National Primary Drinking Water Regulations; Radionuclides;
Notice of Data Availability
[Federal Register: April 21, 2000 (Volume 65, Number 78)]
[Proposed Rules]
[Page 21575-21628]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr21ap00-38]
[[Page 21575]]
-----------------------------------------------------------------------
Part IV
Environmental Protection Agency
-----------------------------------------------------------------------
40 CFR Parts 141 and 142
National Primary Drinking Water Regulations; Radionuclides; Notice of
Data Availability; Proposed Rule
[[Page 21576]]
-----------------------------------------------------------------------
ENVIRONMENTAL PROTECTION AGENCY
40 CFR Parts 141 and 142
[FRL-6580-8]
RIN-2040-AC98
National Primary Drinking Water Regulations; Radionuclides;
Notice of Data Availability
AGENCY: Environmental Protection Agency.
ACTION: Notice of data availability for proposed rules with request for
comments.
-----------------------------------------------------------------------
SUMMARY: The Environmental Protection Agency (EPA) proposed regulations
to limit the amount of radionuclides found in drinking water on July
18, 1991. In general, the proposal revised current National Primary
Drinking Water regulations (NPDWR); a NPDWR was proposed for uranium
which is unregulated. Since that time, new information has become
available which the Agency is considering in finalizing these proposed
regulations. In addition, the 1996 Amendments to the Safe Drinking
Water Act (SDWA) contained provisions which directly affect the 1991
proposed rule.
This document presents additional information relevant to the
Maximum Contaminant Level Goals (MCLGs), the Maximum Contaminant Levels
(MCLs), and monitoring requirements contained in the 1991 proposal. EPA
is seeking public review and comment on these new data. The Agency is
also soliciting comments on several implementation options that are
being evaluated for inclusion in the final regulations.
DATES: Written comments should be postmarked or delivered by hand by
June 20, 2000.
ADDRESSES: Send written comments to the W-00-12 Radionuclides Rule
Comment Clerk, Water Docket (MC-4101), 1200 Pennsylvania Ave., NW,
Washington, DC 20460 or by sending electronic mail (e-mail) to ow-
docket@epa.gov. Hand deliveries should be delivered to: EPA's Drinking
Water Docket at 401 M Street, SW, East Basement (Room EB 57),
Washington, DC 20460. Please submit an original and three copies of
your comments and enclosures (including references). If you wish to
hand-deliver your comments, please call (202) 260-3027 between 9:00
a.m. and 4:00 p.m., Monday through Friday, excluding Federal holidays,
to obtain the room number for the Docket. Please see Supplementary
Information under the heading ``Additional Information for Commenters''
for detailed filing instructions, including electronic submissions.
The record for the proposal has been established under the docket
name: National Primary Drinking Water Regulations for Radionuclides (W-
00-12). The record includes supporting documentation as well as
printed, paper versions of electronic comments. The record is available
for inspection from 9 a.m. to 4 p.m., Monday through Friday, excluding
Federal holidays at the Water Docket, 401 M Street SW, East Basement
(Room EB 57), Washington, DC 20460. For access to the Docket materials,
please call (202) 260-3027 to schedule an appointment.
FOR FURTHER INFORMATION CONTACT: For technical inquiries, contact David
Huber, Standards and Risk Management Division, Office of Ground Water
and Drinking Water, EPA (MC-4607), 401 M Street SW, Washington, DC
20460; telephone (202) 260-9566. In addition, the Safe Drinking Water
Hotline is open Monday through Friday, excluding Federal holidays, from
9:00 a.m. to 5:30 p.m. Eastern Standard Time. The Safe Drinking Water
Hotline, toll free 1-800-426-4791.
SUPPLEMENTARY INFORMATION:
Regulated Entities
Entities potentially regulated by the Radionuclides Rule are public
water systems that are classified as either community water systems
(CWSs) or non-transient non-community water systems (NTNCWSs).
Regulated categories and entities include:
------------------------------------------------------------------------
Category Examples of regulated entities
------------------------------------------------------------------------
Industry............................... Privately-owned CWSs and
NTNCWSs.
State, Tribal, and Local Governments... Publicly-owned CWSs and
NTNCWSs.
------------------------------------------------------------------------
This table lists the types of entities, currently known to EPA,
that could potentially be regulated by the Radionuclides Rule. It is
not intended to be exhaustive, but rather provides a guide for readers
regarding entities likely to be regulated by the Radionuclides Rule.
Other types of entities not listed in the table could also be
regulated. To determine whether your facility is regulated by the
Radionuclides Rule, you should carefully examine the applicability
criteria in Secs. 141.15 and 141.26 of title 40 of the Code of Federal
Regulations, and the definitions of Community Water systems and Non-
Transient, Non-Community water systems in Sec. 141.2 If you have
questions regarding the applicability of the Radionuclides Rule to a
particular entity, consult the person listed in the preceding FOR
FURTHER INFORMATION CONTACT section.
Additional Information for Commenters
To ensure that EPA can read, understand and therefore properly
respond to your comments, the Agency requests that commenters follow
the following format: type or print comments in ink, and cite, where
possible, the paragraph(s) in this document to which each comment
refers. Please use a separate paragraph for each issue discussed and
limit your comments to the issues addressed in today's Document.
If you want EPA to acknowledge receipt of your comments, enclose a
self-addressed, stamped envelope. No facsimiles (faxes) will be
accepted. Comments also may be submitted electronically to ow-
docket@epamail.epa.gov. Electronic comments must be submitted as a
WordPerfect 8.0 or ASCII file avoiding the use of special characters
and forms of encryption and must be transmitted by midnight June 20,
2000. Electronic comments must be identified by the docket name,
number, or title of the Federal Register. Comments and data also will
be accepted on disks in WordPerfect 8.0 or in ASCII file format.
Electronic comments on this document may be filed online at many
Federal Depository Libraries.
Abbreviations and Acronyms Used in This Notice
Organizations
APHA--American Public Health Association
ASTM--American Society for Testing and Materials
AWWA--American Water Works Association
ICRP--International Commission on Radiological Protection
NBS--National Bureau of Standards
NSF--National Sanitation Foundation
ANPRM--Advanced Notices of Proposed Rulemaking
ATSDR--Agency for Toxic Substances and Disease Registry
BNL--Brookhaven National Laboratory
CFR--Code of Federal Regulations
EML--Environmental Measurements Laboratory
ERAMS--Environmental Radiation Ambient Monitoring System
ERD--Environmental Radiation Data
ERIC--Educational Resources Information Center
FGR-13--Federal Guidance Report 13
FR--Federal Register
FRC--Federal Radiation Council
NAS--National Academy of Sciences
NCHS--National Center for Health Statistics
NESHAP--National Emissions Standards for Hazardous Air Pollutants
NIRS--National Inorganic and Radionuclide Survey
[[Page 21577]]
NIST--National Institute of Standards and Technology
NODA--Notice of Data Availability
NPDES--National Pollutant Discharge Elimination System
NPDWRs--National Primary Drinking Water Regulations
NRC--National Research Council
NRC--Nuclear Regulatory Commission
NTIS--National Technical Information Service
ORNL--Oak Ridge National Laboratory
SAB--Science Advisory Board
RADRISK--a computer code for radiation risk estimation
SWTR--Surface Water Treatment Rule
T&C--Technologies and Cost document
UCMR--Unregulated Contaminant Monitoring Rule
USDOE--United States Department of Energy
USDW--underground source of drinking water
USEPA--United States Environmental Protection Agency
USGS--United States Geological Survey
USSCEAR--United Nations Scientific Committee on the Effects of
Atomic Radiation
Units of Measurement
Bq--Becquerel
Ci--Curie
EDE/yr--effective dose equivalent per year
kBq--kiloBecquerels
kBq/m \3\--kiloBecquerels per cubic meter
kg--kilogram
kgpd--kilogram per day
Mgkd--milligram per kilogram per day
L--liter
L/day--liter per day
mg--milligram
mg/L--milligram per liter
mg/kg--milligram per kilogram
mg UN/L--milligram uranyl nitrate per liter
mg/kg/day--milligram per kilogram per day
mg U/kg/day--milligram uranium per kilogram per day
mgd--million gallons per day
mL--milliliter
mrem--millirem
mrem/yr--millirem per year
Sv--Sievert
µCi--microCurie
µCi/kg--microCurie per kilogram
µg or ug--microgram
µg/g or ug/g--microgram per gram
µg/L or ug/L--microgram per liter
µg uranium/L--microgram uranium per liter
µg uranium/kg/day--microgram uranium per kilogram per day
µR/hr--micro Roentgen per hour
µSv/cm--micro Sievert per centimeter
NTU--Nephelometric Turbidity Unit
pCi--picoCurie
pCi/day--picoCurie per day
pCi/g--picoCurie per gram
pCi/L--picoCurie per liter
pCi/µg--picoCurie per microgram
Other Terms
ACA--anticentromere antigen
ALP--alkaline phosphatase
AS--alpha spectrometry
BAT--best available treatment
BEIR--biological effects of ionizing radiation
BMG--b2-microglobulin
CWS--community water systems
DL--detection limit
EDE--effective dose equivalent
FSH--follicle stimulating hormone
GGT--gamma glutamyl transferase
GI--gastrointestinal
IE--ion exchange
LDH--lactate dehydrogenase
LET--low energy transfer
LOAEL--lowest observed adverse effect level
LP--Laser phosphorimetry
MCL--maximum contaminant levels
MCLG--maximum contaminant level goals
MDL--method detection limit
n--number
NAG--N-acetyl-b-D-glucosaminidas
NTNC--non-transient, non-community
NTNCWS--non-transient, non-community water systems
PBMS--performance based measurement system
PE--performance evaluation
POE--point-of-entry
POU--point-of-use
PQL--practical quantitation level
PT--performance testing
PWS--public water systems
RF--risk coefficient
RfD--reference dose
RO--reverse osmosis
RSC--relative source contribution
SM--standard methods
SMF--standardized monitoring framework
SPAARC--Spreadsheet Program to Ascertain Residual Radionuclide
Concentration
SSCTL--``Small Systems Compliance Technology List''
Stnd. Dev.--standard deviation
TR--target risk level
UIC--underground injection control
Table of Contents
I. Purpose and Organization of this Document
II. Statutory Authority and Regulatory Background
A. Safe Drinking Water Act of 1974 and Amendments of 1986 and
1996
B. The 1991 Proposal
C. Court Agreement
D. Statutory Requirements for Revisions to Regulations
III. Overview of Today's Document
A. Health Risk Consistency With Chemical Carcinogens
B. Drinking Water Consumption
C. Risk Modeling and the MCL
D. Sensitive Sub-Population: Children
E. MCL for Beta Particle and Photon Radioactivity
F. Combined Ra-226 and Ra-228
G. Gross Alpha MCL
H. Uranium
I. Inclusion of Non-Transient Non-Community Water Systems
J. Analytical Methods
K. Monitoring
L. Effective Dates
M. Costs and Benefits
IV. References
Appendices
I. Occurence
II. Health Effects
III. Analytical Methods
IV. Treatment Technologies and Costs
V. Economics and Impacts Analysis
I. Purpose and Organization of This Document
In 1976, EPA promulgated drinking water regulations for several
radionuclides. In 1991 (56 FR 33050, July 18, 1991), EPA proposed
revisions to the current radionuclides (i.e. beta and photon emitters,
radium-226 and radium-228, and gross alpha radiation) and proposed
regulations for uranium which is not currently regulated. EPA is
publishing this Notice of Data Availability (NODA) to inform the public
and the regulated community of new information concerning radionuclides
in drinking water. EPA is evaluating these additional data to determine
how they will affect the Agency's decisions relative to final
regulations to control radionuclides in public water systems. The
Agency is under a court agreement to publish these final regulations by
November 2000. Information in today's Document includes data about the
occurrence, health effects, and treatment options for radionuclides in
drinking water, as well as analytical methods, and monitoring
requirements. This Document also presents data concerning the costs and
benefits of several regulatory options. EPA is soliciting public
comment on a number of issues raised by this new information. This
introduction provides an overview of the document, and some of the
information available to EPA and to highlight the risk management
decisions the Agency is contemplating. Subsequent sections will contain
more specific information, with a focus on what is new, relative to
each of the topics listed previously. Finally, to further assist the
public, the Agency has compiled seven appendices, included with this
NODA, with more detailed information on each of these topics in
addition to the public docket of reference materials. EPA seeks comment
on the data and information presented in today's NODA, particularly
where regulatory options or alternatives are discussed. Commenters are
asked to provide their rationale and any supporting data or information
they wish to submit in support of comments offered.
Table I-1 summarizes the major elements of the 1976 rule, the 1991
proposal and the issues being considered in today's NODA.
[[Page 21578]]
Table I-1.--Comparison of the 1976 Rule, 1991 Proposal, and 2000 NODA
----------------------------------------------------------------------------------------------------------------
Provision 1976 Rule (Current Rule) 1991 Proposal 2000 NODA
----------------------------------------------------------------------------------------------------------------
Affected Systems............ CWS....................... CWS + NTNC................ CWS + several NTNC options
based on the 1991
proposal.
MCLG........................ no MCLG................... MCLG of zero.............. MCLG of zero.
Radium MCL.................. Combined Ra-226 + Ra-228 Ra-226 MCL of 20 pCi/L; Ra- Maintain current MCL based
MCL of 5 pCi/L. 228 MCL of 20 pCi/L. on corrected estimates of
risk of current MCL.
Beta/Photon Emitters MCL.... 4 mrem: Methodology for 4 mrem ede (Effective Dose Maintain current MCL based
deriving individual Equivalent). Derived on corrected estimates of
concentration limits concentration limits risk of current MCL.
incorporated by changed to reflect new
reference; MCL = sum of dose limit; Current
the fractions of dose estimate of associated
from one or more risks for these
contaminants; risks concentration limits are
estimated not to exceed between 10-4 and 10-3 for
5.6 x 10-5. most.
Gross alpha MCL............. 15 pCi/L excluding U and ``Adjusted'' gross alpha Maintain current MCL based
Rn, but including Ra-226. MCL of 15 pCi/L, on unacceptable risk
excluding Ra-226, radon, level of 1991 proposed
and uranium. MCL.
Polonium-210................ Included in gross alpha... Included in gross alpha... No changes to current
rule. Monitoring required
under the UCMR rule.
Future action may be
proposed at a later date.
Lead-210.................... Not Regulated............. Included in beta particle No changes to current
and photon radioactivity; rule. Monitoring required
concentration limit under the UCMR rule.
proposed at 1 pCi/L. Future action may be
proposed at a later date
Uranium MCL................. Not Regulated............. 20 µg/L or 30 pCi/ Three options being
L w/ option for 5-80 considered: 20, 40, 80
µg/L. µg/L and pCi/L
Ra-224...................... Part of gross alpha, but Part of gross alpha, but Same as current rule, but
sample holding time too sample holding time too Ra-224 may be addressed
long to capture Ra-224. long to capture Ra-224. in a future proposal.
Radium monitoring........... Ra-226 linked to Ra-228; Measure Ra-226 and -228 Measure Ra-226 and -228
measure Ra-228 if Ra- separately.. separately
226>3 pCi/L and sum.
Monitoring baseline......... 4 quarterly measurements. Annual samples for 3 Implement Std Monitoring
Monitoring reduction years; Std Monitoring Framework as proposed in
based on results: >50% of Framework: >50% of MCL 1991. Four initial
MCL required 4 samples required 1 sample every 3 consecutive quarterly
every 4 yrs; 50% of MCL years; 50% of MCL enabled samples in first cycle.
required 1 sample every 4 system to apply for If initial average level
yrs. waiver to 1 sample every >50% of MCL: 1 sample
9 years. every 3 years; 50% of
MCL: 1 sample every 6
years; Non-detect: 1
sample every 9 years.
(beta particle and photon
radioactivity has a
unique schedule--see
Section III, part K).
Beta monitoring............. Surface water systems Ground and surface water Same as 1991 proposal with
>100,000 population systems within 15 miles clarifications.
Screen at 50 pCi/L/; of source screen at 30 or
vulnerable systems screen 50 pCi/L. Those drawing
at 15 pCi/L. water from a contaminated
source screen at 15 pCi/L.
Gross alpha monitoring...... Analyze up to one year Six month holding time for As proposed in 1991.
later. gross alpha samples; Recommendation to analyze
Annual compositing of within 48-72 hours to
samples allowed. capture Ra-224.
Analytical Methods.......... Provide methods........... Method updates proposed in Current methods with
1991; Current methods clarifications.
were updated in 1997.
----------------------------------------------------------------------------------------------------------------
II. Statutory Authority and Regulatory Background
A. Safe Drinking Water Act of 1974 and Amendments of 1986 and 1996
Regulations for radionuclides in drinking water were first
promulgated in 1976 as interim regulations under the authority of the
Safe Drinking Water Act (SDWA) of 1974. The standards were set for
three groups of radionuclides: beta and photon emitters, radium
(radium-226 and radium-228), and gross alpha radiation. These standards
became effective in 1977.
The SDWA Amendments of 1986 required EPA to establish health-based
regulatory targets, called Maximum Contaminant Level Goals (MCLGs), for
every contaminant ``at the level at which no known or anticipated
adverse effects on the health of persons occur and which allows an
adequate margin of safety.'' The enforceable standard, the Maximum
Contaminant Level (MCL), was required to be established ``as close to
the health-based goal as feasible using the best available technology,
taking costs into consideration.'' EPA proposed an MCLG of zero for the
radionuclides in 1991.
In 1983 and 1986, EPA published an Advanced Notice of Proposed
Rulemaking (ANPRM) requesting additional information and comments on
radionuclides and numerous organic and inorganic contaminants in
drinking water. The 1986 SDWA Amendments identified 83 contaminants for
EPA to regulate, including the currently regulated radionuclides, which
lacked an MCLG, and two additional radionuclides, uranium and radon.
The Amendments also declared the 1976 interim standards to be final
National Primary Drinking Water Regulations.
In 1996, Congress again amended the SDWA. These amendments included
new and revised provisions that must be considered when revising
drinking water regulations. Among these are the
[[Page 21579]]
health protection clause (section 1412(b)(9)) which requires that ``any
revision of a national primary drinking water regulation (NPDWR) shall
be promulgated in accordance with this section, except that each
revision shall maintain, or provide for greater protection of the
health of persons.''
The 1996 Amendments also provide for a cost-benefit analysis when
publishing a proposal for new NPDWRs pursuant to section 1412 (b)(6).
While the EPA had proposed the radionuclides rule prior to these
Amendments, the Agency nevertheless conducted an analysis of the costs
and associated benefits of all of the options described in today's
Document. These analyses serve to update and revise the costs and
benefits estimated for the 1991 proposed rule. For the uranium
standard, the Agency solicits comment on the possible use of its new
discretionary authority at section 1412(b)(6) of the SDWA, which allow
for a proposed regulatory level to be set higher than the feasible
level, after the Agency has made a determination that the benefits do
not justify the costs at the feasible level. Note that section
1412(b)(6) applies to new standards (uranium), not to the revision of
existing standards (combined radium-226 and -228, gross alpha, and beta
particle and photon radioactivity). Where we expect to maintain current
standards at their existing levels, no additional analysis was
undertaken because the rule is already in effect.
B. The 1991 Proposal
In 1991, EPA proposed new regulations for uranium and radon, as
well as revisions to the existing regulations. The proposal included
the following features: (1) an MCLG of zero for all ionizing radiation;
(2) revised MCLs for beta particle and photon radioactivity, radium-
226, radium-228, and gross alpha emitters; (3) proposed MCLs for
uranium and radon; and (4) revisions to the categories of systems
required to monitor, the monitoring frequencies, and the appropriate
screening levels. EPA received comments on the new data and regulatory
options presented in the 1991 proposal. However, the proposal was never
promulgated as a final rule in large part because of controversy
surrounding the proposed MCL for radon. The 1996 Amendments to the SDWA
directed the Agency to withdraw the proposed MCL for radon, which was
subsequently done on August 6, 1997 (62 FR 42221).
Most of the comments EPA received on the proposal related to radon.
Approximately 120 comments related to non-radon radionuclides were
valuable and most are still germane to the Agency's rulemaking efforts.
Those comments are addressed, as appropriate, in today's document.
C. Court Agreement
The SDWA (as amended in 1986) provided a statutory deadline to
promulgate a revised radionuclide rule of June 1989, but EPA failed to
meet this deadline. An Oregon plaintiff brought suit to require EPA to
issue the regulations and EPA entered into a series of consent
agreements setting schedules to issue regulations for the
radionuclides. EPA issued a proposal in 1991. After the SDWA Amendments
in 1996, EPA agreed to publish a final action with respect to the
proposed regulation for uranium by November 21, 2000. EPA also agreed
to either take final action by the same date with respect to radium,
beta/photon emitters, and alpha emitters or publish a notice stating
its reasons for not taking final action on the proposal. This latter
scenario would leave the current rule in effect.
D. Statutory Requirements for Revisions to Regulations
Both the 1986 and the 1996 Amendments to the SDWA state that
revisions be made to existing drinking water regulations periodically.
Section 1412(b)(9) of the 1986 SDWA Amendments directed that ``national
primary drinking water regulations be amended whenever changes in
technology, treatment techniques, and other means permit greater
protection of the health of persons, but in any event, such regulations
shall be revised at least once every 3 years.'' The 1996 SDWA
Amendments provide that EPA `` * * * not less than every 6 years review
and revise, as appropriate, each national primary drinking water
regulation,'' and that ``any revision shall maintain, or provide for
greater, protection of the health of persons.''
The radionuclides emit ionizing radiation and, absent data
indicating that there is a threshold level at which exposure does not
present a risk, EPA uses a linear, non-threshold model to set a zero
MCLG for radionuclides. This means that any exposure can potentially
cause harm and that risk associated with the exposure increases
proportionally to the concentration of the radionculide.
EPA's current estimate of the unit risks posed by many of the
radionuclides covered by today's document has generally increased
relative to the 1991 estimate. In fact, based on the newest science
(Federal Guidance Report 13), the fatal cancer risks associated with
the 1991 proposed MCL changes for combined radium, gross alpha, and
beta particle and photon radioactivity generally exceed the Agency's
risk range of 10-6 to 10-4. This document
discusses and requests comment on the issues EPA has addressed in
determining how to best meet applicable SDWA provisions for each of the
radionuclide categories covered by today's document.
III. Overview of Today's Document
Additional data since the 1991 proposal suggest a need to retain
some portions of the proposal, while retaining much of the current
rule. Any changes that are finalized must meet the provisions for
public health protection in accordance with the 1996 Amendments. EPA
has presented its approach for finalizing the non-radon portions of the
1991 radionuclides proposal at several public meetings.
In December 1997 EPA held a public forum (stakeholder meeting) to
discuss the requirements and limitations of the new Amendments
pertaining to revisions to the radionuclide regulation. The Agency
discussed most of the concepts presented in this document and received
valuable feedback from the public, the regulated community, and other
Federal Agencies. In this Document, EPA is presenting the current
information and options upon which the Agency will make its decisions
regarding revisions to the existing standards. At the same time, the
Agency is requesting additional data and comments on the approach EPA
expects to take in formulating the final rule.
The most significant new information concerns the occurrence,
monitoring, and health effects of radionuclides in drinking water.
Recent data suggest a more widespread occurrence of certain
radionuclides which may point to a need for improved monitoring for
these radionuclides in certain areas of the country. Conversely, a
better understanding of the occurrence patterns may also indicate the
need for less frequent monitoring. The newest health effects models,
which are based on improved age-dependent biokinetic and dosimetric
models of the effects of ionizing radiation on the body and more recent
epidemiological information, reveal that radionuclides generally
present a somewhat greater risk than the estimates of previous models,
including the 1991 RADRISK model. EPA's publication ``Federal Guidance
Report 13'' (FGR-13, EPA 1999b) discusses the newest risk modeling. The
resulting risk estimates based on of the new health effects models are
largely the reason for the publication of this document. The
[[Page 21580]]
following are some aspects of the NODA which the Agency would like to
highlight.
A. Health Risk Consistency With Chemical Carcinogens
The risks associated with exposure to chemical carcinogens are
usually expressed as the risks of illness. It is EPA policy to issue
standards that maintain a risk ceiling in the target risk range of
10-6 (one in one million) up to 10-4 (one in ten
thousand). For consistency between the level of protection between
chemical and radiological drinking water contaminants, EPA is
considering utilizing whichever risk provides the greater protection
for MCL changes, a 1 x 10-4 risk of cancer incidence, or a
mortality risk at half the incidence, 5 x 10-5. The risk of
death at 5 x 10-5 is the more protective if the mortality
rate from a particular radionuclide is more than 50%, which is true for
most of the radionuclides. However, for the thyroid, the mortality rate
from thyroid cancer is at 10%. Protecting at 1 x 10-4
incidence corresponds to a mortality at 1x10-5. Conversely,
protecting at 5 x 10-5 mortality with only a 10% mortality
rate allows an incidence, of 5 x 10-4, a less protective
number.
B. Drinking Water Consumption
EPA received comments in 1991 from the American Water Works
Association (AWWA), the Colorado Water Quality Commission, the Atlantic
Richfield Co., and the Rio Algom Mining Corp. suggesting that
consumption of drinking water was actually 1.2 liters per day, thus EPA
was being too conservative in using two liters per day.
When establishing an MCL for a carcinogen, the risk which the MCL
would represent is considered as well as treatability and costs.
Radionuclides will have an MCLG of zero, with MCLs based on standard
assumptions of two liters intake per person per day (2 L/day), an
average individual weight of 70 kg, and a 70 year life span. EPA now
has data to indicate that the average consumption of tap water is 1.1
liters per day per person and that a consumption rate of 2.2 L/day
represents the 90th percentile consumption level.\1\ Basing the MCL on
a consumption rate higher than the average value is justified since
MCLs are intended to be protective of the persons that comprise the
population and not just ``typical individuals''. Since a consumption
value of 2 L/day is less than the 90th percentile consumption rate, EPA
believes that its assumption of 2 L/day for MCL determinations is not
overly conservative and is justifiable.
---------------------------------------------------------------------------
\1\ If one ranked, from lowest to highest, the average daily
water consumption levels for every CWS customer in the U.S., the
``90th percentile'' value of 2.2 L/day is the best estimate of the
value for which 90 percent of the population would drink that much
or less on an typical day.
---------------------------------------------------------------------------
When computing the national benefits of a regulation and the
estimate of cancer mortality risks or risk reductions, EPA is now using
1.1 liters per person per day (L/day) of water as the estimate of the
average daily consumption rate for individuals. In effect, this reduces
population risk estimates by approximately one half and reduces the
estimate of risk reductions by approximately one half. Since benefits
calculations are based on risk reductions, this reduces monetized
benefits by approximately one half. It should be noted that it is
consistent to set health protection levels based on a subset of
individuals that face the highest risks (sensitive subpopulations and/
or the substantial minority of the population have higher water
consumption levels), while estimating benefits based on average
individuals (average consumption and sensitivity). EPA believes this
approach leads to protective MCLs and realistic benefits calculations.
C. Risk Modeling and the MCL
The Agency's current radionuclides health effects model is based on
Federal Guidance Report 13 (FGR-13, EPA 1999b). The Agency's new health
effects model uses state-of-the-art methods, models and data that are
based on the most recent scientific knowledge. Compared with the
approaches used in 1976 and 1991, the revised methodology includes
substantial refinements (described in appendix II, ``Health Effects'').
While commenters have pointed out the MCLs in the current rule are
based on ``old science'', the newest science indicates that many of the
MCLs proposed in 1991 have corresponding risks that are much greater
than the upper limit of the Agency's acceptable lifetime excess risk
range of approximately 10-6 to 10-4 (one in one
million to one in ten thousand lifetime excess risk of cancer). The
risks associated with each existing and proposed MCL are described in
sections that follow. The risk models are described in detail in
appendix II (Health Effects) and in the Technical Support Document for
the Radionuclides Notice of Data Availability (EPA 2000a).
Between 1976 and the present, different scientific models have been
used to calculate risks from radiation exposure. Each model derives a
different concentration of a particular nuclide for a given level of
risk. For example, in 1991, the RADRISK model indicated that consuming
drinking water with radium-228 at 26 pCi/L would lead to an excess
lifetime cancer risk of 1 x 10-4. However, using today's
model (based on Federal Guidance Report 13), the best estimate of
lifetime risk of Ra-228 at 26 pCi/L is 1 x 10-3, a risk
value ten times greater than thought in 1991.
Likewise, the 1991 proposed MCL for Ra-228 at 20 pCi/L was thought
to correspond to lifetime excess cancer risk of 7.7 x 10-5.
The most current risk estimate for Ra-228 at 20 pCi/L
7.7 x 10-4, again ten-fold greater and much higher than the
Agency's target risk ceiling of 10-4. For individuals
consuming water with 20 pCi/L of both Ra-228 and Ra-226, the risk was
thought to be 1.7 x 10-4 in 1991. However, based on the
newest science, these individuals would be exposed to lifetime excess
risks of 1 x 10-3 risk (one in a thousand), a risk level 10-
fold higher than the Agency's target risk ceiling for drinking water
MCLs. EPA requests comments on these issues.
D. Sensitive Sub-Population: Children
The age-specific, sex-specific models used by EPA for estimating
risk from ionizing radiation implicitly provide for risk
differentiation by gender and age. The computer program suite, DCAL
(FGR-13), uses age-specific metabolic models to calculate the dose from
a unit intake of a radioisotope during each year of life from birth to
120 years of age. Age-specific organ masses are used for all ages up to
adult, and for adult males and adult females. Risk coefficients are
given by age and sex for each year of life from birth to 120 years of
age. The risk is then calculated by combining calculated doses and age-
sex-specific risk coefficients with age-sex-specific intake data and
age-sex-specific survival data.
A separate risk analysis for children was performed and is
described in appendix II (Health Effects), part C. Risks to children
are explicitly considered when setting MCLs for radionuclides. In the
case of the regulated water systems (currently, community water
systems), children are fully protected. In the case of the unregulated
systems of potential concern (non-transient non-community water
systems, NTNCWSs), the analysis is more complicated. Risks to children
served by NTNCWSs are discussed in appendix II, part C, number 3.
[[Page 21581]]
E. MCL for Beta Particle and Photon Radioactivity
1. EPA's Plans for Finalizing the 1991 Proposed MCL for Beta and Photon
Radioactivity
This section presents the important considerations that have led
EPA to consider retaining the current MCL for beta particle and photon
radioactivity when the 1991 proposal is finalized in November of 2000.
EPA is, however, also considering finalizing the 1991 proposed changes
to the monitoring requirements for beta particle and photon
radioactivity, as described later in this section. The current MCL is
(40 CFR 141.16):
(a) The average annual concentration of beta particle and photon
radioactivity from man-made radionuclides in drinking water shall not
produce an annual dose equivalent to the total body or any internal
organ greater than 4 millirem/year.
