[[pp. 59295-59344]] National Primary Drinking Water Regulations; Radon-222

Note: EPA no longer updates this information, but it may be useful as a reference or resource.

[Federal Register: November 2, 1999 (Volume 64, Number 211)]
[Proposed Rules]
[Page 59295-59344]
From the Federal Register Online via GPO Access [wais.access.gpo.gov]
[DOCID:fr02no99-36]

[[Continued from page 59294]]

[[Page 59295]]

In discussions between EPA and the water utility industry, concerns
have been expressed about the difficulties in collecting samples and
the requisite skills that may be required. EPA emphasizes that the
skills required to sample for radon are the same as those required to
sample for other currently regulated drinking water contaminants,
namely volatile organic contaminants. In addition, the 1992 EPA
collaborative study mentioned earlier evaluated four sample collection
techniques and found them all capable of providing equivalent results.
Supplementing this study, EPA has reviewed a sampling protocol for
radon in water developed by the Department of Health Services Division
of Drinking Water and Environmental Management (CA DHS 1998). This
protocol employs one of the four techniques evaluated by EPA, the
immersion technique.
Using the immersion technique, the well is purged for 15 minutes by
running the sampling tap, to ensure that a representative sample is
collected. After the purging period, a length of flexible plastic
tubing is attached to the spigot, tap, or other connection, and the
free end of the tubing is placed at the bottom of a small bucket. The
water is allowed to fill the bucket, slowly, until the bucket
overflows. The bucket is emptied and refilled at least once.
Once the bucket has refilled, a glass sample container of an
appropriate size is opened and slowly immersed into the bucket in an
upright position. Once the bottle has been placed on the bottom of the
bucket, the tubing is placed into the bottle to ensure that the bottle
is flushed with fresh water. After the bottle has been flushed, the
tubing is removed while the bottle is resting on the bottom of the
bucket. The cap is placed back on the bottle while the bottle is still
submerged, and the bottle is tightly sealed. As noted in the California
protocol cited earlier, the choice of the sample container is dependent
on the laboratory that will perform the analysis, and will be a
function of the liquid scintillation counter that is employed. If
bottles are supplied by the laboratory, there is no question of what
container to employ.
Once the sealed sample bottle is removed from the bucket, it is
inverted and checked for bubbles that would indicate headspace. If
there are no visible air bubbles, the outside of the sealed bottle is
wiped dry and cap is sealed in place with electrical tape, wrapped
clockwise. After the sample bottle is sealed, a second (duplicate)
sample is collected in the same fashion from the same bucket. The date
and time of the sample collection is recorded for each sample.
As can be surmised from the description, the sample collection
procedures are not particularly labor intensive. Most of the time is
spent allowing the water to overflow the bucket. Likewise, there are no
significant manual skills required.
(e) Skill Considerations for Laboratory Personnel. While neither of
these techniques is difficult relative to standard drinking water
methods, a discussion of the skills required to employ the methods is
appropriate. Given the long history of successful use of the liquid
scintillation counting technique (it has been used in medical
laboratories and environmental research laboratories for well over 30
years), EPA feels confident that State drinking water laboratories will
be able to adequately use these methods. The skills required are
primarily the ability to transfer and mix aliquots of the sample to a
sealed container for further analysis. The counting process is highly
automated and the equipment runs unattended for days, if needed.
The de-emanation process requires somewhat more manual skill. As
noted in the 1991 proposed rule, EPA expects that this technique would
require greater efforts be made to train technicians than for the
liquid scintillation technique. The technique requires that the
counting cell be evacuated to about 10 mTorr pressure and then a series
of stopcocks or valves are manipulated to transfer the radon that is
purged from the sample into the counting cell. Potential problems with
the analysis, such as a high background level of radon that can develop
over the course of the day, or aspirating water into the counting cell,
can be minimized by a well-trained analyst. However, as EPA concluded
in 1991, the Lucas cell technique is not expected to form the sole
basis of a compliance monitoring program for radon in drinking water.
(f) Cost of Performing Analyses. The actual costs of performing
analysis may vary with laboratory, analytical technique selected, the
total number of samples analyzed by a lab, and by other factors. Based
upon information collected in 1991, the average sample cost for radon
in water was estimated to be $50 per sample. EPA recently updated this
cost estimate to $57 per sample (USEPA 1999b) by conducting a similar
survey of drinking water laboratories. The data from the 1991 and 1998
surveys and the descriptive statistics are summarized in Table
VIII.B.2. There was no clear correlation between the estimated price
and the method cited by the laboratory. The 1998 range of prices
brackets those collected by EPA in 1991. It is expected that the
``market forces'' generated by a radon regulation will tend to lower
per sample costs, especially in light of the fact the LSC is very
amenable to automation, with feed capacities of more than 50 samples/
load possible. However, as will be discussed later, there may be short-
term laboratory capacity issues that resist a lowering of per sample
prices.

