 violations") contain
health effects language specified by EPA
which concisely and in non-technical
terms conveys to the public the adverse
health effects that may occur as a result
of the violation. States and water
utilities remain free to add additional
information to each notice, as deemed
appropriate for specific situations. This
proposed rule contains specific health
effects language for the contaminants
which are in today's proposed
rulemaking. EPA believes that the
mandatory health effects language is the
most appropriate way to inform the
affected public of the health
implications of violating a particular
EPA standard. The proposed mandatory
health effects language in § 141.32(e)
describes in non-technical terms the
health effects associated with the
proposed contaminants. Public comment
is requested on the proposed language.

VII. Economic Impacts and Benefits

  Executive Order 12291 requires EPA
and other regulatory Agencies to
perform a Regulatory Impact Analysis
(RIA) for all "major" regulations. Major
regulations are those which impose a
cost of $100 million or more on the
national economy,  or meet other criteria.
EPA has  determined that this proposed
rule would be a major rule under the
Executive Order, and has accordingly
prepared an RIA which assesses the
costs and benefits of the proposed
regulations (EPA, 199li). This regulation
has also been reviewed by the Office of
Management and Budget and their
comments are available in the public
docket.
  Table 23 presents a summary of the
results of the RIA. Approximately 28,000
public water systems would be required
to install treatment or take other actions
to comply with the proposed MCLs for
these radionuclides. Total national costs
would be approximately $310 million per
year.
                      TABLE 23.—NATIONAL COSTS AND BENEFITS OF PROPOSED RADIONUCLIDES MCLs

Proposed MCL (b) 	 	 	 - 	
Systems affected 	 	 	 	 	 .. .. 	 	 	
Tfnatrrtont cost.. 	
Total caplial (SM) 	
Annual O&M (SM) 	 	
Total annual cost (SM) 	 	
Cancer cases avoWed/yr. 	 . .
Monitoring (SM/Yr) (0 	
Stato Implementation 	
Initial (SM) 	 ... . .
Annual (SM) 	
Annual household cost by system size:
Vary SmaH (25-500) 	 	
Small (501-3,300) 	 	 	
Medium (3,301-10,000) 	
Large (over 10,000) 	

Rn-
222
300
26,000

1,600
70
180
80
5

NA
NA
120
30
7
5

Ra-
226
20
70

190
20
30
3
0.003

NA
NA
630
150
90
60

Ra-228
20
40

40
3
6
0.2
0.89

NA
NA
650
150
90
60

Uranium
20(c)
1,500

350
30
60
0.2
0.003

NA
NA
580
180
80
40

AGA (a)
15
130

230
20
40
(e)
0.64

NA
NA
770
340
200
140

Beta
emitters
4(d)
0

0
0
0
0
0.25

NA
NA
0
0
0
0

Total

28,000

2,400
150
310
84
7

15-28
10-19





   (a) Adjusted gross alpha.
   (b) MCLs are expressed in pCi/L unless otherwise noted.
   icj MCL for uranium is expressed in ug/L.
   (d) MCL for beta emitters is expressed in mrems ede/year.
   (e) Number of cases avoided per year is in the range of 0.2 to 1.4. The low end of the range is based on the risk factor associated with thorium-232; the high end
Is basod on polonium-210 risk. Actual occurrence is likely to be characterized by a mix of several isotopes.
   (0 Gross alpha is used as a screen for radium-226 and uranium.
   Note: Total may not add due to rounding.
  A large proportion of the water
systems affected by this regulation
would be small systems serving fewer
than 500 people. Costs to households
vary considerably over the range of
system sizes that would be covered by
the proposed regulations, with smaller
systems having higher costs, because
these systems  do not benefit from the
engineering economies of scale that
large systems have. In the smallest of
these systems (25 to 100 people), annual
residential water bills could increase by
$700 to $800 for treatment of radium or
uranium. EPA recognizes that these

-------
331U
Federal Register  / Vol. 56,  No.  138 /Thursday, July 18, 1991  /  Proposed Rules
costs could prove very difficult to afford
for small systems. Exemptions may be
available through States to provide
small systems with additional time to
develop financing for water treatment as
described in section V.I.2.

A. Regulatory Flexibility Analysis

  The Regulatory Flexibility Act
reguires EPA to consider the effect of
regulations on small entities, 5 U.S.C.
602 et seq. If there is a significant effect
on a substantial number of small
entities, the Agency must prepare a
Regulatory Flexibility Analysis which
describes significant alternatives that
would minimize the impact on small
entities. An analysis of the impact of the
proposed radionuclides rule on small
water systems is included in the RIA
supporting this rule. The Administrator
has determined that the proposed rule, if
promulgated, will have a significant
effect on a substantial number of small
entities.

B, Paperwork Reduction Act

  The information collection
reguirements  in this proposed rule have
been submitted for approval to the
Office of Management and Budget
(OMB) under the Paperwork Reduction
Act, 44 U.S.C. 3501 et  seq. An
Information Collection Request
document has been prepared by EPA
(ICR No. 0270) and a copy may be
obtained from Sandy Farmer,
Information Policy Branch, (PM-223Y),
U.S. Environmental Protection Agency,
401M Street,  SW., Washington, DC
20460, or by calling  (202) 382-2740.
  The total public reporting burden for
this collection of information is
estimated to be 674,517 hours, with an
average of 4.7 hours per response,
including time for reviewing
instructions, searching existing data
sources, gathering and maintaining the
data needed,  and completing and
reviewing the collection of information.
  Send comments regarding the burden
estimate or any other  aspect of this
collection of information, including
suggestions for reducing this burden,  to
Chief, Information Policy Branch, (PM-
223Y), U.S. Environmental Protection
Agency, 401 M Street, SW., Washington,
DC 20460; and to the Office of
Information and Regulatory Affairs,
Office of Management and Budget,
Washington, DC 20503, marked
"Attention: Desk Officer for EPA". The
final rule will respond to any OMB or
public comments on the information
collection requirements contained in  this
proposal.
                         VIII. References
                         132 Cong. Rec. S6287, 6294 (May 21,1986).
                         Agency for Toxic Substances and Disease
                             Registry (ATSDR). Toxicological Profile
                             for Radium. Atlanta, GA: Agency for
                             Toxic Substances and Disease Registry,
                             U.S. Public Health Service. (December
                             1990). [ATSDR, 1990]
                         American Public Health Association;
                             American Water Works Association;
                             Water Pollution Control Federation.
                             Standard Methods for the Examination
                             of Water and Wastewater, 17th ed.,
                             (1989). [APHA, 1989]
                         American Society for Testing and Materials,
                             1989 Annual Book of ASTM Standards,
                             Water and Atmospheric Analysis, vol.
                             11.02, Philadelphia, PA (1989) [ASTM,
                             1989]
                         Bean, J.A. et al. Drinking water and cancer
                             incidence in Iowa. II. Radioactivity in
                             drinking water. Am.  J. Epidemiol.
                             ll;6:924-932 (1982). [Bean et. al., 1982]
                         Brandom, W.F. et al. Chromosome
                             aberrations as a biological dose-response
                             indicator of radiation exposure in
                             uranium miners. Radiat. Res. 76:159-171
                             (1978). [Brandom, 1978]
                         Brock v. Pierce County, 476 U.S. 253 (1986)
                         Correia, J.A. et al. The kinetics of ingested
                             Rn-222 in humans determined from
                             measurements with Xe-133. MA General
                             Hospital, Boston, MA, unpublished
                             report. (August 1987). [Correia, et al.,
                             1987]
                         Correia, J.A. et al. The kinetics of ingested
                             Radon 222 in humans determined from
                             measurements with Xe-133. Project
                             Summary. PB 88-145297/AS (1988).
                             [Correia et al., 1988]
                         Crawford-Brown, D.J. A Calculation of Organ
                             Burdens, Doses and Health Risks from
                             Rn-222 Ingested in Water (Draft) (Feb.
                             11,1990). [Crawford-Brown 1990].
                         Crawford-Brown, D.J. Cancer Fatalities from
                             Waterborne Radon (Rn-222). Risk
                             Analysis, Vol. 11, No. 1,135-143, (1991).
                             [Crawford-Brown, 1991]
                         Domingo, J.L. et al. The Developmental
                             toxicity of uranium in mice. Toxicology
                             55:143-152. (1989). [Domingo et al., 1989a]
                         Domingo, J.L. et al. Evaluation of the
                             perinatal and postnatal effects of
                             uranium in mice upon oral
                             administration. Arch. Environ. Health
                             (1989) 44:395. [Domingo et al., 1989b]
                         Dunning, D.E. et al. A Combined
                             Methodology for Estimating Dose Rates
                             and Health Effects from Exposure to
                             Radioactive Pollutants (draft). Prepared
                             under Interagency Agreement EPA-78-
                             D-X0231 by Oak Ridge National
                             Laboratory (December 1980). [Dunning et
                             al., 1980]
                         Dunning, D.E. et al. An Assessment of Health
                             Risk From Radiation Exposures. Health
                             Physics 46:1035-1051. [Dunning et al.,
                             1984]
                         Dupree, E.A. et al. Mortality among workers
                             at a uranium processing facility, the
                             Linde Air Products Company Ceramics
                             Plant, 1943-1949. Scand. J. Work Environ.
                             Health Vol. 13:100-107 (1987). [Dupree et
                             al., 1987]
Ershow, A.G. and Cantor, K.P. Total water
    and tapwater intake in the United States:
    Population-based estimates of quantities
    and sources. Report prepared under
    National Cancer Institute Order #263-
    MD-810264. (1989). [Ershow and Cantor,
    1989]
Evans, R.D. et al. Radium metabolism in rats
    and the production of osteogenic
    sarcoma by experimental radium
    poisoning. Am. J. Roetgenol. Radium
    Therapy. 52:353-373 (1944). [Evans et al.,
    1944]
Federal Register. Vol. 40 No. 159. National
    Interim Primary Drinking Water
    Regulations; Final Rule. (Aug. 14,1975)
    pp. 34324-34328 [40 FR 34324, Aug. 14,
    1975]
Federal Register. Vol. 41 No. 133. National
    Interim Primary Drinking Water
    Regulations; Final Rule. (July 9,1976), pp.
    28402-28409. [41 FR 28402, July 9,1976]
Federal Register. Vol. 50, No. 219. National
    Primary Drinking Water Regulations;
    Volatile Organic Chemicals; Proposed
    Rule (Nov. 13,1985), 46902-46932.  [50 FR
    46902, Nov. 13,1985]
Federal Register. Vol. 50, No. 219. National
    Primary Drinking Water Regulations;
    Synthetic Organic Chemicals, Inorganic
    Chemicals and Microorganisms,
    Evaluation of Health Effects and
    Determination of RMCLs; Proposed Rule
    [Nov. 13,1985), 46936-47025. [50 FR
    46936, Nov. 13 1985]
Federal Register. Vol. 51, No. 185. Guidelines
    for Carcinogenic Risk Assessment; Final
    Rule (September 24,1986), 33992-34003.
    [51 FR 33992, September 24,1986]
Federal Register. Vol. 51, No. 185. Guidelines
    for Estimating Exposures; Notices
    (September 24,1986), 34042-34055. [51 FR
    34042, September 24,1986]
Federal Register. Vol. 51, No. 189. Water
    Pollution Control; National Primary
    Drinking Water Regulations;
    Radionuclides; Advance Notice of
    Proposed Rulemaking. (Sept. 30,1986),
    34836-34862. ,[51 FR 34836 Sept. 30,1986]
Federal Register. Vol. 52, No. 130. National
    Primary Drinking Water Regulations for
    Synthetic Organic Chemicals; Monitoring
    for Unregulated Chemicals. (July 8,1987),
    pp. 25690-25717. [52 FR 25690, July 8,
    1987]
Federal Register. Vol. 52, No. 208. Drinking
    Water Regulations; Public Notification.
    (Oct. 28,1987), pp. 41534-41550. [52 FR
    41534, October 28,1987]
Federal Register. Vol. 52, No. 212. Drinking
    Water; National Primary Drinking Water
    Regulation; Total Coliforms; Proposed
    Rule. (Nov. 3,1987), pp. 42224-42245. [52
    FR42224, Nov. 3,1987]           .    ..
Federal Register. Vol. 54, No. 43. National
    Emission Standards for Hazardous Air
    Pollutants; Regulation of Radionuclides;
    Proposed Rule. (March 7,1989), pp. 9612-
    9668. [54 FR 9612, Mar. 7,1989]
Federal Register. Vol. 54, No. 97. National
    Primary and Secondary Drinking Water
    Regulations; Proposed Rule. (May 22,
    1989), pp 22062-22160. [54 FR 22062, May
    22,1989]