(b) Except for the radionuclides listed in Table A, the
concentration of man-made radionuclides causing 4 mrem total body or
organ dose equivalents shall be calculated on the basis of a 2 liter
per day drinking water intake using the 168 hour data listed in
``Maximum Permissible Body Burdens and Maximum Permissible
Concentrations of Radionuclides in Air or Water for Occupational
Exposure,'' NBS Handbook 69 as amended August 1963, U.S. Department of
Commerce. If two or more radionuclides are present, the sum of their
annual equivalent to the total body or to any organ shall not exceed 4
millirem/year.
Table A.--Average Annual Concentrations Assumed To Produce a Total Body
or Organ Dose of 4 mrem/year.
------------------------------------------------------------------------
pCi per
Radionuclide Critical organ liter
------------------------------------------------------------------------
Tritium............................. Total body............. 20,000
Strontium-90........................ Bone marrow............ 8
------------------------------------------------------------------------
Following these instructions leads to a unique list of
concentration limits for 168 other man-made radionuclides. This list is
included in today's document in appendix II, ``Health Effects.''
The 1991 proposed MCL for beta emitter and photon radioactivity was
4 mrem-ede (effective dose equivalents), with the footnote:
``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.''
Following these instructions leads to a unique list of
concentration limits for 230 radionuclides. EPA has determined that
there is no way to update the 4 mrem dose basis (1976) for the beta
particle and photon radioactivity MCL without the extensive process of
a new proposal. While some stakeholders have suggested that reverting
to the existing rule for beta particle and photon radioactivity (``beta
emitters'') is relying on ``old science,'' it should be pointed out the
newest risk estimates, based on the peer-reviewed Federal Guidance
Report 13, indicates that the risks associated with the 1991 proposed
MCL of 4 mrem-ede (effective dose equivalents) are above the
10-4 risk level (10-3 to 10-4) for
many of the beta emitters. Figure 1 shows the most current risk
estimates for the beta emitter concentration limits derived under both
the current and proposed MCLs. As the figure shows, the current MCL
results in concentration limits with risks that fall within the
Agency's risk range goal of 10-6 to 10-4 (while
some are slightly above and some slightly below, all round to values
within these orders of magnitude).
BILLING CODE 6560-50-U
[[Page 21582]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.001
BILLING CODE 6560-50-C
[[Page 21583]]
In summary, the Agency fully recognizes that the dose-based MCL of
4 mrem/year is based on older scientific models. However, the Agency
has decided to retain the current MCL given that:
Federal Guidance Report 13 (FGR-13, EPA 1999b)
demonstrates that the 1991 proposed MCL of 4 mrem-ede/year results in
concentration limits that are outside the 10-6 to
10-4 range;
FGR-13 demonstrates that the current MCL of 4 mrem/year
results in concentration limits that are within the 10-6 to
10-4 range;
the fact that there is no evidence of appreciable
occurrence of man-made beta emitters in drinking water;
the 1996 Safe Drinking Water Act requires EPA to evaluate
all NPDWRs every six years (``Six Year Review'').
EPA believes that Six Year Review is the appropriate vehicle for
updating the beta particle and photon radioactivity MCL.
2. Beta Particle and Photon Radioactivity Monitoring
Currently, surface water systems serving more than 100,000 persons
are required to monitor for beta particle and photon radioactivity
using a screening level of 50 pCi/L, while systems that are determined
to be vulnerable by the State are required to monitor using a screening
level of 15 pCi/L. In 1991, EPA proposed that all ground water and
surface water systems within 15 miles of a potential source, as
determined by the State, be required to monitor using a screening level
of 30 or 50 pCi/L. EPA is considering retaining the current monitoring
requirement of a 15 pCi/L screen for water systems drawing water from
contaminated sources. EPA solicits comment on these issues. EPA is
taking comment on screening levels of 30 or 50 pCi/L for systems within
15 miles of a potential source.
3. Lead-210 and Radium-228
The 1991 proposal included lead-210 (Pb-210) and radium-228 (Ra-
228) in the list of regulated beta and photon emitters, both of which
are naturally occurring. An 1991 the Agency was considering raising the
Ra-228 MCL to 20 pCi/L, which is high enough to significantly
contribute to gross beta levels. However, since the Agency is retaining
the current combined Ra-226 and Ra-228 standard of 5 pCi/L, Ra-228 will
no longer be a significant contributor to gross beta. For the reason,
the Agency sees no value in including Ra-228 in the list of beta/photon
emitters.
New risk analyses indicate that Pb-210 is of concern well below the
current and proposed screening levels for beta and photon emitters. In
order to assess the occurrence of Pb-210 to determine if it is present
at levels high enough to warrant separate monitoring, EPA has included
it on the list published in the Unregulated Contaminant Monitoring Rule
(UCMR) (64 FR 50556, Friday, September 14, 1999). USGS also monitored
for Pb-210 in its study with EPA of 100 locations. The reader is
referred to appendix I and the Technical Support Document (EPA 2000a)
for further information regarding this study. Since Pb-210 specific
monitoring was not proposed in 1991, EPA cannot address this concern
without a new proposal. After occurrence data has been reviewed from
the UCMR, EPA may propose appropriate actions.
F. Combined Ra-226 and Ra-228
1. MCL Considerations
The combined radium-226 and -228 NPDWR has long been a contentious
issue. A number of water systems believe the current MCL is too
stringent and have not installed treatment or taken other measures to
comply. EPA first proposed the possibility of increasing the current 5
pCi/L limit for combined radium-226 and -228 in 1991. The proposal
suggested a new level of 20 pCi/L for Ra-226 and Ra-228 separately
along with a proposed limit of 300 pCi/L for radon-222 . This
combination was proposed in part due to the disproportionate costs of
removing radium compared to radon. The proposal was met with
opposition, largely due to the controversy surrounding the radon
component. In the ensuing deliberations, debates regarding the radon
component of the proposal interfered with promulgation of the proposal.
In the 1996 Amendments to the SDWA, Congress directed EPA to remove the
radon component from the proposal. Consequently, the Agency has once
again considered the issues surrounding the allowable concentration of
radium-226 plus radium-228 in drinking water.
EPA is considering retaining the current MCL for combined radium-
226 and -228 at 5 pCi/L for the following reasons. First, the unit
risks for Ra-226 and Ra-228 are believed to be much greater than
estimated in 1991, such that raising the combined Ra-226 and Ra-228 MCL
up to 20 pCi/L for each radionuclide would result in a lifetime excess
cancer risks that are ten-fold higher than the Agency's acceptable risk
range of 10-6 to 10-4. And second, EPA is
required to consider the MCL for Ra-226 and Ra-228 apart from any NPDWR
for radon, both by the 1996 SDWA Amendments and the later court
stipulated agreement. Both points are discussed further here.
First, in 1976 the estimate of risk from either Ra-226 or Ra-228 at
5 pCi/L was between 5 x 10-5 and 2 x 10 -4, averaging
1 x 10-4. In 1991 the RADRISK model calculated that a
1 x 10-4 risk corresponded to Ra-228 at 26 pCi/L and Ra-226
at 22 pCi/L.
Table III-1 shows the change in estimated risks from 1976 until the
present. ``Current Risk Estimates'' are calculated using the 1999
model, FGR-13 (EPA 1999b). The table allows a comparison between the
calculated risk during each phase of the evolution of the radionuclides
NPDWRs, including the current best estimate of risk based upon FGR-13
(EPA 1999b). Details of why the models have changed and the additional
data taken into consideration are found in the appendix II and the
Technical Support Document (EPA 2000a).
Table III-1.--Changes in Estimated Risks for Various Ra-226 and Ra-228 Levels
--------------------------------------------------------------------------------------------------------------------------------------------------------
Radium-228 Radium-226
----------------------------------------------------------------------------------------------------------------------
Year model used Concentration Previous risk Current risk Concentration Previous risk Current risk
pCi/L estimate estimate pCi/L estimate estimate
--------------------------------------------------------------------------------------------------------------------------------------------------------
2000 FGR-13...................... 5 2 x 10-4 2 x 10-4 5 7.3 x 10-5 7.3 x 10-5
2000 FGR-13...................... 2.5 1 x 10-4 1 x 10-4 6.85 1 x 10-4 1 x 10-4
1994 FGR-11...................... 11 1 x 10-4 4.5 x 10-4 10 1 x 10-4 1.5 x 10-4
1991 RADRISK..................... 26 1 x 10-4 1 x 10-3 22 1 x 10-4 3.3 x 10-4
1991 RADRISK proposed MCL........ 20 7.7 x 10-5 7.7 x 10-4 20 9.1 x 10-5 2.9 x 10-4
1991 RADRISK..................... 5 1.9 x 10-5 2 x 10-4 5 2.3 x 10-5 7.3 x 10-5
[[Page 21584]]
1976*............................ 5 1 x 10-4 2 x 10-5 5 1 x 10-4 7.3 x 10-5
--------------------------------------------------------------------------------------------------------------------------------------------------------
*The risk of either radium-226 or radium-228 at 5 pCi/L was believed to be between 5 x 10-5 and 2 x 10-4 in 1976. The average would have been 1 x 10-4.
The 1991 estimated risk corresponding to 20 pCi/l of Ra-226 in
addition to 20 pCi/l of Ra-228 was thought to be 1.7 x
10-4. However, the current risk estimate based on FGR-13
(EPA 1999b) for 20 pCi/l of Ra-226 in addition to 20 pCi/l of Ra-228 is
1 x 10-3 (one in a thousand), an order of magnitude (ten
times) above the acceptable risk of 1 x 10-4.
In contrast, maintaining the current standard would allow a maximum
lifetime risk of 2 x 10-4 (within the original risk range of
the 1976 regulation). This represents a one in 5,000 lifetime mortality
risk and would only be present if 5 pCi/L in the drinking water were
all radium-228, a relatively rare occurrence situation. If the radium
present were all radium-226, the risk would be 7 x 10-5,
just below EPA's risk ceiling. Since:
the risks associated with the current MCL of 5 pCi/L are
already at the upper end of the Agency's allowable risk range of
10-5; and
the 1991 proposed MCLs for Ra-226 and Ra-228 have risks as
high as 1 x 10-3, ten-fold higher than the Agency's
allowable risk, the Agency believes that maintaining the current MCL
for combined Ra-226 and Ra-228 is the appropriate action.
Regarding treatment feasibility, EPA's determination that water
systems can feasibility treat and quantify combined radium at 5 pCi/L
is supported by case studies of systems that had combined radium levels
in excess of the MCL and that later came into compliance through
treatment. In addition, EPA has case studies of systems that have come
into compliance through purchasing water, blending, and developing new
wells (EPA 2000a).
Since risk estimates for Ra-228 are significantly higher than
thought in 1991, EPA has evaluated the risk reductions, costs, and
benefits of decreasing the allowable level of radium-228 to 3 pCi/L and
has discussed the results in the Technical Support Document (EPA
2000a). The concern is that a system with 5 pCi/L of Ra-228 with
insignificant levels of Ra-226 would be in compliance with the combined
radium MCL, but would have an associated lifetime excess cancer
morbidity risk of 2 x 10-4, which exceeds the risk ceiling
on 1 x 10-4. While this is true, the occurrence data
reported in appendix I suggest that this situation should be rare.
Since EPA did not propose this action in the 1991 proposal, EPA cannot
address this concern in the finalization of this proposal. However, EPA
will consider this situation further and will later determine if a
regulatory action is appropriate.
An unintended effect in the 1991 proposal was that the costs and
benefits were not evenly distributed to all affected persons
(individuals). In the 1991 proposal, an MCL was proposed for radon at
300 pCi/L and a revised MCL for radium from 5 pCi/L combined for both
radium-226 and radium-228 to 20 pCi/L each. Benefits and costs were
considered together for both radon and radium on a national basis.
Compared to radium, radon is easier and cheaper to remove from water
due to its air strippability. Since the risks avoided were higher and
the treatment costs lower for the radon MCL, it was reasoned that the
radon rule was much more cost-effective than the combined radium rule.
However, since radium and radon do not tend to co-occur, individuals
that would have benefitted from the radon rule were not the same
individuals that would have faced higher risks under the proposed
radium MCLs. EPA believes that such a trade-off is no longer
appropriate. Among other considerations, the 1996 Amendments to SDWA
explicitly separated the radon rule from the rule for the other
radionuclides.
In summary, EPA based its proposed increase in the radium standard
on the risk models that existed at that time and on a population risk
trade-off with radon. The models in use in 1991 indicated that radium
posed less of a risk than originally believed in 1976. However, current
risk models (FGR-13, EPA 1999b) suggest that the combined radium
standard of 5 pCi/L presents an even greater health risk than thought
in 1976. Given the much higher current estimate of risks associated
with the proposed Ra-226 and Ra-228 MCLs of 20 pCi/L and the
statutorily required withdrawal of radon-222 from the proposal, the
Agency believes that the MCLs for radium proposed in 1991 are no longer
appropriate. EPA requests public comment on retaining the current
radium standard of 5 pCi/L for combined Ra-226 and Ra-228.
2. Separate Radium Analysis
The 1991 proposal recommended decoupling the monitoring of radium-
228 from radium-226. The current radionuclides rule requires analysis
of Ra-228 only when Ra-226 levels are above 3 pCi/L. The rule
recommends analysis of Ra-226 and/or Ra-228 when gross alpha exceeds 2
pCi/L where Ra-228 may be present, and requires analysis of Ra-226 when
gross alpha exceeds 3pCi.L.
Ra-228 may be present with minor amounts of Ra-226 or in the
absence of Ra-226. In general, the mobility of a parent radionuclide
may be very different from that of a daughter element, depending on the
geochemistry of the elements involved. However, the occurrence of a
radionuclide may still be governed by the occurrence and distribution
of its parent (see EPA 2000a). Since radium-226 arises from the uranium
decay series and radium-228 arises from the thorium series, it is
logical to expect them to occur independently of one another. Also, the
parents of Ra-226 (uranium isotopes) and Ra-228 (thorium isotopes) have
very different geochemical behaviors. Uranium is fairly mobile in
oxidizing ground waters, while thorium is rather insoluble. In
contrast, the daughter radium isotopes are more mobile in reducing
waters and are relatively immobile in oxidizing waters. Since Ra-226 is
part of the uranium series (relatively mobile parent) and Ra-228 is
part of the thorium series (immobile parent), Ra-226 can and does
mobilize in waters containing Ra-228 more frequently than the reverse
situation. These observations indicate that Ra-226 and Ra-228 may be
expected to significantly co-occur, but that the correlation will not
be strong enough to use the occurrence of one to predict the other with
acceptable certainty. Recent
[[Page 21585]]
studies support this conclusion (EPA 2000a).
This conclusion indicates that the current monitoring screen for
Ra-228 based on Ra-226 is not reliable. Therefore as proposed in 1991,
EPA is considering requiring separate monitoring and analysis of both
radium-226 and radium-228 in the final rule. The Ra-228 and Ra-226
results would be summed to determine compliance with the radium MCL.
This will provide a more accurate assessment of systems containing
little or no radium-226, but possessing a significant enough
concentration of radium-228 to exceed the standard.
G. Gross Alpha MCL
The gross alpha standard promulgated in 1976 considered natural and
man-made alpha emitters as a group rather than individually. At the
time, the analytical costs made it impractical to identify each alpha-
emitting nuclide in a given water sample. The existing gross alpha MCL
includes radium-226, but excludes radon-222 and uranium (because these
latter nuclides were to be regulated at a later date). The 1991 risk
estimates indicated that the inclusion of Ra-226 was not warranted.
However, today's risk estimates, based on FGR-13 (EPA 1999b), suggest
that the Ra-226 unit risk is large enough to warrant to include it in
gross alpha, as in the current standard. In today's Document, the
Agency is considering maintaining the current MCL for gross alpha,
believing it to be protective. EPA will consider proposing changes to
the rule in the future.
EPA believes that the term ``gross alpha'' may be confusing.
``Gross alpha'' implies counting the total alpha emissions and is the
appropriate name for that particular analytical method. The standard
excludes uranium and radon from the total or gross count. Just as the
proposal suggested the term ``adjusted gross alpha'' with the exclusion
of radium-226, EPA believes the term ``net alpha'' or ``the alpha
standard'' might better describes the current standard which excludes
such alpha emitters as radon, uranium. EPA requests public comment on
the name change.
The gross alpha MCL was originally established at 15 pCi/L to
account for the risk from radium-226 at 5 pCi/L (the radium regulatory
limit) plus the risk from polonium-210, the next most radiotoxic
element in the uranium decay chain. In 1976, the risk resulting from
exposure to 10 pCi/L of polonium-210 was thought to be equivalent to
the risk resulting from exposure to 1 pCi/L of radium-226. Looked at
another way, the 1976 gross alpha standard equated to 6 pCi/L of
radium-226 (5 pCi/L of radium-226, plus the10 pCi/L of polonium-210
which itself was equal to 1 pCi/L of radium-226). Since the risk
associated with the combined radium standard was believed to be in the
range of 5 x 10-5 to 2 x 10-4, this assumption
placed the gross alpha standard reasonably within that range as well.
The gross alpha standard proposed in 1991 remained at 15 pCi/L, but
excluded radium-226 (because it was proposed at 20 pCi/L). The new
limit was termed ``adjusted gross alpha.'' In effect, it allowed an
increase of 5 pCi/L of non-radium alpha emitters in drinking water from
10 to 15 pCi/L by occupying the 5 pCi/L originally represented by the
radium. In the 1991 proposal, the allowable non-radium gross alpha
contribution in that same water sample i.e. Po-210, would be 15 pCi/L.
Because this latter scenario represents more risk than the scenario
evaluated for the current regulation, EPA no longer supports an
``adjusted gross alpha'' limit of 15 pCi/L.
In the future, EPA may consider a proposal to exclude radium-226
from the gross alpha MCL as proposed in 1991 (because of the existence
of a separate standard for radium-226), but to maintain protection,
limiting the gross alpha standard to 10 pCi/L. Reducing the limit has
the advantage of effectively reducing exposure to polonium-210 and
radium-224. In addition, excluding radium-226 from being in both the
gross alpha and radium standards may avoid confusion. EPA examined the
possibility of this change in the context of the potential for added
treatment costs versus the marginal benefits to be derived. However it
appears that retaining the standard at 15 pCi/L is protective of public
health at a reasonable cost. A picoCurie cap of 15 represents different
risks for various nuclides, but this is not unlike other regulated
carcinogens or the other radionuclides. The risks represented by two
components, namely radium-224 and Polonium-210, are discussed next.
1. Polonium-210
Current risk estimates suggest that the risk resulting from
exposure to polonium-210 is ten times greater than originally believed
in 1976 compared to radium. However, existing occurrence data indicates
that its presence in drinking water is relatively rare. To gain a
better understanding of the public health risk posed by polonium-210 in
drinking water, EPA included this radionuclide in the Agency's
Unregulated Contaminants Monitoring Rule (64 FR 50556, Friday,
September 17, 1999). The Agency may consider a future proposal to
develop a separate limit for polonium-210 within (or separate from) a
potentially revised gross alpha standard.
EPA believes that current technology can limit polonium-210 to 4
pCi/L or below, although precise quantification at this level may
present a challenge. Because of its energetic alpha emissions, a gross
alpha measurement may overestimate the actual concentration of
polonium-210 in the sample by a factor of two. With current gross alpha
measurement, if the total alpha were 15 pCi/l contributed by polonium,
the actual concentration of polonium could be much less, depending on
the calibration standard. At present, since there is no specific
drinking water regulation for polonium-210, there is no EPA-approved
method for measuring polonium to determine compliance with a drinking
water standard. Should EPA decide to develop a separate limit for
polonium-210, the Agency will ensure that the approved analytical
method for demonstrating compliance is in place and includes a
calibration standard appropriate for polonium's energetic alpha,
thereby reducing the possibility of overestimating its presence. EPA
requests information relative to any known occurrence of polonium and
the need for a proposal of a separate limit. Recently, USGS co-operated
with EPA and the American Water Works Association in monitoring for
radionuclides, including Po-210 (103 wells in 27 States). The study and
findings are described in EPA 2000a). USGS will publish the study in
the near future. In this study, Po-210 levels were found above 1 pCi/L
in less than two percent of the wells. Since the wells were targeted
for high radium occurrence, this may not be typical. The reader is
referred to appendix I (Occurrence) and the Technical Support Document
(EPA 2000a) for further information.
2. The Occurrence of Radium-224 and its Impact on Alpha
Recently, the short lived isotope of radium has been found in some
drinking water supplies. Extensive monitoring in the State of New
Jersey over the past several years and follow-on survey by EPA and the
USGS has demonstrated that radium-224 may be present in significant
quantities in ground water, especially where its decay chain ancestor
radium-228 is present. Although it is included in the (gross) alpha
MCL, it was not targeted specifically for several reasons: (1) It was
not believed to be a health risk, (2)
[[Page 21586]]
it was not known to be prevalent and (3) sampling it at a
representative point within the distribution system rather than the
entry point to the system allowed decay. However, newer FGR-13 risk
estimates (EPA 1999b), coupled with the greater occurrence, and the
1991 proposal to sample at the entry point to the distribution system,
now make radium-224 a concern.
Radium-224 is a naturally occurring radioisotope, which is part of
the thorium decay chain. It emits alpha particles and has a half-life
of 3.66 days. The decay of its progeny via alpha and beta decay also
happens very quickly. In approximately 4.1 days, an original radium-224
atom has decayed to stable lead-208 by emission of an equivalent of 4
alpha and 2 beta particles. A gross alpha analysis will detect 3 alpha
particle emissions including daughters in equilibrium with the parent
Ra-224. If a sample analysis is done within 72 hours, preferably 48
hours, an appropriate back-calculation can be performed of the gross
alpha count of the sample water. Otherwise the laboratory will
significantly underestimate the radium-224 and other alpha emitters
that may have been originally present in the sample.
Under the current rule, utilities are allowed to collect quarterly
samples, composite and analyze at the end of the year. In 1991, EPA
proposed a holding time of 6 months for gross alpha. However, neither
the annual composite under the current rule or the proposed holding
time of 6 months can appropriately capture the presence of alpha-
emitting radium-224, or its progeny in a gross alpha analysis. The
Agency intends therefore, to issue a separate proposal to change the
holding time for gross alpha analysis to account for the presence of
radium-224 in the sample.
At this point in time, the Agency strongly recommends to States and
utilities that an alpha analysis be performed within 48 to 72 hour
after sample collection to capture the contribution of the alpha
particles arising from radium-224. In this NODA, the Agency is
reiterating and underscoring its recommendation to that effect as
outlined in a memorandum of January 27, 1999 from Cynthia Dougherty,
Director of the Office of Ground Water and Drinking Water (EPA 1999a).
For systems to whom a rapid analysis might be a burden, a reasonable
screening tool for the presence of Ra-224 under many geochemical
circumstances is the presence of its radiological ancestor, Ra-228.
Since systems will monitor for Ra-228, the result can serve as a
general proxy for the presence or Ra-224 for the purposes of
prioritization. It is not definitive and would not be an acceptable
substitute for a rapid analysis of gross alpha or Ra-224. In the
absence of Ra-228, a system may not need to place as high a priority on
rapid gross alpha or specific Ra-224 analysis. Since, as explained
earlier, each Ra-224 atom contributes approximately three daughter
alpha particles to the gross alpha count, a simple first approximation
of Ra-224's contribution to gross alpha would be three times the Ra-228
concentration in pCi/L. For the purposes of prioritizing monitoring for
Ra-224, grandfathered gross alpha data added to three times the result
of the Ra-228 measurement would be a reasonable first approximation of
the gross alpha including Ra-224 and its daughters available from a
rapid gross alpha test. However, EPA reiterates that this approximation
is not a substitute for rapid analysis of gross alpha or Ra-224.
EPA is not considering requiring a separate MCL or analysis for
radium-224 when the rule is finalized in November of 2000. The
definition of gross alpha will continue to include Ra-224. EPA is
willing to consider comments on the need to apply sub-limits to Po-210
or Ra-224 within the MCL of 15 or as separate standards. Proposing a
separate limit for radium-224 at 10 pCi/L within the alpha MCL of 15
pCi/L is a future possibility, as is a separate MCL for radium-224. The
latter would require a separate, specific, rapid analysis specifically
for radium-224, rather than relying on the gross alpha test and alpha
MCL. Such actions would require a new proposal or proposals.
As part of the alpha standard, the Agency does not consider Ra-224
a significant risk. The lifetime mortality risk associated with
exposure to 10 pCi/L of radium-224 is approximately 5 x
10-5 or one in 20,000. Because radium-224 and its progeny
have very short half-lives, the total alpha count represents the
radium-224 and its progeny. Consequently, there are effectively three
alpha particle counts for every atom of radium-224 present. The health
risk of radium-224 already includes the impact of these progeny in the
body (the committed dose). Therefore while the gross alpha count may be
at 15, the impact of the emissions is approximately related to Ra-224
at 5 pCi/L and the risk of 2.5 x 10-5 or excess mortality
of one in 40,000.
H. Uranium
Uranium is not currently regulated by the 1976 radionuclides
drinking water standards. The 1986 SDWA Amendments included uranium as
one of the 83 contaminants listed to be regulated in drinking water.
Two health effects are associated with exposure to uranium: cancer,
resulting from the radioactive emissions, and kidney toxicity,
resulting from the exposure to the uranium itself. The mass of the
uranium is measured in micrograms (µg) while the radiation
activity is measured in picoCuries. In 1991, EPA proposed a limit on
uranium of 20 µg per liter (µg/L) to protect against
kidney toxicity. The corresponding radioactivity limit was assumed to
be 30 pCi/L. At that time, the Agency also proposed an MCLG of zero,
based on absence of an identifiable dose-response threshold. EPA has
reevaluated both the health impact level for kidney toxicity and the
cancer risks from radiation and costs of regulation. As discussed
briefly next, the best estimate of the cost per cancer case or cancer
death avoided at 20 µg/L is relatively large. However, it
should also be noted that this cost per case avoided excludes the
reduction in kidney toxicity risk. At the present time, kidney toxicity
for uranium must be treated as a non-quantifiable benefit (see appendix
II, ``Health Effects'' and the Technical Support Document, EPA 2000a).
Today's NODA presents new information which supports a regulatory
level of 20 µg/L, based upon protection from kidney toxicity.
The derivation of this number is based on newer, more complete studies
which have also resulted in a lower uncertainty factor, now 100-fold.
In addition, the contribution to ingestion from drinking water relative
to food or inhalation, the relative source contribution (RSC), has been
recalculated. Drinking water is now considered to contribute 80 percent
of a person's total daily uranium intake. This has the effect of
permitting 80% of the reference dose (RfD) to be occupied by the
drinking water component of diet. Both a lower uncertainty factor
coupled with a lower food intake and higher proportional contribution
from drinking water to total intake, might suggest the allowance of a
higher regulatory limit; however, the more recent studies have offset
this by revealing a lower observed effect level for kidney toxicity.
The recalculated ``safe level'' for kidney toxicity remains 20 ug/L.
The derivation of the uncertainty factor is based on the types of
uranium health data available. EPA's policy for uncertainty factors for
estimating LOAELs is summarized in 63 FR 43756 (August 14, 1998,
``Draft Water Quality Criteria Methodology Revisions: Human Health'').
The derivation is described in appendix II.
[[Page 21587]]
Uranium is also classified as a carcinogen because of its
radioactivity, and resulting emissions of ionizing radiation. The two
most prevalent isotopes of uranium, uranium-234 and uranium-238, have
very different half-lives which result in different amounts of
radiation emitted per unit mass. Uranium-234 emits far more
radioactivity than U-238, but is much less abundant in aquifer
materials. Uranium-238 emits less radioactivity, but is far more
prevalent than U-234. The average ratio of uranium activity to mass in
rock is 0.68 picoCuries per µg. Issues involving the activity
to mass ration follow later in this section.
Complicating the Agency's decision making about a uranium standard
is the fact that the monetized benefit of kidney toxicity cannot be
calculated at low concentrations because data are lacking in terms of
the level at which kidney disease is actually manifested. The
calculated 20 µg/L level represents an intake which would
result in no effect over 70 years by drinking two liters per day.
Conclusions based on the toxicity of uranium to the kidney are based
primarily on observed adverse effects at the cellular level, but which
have not necessarily resulted in a recognized disease. It is difficult
to monetize the benefits derived in such a situation, and EPA does not
currently have a methodology for estimating benefits for kidney
toxicity from uranium. In the case of reducing the risk of non-fatal
cancer resulting from uranium, EPA can monetize these benefits based on
avoided ``cost of illness.'' This methodology is discussed in some
detail in the Technical Support Document (EPA 2000a) and elsewhere (EPA
2000b).
Thus, for kidney toxicity, the benefit to society are considered as
``non-quantifiable benefits.'' Kidney toxicity avoidance benefits can
be expressed in terms of ``avoidance of exposure,'' but cannot be
quantified in terms of avoidance of a specified number of cases of
disease or fatalities (and the associated monetized benefits), as with
cancer. In addition, it appears that excess uranium concentrations tend
to be found in small water systems. This suggests that while many
systems will be impacted, the affected populations will be small. In
terms of cancer risk, the number of statistical cases avoided for MCLs
of 20 and 40 µg/L are low (0.2 to 2 cases for 20 µg/L
and 0.04 to 1.5 cases for 40 µg/L). In terms of exposure
avoided for kidney toxicity, around 500 thousand to two million persons
are exposed above 20 µg/L and 50 thousand to 900 thousand
persons are exposed above 40 µg/L. See appendix V and the
Technical Support Document (EPA 2000a) for details.
Although uranium is treatable to levels well below the 1991
proposed MCL of 20 µg/L (5 pCi/L was evaluated), EPA determined
that levels below 20 pCi/l were not feasible under the SDWA, after
taking the costs of treatment into consideration. Section 1412(b)(6) of
the 1996 SDWA permits the Agency to evaluate whether the benefits of
regulating at various MCLs justify the costs. Possible exercise of this
authority is discussed in more detail later in this section.
The MCLG that was proposed for uranium in 1991 was zero because of
concerns about the lack of a known threshold for the carcinogenicity of
ionizing radiation. The MCL that was proposed in 1991 (20 µg/L)
was based on uranium kidney toxicity, as previously described. The
corresponding risk of cancer at a concentration of 20 µg/l is
now estimated to be approximately 5 x 10-5. In terms of
the cost per cancer case avoided and kidney toxicity reduced, the cost
of regulation is still relatively high (see Table V-2 in appendix V).