Table VIII.B.2. Radon Sample Cost Estimate
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Cost Year data
Arbitrary lab No. estimate collected Descriptive statistics for 1991
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1..................................... $30 1991 Mean, $49.80; Median, $47.00; Std. Dev., $18.80; Range, $45; Minimum, $30; Maximum, $75.
2..................................... 44 1991
3..................................... 50 1991
4..................................... 75 1998
Descriptive Statistics for 1998 Data
5..................................... 75 1998 Mean, $56.88; Median, $52.50; Std. Dev., $15.80; Range, $35; Minimum, $40; Maximum, $75.
6..................................... 50 1998
7..................................... 40 1998
8..................................... 75 1998
9..................................... 45 1998
10.................................... 55 1998
11.................................... 75 1998
12.................................... 40 1998
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------

[[Page 59296]]

These cost data are preliminary and may be different in practice
for the following reasons: (a) As the number of experienced
laboratories increases, the costs can be expected to decrease; (b)
analytical costs are determined, to some extent, by the quality control
efforts and quality assurance programs adhered to by the analytical
laboratory; (c) per-sample costs are influenced by the number of
samples analyzed per unit time. EPA solicits comments on its cost
estimates from laboratories experienced in performing these analyses.
(g) Method Detection Limits and Practical Quantitation Levels.
Method detection limits (MDLs) and practical quantitation levels (PQLs)
are two performance measures used by EPA to estimate the limits of
performance of analytic chemistry methods for measuring contaminants in
drinking water. An MDL is the lowest level of a contaminant that can be
measured by a specific method under ideal research conditions. EPA
usually defines the MDL as the minimum concentration of a substance
that can be measured and reported with 99 percent confidence that the
true value is greater than zero. The term MDL is used interchangeably
with minimum detectable activity (MDA) in radionuclide analysis, which
is defined as that amount of activity which in the same counting time,
gives a count which is different from the background count by three
times the standard deviation of the background count. A PQL is the
level at which a contaminant can be ascertained with specified methods
on a routine basis (such as compliance monitoring) by accredited
laboratories, within specified precision and accuracy limits.
The feasibility of implementing an MCL at a particular level is in
part determined by the ability of analytical methods to ascertain
contaminant levels with sufficient precision and accuracy at or near
the MCL. The proposed methods demonstrate good reproducibility and
accuracy at radon concentrations in the range of 150-300 pCi/L (half of
the proposed MCL up to the proposed MCL), as demonstrated in the
results from inter-laboratory studies. In inter-laboratory studies (or
Performance Evaluation studies), prepared samples of known
concentration are distributed for analysis to participating labs, which
have no information on the concentrations of the samples. The results
of the analyses by the participants are compared with the known value
and with each other to estimate the precision and accuracy of both the
methods used and the lab's proficiency in using the method. Table
VIII.B.3 summarizes the statistical results of these inter-laboratory
studies for the proposed methods.
In the 1991 proposed rule, EPA proposed using both the MDL and PQL
as measures of performance for radon analytical methods. EPA also
proposed acceptance limits based on the PQLs that were derived from
these performance evaluation studies. The use of acceptance limits was
confusing to commenters for various reasons. The important issue is the
observation that true analytical method performance is related to
within-laboratory conditions (including counting times in the case of
radiochemicals) and that acceptance limits are based on multi-
laboratory Performance Evaluation studies. For non-radiochemical
contaminants this issue is less troublesome because their PQLs tend to
be ``fixed'' since the MDLs to which they are related reflect optimized
conditions for standard laboratory equipment, whereas for radiochemical
contaminants, counting times can always be increased to increase the
sensitivity and hence lower the appropriate acceptance limits. While