-------
                  Federal  Register / Vol. 56, No. 138 / Thursday,  July 18,  1991  / Proposed Rules
                                                                              33115
Federal Register. Vol. 54, No. 177. National
    Emission Standards for Hazardous Air
    Pollutants; Benzene; Rule and Proposed
    Rule. (September 14,1989), pp. 38044-
    38190. [54 FR 38044, Sept. 14,1989]
Federal Register. Vol. 54, No. 240. National
    Emission Standards for Hazardous Air
    Pollutants; Radionuclides; Final Rule and
    Notice of Reconsideration. (December 15,
    1989), 51654. [54 FR 51654, Dec. 15,1989]
Federal Register. Vol 54, No. 243. National
    Primary Drinking Water Regulations
    Implementation; Primary Enforcement
    Responsibility; Final Rule. (December 20,
    1989), pp. 52126-52140. [54 FR 52126, Dec.
    20,1989]
Federal Register. Vol. 55, No. 191. Variances
    and Exemptions for Primary Drinking
    Water Regulations; Unreasonable Risk to
    Health Guidance. Notice of Availability.
    (Oct. 2,1990). p. 40205 [55 FR 40205, Oct.
    2,1990]
Federal Register. Vol. 56, No. 20. National
    Primary Drinking Water Regulations—
    Synthetic Organic Chemicals and
    Inorganic Chemicals; Final Rule (January
    30,1991), pp. 3526-3597. [56 FR 3526, Jan.
    30,1991]
Federal Register. Vol. 56, No. 106. National
    Primary Drinking Water Regulations
    Implementation Primary Enforcement
    Responsibility; Final Rule (June 3,1991).
    pp. 25046-25050. [56 FR 25046]
Finkel, M.P. Relative biological effectiveness
    of radium and other  alpha emitters in CF
    No. 1 female mice. Proo. Soc. Exp. Biol.
    Mod. 83:494-498. (1953). [Finkel, 1953]
Finkel, M.P. et al. Toxicity of radium-226 in
    mice. In: Erickson A, ed. Radiation
    induced cancer. Vienna, Austria:
    International Atomic Energy Agency, pp.
    369-391, (1969). [Finkel et al., 1969]
H.R. Rep., No. 93-1185, in Legislative History
    of the SDWA, 97 Cong. 2d Sess., p.  549-
    550,1982
Haven, FJL and Hodge, H.C. Toxicity
    following the parenteral administration
    of certain soluble uranium compounds.
    In: Voegtlin, C. and Hodge, H.C. eds.
    Pharmacology and Toxicology of
    Uranium Compounds. National Nuclear
    Energy Series, Div. VI, Vol. 1. New York:
    McGraw-Hill, pp. 281-308, (1949). [Haven
    and Hodge, 1949]
Hess, C.T. et al. The Occurrence of
    Radioactivity in Public Water Supplies in
    the United States. Health Physics. Vol. 48
    No. 5. (May 1985). [Hess et al., 1985].
Hess, C.T. et al. Radon Transferred from
    Drinking Water into House Air. Prepared
    for Health Effects Research Division,
    Cincinnati, OH (January 1991). [Hess et.
    al., 1991]
Hess, C.T. and Brown, W.L. The
    Measurement of the Biotransfer and
    Time Constant of Radon from Ingested
    Water by Human Breath Analysis.
    Prepared under EPA Assistance
    Agreement CR-815156-01-0 for Health
    Effects Research Division, Cincinnati,
    OH (January 1991). [Hess and Brown,
    1991]
Hodge, H.C. A History of Uranium Poisoning
    (1824-1942). In: Hodge, H.C., Stannard,
    J.N., Hursh, J.B., eds. Uranium, Plutonium,
    Transplutonic Elements: Handbook of
    Experimental Pharmacology XXXVI
    (1973), pp. 6-67. New York: Springer-
    Verlag. [Hodge, 1973]
Howland J.W. Studies on human exposures to
    uranium compounds. In: Voegtlin C.,
    Hodge H.C. eds. Pharmacology and
    Toxicology of Uranium Compounds.
    National Nuclear Energy Series, Div. VI,
    Vol. 1. New York: McGraw Hill. (1949).
    [Howland, 1949]
Humphreys et al., The induction by Ra-224 of
    myeloid leukemia and osteosarcoma in
    male CBA mice. Int. J. Radiat. Biol.
    47:239-247 (1985). [Humphreys et al.,
    1985].
International Commission on Radiological
    Protection (ICRP). Report of the Task
    Group on Reference Man. ICRP
    Publication 23. Oxford: Pergamon Press
    (1975). [ICRP, 1975]
International Commission on Radiological
    Protection (ICRP). Recommendations of
    the International Commission on
    Radiological Protection. ICRP Publication
    26. Oxford: Pergamon Press (1977). [ICRP,
    1977]
International Commission on Radiological
    Protection (ICRP). Limits for Intake of
    Radionuclides by Workers. ICRP
    Publication 30, Part 1. Annals of the ICRP
    Vol. 2. (3/4) Oxford: Pergamon Press
    (1979). PCRP, 1979]
International Commission on Radiological
    Protection (ICRP). Limits for Intakes of
    Radionuclides by Workers ICRP
    Publication 30. Part 2. Annals of the ICRP
    Vol. 4 (3/4) Oxford: Pergamon Press
    (1980). PCRP, 1980]
International Commission on Radiological
    Protection (ICRP). Limits for Intakes of
    radionuclides by workers. ICRP
    Publication 30. Part 3. Annals of the ICRP
    Vol. 6 (2/3) Oxford: Pergamon Press
    (1981). PCRP, 1981]
International Commission on Radiological
    Protection (ICRP) Lung Cancer Risk From
    Indoor Exposure to Radon Daughters.
    ICRP Publication 50. Oxford: Pergamon
    Press. (1987). PCRP, 1987]
Keane, A.T. et al. Non-stochastic effects of
    Ra-226 and Ra-228 in the human
    skeleton. In:  Biological effects of low-
    level radiation. Vienna, Austria:
    International Atomic Energy Agency.
    329-350. (1983). [Keane et al., 1983]
Kinner et al., Radon Removal Techniques for
    Small Community Public Water Supplies.
    Prepared under cooperative agreement
    no. CR 812602-01-0 for Risk Reduction
    Engineering Laboratory, Office of
    Research and Development. Cincinnati,
    OH (June 1989). pCinner et. al., 1989]
Kinner et al. Radon Removal Using Point-Of-
    Entry Water Treatment Techniques.
    Prepared under cooperative agreement
    no. CR 812602-01-0 for Risk Reduction
    Engineering Laboratory, Office of
    Research and Development. Cincinnati,
    OH (June 1990). [Kinner et. al., 1990]
Kinner, N.E. Memorandum to Greg Helms,
    ODW from Nancy Kinner. Variability in
    Radon Occurrence Data (July 3,1990).
    [Kinner, 1990]
Leggett, R.W. The Behavior and Chemical
    Toxicity of U in the Kidney: A
    Reassessment. Health Physics, Vol. 57,
    No. 3, 365-383. September, 1989. [Leggett,
    1989]
Lowry, 1991. Measuring Low Radon Levels in
    Drinking Water Supplies. JAWWA April
    1991, pp. 149-153 p,owry, 1991]
Lyman, G.H. et al. Association of leukemia
    with radium groundwater contamination.
    JAMA 254:621-626 (1985). [Lyman et al.,
    1985]
Lyman, G.H., and Lyman, C.G. Leukemia and
    groundwater contamination [letter].
    JAMA 256:2676-2677 (1986). [Lyman and
    Lyman, 1986]
Maletskos, C.J. et al. The metabolism of
    intravenously administered radium and
    thorium in human beings and the relative
    absorption from the human
    gastrointestinal tract of radium and
    thorium in simulated radium dial paints.
    In: Radium and mesothorium poisoning
    and dosimetry and instrumentation
    techniques in applied radioactivity.
    Cambridge, MA: MIT Physics
    Department. 202-317. (1966). [Maletskos
    et al., 1966]
Maletskos, C.J. et al. Retention and
    absorption of Ra-224 and Th-234 and
    some  dosimetric consideration of Ra-224
    in human beings. In: Mays, C.E. et  al. eds.
    Delayed effects of bone-seeking
    radionuclides. Salt Lake City, UT:
    University of Utah Press. 29-49. (1969).
    [Maletskos et al., 1969]
Maynard,  E.A. and Hodge, H.C.  Study of
    toxicity of various uranium compounds
    when fed to experiment animals. In:
    Voegtlin C., Hodge H.C. eds.
    Pharmacology and Toxicology of
    Uranium Compounds. National Nuclear
    Energy Series, Div. VI, Vol.  1. pp. 309-
    376. New York: McGraw Hill. (1949).
Maynard and Hodge, 1949] Maynard, E.A. et
    al. Oral toxicity of uranium compounds.
    In: Voegtlin C., Hodge H.C.  eds.
    Pharmacology and Toxicology of
    Uranium Compounds. National Nuclear
    Energy Series, Div. VI, Vol.  1. New York:
    McGraw Hill. (1953), pp. 1221-1369.
    [Maynard et al., 1953]
Mays, C.W. and Speiss,  H. Bone sarcomas in
    patients given radium-224. In: Boice, J.D.,
    Fraumeni, J.F. eds. Radiation
    carcinogenesis; epidemiology and
    biological significance. New York:  Raven
    Press. 241-252 (1984). [Hays and Spiess,
    1984]
Mays, C.W. et al. Cancer Risk from the
    Lifetime Intake of Ra and U Isotopes.
    Health Physics Vol. 48 No. 5 (May  1985).
    Pvlays et al., 1985]
Moeller, D.W. and Underbill, D.W. Final
    report on the study of the effects of
    building materials on population dose
    equivalent. School of Public Health,
    Harvard University. (1976). pvloeller and
    Underbill, 1976]

-------
33116
Federal  Register / Vol. 56, No. 138 /  Thursday, July  18,  1991 / Proposed Rules
Motor Vehicles Manufacturers 'Assn. v. State
    Farm Mutual Automobile Insurance Co.,
    463 U.S. 29, 42 (1983).
Muller J. et al. Early induction of leukemia
    (malignant lymphoma) in mice by
    protracted low alpha-doses. Health
    Physics (1988) 54:461-463. [Muller et al.,
    1988]
National Academy of Sciences. National
    Research Council. The Effects on
    Populations of Exposure to Low Levels of
    Ionizing Radiation (BEIR HI) (1980).
    [NAS, 1980]
National Academy of Sciences. National
    Research Council. Health Risks of Radon
    and Other Internally Deposited Alpha-
    Emitters (BEIR IV) (1988). [NAS, 1988]
National Academy of Sciences. National
    Research Council. Health Effects of
    Exposure to Low Levels of Ionizing
    Radiation (BEIR V) (1990). [NAS, 1990]
National Academy of Sciences. National
    Research Council. Comparative
    Dosimetry of Radon in Mines  and
    Homes. (1991). [NAS, 1991]
National Council on Radiation Protection and
    Measurements (NCRP). Radiological
    assessment: Predicting the transport,
    bioaccumulation and uptake by man of
    radionuqlides released to the
    environment. NCRP Report No. 76.
    Bethesda, MD. (1984). [NCRP, 1984a]
National Council on Radiation Protection and
    Measurements (NCRP). Exposures from
    the  Uranium Series With Emphasis on
    Radon and Its Daughters. NCRP Report,
    No. 77. Bethesda, MD (1984). [NCRP,
    1984b]
National Institute for Occupational Safety
    and Health (NIOSH). Radon progeny in
    underground mines. Publication No. 88-
    101. (1987). [NIOSH, 1987]
National Radiological Protection Board
    (NRPB). Health Effects Models
    Developed from the 1988 UNSCEAR
    Report. Chilton, Didcot, Oxfordshire, UK.
    (1988) [NRPB, 1988]
Nero, A.V.; Schwehr, M.B.; Nazaroff, K.L. et
    al. Distribution of Airborne Radon 222
    Concentration in U.S. Homes. Vol. 234
    Science (1986). [Nero et al., 1986]
New Jersey State Department of
    Environmental Protection—Division of
    Environmental Quality—Bureau of  :
    Radiation and Inorganic Analytical
    Services. Determination of Radium 228 in
    Drinking Water (August, 1990). [NJ DEQ,
    1990]
New York State Department of Health.
    Determination of Radium 226  and
    Radium 228 (Ra-02), Radiological
    Sciences Institute  Center for Research.
    (June, 1982). [NY State DOH, 1982]
Norris, W.P. et al. Studies of the metabolism
    of radium in man. Am. J. Roentg. Rad.
    Therapy Nuclear Med. 73:785-802 (1955).
    [Norris et al., 1955]
Northern Plains Resources Council v. EPA,
    645  F. 2d 1349 (9 Circ. 1981).
Oakley, D.T. Natural radiation exposure in
    the  United States. ORP/SID 72-1 (1972).
    [Oakley, 1972]
                          Paternian, J.L. et al. The Effects of Uranium
                              on Reproduction, Gestation, and
                              Postnatal Survival in Mice. Ecotoxicol.
                              Environ. Safety 17: 291-296. (1989).
                              [Paternian et al., 1989]
                          Petersen, N.J. et al. An epidemiologic
                              approach to low-level radium-226
                              exposure. Public Health Rep. 81:805-814
                              (1966). [Petersen et al., 1966]
                          Raabe, O.G., et al. Dose-response
                              relationships for bone tumors in beagles
                              exposed to Ra-226 and Sr-90. Health
                              Phys. (1981), 40:863-880. [Raabe et al.,
                              1981]
                          Radiation Effects Research Foundation
                              (RERF). Life Span Study Report 11, Part
                              1. Comparison of Risk Coefficients for
                              Site-Specific Cancer Mortality Based on
                              the DS86 and T65DR Shielded Kerma and
                              Organ Doses.  (November, 1987). RERF
                              TR12-87 [RERF, 1987]
                          Radioactivity in Drinking Water. Proceedings
                              of the National Workshop on
                              Radioactivity in Drinking Water, held
                              May 24-26,1983 at the Tidewater Inn,
                              Easton, MD, published as a special issue
                              of Health Physics, Vol. 48, No. 5, (May,
                              1985) 535-693. [Health Physics, 1985]
                          Rowland et al.  Dose-response relationships
                              for female radium dial workers.
                              Radiation Research. Vol. 76 (1978), 368-
                              383. [Rowland et al., 1978]
                          Rundo, J. et al.  Current (1984) status of the
                              study of Ra-226 and Ra-228 in humans at
                              the Center for Human Radiobiology. In:
                              Gossner W., Gerber G.B., Hagen U., Luz
                              A., eds. The Radiobiology of Radium and
                              Thorotrast. Baltimore: Urban and
                              Schwarzenberg, Baltimore (1986). pp. 14-
                              21. [Rundo et al., 1986]
                          Schlenker, R.A. Risk estimates for bone,. In:
                              Critical issues in setting radiation dose
                           •   limits. Proceedings No.3,  Bethesda, MD:
                              National Council on Radiation Protection
                              and Measurements. (1982), pp. 153-163.
                              [Schlenker, 1982]
                          Schliekelman, R.J.  Determination of Radium
                              Removal Efficiencies in Iowa Water
                              Supply Treatment Processes. EPA-ORP/
                              TAD-76-1 (June 1976). [Schliekelman,
                              1976]
                          Science Advisory Board, Radiation Advisory
                              Committee. Review of ODW's Criteria
                              Documents and Related Reports for
                              Uranium, Radium, Radon, and Manmade
                              Beta-Gamma Emitters (Draft). October
                              13,1990. [SAB/RAC, 1990]
                          Sevc,, J. et al. Cancer in man after  exposure
                              to Rn daughters. Health Physics 54:27-46
                              (1988). [Sevc et al., 1988]
                          Sharpe, W.D. Chronic radium intoxication:
                              Clinical .and autopsy findings in long-
                              term New Jersey survivors. Environ. Res.
                              8:243-383 {1974). [Sharpe, 1974]
                          Sorg, T.J. et al. Removal of Radium-226 from
                              Sarasota County, Florida Drinking Water
                              by Reverse Osmosis. JAWWA (April
                              1980), 230-237. [Sorg et al., 1980]
                          Sorg, T.J. Methods for Removing Uranium
                              from Drinking Water. JAWWA (July
                              1988), Vol. 80, No.7, pp. 105-111. [Sorg,
                              1988]
 Stebbings, J.H., et al. Mortality from cancers
     of major sites in female radium dial
     workers. Am. J. Ind. Med. Vol. 5 (1984)
     435-459. [Stebbings et al., 1984]
 Sullivan, R.E. et al. Estimates of Health Risk
     from Exposure to Radioactive Pollutants
     (draft). Prepared under Interagency
     Agreement 40^-1073-80 by Oak Ridge
     National Laboratory (November, 1981).
     ORNL/TM-7745 [Sullivan et al., 1981]
 Taylor, G.N. et al. Radium induced eye
     melanomas in dogs. Radiat. Res. 51:361-
     373. (1972). [Taylor et al., 1972]
 U.S. Department of Energy (DOE). Radon
     Inhalation studies in animals. U.S.
     Department of Energy. Washington, DC
     DOE/ER-0396 (1988). [DOE, 1988]
 U.S. Department of Energy (DOE),
     Environmental Measurements
     Laboratory. EML Procedures Manual,
     27th edition, (November 1990). [DOE,
     1990]
 U.S. Environmental Protection Agency.
     Interim Radiological Methodology for
     Drinking Water. EPA-600/4-75^008,
     (March, 1976). [EPA, 1976]
 U.S. Environmental Protection Agency.
     Memorandum, Hawkins to Regional
     Administrators, I-X, RE: Guidance for
     Determining BACT under SPD. January 4,
     1979. [EPA, 1979a]
 U.S. Environmental Protection Agency.
     Radiochemical Analytical Procedures for
     Analysis of Environmental Samples.
     EMSL-LV-0539-17. (March, 1979). [EPA,
     1979b]
 U.S. Environmental Protection Agency.
     Prescribed Procedures for Measurement
     of Radioactivity in Drinking Water.
     Environmental Monitoring and Support
     Laboratory, Cincinnati, OH. EPA-600/4-
     80-032, August, 1980. [EPA, 1980]
 U.S. Environmental Protection Agency and
     U.S. Department of Energy, Oak Ridge
     National Laboratory. Uranium in U.S.
     Surface, Ground, and Domestic Waters,
     Vol. 1 (1981) EPA-570/9-81-001 and
.     ORNL/EIS-192, p.8. [EPA/ORNL, 1981]
 U.S. Environmental Protection Agency.
     Determination of Lead-210 in Drinking
     Water: Method 909. Velten, R.J., and
     Jacobs, B.J., Environmental Monitoring
     and Support Laboratory. (March 1982)
     [EPA,  1982]
 U.S. Environmental Protection Agency, Office
     of Radiation Programs. Eastern
     Environmental Radiation Facility
     Radiochemistry Procedures Manual.
     Eastern Environmental Radiation
%    Facility, Montgomery, AL. EPA 520/5-84-
     006, August, 1984. [EPA, 1984a]
 U.S. Environmental Protection Agency.
     Technologies and Costs for the Removal
     of Radium from Potable Water Supplies.
     Prepared by V.J. Ciccone and Associates.
     (July 31,1984). [EPA, 1984b]
 U.S. Environmental Protection Agency.
     Nationwide  Occurrence of Radon and
     Other  Natural Radioactivity in Public
     Water Supplies (The Nationwide Radon
     Survey). Eastern Environmental
     Radiation Facility, Montgomery, AL, EPA
     520/5-85-008, (October 1985). [EPA,
     1985a]