In its current benefit-cost analysis, EPA also evaluated regulatory
options of uranium MCLs of 40 µg/L and 80 µg/L. EPA
estimates that a level of 40 µg/L would correspond to a cancer
risk of approximately a 1 x 10-4, thus providing cancer
risk protection within the Agency's traditional risk range. A level of
40 µg/L would represent a slightly higher risk of kidney
toxicity. At a level of 80 µg/L, the cancer mortality risk is
approximately 2 x 10-4, which is above the Agency's
acceptable risk range. At 80 µg/L, the projected total national
costs decrease significantly, but the estimates of cancer cases avoided
drops to values close to zero (i.e., benefits diminish considerably),
indicating that the cost per cancer case avoided may not be
signficantly lower at an MCL of 80 µg/L than at an MCL of 40
µg/L. From a health effects perspective, the toxic health
effects on the kidneys or other organs or systems in the body at
exposure levels of 80 µg/L is unknown and is four times EPA's
best estimate of the ``safe level'' with respect to kidney toxicity.
In terms of benefits and costs, Table V-2 (appendix V) shows the
range of compliance costs and net benefits for the uranium MCL options
of 20 µg/L, 40 µg/L, and 80 µg/L. While annual
compliance costs drop significantly as the MCL increases from 20 up to
80, the estimate of cancer cases avoided drops considerably also. In
fact, it is not clear whether the cost per case avoided increases or
decreases with increasing MCL because of the uncertainties involved.
The corresponding estimate of cases avoided for MCLs of 20, 40, and 80
pCi/L are 2.1, 1.5, and 1.0 cases annually. Based solely on cancer
incidence, it may be appropriate for EPA to consider using an MCL
higher than 20 µg/L for uranium, since it is arguable that the
benefits do not justify the costs at this level. However, in terms of
kidney toxicity, 20 µg/L may be justified. EPA solicits comment
on this issue.
Health effects from uranium also need to be evaluated in the
context of the effects of various uranium species and their activity
levels. A mortality risk level of 5 x 10-5 translates to 23
pCi/L of U-238, 22 pCi/L of U-235, and 21 pCi/L of U-234 in drinking
water. An ``alpha spec'' analysis of the water would determine the
fractions of each present and a sum of the fractions below 100% would
meet the MCL. However, this level is costly to obtain. Doubling the
radioactivity limit to 46, 44 and 42 pCi/L for U-238, U-235, and U-234
respectively corresponds to a mortality risk level of 1 x
10-4, which may be more acceptable, considering the costs.
Likewise, a doubling of risk to 2 x 10-4 would again
double the picoCurie limits of each isotope to 92, 88, and 84 pCi/L
respectively. However, at these higher risk levels, the calculated
protective limit for toxicity to the kidney may be exceeded, depending
upon the uncertainty factor used.
By contrast, the relative dissolved concentration of the various
isotopes of uranium will differ markedly from one locale to another.
The 1991 proposal utilized a conversion factor of 1.3 picoCuries per
microgram of uranium to convert a 20 µg/L proposed MCL in mass
units to activity units in picoCuries (however, 1991 cost estimates
were based on the more accurate conversion ratio of 0.9). Analysis of
NIRS data suggest that it would have been more appropriate to use the
1.3 pCi/µg conversion factor for total uranium where
concentrations are less than 3.5 pCi/L and a 0.9 conversion factor for
concentrations above 3.5 pCi/L (Telofsky 1999). Converting the derived
MCL option of 20 µg/L from mass to activity using a ratio of
0.9 for levels above 3.5 µg/L yields approximately 18 pCi/L . A
statistical evaluation of uranium data reveals that, based on a linear
regression of the data, the appropriate activity based MCL for 20
µg/L would be 17.3 pCi/L rounded to 17 pCi/L. Coupled with the
knowledge that the concentration of uranium isotopes varies from place
to place, the Agency is led to consider an MCL that
[[Page 21588]]
is protective in any location against both toxicity (µg/L) and
cancer (pCi/L), whichever presents the greatest risk. This can be
determined by conducting isotopic analysis to determine the relative
amounts of each isotope in any one water system. Once the concentration
ratio is known, a regulated entity may choose to measure mass or
activity and select whichever analytical method or methods is most cost
effective.
For example, if the uranium standard were 20 µg/L or pCi/L,
a gross alpha measurement screen for uranium could be used in the
following way (EPA 2000c and 2000d). The analysis breaks out as
follows: if the result is below the detection limit for gross alpha,
neither uranium measurements by mass or activity would be necessary
since neither 20 pCi/L nor 20 µg/L could be exceeded. If the
gross alpha test is between 3 and 5.5 pCi/L, the mass of 20 µg/
L could be exceeded if all the activity were coming from uranium-238.
Therefore a fluroimetric test for uranium mass concentration
(µg/L) or an alpha spectrometry test for the activities (pCi/L,
converted to µg/L using standard isotopic conversion factors)
of the various isotopes present would be necessary to determine the
uranium concentration in µg/L. Because gross alpha tests may
underestimate uranium by a factor of as much as 3.62, if the gross
alpha test exceeded 5.5 pCi/L (203.62), it is indicative that
the 20 pCi/L limit may be exceeded, and an isotopic analysis must be
done. EPA solicits comment on these issues.
EPA is soliciting information and comment on the data and the
appropriate course of action the Agency should pursue, given the
factors of risk levels, national cost, number of cancers avoided, cost
per case, cost per death, and kidney toxicity. EPA is currently
evaluating three regulatory options:
Regulate at 20 µg/L and 20 pCi/L (protective of
kidney toxicity using the Agency standard 100-fold uncertainty factor
for this type of LOAEL with an associated cancer risk of approximately
5 x 10-5 or five in one hundred thousand);
Regulate at 40 µg/L and 40 pCi/L (this is twice
the safe level with respect to kidney toxicity and would reduce the
margin of exposure between the effect level and the proposed regulatory
standard; with an associated cancer risk level of 1 x 10-4
or one in ten thousand, which is the Agency's usual upper cancer risk
target);
Regulate at 80 µg/L and 80 pCi/L (this is four
times the safe level with respect to kidney toxicity and would further
reduce the margin of exposure between the effect level and the proposed
regulatory standard; with an associated cancer risk level of 2 x
10-4 or two in ten thousand, which is above the Agency's
usual upper cancer risk target).
In summary, EPA believes that 20 µg/L is feasible and is
the Agency's preferred option, but may not have benefits that justify
the costs. Were a higher level to be chosen, EPA would be exercising
its discretionary authority under section 1412(b)(6) to select a level
above the feasible level. It should be noted, however, that there may
be considerable non-quantifiable benefits of avoiding exposure to
cancer and kidney toxicity. Also, as discussed previously, there is
little available data or information about the effects of kidney
toxicity at relatively high exposures and thus, the benefits
attributable to avoided illness cannot be quantified. Thus, the costs
may be justified at a more stringent level than would be suggested in
light of the currently quantifiable benefits alone. In addition, the
Agency generally does not establish regulatory levels outside of its
target risk range and, in fact, prefers to set levels at the more
protective end of that range (1 x 10-6), wherever
possible. Further, we usually follow Agency guidelines on use of
uncertainty factors. For these reasons, the Agency does not favor an
MCL option of 80 µg/L, but solicits comment on this and the
previously-described regulatory options, together with any supporting
rationale or data commenters wish to provide.
I. Inclusion of Non-Transient Non-Community Water Systems
Today's document is soliciting comment on several approaches for
covering Non-Transient Non-Community (NTNC) water systems. Although
current radionuclide regulations do not apply to NTNC water systems, in
1991 EPA proposed extending the radionuclides NPDWRs to include them.
Several approaches representing varying degrees of control are being
currently considered for finalization because, although much more has
been learned about NTNC water systems and their customers since 1991,
there is still very little known about the distribution of the highest
levels of radionuclides in their water supplies. Based on the Agency's
occurrence estimates, control of some radionuclides in NTNC water
systems may not present a meaningful opportunity for health risk
reduction. This issue arises as a consequence of the 1996 Amendments to
SDWA which allow the Agency to consider whether the benefits of
extending coverage to this category of water systems would justify the
costs (section 1412(b)(6)(A)) and whether such regulation would provide
a meaningful opportunity for health risk reduction (section
1412(b)(1)(A)(iii)). The Technical Support Document (EPA 2000a)
presents a ``what if'' analysis for costs and benefits for NTNCWSs.
While it is feasible to control radionuclides in NTNC water
systems, extending regulation to these systems needs to be considered
in light of the new SDWA requirements. This analysis requires a
balancing of both quantitative and non-quantitative factors. Based on
the risk modeling discussed in the Technical Support Document (EPA
2000a), the ninetieth (90th) percentile lifetime risk of cancer
incidence in an individual consuming water from a NTNC water system in
the absence of a regulation is not expected to exceed three in 100,000
\2\. The cost per cancer case avoided to achieve reductions in these
risks would considerably exceed the hundred million dollar mark if
coverage of the rule were extended to NTNCs. The associated cost per
case avoided ranges are well above the range of historical
environmental risk management decisions.
---------------------------------------------------------------------------
\2\ Throughout this discussion, exposures and risks were only
considered for populations potentially addressable by regulation,
i.e. systems with radionuclides present in excess of the proposed
MCLs for community water systems.
---------------------------------------------------------------------------
Relative to community water systems, NTNC systems have much lower
associated risk levels because most individuals served by these systems
are expected to receive only a small portion of their lifetime drinking
water exposure from this source \3\. This conclusion holds even using
very conservative assumptions for modeling the NTNC exposure scenarios.
For example, in the case of school children exposure, the Agency has
conservatively assumed all impacted children would attend only schools
served by NTNC water systems, have twelve years of perfect attendance,
and get half of their daily water consumption at school. For the
average thirteen year old, this scenario implies half of a liter (over
sixteen ounces) every school day. Even under this very conservative set
of assumptions, the water consumed by an individual student is
estimated to represent less
[[Page 21589]]
than five percent of lifetime consumption \4\.
---------------------------------------------------------------------------
\3\ It is important to remember that the risk assessment for
NTNC water systems does not consider exposure risk from private
wells which may serve some customers at home. EPA recognizes that
the radionuclide levels in some private wells may exceed the MCLs
for CWSs, but this is a non-controllable factor since private wells
are not regulated by the Safe Drinking Water Act.
\4\ Day care exposure is similarly conservatively estimated by
assuming five years of perfect attendance, fifty weeks per year and
five days per week. Factory workers are assumed to perfectly attend
and work at the same facility for forty-five years. All of these
assumptions are under continuing investigation and will likely be
revised downward in the future as the Agency is able to gather
further information.
---------------------------------------------------------------------------
On the other hand, much remains to be learned about the NTNC water
systems. Little is known about the extent to which users of the
different NTNC water systems use other water systems. It is conceivable
that some areas in the country exist where individuals are subjected to
exposure at a number of different non-community systems (e.g., day care
center plus school plus factory, etc.). In such circumstances,
individuals would be exposed to proportionately higher risks if the
water systems all had elevated levels. For some individuals, the
exposures could approach levels observed in corresponding community
water systems.
This concern is somewhat alleviated by the fact that NTNC systems
generally serve only a very small portion of the total population. For
example, over ninety-five percent (95%) of all school children are
served by community water systems, not NTNC systems. Only a small
percentage of children are served by NTNC water systems and, of that
group, less than one percent (or less than one in 2000 of the overall
student population) would be expected to have individual radionuclides
in their water above the proposed regulatory levels. Likewise, less
than 0.1 percent of the work force population receive water from an
NTNC water system. With such low portions of the total population
exposed to any particular type of NTNC system, the overall likelihood
of multiple exposure cases in the NTNC population should also be small.
Nevertheless, because children are more sensitive to radionuclides
exposure \5\, multiple water system exposure scenarios were considered
in the modeling effort \6\. Tables III-2 and III-3 present individual
risk estimates for average and most sensitive populations among the
NTNC water systems. All of these factors contributed to the Agency's
evaluation of whether or not to extend regulation to NTNC water systems
and are discussed further in the appendix.
---------------------------------------------------------------------------
\5\ As an example, the lifetime risk per pCi/L of Ra-228 to a
child whose exposure begins under the age of five is more than ten
times greater than the lifetime risk of an individual whose exposure
begins between the ages of 25 and 30.
\6\ For example, the possibility that a child spent five years
in a day care center, then twelve years in schools, and then forty-
five years working in a factory served only by NTNC water systems
with high radionuclide levels.
---------------------------------------------------------------------------
Review of Table III-3 shows that 90th percentile individual risk
patterns for NTNC water system users exposed to uranium or radium-226
are relatively low. These 90th percentile figures represent risks
estimated using the previously described conservative exposure
scenarios, maximum water consumption patterns, and what are effectively
99.9th percentile occurrence estimates \7\ from the NIRS data. Even
with these conservative factors, lifetime cancer risks do not exceed
the one in 10,000 level which has traditionally formed the upper bound
of allowable risk in Agency decision-making.
---------------------------------------------------------------------------
\7\ In other words, the expected number of NTNC systems
nationwide would be less than twenty. It is because these levels are
so rare that the level is fairly speculative. As discussed in the
appendix, the Agency believes its estimates of occurrence are
reasonable, based on levels observed in small ground water community
water systems.
Table III-2.--Selected Sector and Overall NTNC, Individual Risk Patterns
[Lifetime cancer risk for individuals using average consumption levels]
--------------------------------------------------------------------------------------------------------------------------------------------------------
Sector Alpha Radium 226 Radium 228 Uranium
--------------------------------------------------------------------------------------------------------------------------------------------------------
School Students.................. 2 x 10-5 0.9-1.1 x 10-5 2-3 x 10-5 0.7-0.9 x 10-5
Day Care Children................ 2-3 x 10-5 0.6-0.7 x 10-5 2-3 x 10-5 0.8-1x10-5
Factory Worker................... 1-2 x 10-5 1 x 10-5 2-3 x 10-5 1 x 10-5
All NTNC Water Systems........... 0.3-0.4 x 10-5 0.2-0.3 x 10-5 0.5-0.7 x 10-5 0.2 x 10-5
--------------------------------------------------------------------------------------------------------------------------------------------------------
Note that Radium 224 is being used as a surrogate for alpha emitters.
Table III-3.--Selected Sector and Overall NTNC, Individual Risk Patterns
[Lifetime cancer risk for individuals using 90th percentile consumption levels]
--------------------------------------------------------------------------------------------------------------------------------------------------------
Sector Alpha Radium 226 Radium 228 Uranium
--------------------------------------------------------------------------------------------------------------------------------------------------------
School Students.................. 0.5-0.6 x 10-4 0.3 x 10-4 0.6-0.7 x 10-4 0.3 x 10-4
Day Care Children................ 0.7-0.8 x 10-4 0.2 x 10-4 0.6 x 10-4 0.3-0.4 x 10-4
Factory Worker................... 0.5-0.6 x 10-4 0.3-0.4 x 10-4 0.7-0.8 x 10-4 0.5-0.6 x 10-4
All NTNC Water Systems........... 0.2 x 10-4 0.1 x 10-4 0.2-0.3 x 10-4 0.1-0.2 x 10-4
--------------------------------------------------------------------------------------------------------------------------------------------------------
Radium-228 and gross alpha pose approximately twice the threat of
the other two radionuclides. While sensitive individual estimates still
fall below the one in ten thousand range, they may not in a scenario in
which other drinking water sources are similarly high. However, as
stated previously, the Agency views it as somewhat improbable that this
system overlap occurs to a significant extent. Nevertheless, it could
be an issue in some rural communities. While such infrequent and highly
site-specific conditions are very difficult to address efficiently in a
National-level regulation, the Agency believes that exempting NTNC
water systems from the radionuclide NPDWRs, given the degree
uncertainty about the occurrence levels and extent of system customer
overlap, may be inappropriate. For these reasons, the Agency believes
it may be appropriate to take a somewhat different approach with
respect to NTNC water systems than previously practiced.
EPA is considering extending partial coverage of the radionuclide
NPDWRs to NTNC water systems under several possible scenarios. Under
the first three options, NTNC systems would be subject to targeted
radionuclide monitoring requirements, in which selected NTNC systems
would follow the radionuclides monitoring requirements for community
water systems. The targeting strategy would be based on small community
water system occurrence for the same radionuclides.
[[Page 21590]]
The States (or primacy agency) would determine which NTNC systems are
likely to be using contaminated water systems, based on CWS monitoring
results. These systems may then be required to monitor and meet CWS
MCLs for gross alpha and combined radium and other relevant
radionuclides. EPA is considering:
Requiring targeted NTNC systems to monitor and meet the
CWS MCLs for all or selected radionuclides, where targeting is
determined by the State based on whether the NTNC system is using
source water for which CWSs have reported MCL violations for
radionuclide in question;
Requiring targeted NTNC systems to monitor and post notice
if the system exceeds the CWS MCL, using the same definition of
targeting as in the first option;
Issuing guidance that recommends that targeted NTNC
systems monitor and meet the CWS MCLs, using the same definition of
targeting as in the first option.
The Agency requests comments on these options and any supporting
rationale for such a decision. The Agency is also interested in
receiving comments on other options such as extending full coverage of
the rule to NTNCs and not extending any aspect of the radionuclides
NPDWRs to NTNC systems. The Agency will decide, as part of the upcoming
finalization of the 1991 proposal, to incorporate what it considers to
be the most appropriate option in view of available the data and
information.
J. Analytical Methods
Today's NODA provides a brief update of the methods-related items
which have occurred since the 1991 proposed rule. For a more thorough
discussion of the analytical methods updates, the public is referred to
appendix III of this NODA and to the Analytical Methods section of the
Technical Support Document for the Radionuclides Notice of Data
Availability (EPA 2000a).
1. Radionuclides Methods Updates
On July 18, 1991 (56 FR 33050; EPA 1991), the Agency proposed to
approve fifty-six methods for the measurement of radionuclides in
drinking water (excluding radon). Fifty-four of the fifty-six were
actually approved in the March 5, 1997 final methods rule (62 FR 10168;
EPA 1997a). In addition to these fifty-four, EPA also approved 12
radiochemical methods, which were submitted by commenters after the
1991 proposed rule. Currently, an overall total of 89 radiochemical
methods are approved for compliance monitoring of radionuclides in
drinking water. These methods are currently listed in 40 CFR 141.25.
The March 5, 1997 Federal Register also approved suitable
calibration standards for the analysis of gross alpha-emitting
particles and gross beta-emitting particles. These specific methods-
related items are addressed in some detail in the Technical Support
Document for the Radionuclides Notice of Data Availability (EPA 2000a)
and in even greater detail in the 1997 final methods rule (62 FR 10168,
EPA 1997a) and the 1991 proposed rule (56 FR 33050; EPA 1991).
This NODA also notifies the public about the use of the gross beta
method for the screening of radium-228. In the 1991 proposed rule (56
FR 33050; EPA 1991), the Agency would have allowed the use of the gross
beta-particle activity method to screen for the presence of radium-228
at the proposed radium-228 MCL of 20 pCi/L. For the combined radium-226
and 228 standard of 5 pCi/L (the current standard), the Agency can not
recommend the use of the gross beta-particle activity method for
screening of radium-228. Instead, a specific analysis for radium-228
would be necessary. Although several methods are currently approved for
the analysis of radium-228 in drinking water, the Agency requests
comments from the public and supporting documentation regarding other
radium-228 methods or method variations which may be able to reach
greater sensitivity at the 2 pCi/L level.
2. The Updated 1997 Laboratory Certification Manual
In the 1991 proposed rule (56 FR 33050; EPA 1991), EPA cited the
1990 laboratory certification manual's guidance for sample handling,
preservation, holding time and instrumentation. In response to the 1991
proposed rule, a commenter questioned why the holding time for
radioactive iodine was six months, when the half-life of iodine-131 is
eight days. The Agency recognized this typographical error and changed
the holding time to eight days in the updated 1997 certification manual
(EPA 815-B-97-001; EPA 1997d). Table III-2 in the appendix shows the
updated guidance for sample handling, preservation, holding times, and
instrumentation that appeared in this manual. Table III-2 in the
appendix also includes additional recommendations for radiochemical
instrumentation (footnoted by the number 6). The Agency is seeking
comment about the additional recommendations found in Table III-2.
3. Recommendations for Determining the Presence of Radium-224
To determine the presence of the short-lived radium-224 isotope
(half life 3.66 days), the Agency recommends using one of
the several options discussed in the appendix III. Although these
measurement options are only recommendations, the Agency strongly urges
water systems to check for the presence of radium-224 in their drinking
water supplies. Comments are solicited from the public about the
options listed in appendix III or any other appropriate methods of
detection.
4. Cost for Radiochemical Analysis
Revised Cost Estimates for Radiochemical Analysis.
In the 1991 proposed rule (56 FR 33050; EPA 1991), EPA cited cost
estimates for radiochemical analyses. The Agency updated these costs
estimates by surveying a small number of radiochemical laboratories (no
more than 9 laboratories) (EPA 2000a). The revised cost estimates are
shown in Table III-3 (appendix III). Because this information is based
on a limited number of laboratories, the slight increase in costs from
1991 to 1999 may be due to either statistical uncertainty or possibly
others factors such as inflation.
After the 1991 proposed rule, there were several comments regarding
analytical costs. One commenter stated the costs of analysis for
radium-226, radium-228, radioactive strontium and total strontium were
unrealistically low. The Agency can neither agree nor disagree. As
noted earlier, EPA revised the cost estimates for radiochemical
analysis. Both the 1991 costs estimates and the revised cost estimates
were from small surveys and may not be truly representative of the
actual costs for some radiochemical analyses. Comparison of the
estimated costs from 1991 with the revised cost estimates indicate the
costs for some analyses to similar, while for other analyses, cost do
appear to be higher. The Agency solicits comments and factual data that
would clarify this matter.
Several commenters stated that small systems, which are likely to
need only a few analyses, cannot take advantage of rates for volume
sample analyses. The Agency agrees that individual small systems may
not be able to take advantage of lower bulk analysis costs. To
alleviate cost burdens, small systems may want to consider pooling
their analytical needs with other small systems to negotiate for bulk
rates.
[[Page 21591]]
5. Externalization of the Performance Evaluation Program
Due to resource limitations, on July 18, 1996 (61 FR 37464; EPA
1996b), EPA proposed options for the externalization of the PE studies
program (now referred to as the Proficiency Testing or PT program).
After evaluating public comment, in the June 12, 1997 final notice EPA
(62 FR 32112; EPA 1997b):
decided on a program where EPA would issue standards for the
operation of the program, the National Institute of Standards and
Technology (NIST) would develop standards for private sector PE (PT)
suppliers and would evaluate and accredit PE suppliers, and the
private sector would develop and manufacture PE (PT) materials and
conduct PE (PT) studies. In addition, as part of the program, the PE
(PT) providers would report the results of the studies to the study
participants and to those organizations that have responsibility for
administering programs supported by the studies.
EPA has addressed this topic in public stakeholders meetings and in
some recent publications. For more information, readers are referred to
the aforementioned Federal Register notices. More information about
laboratory certification and PT (PE) externalization can be accessed at
the OGWDW laboratory certification website under the drinking water
standards heading (www.epa.gov/safewater). At this time, it is
difficult to ascertain how and if externalization of the PT program
will affect radiochemical laboratory capacity and the cost of
radiochemical analyses. In the absence of definitive cost estimates,
the Agency solicits public comments on this subject.
6. The Detection Limits as the Required Measures of Sensitivity
In 1976, the National Primary Drinking Water Regulations defined
the detection limit (DL) as ``the concentration which can be counted
with a precision of plus or minus 100 percent at the 95 percent
confidence level (1.96 s, where s is the standard
deviation of the net counting rate of the sample).'' Table III-4 in the
appendix cites the detection limits or the required sensitivity for the
specific radioanalyses that were listed in the 1976 rule and are also
cited in 40 CFR 141.25. In the 1991 proposal (56 FR 33050; EPA 1991),
EPA proposed using the method detection limit (MDL) and the practical
quantitation level (PQL) as measures of performance for specific
radioanalytical methods. Acceptance limits based on the PQLs, which
were derived from performance evaluation studies, were also proposed in
the 1991 rule. Some commenters found the use of acceptance limits
confusing and the relationship to the actual method performance was not
clear. With perhaps the exception of uranium, the Agency will not go
forward with the proposed acceptance limits, PQL, or MDL. Because
uranium has never been regulated, it did not have a detection limit in
the CFR and one has never been proposed. In 1991, EPA did propose a PQL
of 5 pCi/L with an acceptance limit of +/- 30%. Although it is believed
that a detection limit for uranium would be very similar to the PQL,
because a detection limit has never been proposed, the Agency may have
to adopt the PQL for uranium until a detection limit is proposed. For
the other radionuclides, which are regulated, the Agency believes the
current 1976 detection limit requirements are most appropriate. The
existing definition of the detection limit takes into account the
influence of various factors (efficiency, volume, recovery yield,
background, counting time) that typically vary from sample to sample.
Furthermore, the detection limit is computed for each individual sample
and does not represent an idealized set of measurement parameters.
Therefore, the detection limit reflects the expected random uncertainty
for a given sample analysis.
7. Performance Based Measurement System
On October 6, 1997, EPA published a Notice of the Agency's intent
to implement a Performance Based Measurement System (PBMS) in all of
its programs to the extent feasible (62 FR 52098; EPA 1997c). EPA is
currently determining how to adopt PBMS into its drinking water
program, but has not yet made final decisions. When PBMS is adopted
into the drinking water program, its intended purpose will be to
increase flexibility in laboratories in selecting suitable analytical
methods for compliance monitoring, significantly reducing the need for
prior EPA approval of drinking water analytical methods. Under PBMS,
EPA will modify the regulations that require exclusive use of Agency-
approved methods for compliance monitoring of regulated contaminants in
drinking water regulatory programs. EPA will probably specify
``performance standards'' for methods, which the Agency would derive
from the existing approved methods and supporting documentation. A
laboratory would be free to use any method or method variant for
compliance monitoring that performed acceptably according to these
criteria. EPA is currently evaluating which relevant performance
characteristics under PBMS should be specified to ensure adequate data
quality for drinking water compliance purposes. After PBMS is
implemented, EPA may continue to approve and publish compliance methods
for laboratories that choose not to use PBMS. After EPA makes final
determinations about the implementation of PBMS in programs under the
Safe Drinking Water Act, the Agency would then provide specific
instruction on the specified performance criteria and how these
criteria would be used by laboratories for compliance monitoring of
SDWA analytes.
K. Monitoring
1. Features of Today's NODAA
EPA's 1976 regulations for radionuclides in drinking water
contained separate monitoring requirements for radiums, alpha emitters,
and man-made beta and photon emitters. In 1991, EPA proposed to make
modifications to the 1976 regulations to expand the scope of coverage
to include non-transient non-community water systems, to change the
monitoring location and monitoring frequencies, and to incorporate
monitoring requirements for radon and uranium. A summary of the 1976
requirements and proposed changes in 1991 are presented in this
section.
In today's document EPA is suggesting merging the current
requirements and the 1991 proposed requirements into a unified system
which is consistent with the Standardized Monitoring Framework (SMF),
the current rule, and the proposed changes which are still germane. EPA
is soliciting comment on monitoring at the entry points to the
distribution system, as proposed in 1991, to ensure equal protection
for all customers, under the sampling schedule of the SMF. EPA believes
that this will increase consistency between monitoring requirements for
radionuclides and the other regulated contaminants. As described in
section III, part I (``Inclusion of Non-Transient Non-Community Water
Systems''), EPA is considering several options for NTNC water systems,
some of which would require monitoring. Because some monitoring
provisions of the 1991 proposal were based on the proposed MCLs and not
the current MCLs, their application to the current levels may entail a
slightly different construct than in 1991. To the extent comments
reveal aspects of the framework which need to be addressed separately
from the 1976 rule or 1991 proposal, EPA will return
[[Page 21592]]
to the current rule's framework and propose to correct deficiencies via
a proposal which will address analytical method issues as well, such as
methods for Ra-224, Po-210 and Pb-210.
2. Standardized Monitoring Framework
Per the current rule, once the contaminant concentration in the
water is established by the average results for four consecutive
quarterly samples or by suitable grandfathered data, a system would be
categorized as to whether it was above or below 50% of the MCL for that
contaminant. In accordance with the SMF, as proposed in 1991, EPA is
suggesting a tiered frequency for alpha emitters, combined radium, and
uranium. This would entail one sample every three years for compliant
systems with annual average contaminant levels above 50% of the MCL.
For compliant systems with annual average levels below 50% of the MCL
for these contaminants, one sample would be required every 6 years;
non-detects, one sample every 9 years. EPA believes this system would
align with the standardized monitoring framework, and would provide
regulatory relief for systems with low to very low levels (without
needing a waiver as called for in 1991). It would also provide more
careful screening for systems with multiple sources of water entering
the distribution system, by requiring a sample at each of these points
to be protective of all of the customers within each water system. For
beta particle and photon radioactivity, EPA is considering requiring
four consecutive quarters every four years, the requirement under the
current rule, for vulnerable systems because of their proximity to
contamination sources.
EPA believes this monitoring scheme is less burdensome on systems
in the long term than either the existing or proposed regulations. It
provides slightly more protection than the current rules by more
frequent monitoring for contaminants above half the MCL, and less
frequent monitoring for the vast majority of systems below half the
MCL. EPA believes this is more realistic and less burdensome, while
recognizing the potential for variability of naturally occurring
radionuclide levels in ground water over time. Such variability (e.g.,
a change in pH by nitrogen fertilizer application leading to a higher
solubility of radium) was seen in New Jersey and is further discussed
in appendix I.
Small ground water systems comprise the vast majority of systems
with radionuclide contamination problems. Since most small systems have
only one entry point, an entry point monitoring requirement will not
have an impact. For systems with radium above 50% of the MCL, with
three or fewer entry points to the distribution system, monitoring at
each entry point once every three years would have an equal or smaller
impact (in terms of the number of samples analyzed) than the 1976
requirement of monitoring four times every four years.
3. Entry Point Monitoring
EPA recognizes that sampling conducted at the monitoring location
specified in the current rule may under-represent the risk to some
consumers. Results can vary depending on the usage of each water source
and changes in the monitoring location within the distribution systems.
For systems with more than one water source, monitoring within the
distribution system may yield different results. In the current rule,
sampling is conducted ``at a free flowing tap'' within the distribution
system. The current rule also recognizes the potential problems by
providing that systems with two or more sources of water with different
concentrations of radionuclides monitor the source water, as well as
water from a free flowing tap, when ordered by the State. Entry point
monitoring, a feature of more recent NPDWRs, provides a better measure
of water quality for residents near the start of the distribution
system than monitoring within the distribution system (e.g., the middle
of the system) where water is subject to blending if there are other
sources. Therefore, EPA proposed in 1991 to change the location for
compliance monitoring to the entry points to the distribution system,
consistent with other NPDWRs.
4. Grandfathering Data
In the implementation guidance, which will be available on OGWDW's
home page (http://www.epa.gov/ogwdw), EPA is suggesting that samples
within the latest compliance period, beginning June, 1996, be eligible
for use in determining the baseline for monitoring frequency. While EPA
prefers this approach, others may be possible. Please provide data and
supporting rationale if you comment on this issue. The application of
this provision would extend to all classes of radionuclides for which
data are available.
