[Federal Register Volume 65, Number 78 (Friday, April 21, 2000)]
[Proposed Rules]
[Pages 21576-21628]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 00-9654]



[[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

  Federal Register / Vol. 65, No. 78 / Friday, April 21, 2000 / 
Proposed Rules  

[[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 [email protected]. 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 [email protected]. 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
Ci--microCurie
Ci/kg--microCurie per kilogram
g or ug--microgram
g/g or ug/g--microgram per gram
g/L or ug/L--microgram per liter
g uranium/L--microgram uranium per liter
g uranium/kg/day--microgram uranium per kilogram per day
R/hr--micro Roentgen per hour
Sv/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/g--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--2-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--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 g/L or 30 pCi/  Three options being
                                                           L w/ option for 5-80        considered: 20, 40, 80
                                                           g/L.               g/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 (g) while the radiation 
activity is measured in picoCuries. In 1991, EPA proposed a limit on 
uranium of 20 g per liter (g/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 g/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 g/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 g. 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 g/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 g/L are low (0.2 to 2 cases for 20 g/L 
and 0.04 to 1.5 cases for 40 g/L). In terms of exposure 
avoided for kidney toxicity, around 500 thousand to two million persons 
are exposed above 20 g/L and 50 thousand to 900 thousand 
persons are exposed above 40 g/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 g/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 g/L) 
was based on uranium kidney toxicity, as previously described. The 
corresponding risk of cancer at a concentration of 20 g/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 g/L and 80 g/L. EPA 
estimates that a level of 40 g/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 g/L would represent a slightly higher risk of kidney 
toxicity. At a level of 80 g/L, the cancer mortality risk is 
approximately 2  x  10-4, which is above the Agency's 
acceptable risk range. At 80 g/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 g/L than at an MCL of 40 
g/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 g/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 g/L, 40 g/L, and 80 g/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 g/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 g/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 g/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/g 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 g/L from mass to activity using a ratio of 
0.9 for levels above 3.5 g/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 
g/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 (g/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 g/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 g/L could be exceeded. If the 
gross alpha test is between 3 and 5.5 pCi/L, the mass of 20 g/
L could be exceeded if all the activity were coming from uranium-238. 
Therefore a fluroimetric test for uranium mass concentration 
(g/L) or an alpha spectrometry test for the activities (pCi/L, 
converted to g/L using standard isotopic conversion factors) 
of the various isotopes present would be necessary to determine the 
uranium concentration in g/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 g/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 g/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 g/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 g/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 g/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 , where  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 g/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 g/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 > 1g/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 
g/L; or (3) the establishment of a uranium NPDWR with an 
MCL of 40 g/L; or (4) the establishment of a uranium NPDWR 
with an MCL of 80 g/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 g/L, affecting 33,000 to 460,000

[[Page 21624]]

persons. Implementing an MCL of 20 g/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 g/L is 
predicted to impact 300 to 430 systems, affecting 47,000 to 850,000 
persons; 80 g/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 g/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 
g/L, and 0.01 to 1 case avoided annually for an MCL of 80 
g/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 g/L), and less than 0.1 to 4 
million dollars (80 g/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 g/L, and 5 to 30 million dollars for an MCL of 
80 g/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 g/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 g/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 g/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