-------
                  Federal Register  / Vol.  56, No.  138  /Thursday, July  18, 1991 /Proposed  Rules
                                                                             33117
U.S. Environmental Protection Agency.
    Technologies and Costs for the Removal
    of Uranium from Potable Water Supplies
    (Revised Final Draft). Prepared by V.J.
    Ciccone and Associates (Oct., 30,1985).
    [EPA, 1985b]
U.S. Environmental Protection Agency. Test
    Procedure for Gross Alpha Particle
    Activity in Drinking Water. J.O. 70.12.
    (December, 1985). [EPA, 1985c]
U.S. Environmental Protection Agency.
    Project Summary: Barium and Radium in
    Water Treatment Plant Wastes. Prepared
    by V.J. Snoeyink et al. for EPA Water
    Engineering Research Laboratory,
    Cincinnati OH (March 1985). [EPA, l985d]
U.S. Environmental Protection Agency.
    Statistical Analyses of Analytical
    Methods for Radionuclides in Drinking
    Water (Final Report). Prepared by
    Research Planning Institute, Columbia,
    SC (May, 1986) for ODW under
    Assignment No. 041422. [EPA, 1986a]
U.S. Environmental Protection Agency.
    Technologies and Costs for the Removal
    of Man-Made Radionuclides from
    Potable Water Supplies (Revised Final
    Draft). Prepared by V.J. Ciccone & Assoc.
    (March 10,1986). [EPA, 1986b]
U.S. Environmental Protection Agency.
    Supplement to Technologies and Costs
    for the Removal of Uranium from Potable
    Water Supplies (Draft). Prepared by
    Malcolm Pirnie, Inc. (December 30,1986).
    [EPA, 1986c]
U.S. Environmental Protection Agency.
    Technologies and Costs for the
    Treatment and Disposal of Waste
    Byproducts from Water Treatments for
    the Removal of Inorganic and
    Radioactive Contaminants (Draft).
    (September 1986). [EPA, 1986d]
U.S. Environmental Protection Agency. Air
    Quality Criteria Document for Lead, Vol.
    II, Environmental Criteria and
    Assessment Office. Research  Triangle
    Park, NC, EPA-600/8-S3/028bF (1986).
    [EPA, 1988e]
U.S. Environmental Protection Agency. A
    Citizen's Guide to Radon: What It Is and
    What To Do About It (August 1986).
    [EPA 1988fJ
U.S. Environmental Protection Agency.
    Analysis of Occurrence, Control and/or
    Removal of Radionuclides in Small
    Drinking Water Systems in Virginia.
    Prepared by V.J. Ciccone and Associates
    under EPA contract No. 68-01-7088
    (September 1987). [EPA, 1987a]
U.S. Environmental Protection Agency.
    Technologies and Costs for the Removal
    of Radon from Potable Water Supplies
    (Fourth Draft). Prepared by Malcolm
    Pirnie, Inc. (January 6,1987). [EPA, 1987b]
U.S. Environmental Protection Agency.
    Supplement to Technologies and Costs
    for the Removal of Man-Made
    Radionuclides from Potable Water
    Supplies (First Draft). Prepared by
    Malcolm Pirnie (January 8,1987). [EPA,
    1987c]
U.S. Environmental Protection Agency.
    Supplement to Technologies and Costs
    for the Removal of Radium from Potable
    Water Supplies (First Draft). Prepared by
    Malcolm Pirnie January 7,1987). [EPA,
    1987d]
U.S. Environmental Protection Agency. Two
    Test Procedures for Radon in Drinking
    Water, Interlaboratory Collaborative
    Study. EPA/600/2-87/082, (March 1987).
    [EPA, 1987e]
U.S. Environmental Protection Agency.
    Project Summary: Removal of Uranium
    from Drinking Water by Ion Exchange
    and Chemical Clarification. Prepared by
    Steven W. Hanson et al. for EPA Water
    Engineering Research Laboratory,
    Cincinnati OH (Dec. 1987). [EPA, l987fj
U.S. Environmental Protection Agency.
    Estimates of Co-Occurrence  of
    Radionuclides  in Drinking Water.
    Prepared by Science Applications
    International Corporation (August 17,
    1988) under contract ho. 68-03-3514.
    [EPA, 1988a].
U.S. Environmental Protection Agency.
    Memorandum to Arthur Perler, STB from
    Jon Longtin, WSTB regarding
    Distribution Tables for the National
    Inorganics and Radionuclides Survey
    (NIRS) Results, (February 23,1988). [EPA,
    1988b].
U.S. Environmental Protection Agency.
    Memorandum to Stephen Clark, ODW
    from Warren Peters, PAB and Chris
    Nelson, ORP regarding Preliminary Risk
    Assessment for Radon Emissions from
    Drinking Water Treatment Facilities,
    (June 28, 1988). [EPA, 1988c]
U.S. Environmental Protection Agency.
    Memorandum to Susan MacMullin and
    Michael Morton, ODW, from Frank
    Letkiewicz and Barbara Minkser, Wade
    Miller Assoc., Inc. regarding
    Radionuclide Co-Occurrence Analyses
    (August 12,1988). [EPA, 1988d].
U.S. Environmental Protection Agency.
    Aeration Alternatives for Radon
    Reduction—Addendum to Technologies
    and Costs for the Removal of Radon from
    Potable Water Supplies. Prepared by
    Malcolm Pirnie, Inc.  (April 25,1988).
    [EPA, 1988e]
U.S. Environmental Protection Agency. Letter
    to W. Schull, SAB/RAC from L. Thomas,
    EPA regarding radon risk estimates
    (November 23,1988). [EPA, 1988f]
U.S. Environmental Protection Agency.
    Project Summary: A Study of Possible
    Economical Ways of Removing Radium
    from Drinking Water. Prepared by R.L.
    Valentine et al. for EPA Water
    Engineering Research Laboratory.
    Cincinnati, OH (April, 1988). [EPA, 1988g]
U.S. Environmental Protection Agency.
    Project Summary: Radium Removal for a
    Small Community Water Supply System.
    Prepared by Kenneth A. Mangelson for
    EPA Water Engineering Research
    Laboratory, Cincinnati OH (Sept. 1988).
    [EPA, 1988h]
U.S. Environmental Protection Agency. Risk
    Assessments Methodology.
    Environmental  Impact Statement.
    NESHAPS for Radionuclides.
    Background Information Document—Vol.
    1. Office of Radiation Programs. EPA/
    520/1-89-005 (September 1989). [EPA,
    1989a]
U.S. Environmental Protection Agency.
    Memorandum to Greg Helms, ODW from
    Marc Parrotta, ODW regarding an
    Analysis of Potential Radon Emissions
    from Water Treatment Plants using the
    MINEDOSE Code, (November 22,1989).
    [EPA, 1989b]
U.S. Environmental Protection Agency.
    Occurrence of Man-Made Radionuclides
    in Public Drinking Water Supplies
    (Draft). Prepared by Wade Miller
    Associates, Inc. (November 20,1989).
    [EPA, 1989c]
U.S. Environmental Protection Agency. Letter
    to Greg Helms, ODW from Thomas
    Banks, Wade Miller Assoc. regarding
    The Lucas Cell Method of Testing for
    Radon in Water (November 14,1989).
    [EPA, I989d]
U.S. Environmental Protection Agency.
    Memorandum to G. Helms, ODW from T.
    Banks, Wade Miller Assoc. regarding
    liquid scintillation counting, (1989). [EPA,
    1989e]
U.S. Environmental Protection Agency.
    Drinking Water Criteria Document for
    Uranium. External Review Draft.
    Prepared by Dynamac Corp. (November
    1989). [EPA, I989fj
U.S. Environmental Protection Agency.
    Memorandum from Margo Oge, Director,
    Radon Division. "Current ORP Estimate
    of Annual Radon-Induced Lung Cancer
    Deaths in the General Population."
    (August 17, 1989). [EPA, 1989g]
U.S. Environmental Protection Agency.
    Suggested Guidelines for the Disposal of
    Drinking Water Treatment Wastes
    Containing Naturally-Occurring
    Radionuclides. Office of Drinking Water.
    (July 1990). [EPA, 1990a]
U.S. Environmental Protection Agency.
    Manual for the Certification of
    Laboratories Analyzing Drinking Water.
    Criteria and Procedures. Quality
    Assurance. Third Edition. Office of
    Water, EPA/570/9-90/08 (April 1990).
    [EPA, I990b).
U.S. Environmental Protection Agency.
    Memorandum to Joseph Cotruvo, ODW
    from Jerome Puskin, ORP regarding
    Human Radon-222 Lifetime Risk
    Estimates from Ingestibn (March 28,
    1990). [EPA, 1990c]
U.S. Environmental Protection Agency.
    Memorandum to Folsom from McFarland
    regarding cost revisions for radium,
    radon, and uranium to account for new
    system level treatment design flows.
    adopted by EPA. (May, 1990). [EPA,
    1990d]
U.S. Environmental Protection Agency.
    Memorandum to Greg Helms,.ODW from
    Neal Nelson, ORP regarding draft
    Drinking Water Health Criteria
    Document for Uranium [July 5,199Q).
    [EPA, 1990e]
U.S. Environmental Protection Agency.
    Occurrence and Exposure Assessment