The Agency solicits comment on two different approaches for the
beta monitoring requirements. The first option is to not allow any
reduced monitoring and the second would be to allow reduced monitoring
similar to the alpha emitters. If systems must collect samples on a
quarterly basis (no reduced monitoring) then grandfathering of data is
not necessary. If the Agency decides to allow reduced monitoring, the
Agency believes that States may use historical data to supplement their
vulnerability assessments but should not use grandfathered data to
satisfy the initial monitoring requirements because a sufficient
baseline needs to be established in those systems considered vulnerable
to man-made radioactivity.
Grandfathered data would be used to comply with the initial
monitoring requirements for gross alpha, radium-226/228, and uranium,
under some circumstances. Data collected after June 1996, during the
most recent compliance period, would be considered for grandfathering.
It would be the State's responsibility to determine if grandfathered
data is sufficient to satisfy the initial monitoring requirements
established by this rule. At the State's discretion, systems with one
entry point to the distribution system (EPTDS) could use grandfathered
data to satisfy the initial monitoring requirements. Systems that have
multiple entry points to the distribution system could use
grandfathered data collected after June 1996 to satisfy the initial
monitoring requirements, provided that the data were collected at the
EPTDS.
EPA is also considering that, at the State's discretion, systems
with up to three entry points to the distribution system could also use
grandfathered data to satisfy initial monitoring requirements, even if
not collected from EPTDS, if the State makes a written finding that the
circumstances of the system and their review of historic data justify
such action. While the Agency cannot prescribe every possible scenario
that a State may encounter, an example of circumstances that might
support such a finding could be: a system that has three wells (and
EPTDS), that are simply from different parts of a well-field, using the
same aquifer, with good historical data showing uniform, low to no
radionuclide occurrence from all wells, perhaps from the raw water as
well as distribution system samples.
5. Sample Compositing
In general, compositing of samples is an effective means of
decreasing analytical costs to systems. Compositing is permitted for
alpha emitters and beta and photon emitters in the current rule. It is
also allowed for radium-226 and -228 to the extent gross alpha was used
as a screen for Ra-226 and, in turn, Ra-228. In the 1991 proposal,
gross beta compositing was prohibited. Compositing for other nuclides
was
[[Page 21593]]
allowed for up to five sampling points within one system; if the result
for the composite was more than 3 pCi/l for any nuclide, individual
non-composited samples were to be analyzed. This provision stemmed from
the lowest MCL in the 1991 proposal, adjusted gross alpha at 15 pCi/l.
Because of the possibility that one of the five samples taken might be
at 15 pCi/L even if the other four were at zero, the rule envisioned
one fifth (3 pCi/L) of the MCL as the maximum allowed result to assure
no single well could exceed the MCL. The principle of limiting the
result of five composited samples from separate entry points to one
fifth of the MCL (or four composites to one fourth the MCL etc.) is
still valid as a general matter, and should be followed whenever
compositing is done.
A 3 picoCurie limit in the proposal would have been conservatively
protective for a five sample composited for the proposed separate MCLs
radium-226 and radium-228 at 20 pCi/L each, since \1/5\ of each MCL is
greater than 3 pCi/L. However, because EPA is considering retaining the
current radium standard at 5 pCi/L combined, adding the results of five
composited entry points samples for Ra-228 to the results of 5
composited entry point samples of Ra-226 must yield a result of one
tenth (\1/10\) of the MCL to be assured that the combined Ra-226 and
Ra-228 concentration could not exceed the MCL at any one entry point.
Because one tenth of the MCL (0.5 pCi/L) is below the detection limit
for Ra-226 and Ra-228, compositing of separate entry points cannot
apply in case of Ra-226 or Ra228. However, annual compositing of
samples from the same entry point may apply.
EPA requests comment on the feasibility and practical utility of
compositing separate entry points (spatial compositing) versus
compositing samples over time from the same entry point (temporal
compositing). EPA believes that the use of one or the other (but never
both simultaneously) may be appropriate under some circumstances.
Greater certainty in the analytical result is obtained by taking the
average of four separate (non-composited) results from one sampling
location than by using a single result of composited samples. However,
where an MCL is sufficiently above the detection limit such that
analytical results are not subject to significant error near the MCL,
compositing may be a cost saving measure. Additionally, when historical
data indicate that contaminant levels are negligible (e.g., non-
detects) for a water system, compositing among wells in a system or
between systems having one point of entry may be advisable at State
discretion. However, because of the costs of re-sampling and re-
analysis of all points to confirm an MCL violation, or to qualify for
decreased monitoring, it may not be in the systems best interest to
initially composite in the absence of historical data.
6. Increased and Decreased Monitoring
Additionally, the Agency is considering having the final rule allow
systems that are currently on a reduced monitoring schedule to remain
on that reduced schedule as long as the system qualifies for reduced
monitoring based on the most current analytical result. Systems for
which the most current analytical result indicates a higher level than
allowed for that monitoring schedule would resume monitoring at a
frequency consistent with the most recent result. For example, a system
with an annual average below half of the MCL could reduce monitoring to
one sample every 6 years. If, while on this reduced frequency, the
system collects a sample with an analytical result above half the MCL,
the system would have to increase monitoring again to once every 3
years. It could revert to its previous reduced frequency of once every
six years if the subsequent analytical result (of the sample taken
three years later) was less than half the MCL. EPA also believes it is
prudent to require quarterly samples to be collected at least 60 days
apart, to capture seasonal variations. EPA solicits comment on this and
other monitoring provisions.
7. Compliance Determinations
Compliance would be determined based on the annual average of
quarterly samples collected at each entry point for all classes of
radionuclides. If the annual average of any entry point exceeds an MCL,
the CWS would be in violation. If NTNC systems are subject to MCLs, the
same situation would apply to them. An immediate violation would occur
for any sample analytical result or combination of sample analytical
results that would place the system in violation before four quarters
of data are collected (e.g., the first sample is greater than 4 times
the MCL or the average of the first two samples is greater than twice
the MCL). If a system has a sample that exceeds the MCL while on
reduced monitoring, it would need to begin quarterly monitoring the
following quarter. Compliance would be based on the average of the four
consecutive quarters of data beginning with the initial result that
exceeded the MCL. If a system fails to collect all samples required
during any year, compliance would be calculated based on available
data. Under the current rule, quarterly monitoring is continued until
the annual average concentration no longer exceeds the MCL or until a
monitoring schedule as a condition to a variance, exemption, or
enforcement action becomes effective.
The following is a summary of certain features of the monitoring
requirements for each regulated radionuclide or radionuclide group.
8. Combined Radium-226 and -228
Standardized monitoring: EPA contemplates application of the
standardized 3, 6, 9 year cycles to the combined radium standard
depending on whether analytical results for compliant systems are
greater than (3) or less than (6) half the MCL or are a non-detect (9),
as previously discussed. Decreased and increased monitoring would be
based on the result of the analysis of the most recent required
sample(s).
Entry point monitoring: Monitoring at entry points to the
distribution system would be a requirement per the 1991 proposal unless
EPA receives comments with compelling reasons for not doing so.
Sample Compositing: To decrease the burden of monitoring at
distribution entry points, EPA is contemplating allowance of sample
compositing for radium-226 or radium-228, but only when results will be
indicative of the true level at a single entry point (temporal
compositing). According to the proposal, systems would be required to
analyze for Ra-228 separately from gross alpha or Ra-226. The Agency
sees no reason why four separate samples from a single entry point
(collected 60 days apart) could not be either analyzed and averaged or
composited in the laboratory and analyzed, to determine future
monitoring frequency. Therefore, EPA is suggesting for public comment
that systems take the average analytical results from four individual
samples, or the composite of four samples from each entry point, in
order to determine future frequency.
As discussed previously, EPA does not contemplate allowing
compositing of multiple entry points for derivation of combined radium
results. EPA requests comment on any element of the foregoing
discussion.
9. Alpha Emitters
Standardized monitoring: Same as for combined radium (see previous
discussion).
[[Page 21594]]
Decreased and increased monitoring: Same as for combined radium
(see previous discussion).
Entry point monitoring: Same as for combined radium (see previous
discussion).
Sample Compositing: The current rule allows compositing of four
samples in a laboratory or the averaging of four separate analyses.
Under the 1991 proposal and the current rule, systems would be allowed
to composite annually for samples taken from single entry points
(temporal compositing) and, under the 1991 proposal, to composite
samples representing up to five entry points with a six month holding
time (spatial compositing).
10. Uranium
Standardized monitoring, monitoring frequency, and entry point
monitoring: Same as for combined radium (see previous discussion).
Sample Compositing: For systems with gross alpha levels that are
high enough to warrant uranium monitoring, annual composites for a
single entry point would be allowed. Compositing of five samples
representing five entry points would be permitted. If the result was
greater than one fifth of the MCL, the individual samples would have to
be analyzed or re-sampling and analysis of the new individual samples
would have to occur.
11. Beta and Photon Emitters
Standardized monitoring framework, decreased monitoring: Monitoring
for beta and photon emitters would follow the same schedule as in the
current rule. Decreased monitoring is not envisioned for beta and
photon emitters since only vulnerable systems would monitor, although
EPA is taking comment on the possibility of decreased monitoring
according to the standardized monitoring framework as outlined
previously.
Screening levels: EPA recognizes certain problems with the current
and proposed system. The proposed requirement of a 30 pCi/L screen for
gross beta and photon emitters had the effect of no longer requiring
Sr-90 monitoring because the proposed limit was above the screen of 30
pCi/L. Under the current MCL, there is only one contaminant that has a
concentration limit near the 50 pCi/L screening level (Ni-63). There
are five contaminants with concentration limits at or near 30 pCi/L and
seven with limits below thirty. A screen level of 50 pCi/L would
potentially miss the 12 contaminants with concentration limits below 50
pCi/L and a screening level of 30 pCi/L would potentially miss the 7
contaminants with concentration limits below 30 pCi/L. Systems that are
drawing water from sources with known beta particle and photon
radioactivity are required to use a screening level of 15 pCi/L under
the current rule. The 1991 proposal retained this feature.
EPA thinks it is advisable to retain the proposed monitoring for
sites within 15 miles of a source of beta photon emitters. The
screening level in the original rule only affected surface water
systems serving over 100,000, or other systems at State discretion, and
the screening level for gross beta reflected this limited regulation.
However, a known source of particular beta and photon emitters should
be monitored for the specific radionuclides present at that source
which may be a health concern below the screen, but would not be
triggered by the screen. EPA would give States discretion on requiring
specific monitoring for contaminants from specific sources.
In addition, a 15 pCi/l screening level is currently required for
systems using water contaminated by effluents from nuclear facilities.
These systems may also be required by the State to monitor for
individual nuclides on a case by case basis. Since both screens may
miss radionuclides of concern, EPA believes this issue is important and
may need to be addressed in a future proposal. In addition, since many
beta particle and photon emitters have half-lives that are too short to
be detected under the current holding time, the issue of sample holding
time may have to be re-visited in a future proposal.
Holding time: Another issue has a bearing on the screening level
for which EPA is requesting comment. There are a significant number of
beta and photon emitting radionuclides with short half lives, including
those 13 nuclides of concern below the screening levels being
considered. Because annual sample compositing is allowed under the
current rule for beta and photon emitters, a screen above 30 pCi/L
would detect a greater number of nuclides which (due to decay) may have
been above a screen of 50 at the time of sampling, but are now between
30 and 50 pCi/L by the time of analysis. A screen level at 30 pCi/L
would be more sensitive a screen for beta particle and photon
radioactivity. The Agency requests comment on the selection of
screening levels.
Sample Compositing: Annual compositing is permitted for beta and
photon emitters in the current rule. In addition, for systems utilizing
water contaminated by effluents from nuclear facilities, a quarterly
compositing of five consecutive daily samples was to be analyzed for
iodine-131, with more frequent monitoring at State discretion if it was
detected in the finished water. EPA believes this compositing for
single nuclide determinations is still valid. However, the 1991
proposed rule excluded compositing for beta and photon emitter samples.
It also limited holding times to 6 months for single samples or 12
months for composites per the lab cert manual. A screen above 50 pCi/L
, but with a sample holding time of 6 months without compositing may be
a reasonable approach, considering screening options, holding times,
and compositing issues. EPA solicits comment on these beta and photon
emitter monitoring issues.
Entry point monitoring: EPA solicits opinion on requiring beta
photon monitoring at entry points to the distribution system for
vulnerable systems. EPA believes this is appropriate as it is for other
nuclides, especially as an early warning of contamination from a
localized source of man-made beta photon emitters.
12. Monitoring for Non-Transient Non-Community (NTNC) Systems
If EPA finalizes an option that requires monitoring for some or all
NTNC systems, EPA wishes to make the monitoring requirements consistent
between CWSs and those NTNC systems required to monitor. See the
previous discussion for CWS monitoring for details. As with CWSs,
monitoring under the SMF would be required at entry points to the
distribution system, based on a nine-year cycle, consisting of three,
3-year monitoring periods, with provisions for reduced monitoring as
appropriate. If the radionuclides NPDWRs for CWSs are fully extended to
NTNCWSs, the monitoring frameworks would be the same.
Table III-4 summarizes the monitoring frequencies for CWSs and NTNC
systems, under the options that require monitoring:
[[Page 21595]]
Table III-4. Comparison of the Monitoring Frameworks: The Existing Rule,
the 1991 Proposal, and the Approach Described in the NODA
------------------------------------------------------------------------
Current rule (1976) 1991 proposal 2000 NODA
------------------------------------------------------------------------
Radium Alpha Emitters and Uranium
------------------------------------------------------------------------
Initial baseline: 4 Initial baseline: Initial baseline: 4
consecutive quarterly one sample per year consecutive
samples. for 3 years. quarterly samples
taken within 3
years from
effective date or
grandfathered data
in previous
compliance period.
If average > MCL=treat, etc. Same as 1976........ Same as 1976.
If one or more samples >MCL, Same as 1976........ Same as 1976.
do quarterly sampling until
average MCL.
If >50% of MCL, 4 Quarters If >50% of MCL, one Same as 1991 with no
every 4 years. sample every 3 yrs waiver.
or waiver to every
9 yrs.
If 50% of MCL, 1 sample If 50% of MCL, one If 50% of MCL, one
every 4 years. sample every 3 yrs sample every 6
or waiver to every years.
9 years.
If no detect, 1 sample every If no detect, one If no detect, one
4 years. sample every 3 yrs sample every 9
or waiver to every years.
9 yrs.
------------------------------------------------------------------------
Beta and Photon Emitters
------------------------------------------------------------------------
Quarterly gross beta Vulnerable systems Same as
monitoring. Vulnerable (surface and ground 1991:Vulnerable
systems and surface water water) within 15 systems within 15
systems > 100,000 pop. miles of source of miles of source of
Screen of 50; screen of 15 man made emitters man made emitters
for contaminated water I- do gross beta will monitor with
131 quarterly, Sr-90 and H- screen proposed at screen of 50 or 30.
3 annual Sr-89 and Cs134 if 30. Same as 1976:
above 15. Screen of 15pCi/L
for systems using
contaminated
waters. Same
contaminants as
1976 with
corrections per NBS
HB-69.
------------------------------------------------------------------------
13. Polonium-210 and Lead-210
Risk estimates based on Federal Guidance Report No. 13 indicate
that current screening levels for gross alpha and gross beta may not be
adequate to capture all contaminants of concern. Specifically, based on
the new health-effects information contained in FGR-13 (EPA 1999b), EPA
believes it may be appropriate to require systems to perform isotopic
analyses for additional radionuclides that may present a significant
threat to human health. As a result of this information, EPA is
requiring some systems to do analyses for polonium-210 (a naturally
occurring alpha emitter) and lead-210 (a naturally occurring beta
emitter) under the Revisions to the Unregulated Contaminant Monitoring
Regulation (UCMR) (64 FR 50556, Friday, September 17, 1999), to be
implemented after analytical methods for these contaminants have been
approved.
14. Reporting Requirements
On May 13, 1999, EPA proposed subpart Q (64 FR 25964) to revise the
minimum requirements public water systems must meet for public
notification of violations of NPDWRs and other situations that pose a
risk to public health from the drinking water. EPA anticipates the
final Public Notification Rule (PNR), under part 141, subpart Q to be
published in early 2000. After the final PNR is published, subsequent
EPA drinking water regulations that affect public notification
requirements will amend the PNR as part of each individual rulemaking.
The proposed PNR divides the public notice requirements into three
(3) tiers, based on the type of violation. ``Tier 1'' applies to
violations and situations with significant potential to have serious
adverse effects on human health as a result of short-term exposure.
Notice is required within 24 hours of the violation. ``Tier 2'' applies
to other violations and situations with potential to have serious
adverse effects on human health. Notice is required within 30 days,
with extensions up to three months at the discretion of the State or
primacy agency. ``Tier 3'' applies to all other violations and
situations requiring a public notice not included in Tier 1 and Tier 2.
Notice is required within 12 months of the violation, and may be
included in the consumer confidence report at the option of the water
system.
Today's NODA requests comment on whether community water systems
(CWS) should provide a Tier 2 public notice for MCL violations under
the radionuclide NPDWRs and to provide a tier 3 public notice for
violations of the monitoring and testing procedure requirements. If
NTNC water systems are required to monitor and notify, then they would
be required to provide a Tier 2 notice if the systems exceed the MCLs.
EPA requests comment on the implementation of public notification
requirements by the effective date of the MCL and on the Tier 2 public
notice requirement for quarterly repeat notices for NTNC systems that
continue to exceed the CWS MCL(s) under the ``monitoring and
notification-only'' option. EPA believes States will phase in
monitoring of NTNC systems based on results of CWS systems in the same
proximity. The agency requests comment on whether or not the same
increase or decreased monitoring requirements which pertain to CWSs
should apply to NTNC water systems i.e. the 3, 6 and 9 year monitoring
based on being above 50% of the MCL, below 50%, or non-detect.
As in the current rules, an analytical result that exceeds the MCL
would trigger additional confirmation samples, which in turn could
trigger quarterly monitoring. For man-made beta and photon emitters,
EPA is suggesting to finalize the proposal regarding a screening level
of 30 or 50 pCi/L for ``vulnerable systems,'' which are defined as
being within a 15 mile radius of a source of this class of
radionuclides. For Pb-210, EPA will be collecting data to make a future
determination regarding additional monitoring for this natural beta
emitter.
Tables III-5 and III-6 summarize the current and proposed
monitoring requirements and those suggested by today's document.
[[Page 21596]]
Table III-5.--Initial (Routine) Monitoring Requirements
------------------------------------------------------------------------
1976 1991 proposal 2000 NODA
------------------------------------------------------------------------
GROSS ALPHA
------------------------------------------------------------------------
CWSs: Four consecutive CWSs and NTNCWSs: CWSs and NTNCWSs
quarters at representative Annual monitoring \1\: Four
point(s) within the at each entry point consecutive
distribution system every for first three quarters of
four years. years. monitoring at each
entry point,
anytime during
first 3 years.
RADIUM
------------------------------------------------------------------------
CWSs: Four consecutive CWSs and NTNCWSs: CWSs and NTNCWSs:
quarters at representative Annual monitoring Four consecutive
point(s) within the for each radium quarters of
distribution system every isotope (radium-226 monitoring for each
four years. Initial and radium-228) at radium isotope
monitoring is for radium- each entry point, (radium-226 and
226. If radium-226 exceeds for three years. radium-228) at each
3 pCi/L, analysis for entry point, any
radium-228 is required. A time during first 3
gross alpha measurement can years.\2\
be substituted for radium
226 and/or uranium
monitoring if the gross
alpha measurement is below
the applicable MCL(s).
------------------------------------------------------------------------
URANIUM
------------------------------------------------------------------------
None........................ CWSs and NTNCWSs: CWSs and NTNCWSs:
Annual monitoring Four consecutive
at each entry point quarters of
for three years. monitoring for
uranium to
determine
compliance with
both mass and
activity either by
gross alpha or
specific mass or
activity analysis
at each entry
point, every three
years.\2\
------------------------------------------------------------------------
BETA AND PHOTON EMITTERS
------------------------------------------------------------------------
CWSs serving > 100,000 Vulnerable systems Vulnerable systems
persons and using surface only CWSs and only CWSs and
water (and other systems NTNCWSs: (as NTNCWSs: (as
designated by the State): designated by designated by the
Four consecutive quarters State): Two gross State): Two gross
for gross beta, tritium and beta screening beta 14 screening
strontium-90 at levels were levels are being
representative point(s) discussed in the considered. Using a
within the distribution 1991 Proposal. screen of 50 or 30
system. Determine major Using a screen of pCi/L, quarterly
constituents if exceed 30 pCi/L, quarterly monitoring for
screen of 50pCi/L Systems monitoring for gross beta is
using water contaminated gross beta is required, along
with effluent from nuclear required, along with annual
facilities: Quarterly with annual tritium monitoring for
monitoring 1 for gross beta monitoring. Using a tritium and
and iodine-131, strontium- screen of 50 pCi/L, strontium-90 as in
90 and tritium. If gross quarterly 1976. Vulnerability
beta level is above 15 pCi/ monitoring for based on proximity
L, the same or equivalent gross beta is (15 miles ) to
samples must be analyzed required, along source per 1991.
for strontium-89 and cesium- with annual tritium Screen of 15 for
134. and strontium-90 contaminated waters
monitoring. as in 1976.\3\
------------------------------------------------------------------------
------------------------------------------------------------------------
Note: \1\ This assumes that monitoring will be required at NTNC systems.
If this is not the case, these requirements would not apply to NTNC
systems.
\2\ A gross alpha measurement can be substituted for radium-226 and/or
uranium monitoring if the gross alpha measurement is below the
applicable MCL(s).
\3\ Quarterly monitoring for gross beta would be based on the analysis
of monthly samples or the analysis of a composite of three monthly
samples. For iodine 131, a composite of five consecutive daily samples
shall be analyzed once per quarter. Additional monitoring may be
required to identify specific isotopes if gross beta measurement
exceeds the screening level.
Table III-6.--Reduced Monitoring Requirements
------------------------------------------------------------------------
1976 1991 proposal 2000 NODA
------------------------------------------------------------------------
GROSS ALPHA
------------------------------------------------------------------------
CWSs: One sample every four CWSs and NTNCWSs: CWSs and NTNCWSs:
years if annual average One sample every One sample every
from previous results (four three years, if three years if
consecutive quarterly previous monitoring previous monitoring
samples) is less than \1/2\ results (from three results (``previous
MCL. years of annual results'') are
monitoring) are reliably and
below MCL. If consistently at or
system is reliably below MCL; one
and consistently sample every six
below MCL, the years if previous
system could results are
receive a waiver, reliably and
and monitor once consistently at or
every nine years. below \1/2\ MCL; or
one sample every
nine years if
previous results
are reliably and
consistently at or
below the MDL.
------------------------------------------------------------------------
[[Page 21597]]
RADIUM
------------------------------------------------------------------------
CWSs: One sample every four CWSs and NTNCWSs CWSs and NTNCWSs:
years if annual average using ground water: One sample every
from previous results (four One sample every three years if
consecutive quarterly three years, if previous results
samples) is less than \1/2\ previous monitoring are reliably and
MCL. results (from three consistently at or
years of annual below MCL; one
monitoring) are sample every six
below MCL. If years if previous
system is reliably results are
and consistently reliably and
below MCL, the consistently at or
system could below \1/2\ MCL; or
receive a waiver, one sample every
and monitor once nine years if
every nine years. previous results
are reliably and
consistently at or
below the MDL.
------------------------------------------------------------------------
URANIUM
------------------------------------------------------------------------
None........................ CWSs and NTNCWSs: CWSs and NTNCWSs:
One sample every One sample every
three years, if three years if
previous monitoring previous results
results (from three average below MCL;
years of annual one sample every
monitoring) are six years if
below MCL. If previous results
system is reliably average at or below
and consistently \1/2\ MCL; or one
below MCL, the sample every nine
system could years if previous
receive a waiver, results average
and monitor once below the MDL.
every nine years.
------------------------------------------------------------------------
BETA AND PHOTON EMITTERS
------------------------------------------------------------------------
CWSs serving > 100,000 Vulnerable systems Vulnerable systems
persons and using surface only (as designated only (as designated
water (and other systems by State): Since by the State):
designated by the State): only vulnerable Since only
Every four years, systems systems are vulnerable systems
must collect samples from required to are required to
four consecutive quarters monitor, no reduced monitor, no reduced
for gross beta at monitoring is monitoring is
representative point(s) allowed. allowed.
within the distribution
system. Systems using water
contaminated with effluent
from nuclear facilities: No
reduced monitoring is
allowed.
------------------------------------------------------------------------
------------------------------------------------------------------------
15. Laboratory Capacity Issue `` Possible Extension of Initial
Monitoring Period
As discussed earlier in the analytical methods section (III.J), the
Performance Evaluation Program (now known as the Proficiency Testing
Program) has been externalized. Although the Agency is unsure at this
time how externalization may affect laboratory capacity, EPA recognizes
that it may be an implementation issue for at least three reasons:
The recent externalization of the radionuclides
Performance Evaluation (PE) studies program may cause short-term
disruption in laboratory accreditation;
Requiring NTNCWSs to monitor under the Standard Monitoring
Framework will add approximately 20,000 systems to the universe of
systems that are already required to monitor;
And the radon rule will be implemented simultaneously with
the radionuclides rule.
NIST is in the process of approving a provider for PT samples for
radionuclides. States also have the option of approving their own PT
sample providers. Should laboratory capacity issues related to
externalization present implementation problems for the initial
monitoring period (three years), EPA will consider allowing an
additional year (four years total) for the initial monitoring period.
During the specified time period, systems would be required to analyze
four consecutive quarterly samples to determine compliance. If the
final rule is promulgated in November of 2000, the new monitoring
requirements would begin to be enforced in November of 2003. If EPA
implements a one year extension, water systems would have until
December 31 of 2007 to complete the required initial monitoring. This
scenario would allow the ``one third of systems per year'' strategy
inherent in the Standard Monitoring Framework to be applied, while
allowing one additional year, if necessary, to address any laboratory
capacity issues. EPA solicits public comment on this matter.
L. Effective Dates
Much of the rule that will be finalized in November will involve
retaining current elements of the radionuclides NPDWR. Those portions
of the final rule that are unaffected by the upcoming regulatory
changes are already in effect. MCLs for gross alpha, beta particle and
photon radioactivity, and combined radium-226 and -228 will be
unchanged and are already in effect. Regarding water systems that are
currently out of compliance with the existing NPDWRs for gross alpha,
combined radium-226 and -228, and/or beta particle and photon
radioactivity, States with primacy and EPA will renegotiate enforcement
actions that put systems on compliance schedules as expeditiously as
possible.
Under the Safe Drinking Water Act, final rules become effective
three years after promulgation (November of 2003, assuming that the
rule becomes final in November of 2000). The following discussion
assumes a promulgation date of November 2000. For reasons described in
the monitoring section of the NODA (section III, part K) and the
Appendices (appendix V), initial monitoring will be required to
completed by December 31, 2007. Under the Standard Monitoring
Framework, systems have three years to complete the initial monitoring
cycle of four consecutive quarterly samples. However, for reasons
described in the monitoring section of the NODA (section III, part K) ,
systems will have an additional year to complete the initial monitoring
cycle, which will correspond to an end date of December
[[Page 21598]]
31, 2007. This includes initial monitoring for uranium, the new
monitoring requirements for radium-228, and new initial monitoring
under the requirements for entry points. Compliance determinations and
future monitoring cycle schedules are also discussed in the monitoring
sections cited. MCL violations resulting from the new requirement for
separate Ra-228 monitoring will be treated as ``new violations'' and
will be on the same schedule as other new violations (e.g. uranium).
M. Costs and Benefits
The Safe Drinking Water Act provides for EPA to consider both
public health and the feasibility (taking costs into consideration) in
establishing drinking water MCLs. In addition the new Amendments
require EPA to evaluate the costs and benefits of potential revisions
to the current standards. As noted earlier, the Agency conducted an
analysis of the costs and associated benefits of each of the options
described in today's document. These analyses were performed consistent
with the requirements for a Health Risk Reduction and Cost Analysis set
forth in the 1996 Amendments to the SDWA (section 1412(b)(3)(C)).
First, all public water systems that are currently treating and are
in compliance with the 1976 standards will have no additional cost if
the rule remains the same as it is now. At the same time, EPA
recognizes that it may be costly to systems which have delayed
compliance. However, to the extent the rule remains the same, costs
necessary to comply with the existing rule, as well as public health
benefits associated with it, have accrued to that 1976 rule. If EPA
changes nothing, the existing 1976 requirements must be met. EPA
considers only those costs associated with accommodating revisions to
the current regulations to be new costs. Costs incurred, or those that
should have been incurred to comply with a previous regulation, are not
factored into current considerations.
Second, EPA has reexamined the costs of the 1991 proposal regarding
monitoring for any changes which may be warranted based on new data.
EPA is contemplating several changes which were part of the 1991
proposed regulation and which may increase costs. These include: (1)
Promulgating an NPDWR for uranium; (2) applying the radionuclide NPDWRs
to non-transient, non-community (NTNC) systems; (3) requiring
monitoring at the point of entry to distribution systems, and ; (4)
requiring separate monitoring for radium-226 and radium-228.
EPA is also recommending rapid sample analysis for alpha emitters
to detect the presence of short lived radionuclides such as radium-224,
but is not contemplating requiring it as part of the revision to the
radionuclides rule. The Agency will pursue the issue of a timely
analysis of gross alpha to reflect short half lived Ra-224 in a
separate proposal.
Costs and benefits for the various options are presented in
appendix V of today's document, in the Technical Support Document (EPA
2000a), and in the draft Health Risk Reduction and Cost Analysis (EPA
2000b). Today's NODA solicits comment on whether the incremental risk
reduction may justify the costs for certain of the revisions described
in the NODA. EPA requests public comment on such questions and on the
extent to which its discretionary authority provided by section
1412(b)(6) of the SDWA should be used. This NODA also requests public
input regarding the need for further adjustments to the limits based on
the cost and risk data presented in today's NODA.
IV. References
Parsa, Bahman. ``Contribution of Short-Lived Radionuclides to
Alpha-Particle Radioactivity in Drinking Water and Their Impact on
the Safe Drinking Water Act Regulations''. Radioactivity &
Radiochemistry (Vol. 9, No. 4). pp. 41-50). 1998.
USEPA. Drinking Water Regulations; Radionuclides. Federal
Register, Vol. 41, No. 133, p. 28402. July 9, 1976.
USEPA. National Primary Drinking Water Regulations;
Radionuclides; Proposed Rule. Federal Register, Vol. 56, No. 138, p.
33050. July 18, 1991.
USEPA. Presumptive Response Strategy and Ex-Situ Treatment
Technologies for Contaminated Ground Water at CERCLA Sites: Final
Guidance. EPA 540/R-96/023. (EPA 1996a)
USEPA. Performance Evaluation Studies Supporting Administration
of the Clean Water Act and the Safe Drinking Water Act. Federal
Register, Vol. 61, No. 139, p. 37464. July 18, 1996. (EPA 1996b)
USEPA. National Primary Drinking Water Regulations; Analytical
Methods for Radionuclides; Final Rule and Proposed Rule. Federal
Register, Vol. 62, No. 43, p. 10168. March 5, 1997. (EPA 1997a)
USEPA. Performance Evaluation Studies Supporting Administration
of the Clean Water Act and the Safe Drinking Water Act. Federal
Register, Vol. 62, No. 113, p. 32112. June 12, 1997. (EPA 1997b)
USEPA. Performance Based Measurement System. Federal Register,
Vol. 62, No. 193, p. 52098. October 6, 1997. (EPA 1997c)
USEPA. Manual for the Certification of Laboratories Analyzing
Drinking Water. EPA 815-B-97-001. 1997. (EPA 1997d)
USEPA. 1997. Memorandum from Administrator Carol M. Browner of
EPA to Chairperson Shirley A. Jackson of the Nuclear Regulatory
Commission. February 7, 1997. (EPA 1997e)
USEPA. Office of Solid Waste and Emergency Response (OSWER).
1997. Memorandum from Stephen Ludwig and Larry Weinstock (Office of
Radiation and Indoor Air) to Addressees. ``Establishment of Cleanup
Levels for CERCLA Sites with Radioactive Contamination''. OSWER No.
9200.4-18. August 22, 1997. (EPA/OSWER 1997a).
USEPA. Office of Solid Waste and Emergency Response (OSWER).
1997. Memorandum from Timothy Fields to Regional Administrators.
``The Role of CSGWPP's in EPA Remediation Programs''. OSWER No.
9283.1-09. April 4, 1997. (EPA/OSWER 1997b).
USEPA. Memorandum to Water Management Division Directors,
Regions I-X, from Cynthia C. Dougherty, Director, Office of Ground
Water and Drinking Water regarding Recommendations Concerning
Testing for Gross Alpha Emitters in DrinkingWater (January 27,
1999). (EPA 1999a)
USEPA. Cancer Risk Coefficients for Environmental Exposure to
Radionuclides, Federal Guidance Report No. 13. US Environmental
Protection Agency, Washington, DC, 1999. (EPA 1999b)
USEPA. ``Technical Support Document for the Radionuclides Notice
of Data Availability''. Draft. March, 2000. (EPA 2000a)
USEPA. ``Preliminary Health Risk Reduction and Cost Analysis:
Revised National Primary Drinking Water Standards for
Radionuclides''. Prepared by Industrial Economics, Inc. for EPA.
Draft. January 2000. (EPA 2000b)
U.S. Environmental Protection Agency. Memorandum to David Huber,
Edwin Thomas, OGWDW from Scott Telofski, ORIA regarding Uranium
Analysis Screening Level for Drinking Water Regulations NODA. March
14, 2000. (EPA 2000c)
U.S. Environmental Protection Agency. Memorandum to David Huber,
Edwin Thomas, OGWDW from Scott Telofski, ORIA regarding Gross Alpha
Screening for Uranium. March 20, 2000. (EPA 2000d)
U.S. Geological Survey (USGS). ``Radium-226 and Radium-228 in a
Shallow Ground Water, Southern New Jersey''. Fact Sheet FS-062-98.
June 1998.
Appendix I--Occurrence
In order to estimate the total national costs and benefits of
revising the MCLs it is necessary to develop updated national
estimates of the occurrence and exposure to these radionuclide
contaminants in drinking water. Occurrence data and associated
analyses provide indications of the number of public water supply
systems with concentration of radionuclides above the revised MCL as
well as the population served by these systems. Monitoring and
treatment costs can be estimated from the occurrence data.
A. Background
EPA conducted a nationwide occurrence study of naturally
occurring radionuclides in
[[Page 21599]]
public water supplies called the National Inorganic and
Radionuclides Survey (NIRS) (see EPA 1991, proposed rule). The
objective of NIRS was to characterize the occurrence of a variety of
constituents, including radium-226, radium-228, uranium (mass
analysis), gross alpha-particle activity, and gross beta-particle
activity, present in community ground-water supplies (finished
water) in the United States, and its territories. The survey
included a random sample from 990 collection sites. The public water
supplies were stratified into four size categories, and the samples
were chosen to best represent the same stratification present in the
total population of community water supply in existence at the time,
as shown in ------Table I-1.
Table I-1.--Comparison of NIRS Target Sample With Federal Reporting Data System (FRDS) Inventory
----------------------------------------------------------------------------------------------------------------
Number of FRDS Percentage of Number of NIRS Percentage of
Population category (population range) sites* FRDS sites sites NIRS sites
----------------------------------------------------------------------------------------------------------------
Very small (25-500)......................... 34,040 71.4 716 71.6
Small (501-3,300)........................... 10,155 21.3 211 21.1
Medium (3,301-10,000)....................... 2,278 4.8 47 4.7
Large and very large (10,001->100,000)...... 1,227 2.6 26 2.5
-------------------------------------------------------------------
Total................................. 47,700 100.1 1,000 100.0
----------------------------------------------------------------------------------------------------------------
* Based in FRDS inventory for fiscal year 1985 from Longtin, 1988.
Results of NIRS were used to develop the proposed radionuclide
rule in 1991 (56 FR 33050; EPA 1991). There has not been a
comparable national survey for radionuclides since. Since the
publication of the proposed 1991 revision to the MCLs, the United
States Geological Survey has collected additional data on various
radionuclides in groundwater to augment the data of the NIRS. These
studies are summarized subsequently, and in greater detail in the
Technical Support Document (EPA 2000a).
Szabo and Zapecza (1991) detail the differences in the
occurrence of uranium and radium-226 in oxygen-rich and oxygen-poor
areas of aquifers. Because the chemical behavior of uranium and
radium are vastly different, the degree of mobilization of the
parent and product are different in most chemical environments.
Recently, high concentrations of radium were found to be
associated with ground water that was geochemically affected by
agricultural practices in the recharge areas by strongly enriching
the water with competing ions such as hydrogen, calcium, and
magnesium (Szabo and dePaul, 1998). Radium-228 was detected in about
equivalent concentrations as radium-226 in the aquifer study in New
Jersey (Szabo and dePaul, 1998).
B. USGS Radium Survey
A 1998 USGS survey (see EPA 2000a) was designed to target areas
of known, or suspected, high concentrations of radium-224 as
inferred by associated radium occurrence data, geologic maps, and
other geochemical considerations. Thus, the survey is likely biased
toward the extreme high end of the occurrence distribution for
radium-224 and co-occurring contaminants such as radium-228.
Approximately half of the samples were below the minimum detectable
concentration of radium-226 and radium-228 in spite of the fact that
public water systems were targeted in areas where high
concentrations of radium were expected. Table I-2 shows that, of the
104 samples, 21 exceeded the MCL for combined radium, and about 5
percent exceeded 10 pCi/L of radium-224, though several of these
samples with pH less than 4.0 also contained detectable
concentrations of thorium isotopes as well. Concentrations exceeded
1 pCi/L in about 10 percent of the samples analyzed for lead-210 and
3 percent for polonium-210.
Table I-2.--Percent of Samples Exceeding Specified Concentration
----------------------------------------------------------------------------------------------------------------
Total Percent of samples exceeding given concentration (pCi/L)
Radionuclide number of -----------------------------------------------------------
samples 1 2 3 5 7 10
----------------------------------------------------------------------------------------------------------------
Ra-224................................. 104 30 26 20 15 9 5
Ra-226................................. 104 33 22 17 10 5 2
Po-210................................. 95 3 1 1 1 0 0
Pb-210................................. 96 10 3 1 1 0 0
----------------------------------------------------------------------------------------------------------------
Radium-224 occurs in many of the wells sampled at concentrations
that highlight the limitations of the present monitoring scheme for
the gross alpha-particle standard. In addition, the contribution of
radium-224 and its short-lived daughter products to gross alpha
emissions was estimated with data from a concurrent study of ground-
water supplies by the USGS in cooperation with the state of New
Jersey (Szabo et al., 1998). In that study, gross alpha emissions
were measured before the decay of radium-224 and after sufficient
time had elapsed for radium-224 decay (about 18-22 days). In this
way, the difference between the initial gross-alpha measurement and
the final measurement is indicative of the contribution of radium-
224 and all other alpha emitting isotopes that would decay within
this time frame. The results indicate that the contribution of
radium-224 and its short-lived daughter products is approximately
three times the concentration of radium-224. While this analysis was
developed with a small data set in a restricted geographic range, it
is based on a physical process and has important implications for
such things as projections of radium-224 occurrence in association
with gross-alpha concentrations. These results are also important in
light of both the costliness and difficulty of the radium-224
analysis.
Concentrations of radium-228 were highly correlated with radium-
224. Although this correlation was based on a limited number of data
points, there is a physical basis to the correlation since both
nuclides originate from the same decay chain. Therefore, there is
potential for using radium-228 as a proxy indicator for the much
shorter lived and infrequently sampled radium-224. In addition, the
isotopic ratios of radium-226 to radium-228 were below 3:2 in many
samples indicating that the gross alpha-particle screen that is
currently used for combined radium (radium-226 + radium-228)
compliance would be inadequate in many situations.
Polonium-210 and lead-210 are derived from the uranium-238 decay
series; the decay series that produces radium-226. However, the
survey was designed to assess radium-224; therefore results are
possibly biased to areas that would more likely have isotopes in the
thorium-232 decay series. In addition, the correlations of radium-
226 with radium-224 and radium-228 are only 0.51 and 0.61
respectively; consequently, the wells that were sampled may not be
located in areas expected to have polonium-210 or lead-210. Within
these constraints, the new data help to fill the gap in occurrence
information that existed for these isotopes. Polonium-210 was found
in concentrations exceeding 1 pCi/L in only two wells. At this time,
these
[[Page 21600]]
observations could not be associated with unique geochemical
controls (as has been accomplished in a previous study in Florida;
Harada et al., 1989) and further investigations would be necessary
to infer anything more about the national distribution and
occurrence of polonium-210.
Approximately 12 percent of the samples exceeded a lead-210
concentration of 1 pCi/L; however only one sample was greater than 3
pCi/L. The greatest frequency of detection was in the Appalachian
Physiographic Province of the northeastern United States, especially
in of Connecticut and Pennsylvania. The geochemical mechanism that
controls lead-210 dissolution is also not well established and needs
further study, though lead is less soluble than radium. In addition,
lead-210, like polonium-210, is derived from a different decay chain
than radium-224 and it was therefore not considered in designing the
study. One possible explanation for the frequent detection of lead-
210 in concentrations greater than 1 pCi/L in the Appalachian region
may be the high concentrations of radon-222 in ground water in this
region (Zapecza and Szabo, 1986). As the radon in solution decays
through a series of very short half-lived products to Lead-210, a
small fraction of the lead-210 may not be sorbed onto the aquifer
matrix; thus, the higher the initial radon-222 concentration, the
more likely measurable amounts of lead-210 would be found in the
ground water. This hypothesis could not be tested however because
radon-222 was not analyzed in this study.
C. USGS Beta/Photon Data Collection Effort
The major source of data for man-made radionuclides is the
Environmental Radiation Ambient Monitoring System (ERAMS) which is
published quarterly in the Environmental Radiation Data (ERD)
reports. The ERD reports provide concentration data on gross beta-
particle activity, tritium, strontium-90, and iodine-131 for 78
surface-water sites that are either near major population centers or
near selected nuclear facility environs.
An additional data collection effort was completed by the U.S.
Geological Survey in the summer of 1999 (see EPA 2000a) to analyze
targeted beta-particle emitting radionuclides from a small number of
public water systems that had shown relatively high levels of beta/
photon emitters during the original NIRS survey. Of the 26 public
water systems contacted for this effort none could ascertain which
wells in their systems were originally sampled as part of NIRS.
Consequently, although all efforts were made to include as many of
the original systems as possible, it is presently unknown if the
wells sampled match those in NIRS. The radionuclide analyses for
this data collection effort included; short-term (48 hour) gross
beta-particle and gross alpha-particle activities, long-term (30
days) gross beta-particle and gross alpha-particle activities,
tritium, strontium-89, strontium-90, cesium-134, cesium-137, iodine-
131, uranium-234, uranium-235, uranium-238, radium-228, radium-226,
lead-210, and cobalt-60.
Gross beta-particle activities were all below 50 pCi/L in water
collected from public water systems that were sampled previously
during the National Inorganics and Radionuclide Survey (NIRS) and
had been found to contain gross beta-particle activity in excess of
20 pCi/L. To the extent possible, all samples were collected from
the original public water systems surveyed for NIRS where gross
beta-particle activities were 20 pCi/L or greater. However due to
the amount of time that had elapsed since the NIRS samples were
collected, correlation with the original sampling point could not be
verified for every water supply sampled.
Though the number of samples was limited (26 samples), a few
conclusions can be reached. Concentrations of gross beta-particle
activities will rarely exceed 50 pCi/L in water collected from
public water systems (and did not do so in this study). A
significant percentage (15% or 4 samples) of the 26 samples
analyzed, however, contained gross alpha-particle activities at or
in excess of the 15 pCi/L MCL indicating that concern over the
presence of elevated concentration of gross alpha-particle activity
in ground water is justified. Long-term (30-day) gross beta-particle
activity analyses did not indicate significant ingrowth of beta-
particles in any of the samples, though this result is qualified by
the absence of significant quantities of uranium-238 in any of the
samples collected. Naturally occurring potassium-40 and radium-228
are a significant source of gross beta-particle activity to many of
the samples in agreement with results of Welch et al., 1995., Minor
concentrations of naturally-occurring lead-210 are also detected
occasionally. No manmade radionuclide was detected in concentration
above the maximum detectible concentration (MDC) in any of the
samples. The presence of naturally occurring beta-particle emitting
radionuclides must be taken into account when evaluating the source
of high gross beta-particle activity in ground water as first
suggested by Welch et al., 1995.
D. References
Harada, Kow, William C. Burnett, Paul A. LaRock, and James B.
Cowart, 1989. Polonium in Florida groundwater and its possible
relationship to the sulfur cycle and bacteria. Geochemical et
Cosmochimica Acta Vol. 53, pp. 143-150.
Longtin, Jon, 1988. Occurrence of Radon, Radium and Uranium in
Groundwater. Journal American Water Works Association, pp. 84-93.
Szabo, Z., and V.T dePaul, 1998. Radium-228 and radium-228 in
shallow ground water, southern New Jersey. U.S. Geological Survey
Fact Sheet FS-062-98.
Szabo, Z., V.T. dePaul, and B. Parsa, 1998. Decrease in gross
alpha-particle activity in water samples with time after collection
from the Kirkwood Cohansy aquifer system in southern New Jersey:
Implications for regulations. Drinking Water 63rd annual meeting
American Water Works Association New Jersey Section. Atlantic City,
NJ.
Szabo, Z., and O.S. Zapecza, 1991, Geologic and geochemical
factors controlling uranium, radium-226, and radon-222 in ground
water, Newark Basin, New Jersey: Gundersen, L.C.S. and Wanty, R.B.,
eds., Field studies of radon in rocks, soils, and water, U.S.
Geological Survey Bulleting 1971, p. 243-266.
USEPA. National Primary Drinking Water Regulations;
Radionuclides; Proposed Rule. Federal Register. Vol. 56, No. 138, p.
33050. July 18, 1991.
USEPA. ``Technical Support Document for the Radionuclides Notice
of Data Availability''. Draft. March, 2000. (EPA 2000a)
Welch, A.H., Szabo, Z., Parkhurst, D.L., Van Meter. P.C., and
Mullin, A.H., 1995, Gross-beta activity in ground water: natural
sources and artifacts of sampling and laboratory analysis: Applied
Geochemistry, v. 10, no. 5, p. 491-504.
Zapecza, O. S., and Z. Szabo, 1986. Natural radioactivity in
ground water--a review. U.S. Geological Survey National Water
Summary 1986, Ground-Water Quality: Hydrologic Conditions and
Events, U.S. Geological Survey Water Supply Paper 2325. pp. 50-57.
Appendix II--Health Effects
The following information summarizes the salient changes in risk
assessment information and risk characterization methodology during
the past two decades. The Technical Support Document (EPA 2000a)
also provides additional information.
A. Use of Linear Non-Threshold Assumption
In estimating the health effects from radionuclides in drinking
water, EPA subscribes to the linear, non-threshold model which
assumes that any exposure to ionizing radiation has a potential to
produce deleterious effects on human health, and that the magnitude
of the effects are directly proportional to the exposure level. The
Agency further believes that the extent of such harm can be
estimated by extrapolating effects on human health that have been
observed at higher doses and dose rates to those likely to be
encountered from environmental sources of radiation. The risks
associated with radiation exposure are extrapolated from a large
base of human data. EPA recognizes the inherent uncertainties that
exist in estimating health impact at the low levels of exposure and
exposure rates expected to be present in the environment. EPA also
recognizes that, at these levels, the actual health impact from
ingested radionuclides will be difficult, if not impossible, to
distinguish from natural disease incidences, even using very large
epidemiological studies employing sophisticated statistical
analyses. However, in the absence of other data, the Agency
continues to support the use of the linear, non-threshold model in
assessing risks associated with all carcinogens.
B. Continuous Improvements in Models, Data Base
As various scientific institutions have continued to collect
data on the observed effects of radiation from the cohort of bomb
survivors, patients with medical exposure, and workers with
occupational exposure; continuous improvements have been possible in
models to extrapolate effects and to estimate the risks of small
exposures to radiation from the natural environment or man-made
sources. The data have led to
[[Page 21601]]
changes in risk estimates as summarized here.
1. Basis of 1976 Estimates of Risk
Risk of bone cancer from radium dial painters.
Autopsy radioassay (see EPA 2000a). Body burden from
natural intake or radium, about 1 pCi/day.
Estimate annual dose rate in several organs from
natural radium in rad/year.
BEIR I risk numbers for radium dial painters yields
risk/year per rad/year.
Calculate risk over lifetime.
a. 1976 Estimates of the Risks from Radium-226 and Radium-228.
In general, EPA followed the Federal Radiation Council (FRC)
recommendation that radium ingestion limits for the general
population should be based on environmental studies and not the
models used to establish occupational dose limits (see EPA 2000a).
In setting the MCL, EPA considered bone cancer and other soft tissue
cancers to be the principal health effects associated with radium
ingestion. To calculate body burdens, doses, and risks from
ingestion of radium-226 and radium-228, in 1976, EPA relied on data
from the 1972 report of the United Nations Scientific Committee on
the Effects of Atomic Radiation (see EPA 2000a) and the 1972 the
National Academy of Sciences (NAS) Committee on the Biological
Effects of Ionizing Radiation, BEIR I Report (see EPA 1991, proposed
rule). Additional information and support were found in the
International Commission on Radiological Protection, Publication 20
(see EPA 2000a). The literature suggests that radium-228 was as
toxic as radium-226, and possibly twice as toxic for bone cancers in
dogs. Given this, EPA believed that it was prudent to assume that
the adverse health effects due to chronically ingested radium-228
were at least as great as those from radium-226.
Assuming equal toxicity with radium-226, EPA reasoned that
lifetime ingestion of only radium-228 at 5 pCi/L would yield
lifetime total cancer risks equal to those for a lifetime ingestion
of only radium-226 at the same concentration, i.e., between 0.5 to 2
x 10-4. By setting the MCL at 5 pCi/L for radium-226
and radium-228 combined, rather than individually, EPA sought to
limit the lifetime total cancer risk from the ingestion of both
isotopes in drinking water to 2 x 10-4 or less.
b. Basis for the 1976 MCL for Gross Alpha Particle Activity. One
of the main intentions of the 15 pCi/L MCL for gross alpha particle
activity, which includes radium-226 but excludes uranium and radon,
was to limit the concentration of other naturally-occurring and man-
made alpha emitters relative to radium-226. Specifically, this limit
was based on the fact that EPA estimated that continuous consumption
of drinking water containing polonium-210, the next most radiotoxic
alpha particle emitter in the radium-226 decay chain, at a
concentration of 10 pCi/L might cause the total dose to bone to be
equivalent to less than 6 pCi/L of radium-226.
The 15 pCi/L limit, which includes radium-226 but excludes
uranium and radon, was based on the conservative assumption that if
the radium concentration is limited to 5 pCi/L and the balance of
the alpha particle activity (i.e., 10 pCi/L) is due to polonium-210,
the total dose to bone would be less than that dose associated with
an intake of 6 pCi/L of radium-226.
c. Basis for the 1976 MCL for Beta Particle and Photon
Radioactivity. In 1976, EPA estimated that continuous consumption of
drinking water containing beta and photon emitting radioactivity
yielding a 4 mrem/yr total body dose may cause an individual fatal
cancer risk of 0.8 x 10-6 per year, or a lifetime
cancer risk of 5.6 x 10-5, assuming a 70-year lifetime.
In setting the MCL for man-made beta and photon emitters, EPA used
cancer risk estimates from the BEIR I report for the U.S. population
in the year 1967 (see EPA 1991, proposed rule). For an exposed group
having the same age distribution as the U.S. 1967 population, the
BEIR I report indicated that the individual risk of a fatal cancer
from a lifetime total body dose rate of 4 mrem per year ranged from
about 0.4 to 2 x 10-6 per year depending on whether an
absolute or relative risk model was used. Using best estimates from
both models for fatal cancer, EPA believed that an individual risk
of 0.8 x 10-6 per year resulting from a 4 mrem annual
total body dose was a reasonable estimate of the annual risk from a
lifetime ingestion of drinking water. Over a 70-year period, the
corresponding lifetime fatal cancer risk would be 5.6 x
10-5, with the risk from the ingestion of water
containing less amounts of radioactivity being proportionately
smaller.
Based on 1967 U.S. Vital Statistics (see EPA 1991 and EPA
2000a), the probability that an individual would die of cancer was
about 0.19, and was thought to be increased by 0.1 percent from a
lifetime dose equivalent rate of 15 mrem per year. Therefore, EPA
calculated that the 4 mrem/yr MCL for man-made beta and photon
emitters corresponded to a lifetime risk increase of 0.025 percent
to exposed groups.
EPA knew that partial body irradiation was common for ingested
radionuclides since they are, like radium, largely deposited in a
particular organ, or in a few organs. In such cases, EPA
acknowledged that the risk per millirem varies depending on the
radiosensitivity of the organs at risk. For example, EPA estimated
that cancers due to the thyroid gland receiving 4 mrem per year
continuously ranged from about 0.2 to 0.5 per year per million
exposed persons (averaged over all age groups). Considering the sum
of the deposited fallout radioactivity and the additional amounts
due to releases from other sources existing at that time, EPA
believed that the total dose equivalent from man-made radioactivity
was not likely to result in a total body or organ dose to any
individual that exceeded 4 mrem/yr. Consequently, EPA did not
believe that the 4 mrem/yr standard would affect many public water
systems, if any. At the same time, the Agency believed that an MCL
set at this level would provide adequate public health protection.
2. 1991 Proposal: Basis of Health Risk Estimates
During the years since the publication of the 1976 regulations,
the Agency obtained a great deal of additional data and a better
understanding of the risks posed to human health by ingested
radionuclides. Many of these new studies were presented and
discussed in the Advance Notice of Proposed Rulemaking announcing
EPA's intent to revise the MCLs (51 FR 34836, Sept. 20, 1986) and
the supporting health criteria documents (see EPA 2000a and EPA
1991, the proposed rule).
Among the most important changes made by EPA in developing the
1991 revisions was the adoption of a common calculational framework,
the RADRISK computer code (see EPA 1991, proposed rule), to estimate
the risks posed by ingestion of radionuclides in drinking water. The
RADRISK code consisted of intake, metabolic, dosimetric, and risk
models that integrated the results of a large number of studies on a
variety of radioactive compounds and radiation exposure situations
into an overall model to estimate risks for many different
radionuclides. Radionuclide-specific parameters were based on the
results of individual scientific studies of a specific radionuclide,
such as radium; human epidemiological studies; or experimental
animal studies of groups of chemically-similar radionuclides. To
summarize, the following are some of the salient changes.
Used RADRISK metabolic model instead of natural uptake
equilibrium model. Based on known intakes.
Used ICRP report 20 (see EPA 2000a) on alkaline earth
elements with Oak Ridge modeled exponential fit to that model.
BEIR IV risks for alpha emitters.
Ra-224 data from ankylosing spondylitis, tuberculosis.
Change in results from Ra-228 calculations (Oak Ridge
model of '84) and ICRP 30 (see EPA 2000a) yielded different results
based on retention and distribution of each member of decay chain.
a. Basis for the 1991 MCL for Radium-226 and Radium-228. In
1991, EPA proposed revised MCLs for radium-226 and radium-228
individually at 20 pCi/L each. The Agency thought at that time that
the limit for each of these radium isotopes was within the Agency's
acceptable risk range of 10-6 to 10-4. The
Agency no longer believes the MCLs proposed in 1991 for radium-226
and radium-228 are within the Agency's acceptable risk range.
i. Human and Animal Health Effects Data Considered. In 1991, EPA
based its risk estimates for radium using information from two
epidemiological study groups. The first group consisted of radium
dial painters who had ingested considerable amounts of radium paint
(containing various proportions of radium-226 and radium-228) by
sharpening the point of their paint brush with the lips. The second
group consisted of patients in Europe injected with a short-lived
isotope of radium, radium-224, for treatment of spinal arthritis and
tuberculous infection of the bone (see EPA 2000a). The results of
these studies are described briefly next.
At high levels of exposure to radium, several non-cancer health
effects were observed in radium dial painters, such as benign bone
growths, osteoporosis, severe growth retardation, tooth breakage,
kidney disease, liver disease, tissue necrosis, cataracts, anemia,
immunological suppression and death (see EPA 2000a).
[[Page 21602]]
Exposed radium dial painters also exhibited significantly
elevated rates of two rare types of cancer, bone sarcomas
(osteosarcomas, fibrosarcomas and chondrosarcomas) and carcinomas of
head sinuses and mastoids (see EPA 2000a and EPA 1991, the proposed
rule). The incidence of head carcinomas was associated with exposure
to radium-226, but not radium-228 (see EPA 2000a). This is because
these latter cancers were due to an accumulation of radon gas
(radon-222) in the mastoid air cells and paranasal sinuses caused by
the escape of radon-222 into the air spaces.
ii. Body Burden, Dose, and Risk Calculations. Risk calculations
for ingested radium were made using RADRISK (see EPA 1991, proposed
rule) based on annual dose rates. For this purpose, EPA computed
dose rates for specific organs and tissues at specific ages for an
annual unit intake of each radium isotope (see EPA 2000a).
Calculation of body burdens was based on metabolic models derived
from the radium dial painter studies. Calculations of absorbed doses
in specific organs or tissues included cross irradiation from radium
in all other organs. RADRISK included lifetime cancer risk estimates
for high- and low-LET (linear energy transfer) radiation separately
for leukemia, osteosarcomas, sinus tumors, and other solid tumors.
These estimates were taken from the BEIR III and BEIR IV (see EPA
1991, proposed rule) reports.
Table II-1 compares the methods used by EPA in 1976 and 1991 to
calculate organ burdens, doses, and risks from radium ingestion.
Bone doses calculated for radium-226 in 1991 were about 33 percent
lower than those assumed in 1976, and the soft tissue doses were
about 40 percent lower. Risk estimates for bone per unit dose were
about 65 percent lower in 1991 than in 1976, and the soft tissue
risk estimates were about 9 percent lower.
Table II-1.--Comparison of Derivation of 1976 and 1991 MCLs for Radium
------------------------------------------------------------------------
Model 1976 1991
------------------------------------------------------------------------
Organ and Tissue Burdens.... Calculation of body Calculation of body
burdens based on burdens based on
environmental toxicokinetic
studies and ratio models derived from
of intakes. studies of patients
injected with
radium.
Dosimetry................... Calculation of Calculation of
absorbed dose based absorbed dose based
on organ and tissue on organ or tissue
burden. burden and cross
irradiation terms
from all other
organs.
Risk Coefficients........... Risk estimated using Risk estimated using
the geometric mean the absolute risk
of the absolute and coefficient from
relative risk the 1980 BEIR III
coefficients from report.
the 1972 BEIR I
report.
------------------------------------------------------------------------
b. Basis for the 1991 MCL for Gross Alpha Particle Activity. In
1991, EPA proposed to retain the 15 pCi/L MCL for gross alpha
particle activity, but modify it by excluding radium-226, as well as
uranium and radon. The exclusion of uranium and radon was based on
the fact that the Agency anticipated setting separate NPDWRs for
these contaminants with the finalization of the 1991 proposal. The
proposed exclusion of Ra-226 was based on the 1991 risk estimate
which suggested that its unit risk was small enough not to warrant
regulation within gross alpha. The 1991 limit was intended to limit
the lifetime cancer risk due to ingestion of naturally-occurring and
man-made alpha particle emitters in drinking water to between
10-6 and 10-4, the Agency's target risk range
for carcinogens. Specifically, this limit was based on the following
considerations:
Using RADRISK modeling, EPA estimated that continuous
consumption of 15 pCi/L of most alpha particle emitters in drinking
water at 2 L/day would pose a lifetime cancer risk between
10-6 and 10-4.
EPA performed the risk assessment for the alpha emitters using
RADRISK (EPA 1991, proposed rule). The model was used to estimate
radiation dose to organs, the dose was used to calculate risk to
organs, and the risks to organs were summed to estimate overall
risk. EPA used RADRISK to calculate concentrations of alpha emitters
corresponding to lifetime mortality and incidence risks of
10-4, assuming ingestion of two liters of drinking water
daily, and presented those values in appendix C of the 1991 proposed
rule.
In determining the risks from ingestion of alpha emitters in
drinking water, EPA was particularly interested in polonium-210 and
isotopes of thorium and plutonium, because these radionuclides had
been observed in water and may cause health effects at relatively
low concentrations.
However, the BEIR IV report concluded that there was 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, relied on RADRISK in assessing polonium risk. The model
estimated that continuous ingestion of two liters per day of
drinking water containing 14 pCi/L would pose a lifetime fatal
cancer risk of 1 x 10-4.
EPA also consulted the BEIR IV report for available information
on the adverse effects of thorium. Epidemiological studies of
patients injected with Thorotrast, a contrast agent consisting of
ThO2 and used in medical radiology from the 1920s to
1955, showed clear increases in liver cancer, as well as possible
increases in leukemia and other cancers. However, the BEIR IV report
discussed the limitations of these data for assessing the risk due
to other forms of thorium that might have different metabolic
behaviors and effects. Using RADRISK, EPA estimated that, at a
lifetime fatal cancer risk level of 1 x 10-4, derived
drinking water concentrations for thorium isotopes ranged from 50 to
125 pCi/L, and noted that thorium concentrations in drinking water
were generally near one pCi/L (EPA, 1991f).