-------
 33118
Federal  Register / Vol. 56, No.  138  /  Thursday, July 18, 1991 /  Proposed Rules
     for Radon in Public Drinking Water
     Supplies (draft). Prepared by Wade
     Miller Associates, Inc. (September 25,
     1990). [EPA, 1990f]
 U.S. Environmental Protection Agency.
     Occurrence and Exposure Assessment
     for Radium-226 in Public Drinking Water
     Supplies (draft). Prepared by Wade
     Miller Associates, Inc. (April 30,1990).
     [EPA, 1990g]
 U. S. Environmental Protection Agency.
     Occurrence and Exposure Assessment
     for Uranium in Public Drinking Water
     Supplies (draft). Prepared by Wade
     Miller Associates, Inc. (April 26,1990).
     [EPA, 1990h]
 U.S. Environmental Protection Agency.
     "Estimation of Risks from Indoor Radon
     Exposures" (February 15,1990). [EPA,
     1990i]
 U.S. Environmental Protection Agency.
     Memorandum T. Banks, WMA to G.
     Helms, February 21,1990, on
     Laboratories Conducting Analysis of
     Radon in Drinking Water. [EPA, 1990J],
 U.S. Environmental Protection Agency.
     "Guidance for Developing Health
     Criteria for Determining Unreasonable
     Risks to Health (draft), (1990). [EPA,
     1990k]
 U.S. Environmental Protection Agency.
     Quantitative Risk Assessment for Radon
     in Drinking Water. External Review
     Draft. Prepared by Dynamac Corp. (May
     1990). [EPA, 19901]
 U. S. Environmental Protection Agency.
     Technical Support Document for the  1990
     Citizen's Guide to Radon (Draft; August
     16,1990). [EPA 1990m]
 U.S. Environmental Protection Agency.
     Occurrence and Exposure Assessment
     for Radium-228 in Public Drinking Water
     Supplies (draft). Prepared by Wade
     Miller Associates under contract No. 69-
     03-3514. (November 30,1990). [EPA,
     1990n]
 U.S. Environmental Protection Agency.
     Memorandum from Elliott, Assistant
     Administrator and General Counsel to
     Habicht, Deputy Administrator regarding
     regulation of radionuclides in drinking
     water, (July 30,1990). [EPA, 1990o]   .
 U.S. Environmental Protection Agency.
     Drinking Water Criteria  Document for
     Gross Alpha Emitters (Draft). Prepared
     by Clement International Corp. (June
     1991). [EPA, 1991a]
 U.S. Environmental Protection Agency.
     Criteria Document for Radium in
     Drinking Water (Draft). Prepared by Life
     Systems, Inc. (June 1991). [EPA, 1991b]
 U.S. Environmental Protection Agency.
     Drinking Water Criteria Document for
     Radon in Drinking Water (Draft).
     Prepared'by Life Systems, Inc. (June
     1991). [EPA, 199lc]
 U.S. Environmental Protection Agency.
     Criteria Document for Man-Made
     Radionuclides in Drinking Water (Draft).
     Prepared by Clement International Corp.
     (June 1991). [EPA, 1991d]
U.S. Environmental Protection Agency.
    Drinking Water Criteria Document for
    Uranium (Draft). Prepared by Clement
                              International Corp. (June 1991). [EPA,
                              1991e].
                          U.S. Environmental Protection Agency.
                              Occurrence and Exposure Document for
                              Gross Alpha Radiation (Draft). Prepared
                              by Wade Miller Associates under
                              contract No. 69-03-3514. (1991). [EPA,
                              1991f].
                          U.S. Environmental Protection Agency.
                              Occurrence and Exposure Assessment of
                              Pb-210 in Drinking Water, Food, and Air
                              (Draft)—Addendum to Occurrence of
                              Man-Made Radionuclides in Public
                              Drinking Water Supplies. Prepared by
                              Wade Miller Associates, Inc. (May 20,
                              1991). [EPA, I991g]
                          U.S. Environmental Protection Agency.
                              Radon in Drinking Water: Assessment of
                              Exposure Pathways. Prepared by Life
                              Systems, Inc. (June 1991) [EPA, 1991h].
                          U.S. Environmental Protection Agency.
                              Regulatory Impact Analysis of Proposed
                              National Primary Drinking Water
                              Regulations for Radionuclides  (Draft).
                              Prepared by Wade Miller Associates
                              (June 14,1991). [EPA, 19911]
                          U.S. Environmental Protection Agency.
                              Response To Comments Received on the
                              NPDWRs: Radionuclides in Drinking
                              Water—Advance Notice of Proposed
                              Rulemaking,  September 30,1986. (June
                              1991). [EPA, 1991J]
                          U.S. Environmental Protection Agency.
                              Technologies and Costs for the Removal
                              of Alpha Emitters from Potable Water
                              Supplies. Prepared by Malcolm Pirnie,
                              Inc. (Feb., 1991). [EPA, 1991k]
                          U.S. Environmental Protection Agency.
                              "Proposed Revisions in EPA Estimates of
                              Radon Risks  and Associated
                              Uncertainties" (April,  1991). [EPA, 19911]
                          U.S. Environmental Protection Agency.
                              Memorandum to G. Helms from I.
                              Deloatch regarding estimated cost of
                              analyses for radionuclides, (March 27,
                              1991). [EPA, 1991m]
                          U.S. Environmental Protection Agency.
                              Radiation Research and Methods
                              Validation. Annual Report, FY1990. EPA
                              600/X-91/055, (May 1991).  [EPA,.1991n] .
                          U.S. Environmental Protection Agency.
                              Occurrence and Exposure Assessment
                              for Uranium in Public Drinking Water
                              Supplies. Revision 2. Prepared by Wade
                              Miller Associates, Inc. (June 13,1991)
                              [EPA, 1991o]
                          U.S. Environmental Protection Agency.
                              Health Effects Assessment Summary
                              Tables. Annual FY-1991. OERR 9200.6-
                              303 (91-1) NTIS PB91-921199. [EPA,
                              1991p]
                          U.S. Environmental Protection Agency.
                              Method 913—Radon in Drinking Water
                              by Liquid Scintillation. EMSL/LV, (June
                              1991) [EPA, 1991q]
                          U.S. Environmental Protection Agency.
                              Memorandum Deloatch to Helms, PQL
                              Assessments for Radionuclides. (June 13,
                              1991) [EPA, 1991r]
                          U.S. Environmental Protection Agency. IRIS
                              Print-out for Uranium (soluble salts);
                             Revised 10/1/89. (June 14,1991) [EPA,
                              1991s]
 U.S. Geological Survey (USGS). Methods for
    Determination of Radioactive Substances
    in Water and Fluvial Sediments, Book 5,
    Chapter A5, in Techniques of Water-
    Resources Investigations of the USGS.
    (1989) [USGS, 1989]
 United Nations Scientific Committee on the
    Effect of Atomic Radiation (UNSCEAR).
    Ionizing Radiation: Sources and
    Biological Effects. Report to the General
    Assembly. New York (1982).  [UNSCEAR,
    1982]
 United Nation's Scientific Committee on the
    Effect of Atomic Radiation (UNSCEAR).
    Sources, Effects and Risks of Ionizing
    Radiation. Report to the General
    Assembly. New York, (1988). [UNSCEAR,
    1988]
 Vitz, E. Toward a Standard Method for
    Determining Waterborne Radon. Health
    Physics, vol 60, No. 6 pp. 817-829 (June
    1991). [Vitz, 1991] ,
 Wilkinson, G. Gastric Cancer in New Mexico
    counties with significant deposits of
    uranium. Arch. Environ. Health. (1985),
    40:307. [Wilkinson, 1985]
 Wrenn, M.E. et al. Metabolism of Ingested U
    and Ra. Health Physics. Vol.  48, No. 5
    (1985). [Wrenn et al., 1985]


 Appendix A—Fundamentals of
 Radioactivity in Drinking Water
   To assist commenters, the following section
 provides a summary of concepts and
 definitions involving radioactivity. The
 definitions include those in the Interim
 Regulations along with several additions, one
 of which is being considered (i.e., curie) to be
 added to 40 CFR 141,2,

 Definitions
   (a) Dose equivalent means the product of
 the absorbed dose from ionizing radiation
 and such factors which account for
 differences in biological effectiveness due to
 the type of radiation and its distribution in
 the body as specified by the International
 Commission on Radiological Units and
 Measurements (ICRU).
   (b) Rem means the unit of dose equivalent
 from ionizing radiation to the total body or
 any internal organ or organ system. It is
 equal to the absorbed dose in rads multiplied
 by a quality factor (to account for different
 radiation types). A rem ede (effective dose
 equivalent) is a dose to organs adjusted for
 different radiation types and by an organ
 weighting factor to account for organ
 sensitivity to the effect of radiation. A
 "millirem" (mrem) is 1/1,000 of a rem.
   (c) Curie means a special unit of activity
 equal to a nuclear transformation rate of
 3.7X1010 disintegrations/second. One
picocurie is equal to 10"12 curies, which is
 approximately 2 disintegrations per minute.
  (d) Gross alpha particle emission activity
means the total  alpha particle radioactivity
measured in an  aliquot of an evaporated
water sample.
  (e) Man-made beta particle and photon
emitters means  all radionuclides emitting

-------
                  Federal  Register / Vol. 56, No. 138  /  Thursday, July 18, 1991 /  Proposed Rules
                                                                              33119
beta particles and/or photons that have been
produced artificially and do not exist
naturally.
  (I) Gross beta particle activity means the
total radioactivity due to beta particle
emissions measured in a aliquot of a
evaporated water sample.
  (g) Becquerel (Bq) is a special unit of
radioactivity in the international system of
units (SI). One Becquerel is equal to one
disintegration per second.
  (h) Sievert (Sv) means the unit of dose
equivalent in the international system of
units (SI) from ionizing radiation to the total
body or any internal organ or organ system.
One Sievert equals 100 rem.
  (i) Effective dose equivalent means the sum
of the products of the dose equivalents in
individual organs and the organ weighing
factor.
  (j) Organ weighting factor means the ratio
of the stochastic risk for that organ to the
total risk when the whole body is irradiated
uniformly.
  (k) Natural uranium means uranium with
combined uranium-234 plus uranium-235 plus
uranium-238 which has a varying isotopic
composition but typically is 0.006% uranium-
234, 0.795 uranium-235, and 99.27% uranium-
238.
  (1) Activity means the nuclear
transformations of a radioactive substance
which occur in a specific time interval.
Fundamentals of Nuclear Structure and
Radioactivity
  This section has been included to provide
background information for those not familiar
with nuclear chemistry. It is written in broad
and general terms and some statements may
be simplified.
  An atom consists of a heavy concentration
of mass at the center (the nucleus]
surrounded by shells of electrons in different
orbits. The primary constituents of the
nucleus are neutrons and protons. The
neutrons have no net electric charge while
the protons have a positive charge. The
orbital electrons have a negative charge and
in the un-lonized atoms are  equal in number
to the protons, making the atom neutral in
overall charge.
  The number of protons in the nucleus
determines the chemical element and its
atomic number. A given element can have
more than one particular number of neutrons.
Variation in the number of neutrons does not
change the chemical properties (the element
is the same) but it can produce considerable
change in the stability of the element to
radioactive decay. Atoms with the same
number of protons but different number of
neutrons are called "isotopes." For example,
if an atom has 86 protons, it is radon. There
are three principal isotopes of radon
containing 133,134 and 136 neutrons. The
atomic mass number is the total number of
protons and neutrons in the nucleus and this
sum is usually used to label isotopes. The
three isotopes of radon have atomic masses
of 86+133=219, 88+134=220 and
86+136=222. Symbolically these can be
written as: Radon-219 Radon-220 Radon-222.
  Since the atomic number and the chemical
symbol are synonymous, the number of
protons is usually omitted in the
nomenclature.
  These radionuclides decay by emission of
alpha and beta particles and gamma rays. An
alpha particle, the heaviest nuclear radiation,
consists of two protons and two neutrons. (A
proton or neutron is about 2,000 times as
massive as an electron.) A negative beta
particle is an electron emitted from the
nucleus as a result of neutron decay. An
electron can be "created" and ejected from a
nucleus by a neutron decaying into a proton
(which remains in the nucleus) and an
electron (which is ejected as a beta particle)
and also a neutrino. As a result of this
process the nucleus has one more proton and
thus has become the atom of a different
element with atomic number one greater than
the parent atom.  (There can also be a nuclear
transformation in which a proton emits a
positive beta particle, or positron, and is
transformed into a neutron which remains in
the nucleus). A gamma ray is a form of
electromagnetic radiation. Other forms of
electromagnetic radiation are light, radio
waves, infrared radiation, ultraviolet
radiation and x-rays.
  The process of alpha and beta radioactive
decay leads to a  different element while
gamma ray emission does not. The isotope
that decays is called the parent. The resulting
isotope (if a different element) is called the
progeny. For example, radon-222 decays by
emitting an alpha particle to the progeny
polonium. This reaction is written:
Radon-222	>• Polonium-2l8+helium-4
  The atomic numbers (number of protons)
for radium, polonium and helium (the alpha
particles) are 88, 84 and 2, respectively. Note
that the atomic numbers and atomic mass
numbers balance on the two sides of this
equation. Note that the atomic mass
decreased by 4 due to  the loss of two
neutrons and two protons, and the atomic
number decreased by 2 due to the loss of two
protons.
  Beta decay causes the atomic number to
increase by one. Beta decay can be described
as a neutron in the nucleus converted to a
proton. An example of beta decay is radium-
228 which decays to actinium. This reaction
is written:
Radium-228	> Actinium-228+beta
    particle
  The atomic numbers are 88 for Ra and 89
for Ac (the beta decay described here is the
negative kind). The atomic numbers and
atomic mass numbers  balance in this
equation since the atomic number for an
electron is —1 and its  atomic mass  number is
zero.  Gamma decay changes neither the
atomic number nor the element; it only
involves a loss of energy.
  Not all atoms are equally stable and
different isotopes characteristically decay at
different rates. The concept of half life is
used  to quantitatively describe these
differences. The  half life of an isotope is the
time required for one half of the atoms
present to decay. Half lives can range from
billions of years or more (the half life of
uranium-238 is 4.5 X109 years) to millionths of
a second (the half life of polonium-214 is
164X10"6 sec) and even less. For example,
the half lives of radon-219 and radon-220 are
too short to survive transport through a
drinking water distribution system.
  Atomic fission occurring in a nuclear
reactor can also contribute radioactivity to
drinking water, if by-products are released.
This process, the source of immense energy,
is triggered by adding a neutron to certain
nuclei. The phenomenon occurs for heavy
nuclei, the classical examples being isotopes
of uranium (uranium-235) and plutonium
(plutonium-239). When a neutron is added,
each of'these isotopes breaks into two
roughly equal parts. Each of the parts (called
fission fragments) is itself a radioactive
nucleus and decays through a sequence of
isotopes by beta and gamma decay.
  Generally units such as mg/1, micrograms/
liter or ppm are used to describe the
concentrations in drinking water of
pollutants, toxic and hazardous substances.
However, certain unique properties of
radioactive substances limit the utility of
these units and alternative units are used to
directly compare the health effects of
different radionuclides.
  Two important concepts are needed to
describe radioactivity:
  •  How many nuclear transformations
occur per second.
  •  How much radiation or how much energy
is imparted to tissue (called absorbed dose).
Energy is related to the number of particles
emitted by the radioisotope, per second, and
their energies.
  Damage from radionuclides depends on the
radiation emitted (alpha, beta or gamma) and
not the mass of the radionuclides. Thus it is
essential to have a unit that describes the
number of radioactive emissions per time
period, or activity. The activity is related to
the half life: Longer half lives mean lower
activity. Historically by definition one gram
of radium is said to have 1 curie (1 Ci) of
activity. By comparison, 1 gm of uranium-238
has  an activity of 0.36 millionth of a curie.
One curie is equivalent to 3.7 x 1010
disintegrations per second. The International
System (SI) unit for activity is the Becquerel
(Bq) which is equal to one disintegration/
second
  The effect of radioactivity depends not
only on the activity (decays/time) but on the
kind of radiation (alpha, beta or gamma) and
its energy. These two properties determine
the absorbed dose to tissue when decay
oc'curs internally and the internal organs are
the target.
  A common unit of absorbed dose is called
the rad'and one rad is equivalent to  one
hundred ergs (metric unit of energy) in one
gram of matter (for perspective on the size of
an erg, 10 million erg/sec is one watt). In
general, these units are quite large and
engineering shorthand is used to describe the
activities. Shown below are some commonly
used prefixes.