EPA relied on the BEIR IV report for information on the health
effects of plutonium isotopes and other transuranic radionuclides
that were widely distributed in the environment in very low
concentrations due to atmospheric testing of nuclear weapons from
1945 to 1963. The BEIR IV report concluded that plutonium exposures
caused clear increases in cancers of the bone, liver, and lungs in
animals, but not in humans. At that time, the limited available
epidemiological studies had not demonstrated a clear association
between plutonium exposure and the development of cancer in human
exposure cases. The report recommended that assessing the risks of
plutonium exposure should be based on analogy with other
radionuclides and high-LET radiation exposure risks. Using RADRISK,
EPA estimated that, at a lifetime fatal cancer risk level of 1 x
10-4, derived drinking water concentrations for plutonium
isotopes ranged from about 7 to 68 pCi/L, and noted that plutonium
concentrations in drinking water were generally less than 0.1 pCi/L
(EPA, 1991f).
c. Basis for the 1991 MCL for Beta Particle and Photon
Radioactivity. In 1991, EPA proposed to alter the 4 mrem/yr MCL for
beta particle and photon radioactivity. The Agency modified the
standard by basing the limit on the committed effective dose
equivalent (EDE). (An effective dose equivalent approach adjusts the
dose that an individual organ may receive based on its
radiosensitivity. The less radiosensitive an organ is, the greater
the allowable radiation dose.) The MCL was also modified to include
naturally-occurring beta/photon emitters. The 1991 proposed standard
was intended to limit the lifetime cancer risk due to ingestion of
naturally-occurring and man-made beta particle and photon emitters
in drinking water to between 10-6 and 10-4,
the Agency's target risk range for carcinogens.
Using RADRISK modeling, EPA estimated that continuous
consumption of two liters per day of drinking water containing a
concentration of beta particle or photon emitting radiation
corresponding to 4 mrem EDE/yr would pose a lifetime cancer risk of
about 10-4.
Comparison of the 1976 Regulation and 1991 Proposed Regulation.
In 1976, EPA based the MCL for beta particle and photon emitters on
a target dose rate of 4 mrem/yr. The annual average activity
concentration of
[[Page 21603]]
individual radionuclides and mixtures of radionuclides resulting in
a 4 mrem/yr dose to the total body or any internal organ was then
calculated. This ``critical organ dose'' radiation protection
philosophy was based on the recommendations of ICRP Publication 2
(see EPA 2000a).
The Agency was aware that in 1976, when exposed to equal doses
of radiation, different organs and tissues in the human will exhibit
different cancer induction rates. Consequently, EPA knew that the
lifetime cancer risks for individual radionuclides would vary widely
(from near 10-7 to 5.6 x 10-5 because the
same dose equivalent would be applied to different critical organs,
resulting in different cancer risks. However, at that time, EPA did
not have an accepted method for equalizing risks. In addition, since
no dose could be greater than 4 mrem to every organ, the associated
risk was the ceiling for the risk of beta/photon emitters in
drinking water.
This was addressed in 1991 when EPA proposed to adopt the
effective dose equivalent, or EDE, radiation protection philosophy
recommended by ICRP (1977) (see EPA 1991, proposed rule). The
effective dose equivalent normalizes radiation doses and effects on
a whole body basis for regulation of occupational exposures. The EDE
is computed as the sum of the weighted organ-specific dose
equivalent values, using weighting factors specified by the ICRP
(1977, 1979; see EPA 1991, proposed rule). By changing to a limit of
4 mrem EDE/yr, EPA was able to derive activity concentrations for
individual beta/photon emitters that corresponded to a more uniform
level of risk. Using 4 mrem EDE and the metabolically-based dose
calculations, the derived concentrations for most beta particle and
photon emitters increased in 1991 as compared to the values
calculated in 1976 (shown in Table II-3). As a result of derived
concentrations increasing in 1991, the corresponding risks increased
as well. EPA estimated that, for most of these radionuclides, the
corresponding lifetime fatal cancer risk would be 1 x
10-4, about twice as high as the risk level estimated in
1976.
d. Basis for the 1991 Proposed MCL for Uranium. In 1991, EPA
proposed an MCL of 20 µg/L for uranium based on kidney
toxicity and a corresponding limit of 30 pCi/L based on cancer risk.
The MCLG was proposed at zero because of the carcinogenicity of
uranium, and the MCL was proposed at the most sensitive endpoint,
kidney toxicity. The MCL was based on kidney effects seen in the 30
day study in rats (see EPA 1991, proposed rule).
Using RADRISK modeling, EPA estimated that uranium in water
posed a cancer risk of 5.9 x 10-7 per picoCurie per
liter, assuming continuous intake of water of two liters per day.
Concentrations in water of 1.7 pCi/L, 17 pCi/L and 170 pCi/L
corresponded to lifetime mortality risks of approximately 1 x
10-6, 1 x 10-5 and 1 x 10-4,
respectively. A concentration of 30 pCi/L of uranium-238 was thought
to be equivalent to about 20 micrograms/L, the level considered to
be protective against kidney toxicity (the corresponding mortality
was 5 x 10-5.
In determining the MCL for uranium in 1991, EPA proposed to
regulate uranium at a level that would be protective of both kidney
toxicity, resulting from the element's chemical properties, and
carcinogenic potential due to radioactivity. The carcinogenic
effects of uranium were based on the effects of ionizing radiation
generally, the similarity of uranium to isotopes of radium, and on
the effects of high activity uranium.
C. Today's Methodology for Assessing Risks From Radionuclides in
Drinking Water
1. Background
Since 1991, EPA has refined the way in which it estimates
potential adverse health effects associated with ingestion of
radionuclides in drinking water. The Agency's new approach uses
state-of-the-art methods, models and data that are based on more
recent scientific knowledge. Compared with the approaches used in
1976 and 1991, the revised methodology includes several substantial
refinements. Specifically, the new risk-assessment methodology:
Accounts for age- and gender-specific water-consumption
rates and radionuclide intakes, and for physiological and anatomical
changes with age in quantifying costs and benefits;
Uses Blue Book (see EPA 2000a) for estimating
radiogenic risk: ICRP dosimetry model, 1990 vital statistics instead
of 1980;
Uses the most recent age-dependent biokinetic and
dosimetric models recommended by the ICRP; Federal Guidance Report-
13 dynamic input-output metabolic model;
Incorporates the latest information on radiogenic human
health effects summarized by the National Academy of Sciences and
other national and international radiation-protection advisory
committees;
Includes updated life tables based on data from the
National Center for Health Statistics that are used to adjust
radionuclide risk estimates for competing causes of death; and
Uses an improved computer program to handle the complex
calculations of radiation doses and risks.
Overall, EPA believes that these refinements significantly
strengthen the scientific and technical bases for estimating risks,
and consequently, for deriving MCLs for radionuclides. A brief
overview of this new methodology is summarized later in this
section. Interested individuals are referred to two EPA publications
Estimating Radiogenic Cancer Risks (EPA, 1994) and Federal Guidance
Report No. 13 (EPA, 1999) for detailed discussions on the revised
risk assessment methodology for radionuclides. Electronic copies of
both documents are available for downloading at EPA's web site
(http://www.epa.gov/radiation/rpdpubs.htm).
Federal Guidance Report No. 13: (EPA, 1999) presents the current
methods, models, and calculational framework EPA uses to estimate
the lifetime excess risk of cancer induction following intake or
external exposure to radionuclides in environmental media. The
report presents compilations of risk coefficients that may be used
to estimate excess cancer morbidity (cancer incidence) and mortality
(fatal cancer) risks resulting from exposure to radionuclides
through various pathways.
The risk coefficients for internal exposure represent the
incremental probability of radiogenic cancer morbidity or mortality
occurring per unit of radioactivity inhaled or ingested. For most
radionuclides, Federal Guidance Report No. 13 presents risk
coefficients for seven exposure pathways: inhalation, ingestion of
food, ingestion of tap water, ingestion of milk, external exposure
from submersion in air, external exposure from the ground surface,
and external exposure from soil contaminated to an infinite depth.
For some radionuclides, however, only external exposure pathways are
considered; these include noble gases and the short-lived decay
products of radionuclides addressed in the internal exposure
scenarios.
a. Radium. EPA set the current MCL of 5 pCi/L for radium-226 and
radium-228, combined, based on limiting the lifetime excess total
cancer risk to between 5 x 10-5 and 2 x 10-4.
In 1991, EPA proposed separate, and revised, MCLs for radium-226 and
radium-228 of 20 pCi/L for each. At that time, EPA believed that the
revised MCLs corresponded to lifetime excess fatal cancer risks of
1 x 10-4 each, or 2 x 10-4 combined, assuming
lifetime ingestion. The more sophisticated model used today
calculates a risk for Ra-228 at 5 pCi/l to be 2 x 10-4,
and the risk for 5 pCi/l of Ra-226 to be about
7.3 x 10-5. Retaining a combined MCL at 5 pCi/L would
produce the following risks shown in Table II-2.
Table II-2.--Mortality Risk of Radiums for Concentration Combinations at the MCL
--------------------------------------------------------------------------------------------------------------------------------------------------------
Radium-226 Radium-228 Ra-226 + Ra-228
--------------------------------------------------------------------------------------------------------------------------------------------------------
pCi/L Risk pCi/L Risk pCi/L Risk at 5 pCi/L
--------------------------------------------------------------------------------------------------------------------------------------------------------
0................................. 0 5 2.0 x 10-4 5 2.0 x 10-4
1................................. 1.5 x 10-5 4 1.6 x 10-4 5 1.8 x 10-4
2................................. 2.9 x 10-5 3 1.2 x 10-4 5 1.5 x 10-4
3................................. 4.4 x 10-5 2 8.1 x 10-5 5 1.3 x 10-4
4................................. 5.8 x 10-5 1 4.1 x 10-5 5 9.9 x 10-5
[[Page 21604]]
5................................. 7.3 x 10-5 0 0 5 7.3 x 10-5
--------------------------------------------------------------------------------------------------------------------------------------------------------
b. Alpha Emitters. Both the current and 1991 proposed MCLs for
alpha-emitting radionuclides permit up to 15 pCi/L of alpha particle
radioactivity in drinking water from individual and multiple alpha
emitters. EPA established the current gross alpha MCL of 15 pCi/L
(including radium-226 and excluding radon and uranium) to account
for the risk from radium-226 at 5 pCi/L (the radium regulatory
limit) plus the risk from polonium-210, which the Agency believed
was the next most radiotoxic element in the uranium decay chain. The
current risk estimated (FGR-13) indicates that the unit risk for Ra-
226 is large enough to warrant its inclusion in gross alpha, as
thought in 1976.
In 1991, EPA thought that exposure to 10 pCi/L of polonium-210
posed a lifetime fatal cancer risk comparable to that from
continuous lifetime ingestion of about 1 pCi/L of radium-226, that
is, between 0.5 and 2 x 10-4. In 1991, EPA based the
revised, adjusted gross alpha MCL on revised dose and risk
calculations which indicated that the 15 pCi/L limit posed a
lifetime cancer risk for most alpha emitters that fell within EPA's
acceptable risk range of between 10-6 and
10-4.
The current estimate of risk from polonium-210 at 7.0 pCi/L is
1 x 10-4. The risk for radium-226 at 6.8 p/L is also
1 x 10-4. When the current rule was written, 10 pCi/L of
polonium-210 was believed to be equivalent to 1 pCi/L of radium-226;
however, the risks are now equivalent. Thus polonium is ten times
the risk it was thought to be relative to radium-226. Retaining a 15
pCi/L standard including radium-226 ensures that the risk of 15 pCi/
L will not increase by allowing greater polonium (up to 15 pCi/L) in
addition to the radium-226 in the radium standard. As expected, a
uniform picoCurie limit results in widely differing risks (EPA
2000a).
c. Beta/Photon Emitters. As discussed elsewhere in this
document, EPA is able to calculate the risks from individual beta/
photon emitters using the FGR-13 methodology. It is now possible to
calculate a risk equivalent to the current picoCurie limit for each
beta/photon emitter. Appropriate adjustments are then possible in
keeping with the original risk maximum of 5.6 x 10-5. The
derived concentration values for the beta particle and photon
emitters from 1976 rule and 1991 proposal in comparison to today's
newest risk model using 5.6 x 10-5 mortality are found in
Table II-3.
BILLING CODE 6560-50-U
[[Page 21605]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.002
[[Page 21606]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.003
[[Page 21607]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.004
[[Page 21608]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.005
[[Page 21609]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.006
[[Page 21610]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.007
[[Page 21611]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.008
[[Page 21612]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.009
[[Page 21613]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.010
[[Page 21614]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.011
BILLING CODE 6560-50-C
[[Page 21615]]
d. Uranium. Since the 1991 proposal, a number of new studies
have been published in peer-reviewed journals. A literature search
was conducted and covered the time period between January 1991 to
July, 1998. Databases searched were TOXLINE, MEDLINE, EMBASE,
BIOSIS, TSCATS and Current Contents (see EPA 2000a). The results of
the literature search were reviewed and articles were identified,
retrieved and reviewed and analyzed. Subsequently, the Toxicological
Profile for URANIUM (Update) was published extending the database to
September 1999 (see EPA 2000a).
i. Health Effects in Animals. The potential toxic effects of
uranium following oral exposures have been evaluated in recent
animal studies (see EPA 2000a). In a 28-day range-finding study,
male and female Sprague-Dawley rats (15/sex/group) were administered
concentrations of 0, 0.96, 4.8, 24, 124, or 600 mg uranyl nitrate/L
(UN/L) in drinking water for a period of 28 days. Results of the
study showed no significant dose-related effects on body weight
gain, food intake, fluid consumption, clinical signs, or
hematological parameters of treated animals when compared with
control animals. Histologic examinations indicated no statistically
significant differences in the incidence of a particular lesion in
animals in the 600 mg UN/L treatment group when compared with
animals in the control group. However, a slight increase in the
number of affected animals in the 600 mg UN/L group was observed,
when compared with the control group.
As discussed in the Technical Support Document (EPA 2000a), the
long-term effects of exposure to low-levels of uranium in drinking
water has been demonstrated. Female rabbits and male albino rats
were exposed to 0, 0.02, 0.2, and 1 mg/kg uranyl nitrate for 12
months or 0.05, 0.6, 6, and 60 mg/L uranyl nitrate for 11 months,
respectively. Results of the study indicated a decrease in acid
phosphatase activity in the spleens of rabbits in the 1 mg/kg group,
but not in rats, when compared to controls. A statistically-
significant (p0.05) increase in serum alkaline phosphatase activity
was observed by the eleventh month of exposure in rats in the 6 and
60 mg/L groups, when compared with controls. A statistically-
significant decrease in the content of nucleic acids in the renal
and hepatic tissues was observed in rats in the 60 mg/L group and in
rabbits in the 1 mg/kg group, when compared with controls.
ii. Health Effects in Humans. Recent epidemiological studies
have evaluated the effects observed in humans exposed to uranium in
the drinking water (see EPA 2000a). These studies demonstrate the
relationship between uranium levels in the drinking water and urine
albumin, an indicator of renal dysfunction, was evaluated. Three
sites were selected for the controls (site 1) and the exposed groups
(sites 2 and 3), with mean uranium water levels of 0.71, 19.6 and
14.7 µg/L reported for sites 1, 2 and 3, respectively. An
index of uranium exposure was estimated for each study participant
by multiplying the uranium concentration in the water supply by the
average number of cups consumed at each residence and the total
number of years at that residence. Based on the results of a linear
regression analysis, which included terms for age, diabetes, sex,
smoking, and the use of water filters and softeners, a
statistically-significant association was reported for cumulative
exposure to uranium and urine albumin levels. However, the authors
noted that for most of the study participants, the urine albumin
levels were within the range of normal values.
A recent study of a village in Nova Scotia (see EPA 2000a)
demonstrated the renal effects following chronic exposure to uranium
in the drinking water. Two groups were evaluated, a low exposure
group (uranium levels 1/L) and a high exposure group (uranium
levels > 1µg/L). Twenty-four hour and 8-hour urine samples
were collected and evaluated for uranium, creatinine, glucose,
protein, b2-microglobulin (BMG), alkaline phosphatase
(ALP), gamma glutamyl transferase (GGT), lactate dehydrogenase
(LDH), and N-acetyl-b-D-glucosaminidase (NAG). Statistically
significant positive correlations were reported with uranium intake
for glucose (males, females and pooled data), ALP (pooled data) and
BMG (pooled data). No other statistically significant differences
were reported. Based on these results, the authors concluded that
the proximal tubule was the site of uranium nephrotoxicity.
In June 1998, a workshop was held by the USEPA to discuss issues
associated with assessing the risk associated with uranium exposure
and updating the RfD and MCLG for uranium. The numerous technical
issues associated with the development of a risk assessment for
uranium in drinking water were discussed. Based on these
discussions, it was apparent that there is a range of values for
each factor used in the development of the RfD and MCL for uranium.
However, based upon the input received at the workshop and the most
current information, EPA believes that the LOAEL for renal effects
in male rats of 0.06 mg U/kg/day reported could be used for the
development of an RfD for uranium (see EPA 2000a). The relative
source contribution (RSC) was revised to 80 percent (0.8). The total
uncertainty factor was determined to be about 100 (about 3 for
animal to human extrapolation, about 10 for intraspecies
differences, about 1 for a less than lifetime study, and about 3 for
the use of a LOAEL), with the body weight of 70 kilograms (kg) and
daily water consumption of two liters used in the calculation. These
assumptions are consistent with the data presented at the workshop
and appear to be reasonable and justifiable. EPA believes these
factors allow for the calculation of a safe level of uranium in
drinking water (in terms of chemical toxicity).
The application of the total uncertainty factor of 100 to the
LOAEL of 0.06 mg/kg/day results in an RfD of 0.6 ug uranium/kg/day.
The RfD can be used to determine the MCL by multiplying the RfD by
body weight (70 kg) and RSC (0.8) and dividing by water consumption
(2 L), resulting in a value of 17 ug uranium/L, which can be rounded
off to 20 /L.
2. Consideration of Sensitive Sub-populations: Children's Environmental
Health
In compliance with Executive Order 13045 ``Protection of
Children from Environmental Health Risks and Safety Risks'' (62 FR
19885, April 23, 1997), risks to children from radionuclides have
been considered. There is evidence that children are more sensitive
to radiation than adults, the risk per unit exposure in children
being greater than in adults.
Risk coefficients used by the Agency for radiation risk
assessment explicitly account for these factors. The age-specific,
organ-specific risk per unit dose coefficients used in the lifetime
risk model apply the appropriate age-specific sensitivities
throughout the model. The model also includes age-specific changes
in organ mass and metabolism. The risk estimate at any age is the
best estimate for that age. In developing the lifetime risks, the
model uses the life table for a stationary population. Use of the
life table allows the model to account for competing causes of death
and age-specific survival. These adjustments make the lifetime risk
estimate more realistic.
At the same time, consumption rates of food, water and air are
different between adults and children. The lifetime risk estimates
for radionuclides in water use age-specific water intake rates
derived from average national consumption rates when calculating the
risk per unit intake. Since the intake by children is usually less
than the intake by adults, it tends to partially mitigate the
greater risk in children compared to adults when evaluating lifetime
risk.
D. References
EPA, 1999. Cancer Risk Coefficients for Environmental Exposure
to Radionuclides, Federal Guidance Report No. 13. US Environmental
Protection Agency, Washington, DC, 1999. Uranium Issues Workshop--
Sponsored by United States Environmental Protection Agency,
Washington, DC ; June 23-24, 1998.
USEPA. ``Technical Support Document for the Radionuclides Notice
of Data Availability.'' Draft. March, 2000. (EPA 2000a)
Appendix III--Analytical Methods
Table III-1 briefly summarizes the regulatory events associated
with:
The testing procedures for regulated radionuclides
approved in 1976;
Major analytical additions or changes proposed or
discussed in the 1991 radionuclides rule;
Testing procedures and protocols approved in the March
5, 1997-- radionuclides methods rule (62 FR 10168, cited in 40 CFR
141.25); and
Items discussed in today's NODA.
[[Page 21616]]
Table III-1.--Brief Summary of the Regulatory Events Associated with Radiochemical Methods
----------------------------------------------------------------------------------------------------------------
1976 National primary July 18, 1991-- March 5, 1997-
drinking water Radionuclides proposed rule Radionuclide methods Today's notice of data
regulations final rule availability
----------------------------------------------------------------------------------------------------------------
The 1976 NPDWR approved: The July 18, 1991-- The March 5, 1997 final Updates the public on changes
* Radiochemical methods radionuclides rule rule for radionuclide that have occurred regarding
to analyze for gross proposed: methods: radiochemical methods of
alpha-particle * Fifty-six additional * Approved 66 additional analysis since the 1991
activity, radium-226, methods for compliance radionuclide techniques proposed rule. The updates
total radium, gross monitoring of for gross alpha- discussed in today's NODA
beta-particle activity, radionuclides particle activity, include:
strontium-89 and -90, * Guidance for the sample radium-226, radium-228, * A brief discussion of the
cesium-134 and uranium handling, preservation and uranium, cesium-134, analytical methods updates
* Defined the detection holding times that were iodine-131, and which were promulgated by the
limit (DL) as the cited in the 1990 U.S.EPA strontium-90 Agency on July 18, 1997 final
required measure of ``Manual for the Responded to comments rule.
sensitivity and listed Certification of regarding the * Guidance for the sample
the required DL for Laboratories Analyzing analytical methods handling, preservation and
each regulated Drinking Water'' (excluding radon) holding times listed in the
radionuclide * The use of practical received from the July 1997 U.S.EPA ``Manual for the
quantitation limits (PQLs) 18, 1991 proposed Certification of Laboratories
and acceptance limits as radionuclides rule Analyzing Drinking Water.''
the measures of * Recommendations for the
sensitivity for analysis of short-lived, alpha-
radiochemical analysis emitting radioisotopes (i.e.,
radium-224).
* Revised cost estimates for
radiochemical analysis.
* The Agency's intent to
continue to use the detection
limits defined in the 1976
rule as the required measures
of sensitivity.
* Response to some of the
comments on the 1991 proposed
radionuclides.
* The externalization of the
Performance Evaluation
Program.
The Agency's plans to
implement a Performance Based
Measurement System.
----------------------------------------------------------------------------------------------------------------
A. The Updated 1997 Laboratory Certification Manual
A revised version of the certification manual was published in
1997 (EPA 815-B-97-001, EPA 1997b). Table III-2 lists the guidance
for sample handling, preservation, holding times, and
instrumentation which appeared in this manual. Table III-2 also
includes additional recommendations for radiochemical
instrumentation (footnoted by the number 6), which the Agency is
requesting comment on.
Table III-2.--Sample Handling, Preservation, Holding Times and Instrumentation
--------------------------------------------------------------------------------------------------------------------------------------------------------
Parameter Preservative 1 Container 2 Maximum holding time 3 Instrumentation 4
--------------------------------------------------------------------------------------------------------------------------------------------------------
Gross Alpha........................ Concentrated HCl or P or G 6 months A, B or G
HNO3 to pH 25.
Gross Beta......................... Concentrated HCl or P or G 6 months A or G
HNO3 to pH 25.
Radium-226......................... Concentrated HCl or P or G 6 months A, B, C 6, D or G
HNO3 to pH 2.
Radium-228......................... Concentrated HCl or P or G 6 months A, B 6, C 6 or G
HNO3 to pH 2.
Uranium natural.................... Concentrated HCl or P or G 6 months A 6, F, G 6, or O
HNO3 to pH 2.
Cesium-134......................... Concentrated HCl to pH P or G 6 months A, C or G
2.
Strontium-89 and -90............... Concentrated HCl or P or G 6 months A or G
HNO3 to pH 2.
Radioactive Iodine-131............. None.................. P or G 8 days A, C or G
Tritium............................ None.................. G 6 months E
Gamma/Photon Emitters.............. Concentrated HCl or P or G 6 months C
HNO3 to pH 2.
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ It is recommended that the preservative be added to the sample at the time of collection. It is recommended that samples be filtered if suspended or
settleable solids are present at any level observable to the eye prior to adding preservative. This should be done at the time of collection. If the
sample has to be shipped to a laboratory or storage area, however, acidification of the sample (in its original container) may be delayed for a period
not to exceed 5 days. A minimum of 16 hours must elapse between acidification and start of analysis.
\2\ P = Plastic, hard or soft; G = Glass, hard or soft.
\3\ Holding time 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. If a composite sample is prepared, a holding time cannot exceed 12 months.
\4\ A = Low background proportional system; B = Alpha and beta scintillation system; C = Gamma spectrometer [Ge(Hp) or Ge (Li)]; D = Scintillation cell
system; E = Liquid scintillation system; F = Fluorometer; G = Low background alpha and beta counting system other than gas-flow proportional; O =
Other approved methods (e.g., laser phosphorimetry and alpha spectrometry for uranium).
\5\ If HCl 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.
\6\ Additional instrumentation that was not listed in the USEPA 1997 ``Manual for the Certification of Laboratories Analyzing Drinking Water.''
B. Recommendations for Determining the Presence of Radium-224
To determine the presence of the short-lived radium-224 isotope
(half life 3.66 days), the Agency recommends using one
of the following several options.
1. Radium-224 by Gamma Spectrometry and Alpha Spectrometry
(a) Gamma Spectrometry. Radium-224 can be specifically
determined by gamma spectrometry using a suitably prepared sample.
In this method a precipitate in which the radium isotopes are
concentrated is gamma counted. The primary advantage of this
technique is specificity for radium gamma rays, radium-224 included.
Other advantages of this method include:
[[Page 21617]]
a simple sample preparation were radium isotopes are
concentrated from samples 1 liter or larger;
specificity for the radium-224 isotope based on a
unique gamma energy;
optimal accuracy and precision if the sample is counted
within 72 hours of collection (40 hours is recommended);
and is cost competitive with the gross methods because
a single count rather that three counts (see the gross alpha methods
discussion) is necessary to measure the radium-224 in a routine
sample.
A gamma spectrometry method by Standard Methods is currently
pending but for now the reader is referred to the method used by
Parsa. (Parsa, 1998).
(b) Alpha Spectrometry. The alpha spectrometry method measures
alphas emitted by radium-224 and its alpha emitting daughters. The
alpha spectrometry method, used for the USGS occurrence survey (see
appendix I and EPA 2000a), was a slight modification of an existing
method (see EPA 2000a). Using an appropriate tracer (e.g. Ba-133),
barium and radium isotopes are separated from other radionuclides
and interferences using cation ion exchange chromatography. A
prepared sample, counted for approximately 100 minutes using alpha
spectrometry, can be used to measure the radium-224 in the sample
and is capable of good accuracy and precision. Other alpha
spectrometry techniques, similar to the modified method used for the
USGS occurrence survey, should be sufficient for the detection of
radium-224. It is cost competitive with the gross methods (discussed
next) because a single count rather than three (for gross methods)
is sufficient to for measurement of radium-224.
2. Gross Radium Alpha (Co-precipitation) Within 72 Hours
The presence of radium-224 can be determined indirectly using
the radium-224 half-life decay and the gross radium alpha technique.
Gross radium co-precipitation methods, like EPA 903.0, concentrate
radium isotopes by co-precipitation, separating radium and radium-
like isotopes from potential interferences. Relative to evaporative
methods, the co-precipitation technique can be used for larger (> 1
L) sample sizes with a resulting increase in the method sensitivity.
Initial analysis within 72 hours after sample collection (40 hours
recommended for optimal data quality) using the co-precipitation
methods yield results, reflecting both alpha-emitting radium
isotopes (radium-224 and radium-226). For these to produce
unambiguous results, radium-224 must be the dominant isotope
present, i. e. the ratio of radium-224 to radium-226 must be three
or greater. If this is the prevailing composition, the estimated
contribution of radium-224 to the overall value can be ascertained
by recounting the sample at 4 or 8 days intervals and calculating
the change in the measured activity. The noted change will show a
decrease with a 4 day half-life indicative of Ra-224. Formulas are
available to calculate the initial radium-224 concentration present
in the sample when collected. The advantages of this technique
include:
enhanced sensitivity (1 L samples);
it does not require additional analyst training;
it is specific for radium isotopes; and
the resulting precipitate can be measured by a number
of techniques, including proportional counting, alpha scintillation
counting, or gamma counting.
3. Evaporative Gross Alpha-Particle Analysis Within 72 Hours
The radium-224 isotope, when in equilibrium with its decay
progeny, emits four alpha particles. Three of these alpha particles
equilibrate almost immediately (within 5 minutes) after sample
preparation and add to or amplify the sample count rate. This count
rate amplification can be exploited for the measurement of radium-
224 in a sample at low concentration (15 pCi/L). The presence of the
radium-224 radioisotope in drinking water may be ascertained by
performing an initial evaporative gross alpha-particle analysis
within 72 hours (40 hours recommended) after sample collection. In
the absence of any other alpha-emitting nuclide (e.g., uranium or
radium-226) and if the gross alpha-particle value is above the MCL,
the sample may be re-counted at 4- and 8-day intervals to determine
if the observed decrease in activity follows the 3.66 day half-life
of radium-224. A decrease in the gross alpha value with a 4-day
decay rate indicates the likely presence of radium-224. Formulas are
available to calculate the concentration of radium-224 in the
initial sample. The advantages of this option include:
the method is similar to the general method for
evaporative gross alpha;
it requires no special training of the analyst; and
it can be a definitive test if other alpha-emitting
nuclides are known to be absent.
The Agency recognizes that analysis within the 72-hour time
frame creates difficulties in shipping and handling and may increase
the price of the analysis.
C. Revised Cost Estimates for Radiochemical Analysis
The cost estimates for radiochemical analysis from the 1991
proposed rule and the revised cost estimates are shown in Table III-
3.
Table III-3.--The 1991 and 1999 Estimated Costs of Analyses for
Radionuclides
------------------------------------------------------------------------
Approximate Approximate
Radionuclides cost $ (1991)1 cost $ (1999)2
------------------------------------------------------------------------
Gross Alpha and beta.................... 35 45
Gross alpha--coprecip................... 35 45
Radium-226.............................. 85 90
Radium-228.............................. 100 110
Uranium (total)......................... 45 48 (LP)
Uranium (isotopic)...................... 125 128 (AS)
Radioactive Cesium (-134)............... 100 125
Radioactive Strontium................... 105 144
Total Strontium (-89 and -90)........... ...... 153
Radioactive Iodine -131................. 100 131
Tritium................................. 50 60
Gamma/Photon Emitters................... 110 142
------------------------------------------------------------------------
Source:
\1\ 56 FR 33050; July 18, 1991.
\2\ USEPA, 2000a.
Abbreviations: LP = laser phosphorimetry; AS = alpha spectrometry.
Note: Estimated costs are on a per-sample basis; analysis of multiple
samples may have a lower cost.
[[Page 21618]]
D. The Detection Limits as the Required Measures of Sensitivity
Table III-4 cites the detection limits or the required
sensitivity for the specific radioanalyses that were listed in the
1976 rule and are also cited in 40 CFR 141.25.
Table III-4.--Required Regulatory Detection Limits for the Various
Radiochemical Contaminants (40 CFR 141.25)
------------------------------------------------------------------------
Contaminant Detection limit (pCi/L)
------------------------------------------------------------------------
Gross Alpha............................... 3
Gross Beta................................ 4
Radium-226................................ 1
Radium-228................................ 1
Cesium-134................................ 10
Strontium-89.............................. 10
Strontium-90.............................. 2
Iodine-131................................ 1
Tritium................................... 1,000
Other Radionuclides and Photon/Gamma \1/10\th of the rule NIPDWR
Emitters. 1976 table IV-2A and 2B
------------------------------------------------------------------------
E. References
Parsa, B., 1998. Contribution of Short-lived Radionuclides to
Alpha-Particle Radioactivity in Drinking Water and Their Impact on
the Safe Drinking Water Act Regulations, Radioactivity and
Radiochemistry, Vol. 9, No. 4, pp. 41-50, 1998. USEPA, 1991.
National Primary Drinking Water Regulations; Radionuclides; Proposed
Rule. Federal Register. Vol. 56, No. 138, p. 33050. July 18, 1991.
USEPA, 1997a. National Primary Drinking Water Regulations;
Analytical Methods for Radionuclides; Final Rule and Proposed Rule.
Vol. 62, No. 43, p. 10168. March 5, 1997.
USEPA, 1997b. ``Manual for the Certification of Laboratories
Analyzing Drinking Water.'' EPA 815-B-97-001. 1997.
USEPA, 2000a. ``Technical Support Document for the Radionuclides
Notice of Data Availability.'' 2000.
Appendix IV--Treatment Technologies and Costs
A. Introduction
This section describes updates to EPA's previous evaluations of
the feasibility and costs of treatment technologies for the removal
of radionuclides from drinking water. Prior to this update, the
latest evaluation was the 1992 ``Technologies and Costs document''
for radionuclides in drinking water (EPA 1992). The updates to the
1992 radionuclides Technologies and Costs document comprise an
updated Technologies and Costs Document (EPA 1999a) and a radium
compliance cost study (EPA 1998a), which are described later in this
section. This section also describes other relevant documents,
including the1998 Federal Register notice of the ``Small Systems
Compliance Technology List'' (SSCTLs) for the currently regulated
radionuclides (63 FR 42032) and its supporting guidance document
(EPA 1998b). Both of the documents supporting the SSCTs can be
obtained on-line at ``http://www.epa.gov/OGWDW/standard/
tretech.html''.
The SSCTLs for the meeting the MCLs for combined radium-226 and
radium-228, gross alpha emitters, and combined beta and photon
emitters are included in ``Announcement of Small System Compliance
Technology Lists for Existing National Primary Drinking Water
Regulations and Findings Concerning Variance Technologies,''
published in the Federal Register on August 6, 1998 (63 FR 42032).
The supporting guidance document cited previously includes
information regarding small systems treatment and waste disposal
concerns relevant to radionuclide contaminants and was made publicly
available on September 15, 1998. Further evaluations of small
systems treatment technology applicability and affordability have
been done since the SSCTLs for radionuclides were published,
including an analysis of SSCTs for uranium (EPA 1999b). These
evaluations are summarized later in this section.
B. Treatment Technologies Update
1. Updates on Performance of Technologies for Removal of Regulated
Radionuclides and Uranium
One of the purposes of the update to the radionuclides
Technologies and Costs (T&C) document (EPA 1999a) was to update the
treatment technology performance sections of the 1992 radionuclides
T&C document. The peer-reviewed literature revealed no new
significant sources of information regarding performance for the
previously described technologies, nor did it reveal literature
regarding any new treatment technologies for radionuclides in
drinking water. Both the 1992 and 1999 radionuclides T&C documents
include performance evaluations of the BATs proposed in 1991 for the
regulated radionuclides and uranium (56 FR 33050, Jul. 18, 1991) and
additional technologies that were reviewed as potential BATs for the
1991 proposed rule, but that were not proposed as BAT for various
reasons.
Although the 1999 T&C document concludes that the peer-reviewed
literature describes no new technologies since the 1992 T&C document
was completed, there have been some developments that are
significant. In particular, both package plant \1\ technologies,
including those equipped with remote control/communication
capabilities, and point-of-entry (POE)/point-of-use (POU) versions
\2\ of existing technologies have become more widely applicable for
use for compliance. This is true both because of improvements in
these technologies themselves (NRC 1997) and since the 1996 SDWA
explicitly allows package plants and POE/POU devices to be used as
compliance technologies for small systems (section 1412.b.4.E).
Package plant technologies and POE/U technologies are discussed in
more detail in the Technical Support Document (EPA 2000a).
---------------------------------------------------------------------------
\1\ Package plants are skid mounted factory assembled
centralized treatment units that arrive on site ``virtually ready to
use''. Package plants offer several advantages. First, since they
combine elements of the treatment process into a compact assembly
(such as chemical feeders, mixers, flocculators, basins, and
filters), they tend to require lesser construction and engineering
costs. Another advantage is that many package plant technologies are
becoming more automated and thus can be less demanding of operators
than their fully engineered counter-parts (EPA 1998b).
\2\ Point-of-entry (POE) treatment units treat all of the water
entering a household or other building, with the result being
treated water from any tap. Point-of-use (POU) treatment units treat
only the water at a particular tap or faucet, with the result being
treated water that one tap, with the other taps serving untreated
water. POE and POU treatment units often use the same technological
concepts employed in the analogous central treatment processes, the
main difference being the much smaller scale of the device itself
and the flows being treated (EPA 1998b).
---------------------------------------------------------------------------
2. Treatment Technologies Evaluated as Compliance Technologies for
Radionuclides
The following technologies are reviewed in the 1999
radionuclides T&C document: (1) for radium, the 1991 proposed Best
Available Technologies (BATs), which are lime softening, ion
exchange, and reverse osmosis; and two other applicable technologies
with significant radium removal data, electrodialysis reversal and
greensand filtration; (2) for uranium, the 1991 proposed BATs, which
are coagulation/filtration, ion exchange, lime softening, and
reverse osmosis; and two other applicable technologies,
electrodialysis reversal and activated alumina; (3) for gross alpha
particle activity, the 1991 proposed BAT, which is reverse osmosis;
and one other applicable technology, ion exchange; and (4) for beta
particle activity and photon radioactivity, the 1991 proposed BATs,
which are ion exchange and reverse osmosis. No other technology
studies pertinent to total beta and photon activity were found, but
this is largely due to the fact that treatment applicability depends
on what specific beta and photon emitters are present and so should
be evaluated on a case-by-case basis. This consideration also
applies to gross alpha activity. It is likely that reverse osmosis,
being applicable to a broad range of inorganic contaminants,
including radionuclide contaminants, is the best alternative for
situations where multiple radionuclides occur.
3. Data on Additional Treatment Technologies
The 1999 radionuclides T&C document does not identify any new
treatment technologies for radionuclides, but does provide
information on two additional variants of coagulation/filtration for
uranium removal: direct filtration and in-line filtration.
4. Small Systems Compliance Technology List and Guidance Manual for the
Regulated Radionuclides and Uranium
The 1996 SDWA identifies three categories of small drinking
water systems, those serving populations between 25 and 500, 501 and
3,300, and 3,301 and 10,000. In addition to BAT determinations, the
SDWA directs EPA to make technology assessments for each of the
three small system size categories in all future regulations
establishing an MCL or
[[Page 21619]]
treatment technique. Two classes of small systems technologies are
identified for future National Primary Drinking Water Regulations
(NPDWRs): compliance technologies and variance technologies.
Compliance technologies may be listed for NPDWRs that promulgate
MCLs or treatment techniques. In the case of an MCL, ``compliance
technology'' refers to a technology or other means that is
affordable (if applicable) and that achieves compliance. Possible
compliance technologies include packaged or modular systems and
point-of-entry (POE) or point-of-use (POU) treatment units, as
described previously.
Variance technologies are only specified for those system size/
source water quality combinations for which no technology meets all
of the criteria for listing as a compliance technology (section
1412(b)(15)(A)). Thus, the listing of a compliance technology for a
size category/source water combination prohibits the listing of
variance technologies for that combination. While variance
technologies may not achieve compliance with the MCL or treatment
technique requirement, they must achieve the maximum reduction that
is affordable considering the size of the system and the quality of
the source water. Variance technologies must also achieve a level of
contaminant reduction that is ``protective of public health''
(section 1412(b)(15)(B)).
In the case of the currently regulated radionuclides, i.e.,
combined radium-226 and -228, gross alpha activity, and total beta
and photon activity, there are no variance technologies allowable
since the SDWA (section 1415(e)(6)(A)) specifically prohibits small
system variances for any MCL or treatment technique which was
promulgated prior to January 1, 1986. The Variance and Exemption
Rule describes EPA's interpretation of this section in more detail
(see 63 FR 19442; April 20, 1998).
Small systems compliance technologies for the currently
regulated radionuclides, combined radium-226 and -228, gross alpha
emitters, and total beta and photon activity, were listed and
described in the Federal Register on August 6, 1998 (EPA 1998a) and
in an accompanying guidance manual (EPA 1998b). Small systems
compliance technologies for uranium were also evaluated (EPA 1999a).
Small systems compliance technologies (SSCTs) for uranium were
evaluated in terms of each technology's removal capabilities,
contaminant concentration applicability ranges, other water quality
concerns, treatment costs, and operational/maintenance requirements.
The SSCT list for uranium is technology specific, but not product
(manufacturer) specific. Product specific lists were determined to
be inappropriate due to the potential resource intensiveness
involved. Information on specific products will be available through
another mechanism. EPA's Office of Research and Development has a
pilot project under the Environmental Technology Verification (ETV)
Program to provide treatment system purchasers with performance data
from independent third parties.
Tables IV-1 and IV-2 summarize the small systems compliance
technologies listed in the 1998 SSCTL for combined radium-226, and -
228, gross alpha emitters, total beta and photon activity. Table IV-
1 is shown as it will be updated when uranium is regulated. Table
IV-1 describes limitations for each of the listed technologies and
Table IV-2 lists SSCTs for each contaminant.
Table IV-1.--List of Small Systems Compliance Technologies for Radionuclides and Limitations to Use
----------------------------------------------------------------------------------------------------------------
Limitations
Unit technologies (see Operator skill level Raw water quality range and
footnotes) required1 considerations1
----------------------------------------------------------------------------------------------------------------
1. Ion Exchange (IE).................... (a) Intermediate............... All ground waters.
2. Point of Use (POU) IE................ (b) Basic...................... All ground waters
3. Reverse Osmosis (RO)................. (c) Advanced................... Surface waters usually
require pre-filtration.
4. POU RO............................... (b) Basic...................... Surface waters usually
require pre-filtration.
5. Lime Softening....................... (d) Advanced................... All waters.
6. Green Sand Filtration................ (e) Basic...................... ...........................
7. Co-precipitation with Barium Sulfate. (f) Intermediate to Advanced... Ground waters with suitable
water quality.
8. Electrodialysis/Electrodialysis ............ Basic to Intermediate...... All ground waters.
Reversal.
9. Pre-formed Hydrous Manganese Oxide (g) Intermediate............... All ground waters.
Filtration.
10. Activated alumina................... (a), (h) Advanced................... All ground waters;
competing anion
concentrations may affect
regeneration frequency.
11. Enhanced coagulation/filtration..... (i) Advanced................... Can treat a wide range of
water qualities.
----------------------------------------------------------------------------------------------------------------
1 National Research Council (NRC). Safe Water from Every Tap: Improving Water Service to Small Communities.
National Academy Press. Washington, D.C. 1997.
Limitations Footnotes to Table IV-2: Technologies for Radionuclides:
a The regeneration solution contains high concentrations of the contaminant ions. Disposal options should be
carefully considered before choosing this technology.
b When POU devices are used for compliance, programs for long-term operation, maintenance, and monitoring must
be provided by water utility to ensure proper performance).
c Reject water disposal options should be carefully considered before choosing this technology. See other RO
limitations described in the SWTR Compliance Technologies Table.
d The combination of variable source water quality and the complexity of the water chemistry involved may make
this technology too complex for small surface water systems.
e Removal efficiencies can vary depending on water quality.
f This technology may be very limited in application to small systems. Since the process requires static mixing,
detention basins, and filtration, it is most applicable to systems with sufficiently high sulfate levels that
already have a suitable filtration treatment train in place.
g This technology is most applicable to small systems that already have filtration in place.
h Handling of chemicals required during regeneration and pH adjustment may be too difficult for small systems
without an adequately trained operator.
i Assumes modification to a coagulation/filtration process already in place.
Table IV-2 lists the Small Systems Compliance Technologies for
the currently regulated radionuclides. Technology numbers refer to
the technologies listed in Table IV-1.
[[Page 21620]]
Table IV-2.--Compliance Technologies by System Size Category for Radionuclide NPDWRs (Affordability Not
Considered, Except for Uranium, Due to Statutory Limitations)
----------------------------------------------------------------------------------------------------------------
Compliance technologies1 for system size
categories (population served)
Contaminant -----------------------------------------------
25-500 501-3,300 3,300-10,000
----------------------------------------------------------------------------------------------------------------
Combined radium-226 and radium-228.............................. 1, 2, 3, 4, 5, 1, 2, 3, 4, 5, 1, 2, 3, 4, 5,
6, 7, 8, 9 6, 7, 8, 9 6, 7, 8, 9
Gross alpha particle activity................................... 3, 4 3, 4 3, 4
Total beta particle activity and photon activity, average annual 1, 2, 3, 4 1, 2, 3, 4 1, 2, 3, 4
concentration..................................................
Uranium......................................................... 1, 2, 4, 10, 1, 2, 3, 4, 5, 1, 2, 3, 4, 5,
11 10, 11 10, 11
----------------------------------------------------------------------------------------------------------------
Note: 1 Numbers correspond to those assigned to technologies found in the table ``List of Small Systems
Compliance Technologies for the Currently Regulated Radionuclides.''
C. Waste Treatment, Handling and Disposal Guidance
In the proposed radionuclides rule of July 1991, EPA referenced
a 1990 EPA draft report entitled ``Suggested Guidelines for Disposal
of Drinking Water Treatment Wastes Containing Naturally-Occurring
Radionuclides'' (EPA 1990). That 1990 report offered guidance to
system managers, engineers, and State agencies responsible for the
safe handling and disposal of treatment wastes that, in many cases,
were not specifically addressed by any statute. That guidance report
was later updated in 1994 (EPA 1994).
The guidance provided information on the following: (1)
Background on water treatment processes and characteristics of
wastes generated; (2) rationale for radiation protection, including
citation of programs and regulations affecting other sources of such
waste; (3) guidelines for several methods of disposal of solid and
liquid type wastes containing the subject radionuclides; and, (4)
the specification of practical guidance to protect workers and
others who may handle or be exposed to water-treatment wastes
containing radiation above background levels.
The Technical Support Document (EPA 2000a) discusses disposal
methods and issues, including comments received in reference to the
1990 ``Suggested Guidelines for Disposal of Drinking Water Treatment
Wastes Containing Naturally-Occurring Radionuclides,'' and the 1994
update to this report.
D. Unit Treatment Cost Updates
Treatment costs for coagulation/filtration (including direct
filtration and in-line filtration), lime softening, ion exchange,
reverse osmosis, electrodialysis reversal, greensand filtration,
point-of-use (POU) reverse osmosis, POU ion exchange, and point-of-
entry cation exchange were updated in the appendix to the 1999
radionuclides T&C document. This update includes land-cost
considerations and waste-disposal cost estimates. Cost estimates
were made using standard EPA treatment technology costing models.
Outputs were updated to current dollars using standard engineering
costing indices, e.g., the Bureau of Labor's Chemical and Allied
Products Index. Costs for individual technologies were analyzed in
terms of water usage, removal efficiency, interest rate, and other
variables.
In addition to cost model updates, EPA has performed a study of
the actual costs of treatment and other compliance measures for the
radium standard (EPA 1998c), which provided a ``snapshot'' of the
costs incurred by water systems in complying with the existing
combined radium-226 and radium-228 MCL. Studies of this nature allow
EPA to compare modeled costs used in regulatory impact assessments
with real-world data for the purposes of model validation and cost
estimate amendments. They also allow EPA to check assumptions about
the prevalence of use of particular water-treatment technologies.
The study comprises data compiled from contacts with water-
treatment personnel, State representatives, and EPA Regional
representatives within EPA Regions 5 (IL, IN, MI, MN, OH, and WI)
and 8 (CO, MT, ND, SD, UT, and WY). Specifically, data were obtained
regarding water systems in California, Florida, Idaho, Illinois,
Indiana, Ohio, Wisconsin, and Wyoming. State Agencies and EPA
Regional offices identified 136 systems as having water sources with
combined radium-226 and radium-228 above the MCL of 5 pCi/L. Of
these, 55 of the systems were contacted, of which 29 were either
treating for radium or were in the process of selecting a treatment
method. The remaining systems were either further behind in
treatment selection plans or pursuing other compliance measures. All
of the systems that were currently treating for radium were in
compliance with the MCL. Twenty-six of these systems responded with
cost data, of which 17 were small systems (design flow 1 mgd).
Thirty-five percent of the small systems reported were using reverse
osmosis which, at an average total treatment cost of $3.02 per
thousand gallons, was the most expensive treatment technology
identified. Other treatment options used were lime softening and ion
exchange. These had average total treatment costs of $2.36 and $0.73
per thousand gallons, respectively. Unit costs are discussed in more
detail in the Technical Support Document (EPA 2000a).
EPA requests comments on its analysis of treatment technologies,
costs, and treatment residuals disposal.
E. References
National Research Council (NRC). Safe Water From Every Tap:
Improving Water Service to Small Communities. National Academy
Press. Washington, DC. 1997.
USEPA. Office of Drinking Water. Suggested Guidelines for
Disposal of Drinking Water Treatment Wastes Containing Naturally-
Occurring Radionuclides (July 1990 draft).
USEPA. National Primary Drinking Water Regulations;
Radionuclides; Proposed Rule. Federal Register. Vol. 56, No. 138, p.
33050. July 18, 1991.
USEPA. Technologies and Costs for the Removal of Radionuclides
from Potable Water Supplies. Prepared by Malcolm Pirnie, Inc. July
1992.
USEPA. Office of Ground Water and Drinking Water. Suggested
Guidelines for Disposal of Drinking Water Treatment Wastes
Containing Radioactivity (June 1994 draft).
USEPA. Announcement of Small Systems Compliance Technology Lists
for Existing National Primary Drinking Water Regulations and
Findings Concerning Variance Technologies. Federal Register. Vol.
63, No. 151, p. 42032. August 6, 1998. (EPA 1998a).
USEPA. ``Small System Compliance Technology List for the Non-
Microbial Contaminants Regulated Before 1996.'' EPA-815-R-98-002.
September 1998. (EPA 1998b).
USEPA. ``Actual Cost for Compliance with the Safe Drinking Water
Act Standard for Radium-226 and Radium-228.'' Final Report. Prepared
by International Consultants, Inc. July 1998. (EPA 1998c).
USEPA Technologies and Costs for the Removal of Radionuclides
from Potable Water Supplies. Draft. Prepared by International
Consultants, Inc. April, 1999. (EPA 1999a).
USEPA. ``Small System Compliance Technology List for the
Radionuclides Rule.'' Prepared by International Consultants, Inc.
Draft. April 1999. (EPA 1999b).
USEPA. ``Technical Support Document for the Radionuclides Notice
of Data Availability.'' Draft. March, 2000. (EPA 2000a)
Appendix V--Economics and Impacts Analysis
A. Overview of the Economic Analysis
1. Background
Analysis of the costs, benefits, and other impacts of
regulations is required under the Safe Drinking Water Act Amendments
of 1996, Executive Order 12866 (Regulatory Planning and Review), and
EPA's internal guidance for regulatory development. These
[[Page 21621]]
requirements are new relative to the 1991 proposal for revisions to
the existing National Primary Drinking Water Regulations (NPDWRs)
for radionuclides.
The actions that are anticipated to have regulatory impacts are
evaluated in this section. These actions are: (1) the correction the
monitoring deficiency for combined radium-226 and radium-228; and
(2) the establishment of a uranium NPDWR with an MCL of 20
µg/L; or (3) the establishment of a uranium NPDWR with an
MCL of 40 µg/L; or (4) the establishment of a uranium NPDWR
with an MCL of 80 µg/L. See ``Combined Ra-226 and Ra-228''
in the today's NODA (section III, part F) for a discussion of the
monitoring corrections that will be finalized for the combined
radium-226 and radium-228 (``combined radium'') NPDWR. See
``Uranium'' in the NODA (section III, part H) for a discussion of
the options being considered for finalization for the uranium NPDWR.
2. Economic Analysis of the Regulatory Actions Being Considered for
Radionuclides in Drinking Water
The economic analysis summarized here supports the finalization
of the 1991 Radionuclides proposal. The more detailed economic
analysis (the Health Risk Reduction and Cost Analysis, EPA 2000b)
may be obtained from the Water Docket, as described in the
Introduction to today's NODA (see ADDRESSES). It provides central-
tendency estimates of national costs and benefits and presents
information on the data sources and analytic approaches used,
including a qualitative discussion of the analytical limitations and
uncertainties involved. Further uncertainty analyses will be
performed to support the analyses summarized here and will be
reported in the preamble to the final rule. It should be noted that
these additional uncertainty analyses are not expected to alter
regulatory decisions.
The basic steps in a comprehensive economic analysis include:
(1) Estimating baseline conditions in the absence of revisions to
the regulations; (2) predicting actions that water systems will use
to meet each regulatory option (the ``decision tree''); (3)
estimating national costs resulting from compliance actions; (4)
estimating national benefits resulting from compliance; and (5)
assessing distributional impacts and equity concerns. In today's
NODA, we present preliminary estimates of national costs and
benefits for the options evaluated, focusing on monitoring and
compliance costs and reductions in cancer risks. Other national
costs and benefits (e.g., state administrative costs and risk
reductions from incidental treatment of co-occurring contaminants)
and potential distributional impacts are described qualitatively
(see EPA 2000a and EPA 2000b).
The first step in the economic analysis, defining the analytical
baseline, requires that water systems be apportioned into several
groups based on their predicted levels of radionuclides and the
current monitoring scheme. In the case of the radionuclides NPDWRs,
this provides unusual challenges. This is partly due to the fact
that several community water systems are not complying with the
existing regulations, which is reflected in the occurrence database
used for this work (the National Inorganics and Radionuclides
Survey, ``NIRS'; see EPA 1991, proposed rule and EPA 2000a). Also,
as discussed in the Introduction to today's NODA, there are
weaknesses in the current monitoring requirements that has lead to a
situation in which some water systems having combined radium levels
greater than the MCL of 5 pCi/L will not have knowledge of this fact
(and hence are not presently in violation of the combined radium
NPDWR). Both of these influences, the existing unresolved
radionuclides NPDWR violations and the monitoring deficiencies, must
be accounted for in the analytical baseline.
The regulatory baseline and other analytical baselines are
benchmarks to measure regulatory impacts against. Generating a
national-level contaminant occurrence profile is an important part
of this benchmarking process. The database used as the basis for
this model, NIRS, is described in appendix I of today's NODA
(Occurrence). The analysis of regulatory impacts uses this system-
size stratified baseline occurrence model \1\ to estimate the
percentages of water systems with contaminant levels within
specified values (e.g., 30 to 50% above the MCL). This information
is then combined with other models to estimate the compliance costs
and benefits associated with each option. Examples of models
relevant to national costs estimation include ``model systems \2\,''
compliance cost equations \3\, and the compliance action prediction
model or ``decision trees \4\.'' Examples of models relevant to risk
reduction and benefits estimation include the risk models described
in appendix II and the risk reduction valuation models described in
the Technical Support Document (EPA 2000a).
---------------------------------------------------------------------------
\1\ The NIRS database is stratified into four categories:
systems serving between 25-500 persons, 501--3,300 persons, 3,301-
10,000 persons, and 10,001-1,000,000 persons. Because of the small
sample size used to describe the larger systems, our model uses only
three categories: we combine the two categories for systems serving
greater than 3,301 persons into a single category.
\2\ Model systems describe the universe of drinking water
systems by breaking it down into discrete ``system size categories''
by population served. There are nine size categories: 25-100 persons
served; 101-500; 501-1,000; 1,001-3,300; 3,301-10,000; 10,001-
50,000; 50,001-100,000; 100,000-1,000,000; > 1,000,000. Within each
size category, the systems are described by a single set of
``typical characteristics'' by source water type (ground versus
surface water) and ownership type (public versus private ownership).
These characteristics include the average and design flows and the
distribution of numbers of entry points per system.
\3\ Unit compliance costs models include water treatment cost
models (e.g., W/W Cost and the WATER model) and models for other
compliance options, like alternate water well sources and purchasing
water. For a discussion of the standard EPA water treatment cost
models, see EPA 1999d.
\4\ Decision trees are models of the relative probabilities that
water systems will choose particular compliance actions when in
violation. The probabilities are estimated based on considerations
of source water type, system size, water quality, required removal
efficiency, unit costs, treatment issues (e.g., co-treatment and
pre-/post-treatment requirements), and residuals disposal costs and
issues.
---------------------------------------------------------------------------
The analytical baseline for combined radium reflects full
compliance with the existing regulations as written, which have been
fully enforceable since the 1986 reauthorization of the SDWA. This
approach assumes that, in the absence of any changes to the
radionuclides NPDWRs, EPA and the States will eventually ensure that
all systems fully comply with the existing regulations. This
approach allows us to separate out the predicted number of systems
with combined radium levels in excess of the MCL that have knowledge
of the violation (``systems in violation'') from the predicted
number of systems that have levels in excess of the MCL, but that
would not have knowledge of this under the current monitoring
requirements. Since uranium is not currently regulated, no such
corrections are necessary. It was also determined that treatment
installed to remove the other radionuclides should not significantly
impact the uranium analytical baseline.\5\
---------------------------------------------------------------------------
\5\ While the treatments installed to eliminate gross alpha and
combined radium may also reduce uranium levels, we do not quantify
these impacts in this analysis. We make no adjustment for three
reasons. First, the NIRS data suggest that systems with elevated
levels of gross alpha or combined radium rarely report uranium
concentrations above levels of concern. Second, some types of
treatment used to remove gross alpha or radium are less effective in
removing uranium. Lastly, radium and uranium occur at higher levels
under very different aquifer conditions: radium tends to occur at
high levels in water with low dissolved oxygen and high total
dissolved solids, while uranium occurs at higher levels in oxygen-
rich waters with low total dissolved solids (see the Technical
Support Document, EPA2000a).
---------------------------------------------------------------------------
B. Approach for Assessing Occurrence, Risks and Costs for Community
Water Systems
1. Assessing Occurrence
To develop estimates of the baseline radionuclides occurrence
profile for community water systems, we began by extrapolating from
data obtained through EPA's National Inorganics and Radionuclides
Survey (NIRS). This survey measured radionuclide concentrations at
990 community ground water systems between 1984 and 1986. For
detailed information on the design of NIRS, see Longtin 1988. For
detailed information on how NIRS was used in this work, see the
background documents (EPA 2000a and 2000b).
We made adjustments to the NIRS data to address certain
limitations, including (1) the small size of the sample of systems
serving populations greater than 3,300 persons; (2) the decay of
radium-224 prior to analysis of the NIRS water samples; (3) the need
to convert mass measurements of uranium to activity levels; and, (4)
the lack of information on surface water systems. The analyses and
discussions that follow concentrate on CWSs serving retail
populations of less than one million persons. Discussions of
preliminary and future economic impacts analyses of Non-Transient
Non-Community Water Systems (NTSC systems) and the largest CWSs
follow later in this section. The two occurrence approaches we
examined are described next. For a discussion of the relative
strengths and weaknesses of the two approaches to estimating
occurrence, see the Technical Support Document (EPA 2000a).
[[Page 21622]]
2. ``Direct Proportions Approach'' to Estimating Occurrence
Because of uncertainties related to extrapolating from the NIRS
database to national-level estimates, we applied two approaches for
estimating the national-level central-tendency occurrence estimate.
First, we assumed that national occurrence is directly proportional
to the occurrence levels measured in NIRS. For example, if the
radionuclide concentration in one percent of the samples from NIRS
representing a particular water system size category are greater
than the MCL, we assumed that one percent of all systems in that
size class would be out of compliance at the national level (It is
worth noting that using NIRS to extrapolate to the State or regional
level is not valid, since NIRS was designed to be representative at
the national-level, but not at these other levels). In cases where
this approach predicts ``zero probability'' of non-compliance for a
system size category (i.e., no samples in NIRS were above the MCL
being considered), this approach is flawed, since the expectation is
that this finding actually reflects a small probability, not ``zero
probability.'' In other words, in situations where ``zero impact''
is predicted, it is much more likely that a very small number of
water systems will be impacted compared to true ``zero impact.'' For
this reason, we also used a mathematical model to simulate the
occurrence distribution, in which these ``zero probabilities'' are
replaced by estimated small probabilities.
3. ``Lognormal Model Approach'' to Estimating Occurrence
The second approach recognizes that ``true'' radionuclides
occurrence will most likely be spread over a range wider than that
observed in the survey. This approach assumes that ``probability
plots'' of the NIRS data are lognormally distributed. A probability
plot compares the radionuclide concentration for the various samples
to the probability of a given sample having that level or less,
where this probability is estimated from the actual occurrence data
from NIRS. An assumption of lognormality means that a probability
plot for the logarithms of the radionuclide levels would be expected
to be linear (fall on a straight line).
Inspection of the NIRS data suggests that it is distributed in a
roughly lognormal pattern, with most systems reporting concentration
levels well below the MCLs of concern. Several other studies also
suggest that the distribution of radionuclide occurrence in drinking
water systems is likely to follow a lognormal distribution \6\, so
this assumption should be robust in most cases. If the NIRS data
were perfectly lognormally distributed, both approaches would lead
to similar estimates of occurrence. This is usually the case.
However, it should be noted that there instances of significant
deviations between the two approaches. For example, the direct
proportions approach predicts that 0.4 % of the systems serving more
than 500 persons will be impacted (61 systems) by an MCL of 20 pCi
pCi/L for uranium, whereas the lognormal model approach predicts
that 1.8% of systems will be impacted (255 systems), amounting to a
difference in prediction of almost 200 impacted water systems in
this size category. There are several possible explanations for this
deviation, but the important point is that the use of both
approaches allows the data gap to be recognized and fully
considered.