-------
33120
Federal Register  /  Vol.  56,  No.  138 /  Thursday, July 18,  1991  /  Proposed Rules

Greek, prefix & abbreviation
mill— m 	
micro— Greek m 	
nano— n 	
pico— p 	
femto— f 	 	 	 	
atto— a 	


Value








Shorthand
exponen-
tial
notation








Description

thousand.





  Thus 1 picocurie is a millionth millionth of
 a curie and is abbreviated 1 pCi. Also 1
 millirad [1 mrad) is one thousandth of a rad.
  Because of the particle mass and charge, 1
 rad deposited in tissue by alpha particles
 creates a more concentrated biological
 damage than 1 rad of gamma rays. To
 compensate for this difference in damage and
 subsequent effect, a new unit was created—
 the rem. This is called the dose equivalent.
 The absorbed dose is measured in rads and
 the dose equivalent is measured in rems.
  The rad and rem are related by a quality
 factor as follows:
 Number of rems=Q times the number of rads
  Where Q is the quality factor which has
 been assigned the following value:
 Q=l for beta particles and all
    electromagnetic radiations (gamma rays
    and x-rays)
 Q=10 for neutrons from spontaneous fission
    and for protons
 Q=20 for alpha particles and fission
    fragments
  The quality factor is meant to
approximately account for the relative harm
caused by various types of radiation. The
International System (SI) unit corresponding
to the rem is the Sievert (Sv). One Sievert
equals 100 rem.

    APPENDIX B—BETA PARTICLE AND
           PHOTON EMITTERS
Nuclide
H-3 	
BE-7 	
N-13 	
C-11 	
C-14 	
C-15 	
O-15 	
F-18 	
NA-22 	
NA-24 	
SI-31 	
P-32 	
P-33.
S-35 	
CL-36 	 	
CL-38 	
K-42 	 	 	
CA-45 	
CA-47 	
SC-46 	
SC-47 	
SC-48 	
V-48 	
CR-51 	
MN-52 	
MN-54 	
MN-56 	
FE-55 	
Ch (pCi/liter)
6 09E+04
4 35E+04
1 52E+05
9 92E+04
3 20E+03
669E+06
4 95E+05
3 95E+04
466E+02
3 35E+03
1 02E+04
6 41 E+02
1 87E+03
1 29E+04
1 85E+03
2 12E+04
3 90E+03
1 73E+03
8 46E+02
8 63E+02
2 44E+03
766E+02
6 44E+02
3 80E+04
7 33E+02
2 01 E+03
5 64E+03
9.25E-4-03
APPENDIX B— BETA PARTICLE AND
PHOTON EMITTERS— Continued
Nuclide
FE-59 	
CO-57 	
CO-58 	
CO-58M 	
CO-60 	 	 	
Nl-59 	
NI-63 	
NI-65 	 	 	
CU-64 	
ZN-65 	 	 	
ZN-69 	 	 	
2N-69M 	 	 	 	 	
GA-67 	
GA-72 	
GE-71 	
AS-73 	
AS-74 	
AS-76 	
AS-77 	
SE-75 	 	
BR-82 	
RB-82 	
RB-86 	
RB-87 	
RB-88 	
RB-89 	
SR-82 	
SR-85 	
SR-85M 	
SR-89 	
SR-90 	
SR-91... 	
SR-92 	
Y-90 	
Y-91 	
Y-91M 	
Y-92 	
Y-93 	
ZR-93 	
ZR-95 	
ZR-97 	
NB-93M 	
NB-94 	
NB-95 	
NB-95M 	
NB-97 	 	 	
NB-97M 	
MO-99 	
TC-95 	
TC-95M 	
TC-96 	
TC-96M 	
TC-97 	
TC-97M 	
TC-99 	
TC-99M 	
RU-97 	
RU-103 	
RU-105 	
RU-106 	
RH-1 03M 	
RH-105 	
RH-105M 	
Ch (pCi/liter)
8.44E+02
4.87E+03
1.59E+03
6.49E+04
2.18E+02
2.70E+04
9.9'1E+03
8.81 E+03
1.19E+04
3.96E+02
6.31E+04
4.22E+03
7.02E+03
1.19E+03
4.36E+05
7.85E+03
1.41 E+03
1.06E+03
4.33E+03
5.74E+02
3.15E+03
4.36E+05
4.85E+02
5.01 E+ 02
2.91 E+04
5.27E+04
2.41 E+02
2.83E+03
2.37E+05
5.99E+02
4.20E+01
2.16E+03
3.10E+03
5.10E+02
5.76E+02
1.32E+05
2.87E+03
1.20E+03
5.09E+03
1.46E+03
6.50E+02
1.05E+04
7.07E+02
2.15E+03
2.39E+03
2.35E+04
1.37E+06
1.83E+03
6.97E+04
3.12E+03
2.05E+03
1.76E+05
3.25E+04
4.45E+03
3.79E+03
8.96E+04
7.96E+03
1.81E+03
4.99E+03
2.03E+02
4.71E+05
3.72E+03
5.51E-I-06
APPENDIX B— BETA PARTICLE AND
PHOTON EMITTERS— Continued
Nuclide
RH-1 06 	 	 	
PD-100..... 	
PD-101 	
PD-103 	
PD-107 	
PD-1 09 	
AG-1 05 	
AG-108 	
AG-108M 	
AG-109M 	
AG-1 10 	
AG-110M 	
AG-1 1 1 	
CD-109 	
CD-1 15 	
CD-115M 	
IN-113M 	
IN-1 14 	
IN-114M 	
IN-1 15 	
IN-115M 	
SN-113 	
SN-1 21 	
SN-1 21 M 	
SN-1 25 	
SN-1 26 	
SB-1 22.. 	
SB-1 24 	
SB-1 25 	
SB-126 	
SB-1 26M 	
SB-127 	 : 	
SB-1 29. ,
TE-1 25M 	
TE-127 	 	 	
TE-127M 	
TE-129 	
TE-1 29M 	
TE-1 31
TE-1 31 M 	
TE-132 	 	 	
1-1 22 	
1-123 	 	 	
1-125 	
1-126.... 	
1-129 	
1-130 	
1-131 	
1-132 	
1-133 	
1-134 	
1-135 	
CS-1 31 	
CS-134 	 	 	
CS-134M 	 	
CS-1 35 	
CS-1 36 	
CS-1 37 	
CS-1 38 	
BA-131 	
BA-133 	
BA-133M 	
BA-1 37M 	
Ch (pCi/liter)
1.24E+06
1.30E+03
1.34E+04
6.94E+03
3.66E+04
2.12E+03
2.70E+03
6.26E+05
7.23E+02
1.67E+07
1.84E+06
5.12E+02
1.08E+03
2.27E+02
9.58E+02
3.39E+02
5.24E+04
9.76E+05
3.23E+02
3.51E+01
1.64E+04
1.74E+03
6.06E+03
2.26E+03
4.46E+02
2.93E+02
8.10E+02
5.63E+02
1.94E+03
5.44E+02
5.85E+04
8.18E+02
3.09E+03
1.49E+03
7.92E+03
6.63E+02
2.72E+04
5.24E+02
2.68E+04
9.71 E+02
5.80E+02
2.11E+05
1.07E+04
1.51 E+02
8.10E+01
2.10E+01
1.19E+03
1.08E+02
8.19E+03
5.49E+02
2.14E+04
2.34E+03
2 28E+04
8.13E+01
1.01E+05
7.94E+02
5.18E+02
1.19E+02
2 56E+04
2 95E+03
1.52E+03
2.62E+03
2.15E+06

-------
Federal Register / Vol. 56. No. 138 / Thursday, July 18, 1991 / Proposed Rules         33121
APPENDIX B— BETA PARTICLE AND
PHOTON EMITTERS— Continued
Nuclida
8A-139
BA-140
LA-140
CE-141 	 	 ,„.,,... 	 	 	 	 	 	
CE-1 43
CE-144
PR-142
PR-143





PM-148 	
PM-148M 	
SM-151 	
SM-1 53 .~ 	
EU-1 52 	
EU-154 	
EU-155
EU-1S6 	 	 	
GD-153.....™. 	
GD-1S9, 	
TB-1 58 	 	 	
TB-160..... 	
DY— 165
DY-1 66 	 	 	 , 	 	 	 .,
HO-168 	 	 ~ 	
ER-169.,™ 	
ER-171 	 	 	
TM-170 	
TM-171 	 	 , 	
YB-169..,., 	 ~ 	
YB-175 	
LU-177..,™ 	
HF-181 	 , 	 	 	
TA-182 	
W-1 81 ....... 	
W-1 85
W-1 87
RE-183. 	 ......
RE-186 	 „ 	 ......
RE-187 	
RE-188
OS-185
OS-191
OS-191M
OS-193
IR-100
IR-192
IR-194
PT-1 91 ......... 	 	 	
PT-193 	 	 	
PT-193M 	
PT-197 , „ . „ 	
PT-197M 	
AU-196 	
AU-1 98 	 « 	
HQ-197 	
HG-203.. 	 „ 	
TL-202 „ . . ...
TL-204 	
TL-207 ..™™ 	 « 	
TU-208 ™ 	 	 ~ ...
TL-209 	
PB-203 .. 	
PB-209 	
PB-210 	
PB-21 1 ..„.„„. 	
PB-212... 	 	 	 	 	
PB-214™ 	
BI-206.. 	
BI-207 .......... 	 	
BI-212.... 	
Bl-213...., 	
BI-214 	
PR-223 	
Ch (pCi/liter)
1.38E+04
5.82E+02
6.52E+02
1.89E+03
1.21E+03
2.61 E+02
1.04E+03
1.17E+03
4.70E+04
1.12E+05
1.25E+03
1.17E+04
5.24E+03
5.05E+02
5.75E+02
1.38E+03
1.41 E+04
1.83E+03
8.41 E+02
5.73E+02
3.59E+03
6.00E+02
4.68E+03
2.76E+03
1.25E+03
8.15E+02
1.51 E+04
8.30E+02
9.81 E+02
3.64E+03
3.80E+03
1.03E+03
1.27E+04
1.83E+03
3.11E+03
2.55E+03
1.17E+03
8.42E+02
1.90E+04
3.44E+03
2.66E+03
5.40E+03
1.88E+03
5.82E+05
1.79E+03
2.46E+03
2.38E+03
1.43E+04
1.69E+03
1.01E+03
9.57E+02
1.04E+03
3.81 E+03
4.61 E+04
3.02E+03
3.40E+03
1.75E+04
3.66E+03
1.31E+03
5.76E+03
2.39E+03
3.84E+03
1.68E+03
4.00E+05
2.83E+05
3.58E+05
5.06E+03
2.53E+04
1.01E+00
1.28E+04
1.23E+02
1.18E+04
6.56E+02
1.01E+03
5.20E+03
1.50E+04
1.89E+04
3.41 E+03
APPENDIX B— BETA PARTICLE AND
PHOTON EMITTERS— Continued
Nucllde
RA-225 	
RA-228 	
AC-227 	
AC-228 	
TH-231 	
TH-234 	
PA-233 	
PA-234 	
PA-234M 	
U-237 	
U-240 	
NP-236 	
NP-238 	
NP-239 	
NP-240 	
NP-240M 	
PU-241 	


Ch— Concentratic
assuming 2 liters d£
APPENDIX
NUCLIDE
0\Jt -\A-7
BI-210
RI-91 1
pn_9in
PO-212 	
PO-21 3
PO-214
PO-21 5
PO-21 6 	
PO-21 8
AT-217
FR-221 	
RA-223 	
RA-224 	
RA-226 	
AC-225 	
TH-227 	
TH-228 	
TH-229 	
TH-230 	
TH-232 	
PA-231 	
U-232 	
U-233 	
U-234 	
U-235 	
(J-236 	
U-238 	
NP-237
PU-236 	
PU-238 	
PU-239 	
PU-240 	
PU-242 	
PU-244 	
AM-241 	
AM-242 	
AM-243 	
CM-242 	
CM-244 	
CM-245 	




n in water for
i ly intake.
C — ALPHA EN/
Cm (pCi/liter)
1.06E+02
1.94E+03
2.05E+05
1.40E+01
1.15E+14
8.03E+12
2.43E+11
9.17E+09
7.38E+07
9.50E+04
5.74E+08
4.50E+04
3.21E+01
5.46E+01
2.07E+01
1.85E+02
6.62E+02
1.53E+02
5.15E+01
8.27E+01
9.18E+01
1.02E+01
1.02E+01
2.56E+01
2.59E+01
2.65E+01
2.74E+01
2.62E+01
7.19E+00
3.33E+01
7.15E+00
6.49E+01
6.49E+01
6.83E+01
7.02E+00
6.45E+00
8.66E+03
6.49E+00
1.45E+02
8.47E+00
1.00E+01
6.35E+00
6.38E+00
6.93E+00
1.71E+00
1.70E+01
Ch (pCi/liter)
9.14E+00
7.85E+00
1.27E+00
3.27E+03
4.07E+03
4.01 E+02
1.51 E+03
2.56E+03
9.30E + 05
1.78E+03
1.54E+03
5.96E+03
1.39E+03
1.68E+03
2.31 E+04
1.74E+05
6.26E+01
1.64E+04
1.27E+00
4 mrem ede/y,
UTTERS
Ci (pCi/liter)
1.04E+02
1.01 E+03
1.56E+05
7.46E+00
8.78E+13
6.06E+12
1.86E+11
6.84E+09
5.30E+07
6.91 E+04
4.27E+08
3.26E+04
2.41 E+01
4.06E+01
1.57E+01
1.13E+02
4.03E+02
1.25E+02
4.93E+01
7.92E+01
8.80E+01
1.02E+01
5.72E+00
1.38E+01
1.39E+01
1.45E+01
1.47E+01
1.46E+01
7.06E+00
3.23E+01
7.02E+00
6.21 E+01
6.22E+01
6.54E+01
6.87E+00
6.34E+00
5.34E+03
6.37E+00
1.33E+02
8.30E+00
9.84E+00
6.23E+00
6.27E+00
6.79E+00
1.67E+00
1.62E+01
Cm = Concentration in water for lifetime mortality
risk=1x10-1
Ci= Concentration in water for lifetime incidence
risk=1x10'4
Both assume 2 liters daily intake of water.
List of Subjects in 40 CFR Parts 141 and
142
Chemicals, Reporting and record
keeping requirements, Water supply,
Administrative practice and procedure.
Dated: June 17, 1991.
William K. Reilly,
Administrator, Environmental Protection
Agency.
For the reasons set forth in the
preamble, title 40 of the Code of Federal
Regulations is proposed to be amended
as follows:
PART 141— NATIONAL PRIMARY
DRINKING WATER REGULATIONS
1. The authority citation for part 141
continues to read as follows:
Authority: 42 U.S.C. 300f, 300g-l, 300g-2,
300g-3, 300g-4, 300g-5, 300g-6, 300J-4 and
300J-9.
2. Section 141.2 is amended by adding,
in alphabetical order, a definition for
"adjusted gross alpha" as follows:
§ 141.2 Definitions
*****
Adjusted gross alpha: Adjusted gross
alpha is defined as the result of a gross
alpha measurement, less radium-226 and
less uranium. Radon is not included in
adjusted gross alpha.
*****
3. Section 141.15 is amended by
revising the introductory text to read as
follows:
§ 141.15 Maximum contaminant levels for
radium-226, radium-228, and gross alpha
particle radioactivity in community water
systems.
The following are the maximum
contaminant levels for radium-226,
radium-228, and gross alpha particle
radioactivity, which shall remain
effective until [insert date 18 months
after publication of the final rule in the
Federal Register];
*****
4. Section 141.16 is proposed to be
amended by adding introductory text to
read as follows:
§ 141.16 Maximum contaminant levels for
beta particle and photon radioactivity from
man-made radionuclides in community
water systems.
The following maximum contaminant
levels shall remain effective until [insert
date 18 months after publication of the
final rule in the Federal Register];
*****
5. Section 141.25 is amended by
revising the section to read as follows:

-------
 33122
Federal  Register / Vol. 56, No. 138 / Thursday, July 18, 1991 /  Proposed Rules
 § 141.25  Sampling and analytical methods
 for radionuclides.
   The current analytical methods
 outlined in § 141.25 and the monitoring
 requirements in § 141.26 shall remain
 effective until [insert date 18 months
 after promulgation of the final rule].
 After that date, the monitoring and
 analytical methods specified below will
 be effective. Community water systems
 and non-transient, non-community
 water systems shall conduct monitoring
 to determine compliance with the
 maximum contaminant levels specified
 in § 141.64 in accordance with this
 section.
   (a) Monitoring shall be conducted as
 follows:
   (1) Groundwater systems shall take a
 minimum of one sample at every entry
 point to the distribution system which is
 representative of each well after
 treatment (hereafter called a sampling
 point) beginning in the compliance
 period starting January 1,1996. The
 system shall take  each sample at the
 same sampling point unless conditions
 make another sampling point more
 representative of each source or
 treatment plant.
  (2) Surface water systems shall take a
 minimum of one sample at every entry
 point to the distribution system after
 any application of treatment or in the
 distribution system at a point which is
 representative of each source after
 treatment (hereafter called a sampling
 point) beginning in the compliance
 period starting January 1,1996. The
 system shall take each sample at the
 same sampling point unless conditions
 make another sampling point more
 representative of each source or
 treatment plant.
  (3) If a system draws water from more
 than one source and the sources are
 combined before distribution, the
 system must sample at an entry point to
 the distribution system during periods of
normal operating conditions (i.e., when
water is representative of all sources
being used).
  (4) The State may reduce the total
number of samples which must be
analyzed by allowing the use of
compositing, except for radon and gross
beta samples, which may not be
composited.
  (i) Composite samples from a
maximum of five sampling points within
one system are allowed. Compositing of
samples must be done in the laboratory.
  (ii) If the concentration hi the
composite sample  is greater than or
equal to 3 pCi/1 of any radionuclide, the
individual non-composited samples, or if
these are not available, follow-up
samples must be analyzed to identify
the sampling points which may violate
                       one of the MCLs. Any follow-up samples
                       must be taken within 14 days at each
                       sampling point included in the
                       composite. Samples must be analyzed
                       for the contaminants which were
                       detected in the composite sample.
                         (5) The frequency of monitoring for
                       radon shall be in accordance with
                       paragraph (b) of this section; the
                       frequency of monitoring for radium-226,
                       radium-228, uranium, and adjusted gross
                       alpha shall be in accordance with
                       paragraph (c) of this section; and the
                       frequency of monitoring for beta and
                       photon emitters shall be in accordance
                       with paragraph (d) of this section.
                         (b) The frequency of monitoring
                       conducted to determine compliance with
                       the maximum contaminant level for
                       radon specified in § 141.64 shall be
                       conducted as follows:
                         (1) Groundwater systems or systems
                       using both ground and surface water are
                       required to take four consecutive
                       quarterly samples during the first year
                       of each three-year compliance period of
                       each nine-year compliance cycle.
                       Annual samples are required in the
                       second and third years of each
                       compliance period. The initial
                       monitoring for radon must be completed
                       by January 1,1999.
                         (2) Surface water systems are not
                       required to monitor for radon. The State
                       may require it.
                         (3) The State may grant a waiver to
                       ground water systems or systems that
                       use both ground and surface water for
                       monitoring requirements in paragraph
                       (b)(l) of this section, provided that they
                       have monitored quarterly in the initial
                       year, and completed annual testing in
                       the second and third year of the first
                       compliance period (at least one sample
                       shall have been taken since January 1,
                       1990). Groundwater systems shall
                       demonstrate that all previous analytical
                       results were less than the maximum
                       contaminant level. Systems that use a
                       new water source are not eligible for a
                       waiver until 4 quarters of monitoring
                       and two subsequent years of a single
                       annual sample of the new source has
                       been completed.
                         (4) The State may grant a waiver if the
                       State determines that the system is
                       reliably and consistently below the
                       MCL, based on a consideration of the
                       following factors:
                         (i) Potential radon contamination of
                       the water source due to the geological
                       characteristics  of the area where the
                       water source is located.
                         (ii) Previous analytical results.
                         (5) A condition of the waiver shall
                       require that a system take a minimum of
                       1 sample every three-year compliance
                       period.
   (6) A waiver remains in effect until
 the completion of the nine-year
 compliance cycle. Systems not receiving
 a waiver must monitor in accordance
 with the provisions of paragraph (b)(l)
 of this section.
   (7) A decision by the State to grant a
 waiver shall be made in writing and
 shall set forth the basis for the
 determination. The determination may
 be initiated by the State or upon an
 application by the public water system.
 The public water system shall specify
 the basis for its request. The State shall
 review and, where appropriate, revise
 its determination of appropriate
 frequency.
   (8) A system which exceeds the
 maximum contaminant level in § 141.64
 of this part shall monitor quarterly
 beginning in the next quarter after the
 violation occurred. Quarterly monitoring
 must continue until the average of 4
 consecutive quarterly samples is below
 the MCL.
   (9) If monitoring data collected after
 January 1,1990 are generally consistent
 with the requirements of § 141.25, then
 the State may allow systems to use
 those data to satisfy the monitoring
 requirement for the initial compliance
 period.
   (c) The frequency of monitoring
 conducted to determine compliance with
 the maximum contaminant levels in
 § 141.64 for radium-226, radium-228,
 uranium, and adjusted gross alpha shall
 be as follows:
   (1) Groundwater systems, surface
 water systems and systems using both
 ground and surface water shall take one
 sample annually at each sampling point
 during each compliance period starting
 in the compliance period beginning
 January 1,1996. If all samples are less
 than the MCL, then monitoring can be
 reduced to one sample per compliance
 period, in accordance with paragraphs
 (c) (2) through (6) of this section.
   (2) Systems may apply to the State for
 a waiver from the monitoring
 frequencies specified in paragraph (c)(l)
 of this section, if they have completed
 the required three annual samples in the
 first three-year compliance period.
 Systems that use a new water source
 are not eligible for a waiver until three
 years of monitoring of the new source
 has been completed.
   (3) The State may grant a waiver if it
 finds that the system is reliably and
 consistently below the MCL, based on a
 consideration of the following factors:
   (i) Potential contamination of the
water source; and
   (ii) Previous analytical results.
   (4) A condition of the waiver shall
require that a system take a minimum of

-------
                Federal Register / Vol. 56,  No. 138  /  Thursday, July  18, 1991 / Proposed Rules
                                                                                                             33123
one sample during the effective period
of the waiver. The term during which the
waiver is effective shall not exceed one
nine-year compliance cycle.
  (5) The State may grant a waiver
provided water systems have monitored
annually for at least three consecutive
years. (At least one sample shall have
been taken since January 1,1990.) Both
surface and groundwater systems shall
demonstrate that all previous analytical
results were less than the maximum
contaminant level. Systems that use a
new water source are not eligible for a
waiver until three consecutive annual
samples from the new source have been
collected and analyzed.
  [6) A decision by the  State to grant a
waiver shall be made in writing and
shall set forth the basis for the
determination. The determination may
be initiated by the State or upon an
application by the public water system.
The public water system shall specify
the basis for its request. The State shall
review and, where appropriate, revise
its determination of the appropriate
monitoring frequency when the system
submits new monitoring data or when
other data relevant to the system's
appropriate monitoring frequency
become available.
  (7) Systems which exceed the
maximum contaminant levels in § 141.64
of this part shall monitor quarterly
beginning in the next quarter after the
violation occurred. Quarterly monitoring
must continue until 4 consecutive
quarterly samples are below the MCL.
  (8) If monitoring data collected after
January 1,1985 are generally consistent
with the requirements of § 141.25, then
the State may allow systems to use
these data to satisfy the monitoring
requirements for the initial compliance
period beginning January 1,1996, except
at least one sample shall have been
collected since January 1,1990.
   (d) The frequency of monitoring
conducted to determine compliance with
the maximum contaminant levels in
 § 141.64 for beta and photon emitters
shall be as follows:
   (1) Only systems (both surface and
ground water) determined by the State
to be vulnerable need to sample for beta
and photon emitters. Vulnerability shall
be based on the proximity of the water
source(s) to facilities using or producing
radioactive materials. Vulnerable
systems shall monitor quarterly for beta
and annually for tritium and strontium,
beginning in the compliance period
 starting January 1996. Systems must
begin monitoring within one quarter
after being notified by the State that the
system is vulnerable. Existing
vulnerability determinations by the
State shall remain effective until the
State reviews and either reaffirms them
or revises them.
  (2) Systems determined to be
vulnerable may not apply to the State
for a waiver from the monitoring
frequencies specified in paragraph (d)(l)
of this section.
  (3) If the gross  beta particle activity
exceeds 50 pCi/1, an analysis of the
sample must be performed to identify
the major radioactive constituents
present in the sample and the
appropriate doses shall be calculated
and summed to determine compliance
with § 141.64, using appendix B,  [insert
citation for final  Federal Register].
Measured levels  of tritium and strontium
shall be included in this calculation.
Doses shall also  be calculated and
combined for measured levels of tritium
and strontium to determine compliance.
  (4) Suppliers of water shall conduct
additional monitoring as directed by the
State, to determine the concentration of
man-made radioactivity in principal
watersheds designated by the State.
  (5) Vulnerable systems which exceed
the maximum contaminant levels in
§ 141.64 shall monitor monthly beginning
in the next month after the violation
occurred. Monthly monitoring shall
continue until the system has
established, by a rolling average of 3
monthly samples, that the MCL is being
met.
  (e) Confirmation samples:
  (1) Where the  results of sampling for
radon, radium-226, radium-228, adjusted
gross alpha, uranium, and beta and
photon emitters  indicate an exceedence
of the maximum contaminant level, the
State may require that one additional
sample be collected as soon as possible
after the initial sample was taken (but
not to exceed two weeks) at the same
sampling point.
  (2) If a State-required confirmation
sample is taken  for any contaminant,
then the results of the initial and
confirmation sample shall be averaged.
The resulting average shall  be used to
determine the system's compliance in
accordance with paragraph (h) of this
section. States have the discretion to
delete results of obvious sampling or
analytic errors.
   (f) The State may require more
frequent monitoring than specified in
paragraphs  (b), (c), and (d) of this
section or may require confirmation
samples for positive and negative results
at its discretion.
  (g) Systems may apply to the State to
conduct more frequent monitoring than
the minimum monitoring frequencies
specified in this section.
  (h) Compliance with §§ 141.15,141.16,
and 141.64 (as appropriate) shall be
determined based on the analytical
result(s) obtained at each sampling
point.
  (1) For systems which are conducting
monitoring at a frequency greater than
annual, compliance with the maximum
contaminant levels for radon, radium-
226, radium-228, adjusted gross alpha,
uranium, and beta and photon emitters
is determined by a running annual
average at each sampling point. If the
average at any sampling point is greater
than the MGL, then the system is out of
compliance. If any one sample would
cause the annual average to be
exceeded, then the system is out of
compliance immediately. Any sample
below the detection limit shall be
calculated at one-half the detection limit
for the purpose of determining the
annual average.
  (2) For systems which are monitoring
annually, or less frequently, the system
is out of compliance with the maximum
contaminant levels for radon, radium-
226, radium-228, adjusted gross alpha,
uranium, and beta and photon emitters
if the level of a contaminant at any
sampling point is greater than the MCL.
If a confirmation sample is required by
the State, the determination of
compliance will be based on the average
of the two samples.
  (3) If a public water system has a
distribution system separable from other
parts of the distribution system with no
interconnections, only those parts of the
system that exceed the MCL need to
conduct increased monitoring.
   (4) If a public water system has a
distribution system separable from other
parts of the distribution system with no
interconnections, the State may allow
the system to give public notice to only
the area served by that portion of the
system which is out of compliance.
   (i) Each public water system shall
monitor at the time designated by the
State during each compliance period.
   (j) Radionuclides analysis:
   (1) Analysis for radon, radium-226,
radium-228, adjusted gross alpha,
uranium, and beta and photon emitters
shall be conducted using the following
methods:

-------
 33124
Federal Register /  Vol.  56, No. 138  / Thursday, July 18,  1991 /  Proposed  Rules
                                     PROPOSED METHODOLOGY FOR RADIONUCLIDE CONTAMINANTS
Contaminant
Naturally
occurring
Gross alpha and
beta.
Gross alpha 	
Radium 226 	
Radium 228 	
Radon 222 ..„ 	
Uranium 	
Man-made
Radioactive
cesium.
Radioactive
iodine.
Radioactive
strontium 89,
90.
Tritium 	
Gamma and
photon
emitters.
Methodology
Evaporation 	 	
Co-precipitation ...
Radon
emanation.
Radiochemical 	
Radiochemical 	
Liquid
scintillation.
Lucas cell 	
Radiochemical 	
Fluorometric 	 	
Alpha
spectrometry.
Precipitation 	
Precipitation 	
Precipitation 	
Radiochemical 	
Liquid
scintillation.
Gamma ray
Spectrometry.
Reference (method or page number)
EPA1
900.0
903.1
903.0
904.0

908.0
908.1
901.0
902.0
905.0
906.0
901.1
EPA2
pp. 1-3 	

nn 16-23

pp 24-28





pp. 4-5 	 	
pp. 29-33. 	 	
pp. 108-114 	
pp. 34-40 	

EPA3
00-01 ....
00-02....
Ra-03 ...
Ra-05 ...
Ra-05 ...



00-07....
1-01 .......
Sr-04 	
H-02.::...

EPA4
p. 1 	

p. 19 	
p. 19 	



p. 33 	
p. 65 	 	

p. 87 ......

SM6
7110 B 	

7500-RaB 	
7500-Ra D*...

7500-U B 	
7500-UC 	 	
7500-Cs B
7500-I B... 	
7500-Sr B

7500-3H B 	

ASTM6
D 1943-81

D 3454-86 	


D 3972-82
D 2907-83 	
D 2334-88


D 2476-81 (87) 	
D-3649-85 	
USGS7
R-1 120 76



R-1 142-76 	


R-1 180-76 	
R-1181-76 	
R-1 182-76......
R-1 110-87 	
R-1 160-76



DOE8






E-U-03 	
E-U-04 	
E-Cs-01 	

E-Sr-01
4.5.2.3 	
Other
N.Y.9
N.Y.9
N.J.10
913"
LS>2
LC12
         A* Kon8d/S!f to Measurement of Radioactivity in Drinking Water," EPA Environmental Monitoring and Support Laboratory, Cincinnati, OH (EPA-600/
         AUgUSt 10oU. (trA, 1980).                  .  -                      -       -
    2 "Interim Radiochemical Methodology for Drinking Water," EPA-600/4-75-008, March 1976. (EPA, 1976)
    'Eastern Environmental Radiation  Facility, Montgomery, AL 36109, "Radiochemical Procedures Manual,"  EPA 520/5-84-006, August 1984. (EPA, 1984a).
    I ,.S?  2P f m.'u Analvtical Procedures for Analysis of Environmental Samples," EMSL-LV-0539-17, March 1979. (EPA, 1976b).
Pollution Control Federato^iglg^rAPHA^gs'g)^'3161' ^ Wastewater'" 17th edition- American Public Health Association, American Water Works Association, Water
    "Iteth A.nn«ualIB°ok 9' ASTM Standards, Vol.  11.02, American Society for Testing  and Materials, 1916 Race Street, Philadelphia, Pa. 19103. (ASTM, 1989)
United aatesGeolofca'i'surve0nch  ?  AS TuSGS8'3"068 '" Wat6r 3nd F'UV'al Sediments'" Book 5- 1989- Techniques of Water-Resources Investigations of the
    8 Environmental Measurements Laboratory, U.S. Department of Energy, "EML PROCEDURES MANUAL, 27th edition." (DOE, 1990).
/Bowi.JiiMnT1?!!!?? XivM^iSRi ^JJ? (Ra-°2)- Radiological Sciences Institute Center for Research—New York State Department of Health, January 1980
^nevisea June laoz.). (NY otcltG DOH, 1982).
„*  ".'^eterrojnat'on of.Radiu.m,228Jn Prinkin3 Water," State of New Jersey—Department of Environmental Protection—Division of Environmental Quality—Bureau
of Radiation and Inorganic Analytical Services, August 1990. (NJ DEQ, 1990)
      Method91~Rad0 indnnking water by liquid scintillation, Environmental Monitoring and Support Laboratory, Las Vegas, NV.  (EPA 1991q).
                                                                      Drinkln9  Water'" * 22> Tw° Test *«»***>.«* Radon In Drinking Water.
   (2) Sample collection for radon,
radium-226, radium-228, adjusted gross
alpha, uranium, and beta and photon
                           emitters under this section shall be
                           conducted using the sample
                           preservation, container, and maximum
holding time procedures specified in the
table below:
Sampling handling, preservation, holding times
Parameter
Gross alpha 	
Gross beta 	 	 	
Radium-226 	
Radium-228 	
Radon-222 6 	
Uranium natural 	
Radioactive Cesium 	 	
Radioactive Strontium 	
Radioactive Iodine 	
Tritium 	
Photon emitters 	

Preservative >
Cone. HCI or HNO3 to pH <2 *
Cone. HCI or HNO3 to pH <2 *
Cone. HCI or HNOs to pH <2
Cone. HCI or HNO3 to pH <2
Cool 4°C 	
Cone. HCI or HNO3 to pH <2 	
Cone. HCI to pH <2 	
Cone. HCI or HNQ, to pH <2 	
None 	
None 	
Cone. HCI or HNO3 to pH <2

Container 2
P or G
P or G
P or G
P or G

P or G
P or G


Q
P or G

Maximum
holding
time3
6 months.
6 months.
6 months.
6 months.
4 days
6 months!
6 months.
6 months.
6 months.
6 months.
6 months.
    »(All except radon-22 samples). It is recommended that the preservative be added to the sample at the time of collection unless suspended solids activity is to
SLEfS!???;J±^l'hl^±J]^                                              °f t"e sample (in its original container) may be delayed for a
period not to exceed 5 days. A minimum of 16 hours'must elapse between'addfficaUoTTmdTma'iysisr'
    2 P=Plastic, hard or soft; G=Glass, hard or soft.

-------
                Federal Register  /  Vol. 56. No.  138 / Thursday,  July  18, 1991 / Proposed Rules          33125
    3 Holding lima is defined as the period from time of sampling to time of analysis. In all cases, samples should be analyzed as soon after collection as possible.
    4 If HCI Is used to acidify samples which are to be analyzed for gross alpha or gross beta activities, the acid salts must be converted to nitrate salts before
transfer of the samples to planchets.                                               •                             .  _  „„„„ „   ,.
    *Tht procedure  of a positive pressure collection in 60-ml glass bottles is to be followed. This procedure is described in appendix C,  NIRS Sampling
Instructions—Radon, p. 26, Two Test Procedures For Radon In Drinking Water, Interlaboratory Collaborative Study, EPA/600/2-87/082, March 1987.
  (3) Analysis under this section shall
only be conducted by laboratories that
have received approval by EPA or the
State. To receive approval to conduct
analyses for radon, radium-226, radium-
228, adjusted gross alpha, uranium, and
beta and photon emitters the laboratory
must:
  (i) Analyze Performance Evaluation
samples which include those substances
provided by EPA Environmental
Monitoring and Support Laboratory or
equivalent samples provided by the
State.
  (ii) Achieve quantitative results on the
analyses that are within .the following
acceptance limits:
Contaminant
Radium-226............ 	
RadKim-228.*,.. 	
Radon-222 *.„ 	
Gross alpha emitters 	
Gross beta emitters 	
Radioactive Cesium. 	
Radioactive Strontium
total, 69 and 90.
THtiuffit<»>»!» 	 	 	
Acceptance Limits '
±30% at £5 pCi/l.
±50% at £5 pCi/l.
±30% at ^5 pCi/l.
±30% at £300 pCi/l.
±50% at £l5pCi/l.
±30% at £30 pCi/l.
±30% at £lOpCi/l.
±20% at ^20 pCi/l.
±30% at £5pCi/l.
±20% at £l200pCi/l.
  1 Acceptance limits based on 100 minute count.
  * Radon acceptance  limits  based  on  4 day
elapsed time from sample  collection  to analysis.

  6. Section § 141.32 is amended by
adding paragraphs (e)(77) through (82],
to read as follows:

§ 141.32  Public notification.
*****
  (e)  *  * *
  (77] Radon: The United States
Environmental Protection Agency (EPA)
sets drinking water standards and has
determined that radon is of health
concern at certain levels of exposure.
Radon is a naturally occurring
radioactive contaminant that occurs hi
ground water. It is a gas, and is released
from  water into household air during
water use. Radon has been found in
epidemiology studies to cause lung
cancer hi humans at high exposure
levels; at lower exposure levels the risk
of lung cancer is reduced. EPA has set
the drinking water standard for radon in
public water supplies at 300 picocuries
per liter (pCi/l) to protect against lung
cancer risk. Drinking water that meets
the EPA standard is associated with
little  of this risk and is considered safe
for radon.
  (78) Radium 226: The United  States
Environmental Protection Agency (EPA)
sets drinking water standards and has
determined that radium 226 is of health
concern at certain levels of exposure.
Radium 226 is a naturally occurring
radioactive contaminant that occurs
primarily in ground water. Radium 226
has been found in epidemiology studies
to cause bone cancer in humans at high
exposure levels, and is believed to cause •
other cancers as well; at lower exposure
levels the risk of cancer is reduced. EPA
has set the drinking water standard for
radium 226 at 20 picocuries per liter
(pCi/l) to protect  against cancer risk.
Drinking water that meets the EPA
standard is associated with little of this
risk and is considered safe for radium
226.
  (79) Radium 228: The United States
Environmental Protection Agency (EPA)
sets drinking water standards and has
determined that radium 228 is of health
concern at certain levels of exposure.
Radium 228 is a naturally occurring
radioactive contaminant that occurs
primarily in ground water. Radium 228
has been found in epidemiology studies
to cause bone cancer in humans at high
exposure levels and is believed to cause
other cancers as well; at lower exposure
levels the risk of bone cancer is reduced.
EPA has set the drinking water standard
for radium 228 and 20 picocuries per
liter (pCi/l) to protect against cancer
risk. Drinking water that meets the EPA
standard is associated with little of this
risk and is considered safe for radium.
   (80) Uranium: The United States
Environmental Protection Agency (EPA)
sets drinking water standards and has
determined that uranium is of health
concern at certain levels of exposure.
Uranium is a naturally occurring
radioactive contaminant that occurs in
both ground and surface water. Uranium
is believed to cause bone cancer and
other cancers in humans at high
exposure levels; at lower exposure
levels the risk of cancer is reduced. EPA
also believes uranium can be toxic to
the kidneys. EPA has  set the drinking
water standard for uranium at 20
micrograms per liter (jugl) to protect
against both cancer risk and risk of
kidney damage. Drinking water that
meets the EPA standard is associated
with little of this risk and is considered
safe for uranium.
   (81) Gross Alpha: The United States
Environmental Protection Agency (EPA)
sets drinking water standards and has
determined that alpha emitting
radionuclides may be of health concern
at certain levels of exposure. Alpha
emitters are primarily naturally
occurring radioactive contaminants, but
several derive from man-made sources.
They may occur in either ground or
surface water. Alpha emitters are
believed to cause cancer in humans at
high exposure levels because they emit
ionizing radiation. At lower levels, the
risk of cancer is reduced. EPA has set
the drinking water standard for alpha
emitters at 15 picocuries per liter (pCi/l)
to protect against cancer risk. Drinking
water that meets the EPA standard is
associated with little of this risk and is
considered safe for alpha emitters.
   (82) Beta and photon emitters: The
United States Environmental Protection
Agency (EPA) sets drinking water
standards and has determined that beta
and photon emitting radionuclides may
be of health concern at certain levels of
exposure. Beta and photon emitters are
primarily man-made radioactive
contaminants associated with the
operation of nuclear power facilities,
facilities using radioactive material for
research or manufacturing, or facilities
where these materials are disposed.
Some beta emitters are naturally
occurring. Beta and photon emitters are
expected to occur primarily in surface
water. Beta and photon emitters are
believed to cause cancer in humans at
high exposure levels because they emit
ionizing radiation. At lower levels, the
risk of cancer is reduced. EPA has set
the drinking water standard for beta and
photon emitters at 4 millirems effective
dose equivalent per year (mrem ede/yr)
to protect against cancer risk. Drinking
water  that meets the EPA standard is
associated with little of the risk and is
considered safe for beta and photon
emitters.
*****
   7. A new section § 141.44 is added to
subpart E to read as follows:

§ 141.44  Special monitoring for
radionuclides.
   (a) Each community and non-
transient, non-community water system
shall take one sample at each sampling
point for lead-210 and report the results
to the  State. Monitoring must be
completed by December 1996.
   (b) Groundwater systems shall take a
minimum of one sample at every entry
point to the distribution system which is
representative of each well after
treatment (hereafter called a sampling
point). Each sample must be taken at the
same sampling point unless conditions

-------
 33126
Federal Register  /  Vol. 56, No. 138  /  Thursday, July 18, 1991 / Proposed Rules
 make another sampling point more
 representative of each source or
 treatment plant.
   (c) Surface water systems.
   Note: For purposes of this paragraph,
 surface water systems include systems with a
 combination, of surface and ground sources.
 shall take a minimum of one sample at
 points in the distribution system that are
 representative of each'source or at each
 entry point  to the distribution system
 after treatment (hereafter called a
 sampling point). Each sample must be
 taken at the same sampling point unless
 conditions make another sampling point
 more representative of each source or
 treatment plant.
   (d) If the system draws water from
 more than one  source and the sources
 are combined before distribution, the
 system must sample at an entry-point to
 the distribution system during periods  of
 normal operating conditions (i.e., when
 water representative of all sources is
 being used).              '  •
   (e) The State may require a
 confirmation sample for positive or
 negative results.
   (f) Instead of performing the
 monitoring required by this section, a
 community water system or non-
 transient non-community water  system
 serving fewer than 150 service
 connections may send a letter to the
 State stating that the system is available
 for sampling. This letter must be sent to
 the State by January 1,1996. The system
 shall not send such samples to the State,
 unless requested to do so by the State.
   8. A new § 141.53 is added to subpart
 F to .read as follows:

 § 141.53  Maximum contaminant level
 goals for Radionuclides.
  MCLGs for the radionuclides are as
 follows:
Contaminant
Radon 222 	 	 	
Radium-226 	 	 	 	
Radium-228 .., 	
Uranium. 	 , 	 „„•.
Gross alpha emitters 	 , 	 	 	 .'. 	
Beta arid photon emitters 	 	 	 	 	 	

MCLG




Zero
Zero

  9. A new/section 141.64 is added to
subpart G to read as follows:

§ 141.64  Maximum contaminant levels for
radionuclides.      '
  (a) The following maximum
contaminant levels for Radionuclides
apply to community and.non-transient,
non-community water systems. The
effective date for these MCLs is [insert
date 18 months a'fter publication of the
final rule in the Federal Register].
Contaminant
(1) Radon-222 	 	 	
(2) Radium-226 	
(3) Radium-228 	 	
(4) Uranium ..„ 	
(5) Adjusted gross alpha. .
(6) Beta particle and photon
emitters.
MCL
300 pCi/l
20 pCi/l
20 pCi/l
20 u.q/1 i
15 pCi/l
4 mrem ede/yr.2
                          1 NOTE 20 ug/l uranium is approximately equal to
                        30 pCi/l, using an activity-to-mass conversion of 1.3
                        pCi/ug. The activity-to-mass ratio can vary depend-
                        ing on the relative amounts  of uranium-234, -235
                        and  -238 that  are present in a sample. The MCL
                        applies to the total mass of uranium in the sample.
                          2 NOTE: The  unit mrem ede/yr refers to the dose
                        committed over a period of 50 years to reference
                        man (ICRP 1975) from an annual intake at the rate
                        of 2 liters of drinking water per day.

                          (b) The Administrator, pursuant to
                        section 1412 of the Act, hereby identifies
                        as indicated in the table below the best
                        technology,  treatment technique, or
                        other means available for achieving
                        compliance  with the maximum
                        contaminant level for Radionuclides
                        identified in paragraph (a) of this
                        section:

                           BAT FOR RADIONUCLIDES LISTED IN
                                    SECTION 141.64
                         Contaminant
                        Radon 222.......

                        Radium 226	

                        Radium 228	

                        Uranium (N)	
                        Alpha particle
                         emitters.
                        Beta particle
                         and photon
                         emitters.
                                                BAT
Aeration: Packed tower, spray, slat
  tray and other forms.
Ion exchange, Reverse osmosis,
  Lime softening.
Ion exchange, Reverse osmosis,
  Lime softening.
Ion exchange, reverse osmosis,
  Lime softening, coagulation/fil-
  tration.
Reverse osmosis.

Mixed bed ion exchange, Reverse
  osmosis.
                        PART 142—NATIONAL PRIMARY
                        DRINKING WATER REGULATIONS
                        IMPLEMENTATIONS

                          1. The authority citation for part 142
                        continues to read as follows:
                          Authority: 42 U.S.C. 300g, 300g-l, 300g-2,
                        300g-3, 300g-4, 300g-5, 300g-6, 300J-4, and
                        300J-9.
                          2. Section 142.14 is amended by •
                        adding paragraphs (d)(12) and (d)(13).

                        § 142.14  Records kept by States.
                        *...   *•    *'.-   *    *
                          (d) * * *
                          (12) Records of any determination of a
                        system's vulnerability to contamination
                        from photon and beta emitters due to
                        their proximity to an emitting source or
                        use.of source water influenced by a
                        source of radiation.  The records shall
                        also include the basis for such .
                        determination.
   (13) Records of all current monitoring
 requirements and the most recent  •
 monitoring frequency decision
 pertaining to each contaminant,
 including the monitoring results and
 other data supporting the decision, the
 State's findings based on the supporting
 data and any additional bases for such
 decision; records shall be kept in
 perpetuity or until a more recent
 monitoring frequency decision has been
 issued.
 *     *<•*>.*,   p *
   3. In § 142.15 is amended by adding a
 new paragraph  (c)(5) to read as follows:

 § 142.15 Reports by States.
 *     *    ,*^ .   •*     *
   (c) * * *
 .  (5) The results of monitoring for the
 unregulated contaminants in § 141.44
 shall be reported within one quarter
 after the December 1996 completion date
 for monitoring lead-210.
 *     *    *     *     * _
   4. Section 142.16 is amended by
 adding a new paragraph (f) to read as
 follows:

 § 142.16 Special primacy requirements.
 *     *    *   -  *     *
   (f) An application for approval of a
 State program revision for
 Radionuclides which adopts the
 requirements specified in § § 141.25,
 141.32,141.44, and 141,64.must contain
 the following (in addition to the general
 primacy requirements enumerated
 elsewhere in this part, including the
 requirement that state regulations be  at
 least as stringent as the federal
 requirements):
   (1) If a State chooses to issue waivers
 from the monitoring requirements in
 §§141.25 and 141.44, the State shall
 describe the procedures and criteria
 which it will use to review waiver
 applications and issue waiver
 determinations.
   (i) The procedures for each
 contaminant or class of contaminants
 shall include a description of:
   (A) The waiver application
requirements;        .
   (B) The State review process for
reviewing waiver applications;
   (ii) The State decision criteria,
including the factors that will be
considered in deciding to grant or deny
waivers. The decision criteria must
include the factors specified in
§§ 141.25(b)(4) and 141.25(c)(3).
  (2) A State shall determine what
systems are vulnerable to beta and
photon emitting  sources. States shall
specify the procedures they will use to
decide'which systems are vulnerable.
Vulnerability of each public water

-------
                Federal Register  /  Vol. 56,  No. 138 / Thursday, July 18, 1991  / Proposed Rules
                                                                                                               33127
system shall be determined by the State
based upon an assessment of the
following factors:
  (i) Proximity of water system to a
potentially discharging source, such as a
nuclear power facility, or where there is
a commercial or industrial use, disposal,
or storage of the materials;
  (ii) Previous monitoring results; and
  (iii) Use of water influenced by a
nuclear power facility or other potential
discharger.
  5. A new § 142.65 is added to subpart
G to read as follows:
§ 142.65  Variances and Exemptions from
the maximum contaminant levels for the
radlonucllde contaminants listed In
§ 141.64.
  (a) The Administrator, pursuant to
section 1415(a)[l)(A) of the Act, hereby
identifies the following as the best
technology, treatment techniques, or
other means available for achieving
compliance with the maximum
contaminant levels for the radionuclides
listed in  § 141.64, for the purpose of
issuing variances and exemptions.

    BAT FOR RADIONUCLIDES LISTED IN
               §141.64
Contaminant
Radon 222 	
Radium 226 	
Radium 228 	 	 	
Uranium (N) 	
Gross alpha particle emitters 	
Gross beta particle and photon
emitters.
BAT
1.
2,3,4.
2,3,4.
2,3,4,5.
3.
3,6.
  Key to BATs in table:
  1» Aeration: Packed Tower, spray, slat tray and
other forms.
  2 »lon exchange.
  3«Rovors9 osmosis.
  4-Limo softening; except for systems serving 500
or fewer connections.
  5=Coagulation/filtration; except for systems serv-
ing 500 or fawor connections.
  6=Mixed bed ton exchange.

  (b) A State shall require community
water systems and non-transient, non-
community water systems to install
and/or use any treatment method
identified in § 141.64 as a condition for
granting a variance except as provided
in paragraph (c) of this section. If, after
the system's installation of the treatment
method, the system cannot meet the
MCL, that system shall be eligible for a
variance under the provisions of section
1415(aHlHA) of the Act.
  (c) If a system can demonstrate
through comprehensive engineering
assessment, which may include pilot
plant studies that the treatment methods
identified in § 141.64 would only achieve
a de minimis reduction in contaminants,
the State may issue a schedule of
compliance that requires the system
being granted the variance to examine
other treatment methods as a condition
of obtaining the variance.
  (d) If the State determines that a
treatment method identified in
paragraph [c] of this section is
technically feasible, the Administrator
or primary State may require the system
to install and/or use that treatment
method in connection with a compliance
schedule issued under the provisions of
section 1415(a)(l)(A) of the Act. The
State's determination shall be based
upon studies by the system and other
relevant information.
  (e) The State may require a public
water system to use bottled water
(except for radon) or other means as a
condition of granting a variance or an
exemption from the requirements  of
§ 141.64, to avoid an unreasonable risk
to health. Granular activated carbon
point-of-use devices,cannot be used as a
means of being granted a variance or an
exemption for radon.
   (fj Public water systems that use
bottled water as a condition for
receiving a variance or an exemption
from the requirements of § 141.64  must
meet the following requirements. Bottled
water cannot be used as a means of
being granted a variance or an
exemption for radon.
   (1) The Administrator or primacy
State must require and approve a
monitoring program for bottled water.
The public water system must develop
and put in place a monitoring program
that provides reasonable assurances
that the bottled water meets all MCLs.
The public water system must monitor a
representative sample of the bottled
water for all contaminants under
regulated § 141.64 the first quarter that it
supplies that bottled water to the  public,
and annually thereafter. Results of the
monitoring program shall be provided to
the State annually; or
   (2) The public water system must
receive a certification from the bottled
water company that the bottled water
supplied has been taken from an
"approved source" as defined in 21 CFR
129.3(a); the bottled water company has
conducted monitoring in accordance
with 21 CFR 129.80(g) (1) through  (3);
and the bottled water does not exceed
any MCLs or quality limits as set  out in
21 CFR 103.35,110, and 129. The public
water system shall provide the
certification to the State the first quarter
after it supplies bottled water and
annually thereafter; and
  (3) The public water system is fully
responsible for the provision of
sufficient quantities of bottled water to
every person supplied by the public
water system, via door-to-door bottled
water delivery.
  (g) Public water systems that use
point-of-use devices as a condition for
obtaining a variance or an exemption
from NPDWRs for Radionuclides
(except radon, as POU treatment is not
allowed for variances to the radon MCL)
must meet  the following requirements:
  (1) It is the responsibility of the public
water system to operate and maintain
the point-of-use device.
  (2) The public water system must
develop a monitoring plan and obtain
State approval for the plan before point-
of-use devices are installed for
compliance. This monitoring plan must
provide health protection equivalent to  a
monitoring plan for central water
treatment.
   (3) Effective technology must be
properly applied under a plan approved
by the State.
   (4) The State must require adequate
certification of performance, field
testing, and if not included in the
certification process, a rigorous
engineering design review of the point-
of-use devices.
   (5) The design and application of the
point-of-use devices must consider the
tendency for an increase in
heterotrophic bacteria concentrations in
water treated with activated carbon. It
may be necessary to use frequent
backwashing, post-contactor
disinfection, and Heterotrophic Plate
Count monitoring to ensure that the
microbiological safety of the water is
not compromised.
   (6) All consumers shall be protected.
Every building connected to the system
must have a point-of-use device
installed, maintained, and adequately
monitored. The State must be assured
that every building is subject to
treatment and monitoring and that the
rights and  responsibilities of the public
water system customer convey with title
upon sale of property.
[FR Doc. 91-16523 Filed 7-17-91; 8:45 am]
BILLING CODE 6560-50-M

-------