---------------------------------------------------------------------------
\6\ See the Technical Support Document (EPA 2000a) and the HRRCA
(EPA 2000b).
---------------------------------------------------------------------------
A statistical software package (``Stata'') was used to estimate
a lognormal distribution that best fits the data for systems in each
size class. We then used the fitted log means and log standard
deviations of the resulting distributions to estimate the number of
systems out of compliance with each regulatory option using standard
statistical equations. More detail regarding the occurrence models
and the estimation of the numbers of impacted systems can be found
elsewhere (EPA 2000a and 2000b).
4. Assessing Risk
After determining the number of systems out of compliance with
each regulatory option under consideration, we assessed the risk
reductions that would result from these systems taking actions to
come into compliance. The approach for the risk analysis begins with
the development of intrinsic ``risk factors'' for each group of
radionuclides. These risk factors are composites that involve
multiplying EPA's best estimates of unit mortality and morbidity
cancer risk coefficients (risk per pCi) for each group of
radionuclides by standard assumptions regarding drinking water
ingestion to determine the risk factors associated with drinking
water exposure (risk per pCi/L). We then applied the individual risk
factors \7\ to the estimates of the reduction in exposure associated
with each regulatory change under consideration, taking into account
the population exposed. The calculation of risk factors from risk
coefficients and a discussion of exposure assumptions are detailed
elsewhere (EPA 2000a). The risk factors (per pCi/L in drinking
water) used in the risk reduction analyses are summarized in Table
V-1.
---------------------------------------------------------------------------
\7\ This analysis focuses on changes in cancer risks from tap
water ingestion. Individuals may be exposed to radionuclides in
drinking water through other pathways (e.g., inhalation while
showering), and uranium may have toxic effects on the kidneys;
however, we expect that any changes in these types of risks will be,
while not insignificant, much smaller than the changes in cancer
risks from ingestion, and hence discuss them only qualitatively in
this analysis.
---------------------------------------------------------------------------
The unit \8\ risk factors applied in this analysis refer to the
aggregated small changes in the probability of incurring cancer over
a large population. These unit probabilities can be interpreted in
two ways: as the unit lifetime excess probability of cancer
induction averaged over age and gender for all individuals in a
population or as the risk for a statistically ``averaged
individual.'' It should be noted that no one individual is truly
average, since the averaging also occurs over gender. Given a model
of radionuclide occurrence, the population risks of excess cancer
incidence can be estimated before and after a given regulatory
option for the individuals comprising the population, with the
difference being equal to the reduced risk. These reductions in
individual cancer incidence probabilities may then be summed over
the population to indicate the central-tendency number of
``statistical cancer cases avoided'' annually. However, it should be
kept in mind that for many reasons, including the large variance
associated with such risk factors, it is impossible to ``check this
prediction'' in any meaningful way. In interpreting reduced risks
for given options, it is arguably best to think of them in terms of
reduced average ``individual excess risk,'' rather than ``cases
avoided,'' for the reasons just described. For example, it is much
easier to understand the idea that an individual's average lifetime
risk of developing cancer due to exposure to radionuclides in
drinking water has been reduced from three in ten thousand to one in
ten thousand for a number of water systems under a given option then
to understand that an average of 0.5 cancer cases are avoided
annually at the national level for that option. The use of
``individual excess risk'' avoids much the confusion about
``statistical cases,'' which are conceptually difficult to
understand.
---------------------------------------------------------------------------
\8\ ``Unit risk factors and ``unit risks'' refer to the risk per
pCi/L in drinking water. They are not estimates of cancer incidence
per se, but rather are indicators of the ``potency'' of a
radionuclide. To get estimates of the risks of cancer incidence for
an exposed population, the unit risk factors must be used in
conjunction with a radionuclide drinking water occurrence model.
These population risks refer to the estimated numbers of excess
statistical cases of cancer that a population will face under a
given set of exposure assumptions.
Table V-1.--Average Individual Risk Factors, Average Water Consumption (1.1 L/person/day) (per pCi/L)
----------------------------------------------------------------------------------------------------------------
Morbidity Mortality
---------------------------------------------------
Regulatory option Lifetime Annual Lifetime Annual
ingestion ingestion ingestion ingestion
----------------------------------------------------------------------------------------------------------------
Gross Alpha: changes in monitoring requirements (weighted 5.24E-06 7.48E-08 3.26E-06 4.65E-08
average of Ra-224 and Ra-226)..............................
Gross Alpha: changes in MCL (Ra-224 only)................... 4.77E-06 6.81E-08 2.90E-06 4.15E-08
[[Page 21623]]
Combined Radium: changes in monitoring requirements 2.30E-05 3.28E-07 1.63E-05 2.32E-07
(weighted average of Ra-226 and Ra-228)....................
Combined Radium: changes in MCL (Ra-228 only)............... 2.98E-05 4.26E-07 2.12E-05 3.03E-07
Uranium: establish MCL (simple average of U-234, U-235, and 1.95E-06 2.79E-08 1.26E-06 1.81E-08
U-238).....................................................
----------------------------------------------------------------------------------------------------------------
5. Estimating Monetized Benefits
In this section, we summarize the information used in estimating
monetized benefits. A description of the methodology used for these
estimates is found in the Technical Support Document (EPA 2000a),
which provides background information on: (1) The economic concepts
that provide the foundation for benefits valuation; (2) the methods
that are typically used by economists to value risk reductions, such
as wage-risk, cost of illness, and contingent valuation studies; (3)
the approach for valuing the reductions in fatal cancer risks and
nonfatal cancer risks; (4) the use of these techniques to estimate
the value of the risk reductions attributable to the regulatory
options for radionuclides in drinking water; and (5) the limitations
and uncertainties involved in the estimation. For more detail on the
methodology employed, see the Health Risk Reduction and Cost
Analysis (HRRCA, EPA 2000b).
This benefits analysis is based on two basic types of valuation:
fatal cancer risk reductions and non-fatal cancer risk reductions.
Fatal cancer risk reductions are valued in terms of the ``value of a
statistical life'' (VSL), which does not refer to the value of an
identifiable individual, but rather refers to the value of small
reductions in mortality risks over a large population. For example,
let us assume that a regulatory option results in a risk reduction
of ``one statistical fatal cancer case.'' This refers to the
summation of small risk reductions over a large number of persons
such that the summation equals ``one case'' (say, one hundred
thousand persons each face a risk reduction of 1/100,000). Using our
methodology, the resulting benefits would be equal to ``one
statistical life.'' Continuing the example, if each person were
willing to pay $20 for such a risk reduction (1/100,000), the
resulting VSL would be $2 million ($20 times 100,000 persons).
However, since there is no direct information on what persons are
willing to pay for the risks we are interested in, we must use
indirect methods for estimating the VSL. The currently accepted
methodology involves transferring the VSL from studies of the wage
increases that persons ``demand'' in exchange for accepting jobs
with slightly higher chances of accidental fatality (``wage-risk
studies''). There are a number of assumptions involved in making
this transfer, which are discussed in more detail in the background
documentation (EPA 2000a and 2000b).
Valuing nonfatal cancer risk reductions is often done with
``cost of illness studies,'' which examine the actual direct (e.g.,
medical expenses) and indirect (e.g., lost work or leisure time)
costs incurred by affected individuals. Unfortunately, this
valuation does not measure the ``willingness to pay'' to avoid
nonfatal cancers, but rather assumes that benefits are equal to the
avoided costs. The studies used and assumptions involved are
discussed elsewhere (EPA 2000a and 2000b).
Because of the uncertainties involved in valuations, we used an
estimate of the range of values of reductions in fatal and non-fatal
risks attributable to the radionuclides regulations using the
following estimates (1998 dollars):
Fatal Risk Reduction Valuations (``Value of a Statistical
Life'', VSL):
Best Estimate: Value of fatal risk reductions = Statistical
lives saved * $5.9 million per statistical life.
Low End Estimate: Value of fatal risk reductions = Statistical
lives saved * $1.5 million per statistical life.
High End Estimate: Value of fatal risk reductions = Statistical
lives saved * $11.5 million per statistical life.
Non-Fatal Risk Reduction Valuations
Best Estimate: Value of nonfatal risk reductions (medical costs
only) = Statistical cases averted * $0.10 million.
Low End Estimate: Value of nonfatal risk reductions (medical
costs only) = Statistical cases averted * $0.09 million.
High End Estimate: Value of nonfatal risk reductions (medical
costs only) = Statistical cases averted * $0.11 million.
6. Estimating the Costs of Compliance
The last component of the analysis involves estimating the costs
of compliance for each regulatory option. The options under
consideration will increase the costs of monitoring for all
regulated systems, as well as require a small fraction of the
systems to take action to reduce the contaminant levels in their
finished water to achieve compliance. Examples of compliance actions
include installing treatment, purchasing water from another system,
changing the water source used (e.g., installing a new well),
blending the contaminated water with other source water that is
below the MCL, and, in cases where the contaminated well is not
essential to meet capacity, stopping production from the
contaminated well. The cost analysis models both new capital costs
and, and when appropriate, incremental operations and maintenance
costs for this variety of compliance options. The inputs used in the
cost analysis and a comparison of the modeled costs for treatment,
alternate source, purchased water to case studies can be found in
the Technical Support Document (EPA 2000a) and elsewhere (EPA
1998a).
C. Summary of Annual Costs and Benefits
1. Estimates of Costs and Benefits for Community Water Systems
The following results reflect the regulatory options that are
currently being considered. Results for the other options that were
analyzed (correction of monitoring deficiencies for gross alpha and
changes to MCLs for gross alpha and Ra-228), but that EPA does not
plan to adopt, are located in the Technical Support Document (EPA
2000a). In addition to EPA's preferred options, we have included all
results in the Technical Support Document to allow interested
stakeholders to comment on these other options, if desired.
Table V-2 shows the summarized results for EPA's analysis of
risk reductions, benefits valuations, and costs of compliance (see
EPA 2000b for a break-down of the summary by water system size). The
risk reductions and cost estimates are based on the estimated range
of numbers of community water systems predicted to be out of
compliance with each of the regulatory options assessed. The ranges
shown reflect the two occurrence model methodologies previously
described, the ``direct proportions'' and ``lognormal model''
approaches. The ranges in occurrence predictions necessarily result
in ranges of estimates for risk reductions, benefits valuations, and
compliance costs. There are two ranges shown for values of cancer
cases avoided, the ``best-estimate range,'' based on the best-
estimate of risk reduction valuations, and the ``low/high-estimate
range,'' which reflects the use of the two occurrence models and the
uncertainty in the risk reduction valuations (``low-end'' versus
``high-end'' estimates). These ranges do not reflect uncertainty in
other model inputs, like risk factors in the case of risk reduction
estimates and treatment unit costs in the case of compliance costs.
Quantitative uncertainty analyses for risk reductions, benefits, and
compliance costs will be conducted and reported in the preamble to
the final rule. EPA expects that these uncertainty analyses will not
impact final decisions.
Eliminating the combined radium-226/-228 monitoring deficiency
\9\ is predicted to lead to 210 to 250 systems out of compliance
with an MCL of 5 µg/L, affecting 33,000 to 460,000
[[Page 21624]]
persons. Implementing an MCL of 20 µg/L for uranium is
predicted to impact 830 to 970 systems, affecting 470,000 to
2,100,000 persons. An MCL for uranium of 40 µg/L is
predicted to impact 300 to 430 systems, affecting 47,000 to 850,000
persons; 80 µg/L is predicted to impact 40 to 170 systems,
affecting 7,000 to 170,000 persons. These estimates for uranium are
based on the assumption that the activity-to-mass ratio in drinking
water is 1:1. EPA's current best-estimate for the average activity-
to-mass ratio for the various uranium isotopes in drinking water is
0.9. EPA will update this assumption for the uranium options in the
Regulatory Impact Assessment supporting the rule finalization.
However, the impact is expected to be small. For example, using the
lognormal occurrence distribution model for the 40 µg/L
option, an assumption of an activity-to-mass ratio of 0.9 results in
an estimated number of impacted systems of 370, a decrease of only
12-13%.
---------------------------------------------------------------------------
\9\ The monitoring deficiency will be corrected by requiring the
separate analysis of Ra-228 for systems with gross alpha levels
below 5 pCi/L.
---------------------------------------------------------------------------
The estimated risk reduction range for the option addressing the
combined radium monitoring deficiency is 0.3 to 0.5 cancer cases
avoided annually, with an associated annual monetized benefits range
of one to two million dollars. The risk reductions estimated for the
uranium options range from 0.2 to 2 cases avoided annually for an
MCL of 20 mg/L, 0.04 to 1.5 cases avoided annually for an MCL of 40
µg/L, and 0.01 to 1 case avoided annually for an MCL of 80
µg/L. The associated annual monetized benefits for the
uranium options range from 0.6 to 8 million dollars (20 mg/L), 0.1
to 6 million dollars (40 µg/L), and less than 0.1 to 4
million dollars (80 µg/L).
Annual compliance costs range from 20 to 30 million dollars for
the option addressing the combined radium monitoring deficiencies.
Annual compliance costs for the uranium options range from 30 to 140
million dollars for an MCL of 20 mg/L, 6 to 60 million dollars for
an MCL of 40 µg/L, and 5 to 30 million dollars for an MCL of
80 µg/L.
As demonstrated by this analysis the estimated range of central-
tendency annual compliance costs exceed the ranges of central-
tendency annual monetized benefits for all options. This is not
surprising given that most of the systems impacted are small water
systems, which tend to have much higher per customer compliance
costs relative to large systems, while the per customer risk
reduction is independent of water system size. Except in cases where
risk reductions are quite large, it is predictable that estimated
annual costs will outweigh estimated annual benefits for small water
systems (given the current methodologies for estimating benefits).
However, it should be pointed out that all of the regulatory options
being considered have associated lifetime morbidity risks near or in
excess of one in ten thousand, which is the upper bound on the
preferred risk range according to EPA's policies on regulating
drinking water contaminants. In the case of uranium, it is also
important to recognize that there may be considerable non-quantified
(not monetizable) benefits associated with reductions in kidney
toxicity risks. If such benefits were quantified, it is likely that
the net benefits would be more favorable for all uranium options.
Some commenters may argue that costs and benefits considerations
should lead to the conclusion that the finalization of the
correction of the combined radium monitoring deficiencies and/or the
establishment of a NPDWR for uranium are not warranted. However,
this conclusion would lead to a situation where customers of many
ground water systems face lifetime morbidity risks greatly in excess
of the acceptable risk upper limit of one in ten thousand. According
to EPA's policies, the proper use of this flexibility should lead to
regulatory decisions that have associated risks that are within or
acceptably close to EPA's longstanding goals of limiting excess
lifetime morbidity risks to the range of one in a million to one in
ten thousand, except under unusual circumstances. EPA solicits
comment on this interpretation of costs and benefits for the
finalization of the 1991 radionuclides proposal.
BILLING CODE 6560-50-U
[[Page 21625]]
[GRAPHIC] [TIFF OMITTED] TP21AP00.012
BILLING CODE 6560-50-C
[[Page 21626]]
2. Uncertainties in the Estimates of Benefits and Costs
The models used to estimate costs and benefits related to
regulatory measures have uncertainty associated with the model
inputs. The types and uncertainties of the various inputs and the
uncertainty analyses for risks, benefits, and costs are
qualitatively discussed later in this section.
a. Uncertainties in Risk Reduction Estimates. For each
individual radionuclide, EPA developed a central-tendency risk
coefficient that expresses the estimated probability that cancer
will result in an exposed individual per unit of radionuclide
activity (e.g., per pCi/L) over the individual's lifetime (assumed
to be 70 years). Two types of risks are considered, cancer
morbidity, which refers to any incidence of cancer (fatal or non-
fatal), and cancer mortality, which refers to a fatal cancer
illness. For this analysis, we used the draft September 1999 risk
coefficients developed as part of EPA's revisions to Federal
Guidance Report 13 (FGR-13, EPA 1999e). FGR-13 compiled the results
of several models predicting the cancer risks associated with
radioactivity. The cancer sites considered in these models include
the esophagus, stomach, colon, liver, lung, bone, skin, breast,
ovary, bladder, kidney, thyroid, red marrow (leukemia), as well as
residual impacts on all remaining cancer sites combined.
There are substantial uncertainties associated with the risk
coefficients in FGR-13 (EPA 1999e): researchers estimate that some
of the coefficients may change by a factor of more than 10 if
plausible alternative models are used to predict risks. While the
report does not bound the uncertainty for all radionuclides, it
estimates that the central-tendency risk coefficients for uranium-
234 and radium-226 may change by a factor of seven depending on the
models employed to estimate risk.\10\ Ranges that reflect
uncertainty and variability in the risk coefficients will be used in
a Monte Carlo analysis of risk reductions and benefits, the results
of which will be reported in the preamble to the final rule.
---------------------------------------------------------------------------
\10\ Table 2.4, Uncertainty Categories for Selected Risk
Coefficients. Federal Guidance Report 13 (1999).
---------------------------------------------------------------------------
In addition, as previously described in appendix I,
``Occurrence,'' the available occurrence data do not provide
information on the contribution of individual radionuclides or
isotopes to the total concentrations of gross alpha or uranium.
Therefore, there is uncertainty involved in the assumptions about
which radionuclides comprise the reported gross alpha or uranium
activity. These and other uncertainties related to occurrence
information (e.g., uncertainty in extending the NIRS database
results to the national level) will also be incorporated in a Monte
Carlo analysis of benefits to estimate the range of uncertainty
surrounding the central-tendency estimates. Other inputs that will
be used in the Monte Carlo analysis of benefits are the age- and
gender-dependent distributions of water ingestion, which are used in
estimating lifetime exposure, and the credible range for the ``value
of a statistical life.'' This uncertainty analysis is not expected
to alter the regulatory options discussed in today's NODA.
b. Effects of the Inclusion of a Latency Period and Other
Factors on the Estimate of Benefits. The expected analytical impacts
of the inclusion of other factors, e.g., a cancer-latency period,
cancer premiums, and non-quantifiable benefits have been discussed
in the recent radon proposed NPDWR (64 FR 59295). The relevant
points are summarized briefly here and in more detail in the
Technical Support Document for the Radionuclides NODA (USEPA 2000a).
There are several potentially important sources of uncertainty
related to the valuations of risk reductions for the regulatory
options examined. Since the mortality valuations dominate the
estimated benefits, factors that affect the VSL are most important.
Factors that may affect the VSL include discounting due to cancer
latency periods,\11\ cancer-related premiums that may raise the
value of statistical life, and other currently non-quantifiable
benefits. Cancer latency-related discounting would be expected to
decrease the present VSL, while cancer premiums would tend to
increase the present VSL. It is not clear whether an inclusion of
all of these factors would be expected to result in a lower or
higher present VSL. However, EPA is currently working with the
Science Advisory Board (SAB) to determine how to best include these
factors, whether the inclusion is quantitative or qualitative.
---------------------------------------------------------------------------
\11\ A latency period refers to the average amount of time that
passes between the beginning of exposure to a carcinogen or multiple
carcinogens and the on-set of fatal cancer. There is considerable
uncertainty in estimating a ``typical latency period'' for the
options studies here for many reasons, including the large ranges in
estimated latency periods for given cancer types and the large
uncertainty involved in predicting which type or types of cancer
will result from exposure to a given radionuclide in drinking water.
It is also uncertain what discounting rate would be appropriate in
this situation. Some may argue that discounting is entirely
inappropriate (a rate of zero) and others may argue that typical
financial discount rates are appropriate (3 to 7%).
---------------------------------------------------------------------------
c. Uncertainty in Compliance Cost Estimates. Regarding
uncertainty in the compliance cost estimates, these estimates assume
that most systems will install treatment to comply with the MCLs,
while recent research suggests that water systems usually select
compliance options like blending (combining water from multiple
sources), developing new ground water wells, and purchasing water
(EPA 1998a and c, EPA 2000a). Preliminary data (202 compliance
actions from 14 States) on nitrate violations suggest that only
around a quarter (25%) of those systems talking action in response
to a nitrate violation installed treatment, while roughly a third
developed a new well or wells. The remainder either modified the
existing operations (10-15%), blended (15%), or purchased water (15-
20%). Similar data for radium violations from the State of Illinois
(77 compliance actions) indicate that around a quarter of systems
taking action installed treatment, while the majority (50-55%)
purchased water, with the remainder (20-25%) either installing a new
well, blending, or stopping production from the contaminated well or
wells. The prevalence of the use of these non-treatment options is a
cross-cutting issue for future Regulatory Impact Assessments and
probably will not be resolved before the radionuclides NPDWR is
finalized. EPA is following up with this study and will report the
results at a later date.
While these ``other than treatment'' options may cost as much as
or more than treatment in some cases, they are expected to be less
expensive on average, which largely explains their prevalence as
compliance options. For example, EPA has recently estimated the
costs associated with developing municipal wells to range from
$0.08/kgal to $0.46/kgal, depending on system size, geologic
setting, and other site specific parameters (EPA 1999b), with an
average of $0.23/kgal for systems serving between 501 and 1,000
persons and $0.17/kgal for systems serving between 10,001 and 50,000
persons.\12\ These costs include testing and drilling, steel casings
with cement lining, pumps, including electrical connections and
controls, and a pump shelter. For smaller, non-municipal PWS
systems, we estimate that wells could cost from 10 to 80 percent of
the costs presented for municipal systems. As shown in the Technical
Support Document (EPA 2000a), these production costs are much lower
than those for typical treatment, especially for small systems. When
feasible, selection of such options may reduce compliance costs
significantly. The Technical Support Document includes data on other
non-treatment options like purchasing water and blending.
---------------------------------------------------------------------------
\12\ This estimate is based on total capital costs ranging from
approximately $135,000 to $550,000 per MGD of flow. The estimate
assumes typical relationships between design and average daily flows
and a capital discount rate of 3 or 7%.
---------------------------------------------------------------------------
Preliminary uncertainty analyses suggest that variability in the
unit compliance costs and decision tree assumptions dominate the
over-all cost variability. To evaluate the potential variability in
the compliance cost estimates, a Monte Carlo analysis will support
the Regulatory Impact Assessment for the final rule. Inputs that
influence cost variability include:
Numbers of total systems in the various system size
categories.
Distributions of entry points per system in the various
system size categories.
Distributions of populations served by size category.
Flow sizes as a function of population served.
Daily household water consumption.
Proportions of systems and sources exceeding regulatory
limits.
Unit costs (capital and O&M) of treatment technologies
and annual costs of alternate source and regionalization.
Proportions of non-compliant systems choosing between
treatment, alternate source, and regionalization.
Since per system costs are much higher for very large systems,
the assumptions used in the larger water system size categories can
be expected to dominate the variability in national costs. Each of
these inputs will be modeled using probability distributions that
[[Page 21627]]
reflect the state of the available data. In some cases, input
variability will be estimated from SDWIS, the CWSS, or other sources
(e.g., distributions of populations served, daily household water
consumption, unit costs) . In other cases, input variability will
have to be based on best professional judgement. Again, this
uncertainty study is expected to provide useful information, but is
not expected to result in changes to the regulatory decisions
described in today's NODA.
D. Estimates of Costs and Benefits for Non-Transient Non-Community
Water Systems
The available data are not sufficient to allow EPA to predict a
central-tendency impact of the regulatory options on non-transient
non-community water systems (NTSC systems). Instead, EPA conducted a
``what-if'' analysis of potential costs and benefits based on
reasonable assumptions of the percentage of NTSC systems impacted by
the various options (EPA 2000b). A ``what-if'' analysis allows us to
pose hypothetical occurrence scenarios and to estimate costs and
benefits for these scenarios. If the scenarios are chosen properly,
they should bound the reasonable set of potential costs and benefits
for NTSC systems. However, the estimates should not be interpreted
as representing ``best estimates,'' which would be based on an
occurrence survey of radionuclides occurring at NTSC systems. The
Technical Support Document (EPA 2000a) provides details on the
inputs and assumptions used for estimating regulatory impacts for
NTSC systems. The resulting estimates of the percentage of systems
out of compliance are provided in Table V-3.
Table V-3.--Assumptions for Hypothetical ``What-If'' Analysis for Non-Transient Non-Community Water Systems
(Approximately 19,300 Systems Nationwide)
----------------------------------------------------------------------------------------------------------------
Percent of
national
systems in Upper bound: Lower bound:
Regulatory option states with 10% of col. 1% of col. (1)
elevated (1) (percent) (percent)
levels (1)
(percent)
----------------------------------------------------------------------------------------------------------------
Gross Alpha at 15 pCi/L......................................... 60 6 1
Combined Radium at 5 pCi/L...................................... 79 8 1
Uranium at 20 pCi/L:
Ground water................................................ 54 5 1
Surface water............................................... 29 3 0
----------------------------------------------------------------------------------------------------------------
We calculated risk reductions associated with each set of
assumptions using the same analytic approach as outlined for the
community water systems. However, we use lower water intake
assumptions because the population affected generally is not at the
location served full-time or year-round. The risk factors were
estimated using the same risk coefficients as a starting point (risk
per pCi), but use different water consumption assumptions to
calculate lifetime excess risk factors (risk per pCi/L). A cost
model is used to predict the annual compliance costs for these
systems based on their size classes (EPA 2000); in general, non-
transient non-community systems tend to use ground water and serve
small populations.
The results of the analysis are summarized in Table V-4. If EPA
requires non-transient non-community systems to comply with the
gross alpha standard of 15 pCi/L, under the assumptions used in the
analysis the number of systems out of compliance could range from
110 to 1,100 systems. The associated annual costs range from $1
million to $4 million and the statistical cancer cases (fatal and
nonfatal) avoided annually range from 0.01 cases to 0.1 cases. For
combined radium, the resulting number of impacted systems ranges
from 150 to 1,500 systems with annual costs ranging from $1 million
to $6 million and an associated number of annual statistical cancer
cases avoided ranging from 0.02 cases to 0.2 cases. For a uranium
MCL of 20 µg/L, the results suggest a range of impacted
ground water systems from 100 up to 1,000 systems with annual costs
ranging from $1 million to $4 million and an associated number of
annual statistical cancer cases avoided ranging from less than 0.01
cases up to 0.04 cases. The resulting number of surface water
systems impacted by a uranium MCL of 20 µg/L ranges from
less than 10 to less than 20 systems. The associated national annual
costs for surface water systems is less than $0.1 million up to 0.1
million with annual risk reductions of less than 0.01 statistical
cancer cases.
Table V-4.--Hypothetical ``What-If'' Results for Non-Transient Non-Community Water Systems
--------------------------------------------------------------------------------------------------------------------------------------------------------
Regulatory option Lower Bound Estimate Upper Bound Estimate
--------------------------------------------------------------------------------------------------------------------------------------------------------
Number of Statistical Number of
systems out Annual costs cancer cases systems out Annual costs Statistical
of (million avoided of (million cancer cases
compliance dollars) (cases) compliance dollars) avoided
--------------------------------------------------------------------------------------------------------------------------------------------------------
Gross Alpha at 15 pCi/L............................................ 110 1 0.01 1,100 4 0.1
Combined Radium at 5 pCi/L......................................... 150 1 0.02 1,500 6 0.2
Uranium at 20 pCi/L:
Ground water................................................... 100 1 0.01 1,000 4 0.04
Surface water.................................................. 10 0.03 0.01 20 0.1 0.01
--------------------------------------------------------------------------------------------------------------------------------------------------------
Note: These results are based on hypothetical assumptions regarding the percent of systems likely to be out of compliance with each regulatory option as
discussed in the preceding text. These are not estimates of actual compliance costs or risk reductions, and are provided for illustrative purposes
only.
E. Impacts for Systems Serving Greater Than One Million Persons
Based on an Internet search of the available water quality
information for water systems serving greater than one million
persons (very large systems), there is no direct evidence that
closing the monitoring deficiencies for radium will impact these
systems. However, the internet search was not conclusive in ruling
out the possibility that one or more systems serving greater than
one million persons would be impacted by these options. For this
reason, EPA has followed up with the few systems in question to
determine the likelihood of impact. The follow-up confirmed that
there were no impacts expected for these systems. Uranium occurrence
data for these systems was collected to the extent feasible and
there is no evidence of an impact at 20 or 40 µg/L.
[[Page 21628]]
F. References
Longtin, J.P. ``Occurrence of Radon, Radium, and Uranium in
Groundwater.'' JAWWA. Vol. 80, No. 7, pp. 84-93. July 1988.
National Research Council. Risk Assessment of Radon in Drinking
Water. National Academy Press. Washington, DC. 1999. (NAS 1999).
USEPA. National Primary Drinking Water Regulations; Radionuclides;
Proposed Rule. Federal Register. Vol. 56, No. 138, p. 33050. July
18, 1991.
USEPA. ``Actual Cost for Compliance with the Safe Drinking Water Act
Standard for Radium-226 and Radium-228--Final Report.'' Prepared by
International Consultants, Inc. for the Office of Ground Water and
Drinking Water. Draft. July 1998. (EPA 1998a)
USEPA. ``Evaluation of Full-Scale Treatment Technologies at Small
Drinking Water Systems: Summary of Available Cost and Performance
Data.'' Prepared by ICF, Inc. and ISSI, Inc. for the Office of
Ground Water and Drinking Water. December 10, 1998. (EPA 1998b)
USEPA. Results from survey of compliance actions taken in the State
of Illinois in response to the NPDWR for combined radium. Submitted
from EPA Region 5. 1998. (EPA 1998c)
USEPA. ``Guidelines for Preparing Economic Analysis.'' Review Draft.
November 3, 1998. (EPA 1998d)
USEPA. ``Regional Variation of the Cost of Drinking Water Wells for
Public Water Supplies.'' Prepared by Cadmus for the Office of Ground
Water and Drinking Water. Draft. October 1999. (EPA 1999b)
USEPA. ``Drinking Water Baseline Handbook: First Edition.'' Draft
dated March 2, 1999. (EPA 1999c)
USEPA. ``Evaluation of Central Treatment Options as Small System
Treatment Technologies.'' Prepared by SAIC for EPA. Draft dated
January 28, 1999. (EPA 1999d)
USEPA. Cancer Risk Coefficients for Environmental Exposure to
Radionuclides, Federal Guidance Report No. 13. US Environmental
Protection Agency, Washington, DC, 1999. (EPA 1999e)
USEPA. ``Technical Support Document for the Radionuclides Notice of
Data Availability.'' Draft. March, 2000. (EPA 2000a)
USEPA. ``Preliminary Health Risk Reduction and Cost Analysis:
Revised National Primary Drinking Water Standards for
Radionuclides.'' Prepared by Industrial Economics, Inc. for the
Office of Ground Water and Drinking Water. Draft. January 2000. (EPA
2000b)
Dated: April 7, 2000.
Charles J. Fox,
Assistant Administrator, Office of Water.
[FR Doc. 00-9654 Filed 4-20-00; 8:45 am]
BILLING CODE 6560-50-U
![[logo] US EPA](http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}