the fifty minute counting time in Standard Method 7500-Rn reflects a
balanced trade-off between time of analysis (and hence the cost of
analysis) and sensitivity, it can obviously be adjusted as needed to
adjust sensitivity. For this reason, commenters objected to the use of
acceptance limits (and, relatedly, PQLs) for radiochemical
contaminants.
EPA agrees that these comments have merit and has decided to seek
comment on two proposals regarding the use of acceptance limits and
PQLs for radon. The first proposal, and the preferred option, is to not
use acceptance limits or PQL for radon, and to adopt the detection
limit as the measure of sensitivity, as done in the 1976 Radionuclides
rule. The existing definition of the detection limit takes into account
the influence of the various factors (efficiency, volume, recovery
yield, background, counting time) that typically vary from sample to
sample. Thus, the detection limit applies to the circumstances specific
to the analysis of an individual sample and not to an idealized set of
measurement parameters, as with acceptance limits and PQLs. The
proposed detection limit is 12 +/- 12 pCi/L, which is based on the
detection limit described in SM 7500-Rn (50 minute counting time, 6 cpm
background, 2.7 cpm/dpm efficiency, and under the energy window
optimization procedure as described in the method). This detection
limit should be applicable to all three approved methods.
One of the reasons for setting a sensitivity standard is to ensure
that laboratories will perform acceptably well on a routine basis at
contaminant levels near the MCL. Internal quality control/quality
assurance procedures are of paramount importance. In addition,
Proficiency Tests are administered by laboratory certifying authorities
to ensure that laboratory performance is acceptable. Currently, the
system for administering proficiency tests and certifying laboratories
is in a state of transition. Up to the recent past, all primacy
entities evaluated laboratory performance based on EPA's Performance
Evaluation (PE) studies program, the National Exposure Research
Laboratory (NERL-LV) Performance Evaluation (PE) Studies program for
radioactivity in drinking water. Currently, the Proficiency Testing
(PT) program for radionuclides is being privatized, i.e., operated by
an independent third party provider accredited by the National
Institute of Standards and Technology (NIST). A lack of uniformity in
state PT requirements may limit laboratory availability for a given
public water system to laboratories that use PT samples approved by the
state. It should be noted that this issue is general and is not
specific to the proposed radon regulation. Efforts to encourage
uniformity in state PT requirements are described in more detail in the
laboratory capacity section.
Under the alternative of using the MDL as the measure of
sensitivity, standard statistical procedures would be used to ensure
that a laboratory has analyzed PT samples acceptably. Since the
national PT program will still be overseen by EPA, the exact procedures
for determining acceptable performance will be developed by EPA and
NIST as the PT program develops. The respective roles of EPA and NIST
in the PT program and discussed further in the Laboratory Approval and
Certification section.
The second proposal is to use the concepts of the acceptance limit
and PQL for radon. Using the standard relationship that PQLs are equal
to 5 to 10 times the MDL yields a PQL for radon in the range of 60 to
240 pCi/L. EPA is proposing a PQL of 100 pCi/L and is seeking comment
on this value. The proposed acceptance limit for a single sample is
5 %. The proposed acceptance limits for triplicate analyses
at the 95th and 99th percent confidence intervals are 6 %
and 9 %, respectively. All of these acceptance limits are
based on the inter-laboratory studies used for the precision and
accuracy results reported in Table

[[Page 59297]]

VIII.B.3. EPA seeks comments on the relative merits between the first
option (the preferred option) of using only an MDL as the measure of
sensitivity and the second option of using a PQL with prescribed
acceptance limits.

Table VIII.B.3.--Inter-laboratory Performance Data for Proposed Radon Analytical Methods \1\
----------------------------------------------------------------------------------------------------------------
Sample
Method Conc. pCi/ Accuracy % Repeatability Reproducibility Bias %
L pCi/L pCi/Ls
----------------------------------------------------------------------------------------------------------------
SM 7500-Rn............................... 111 101-102 9 12 0.7-2.3
SM 7500-Rn............................... 153 102-103 10 16-18 2.3-3.4
De-Emanation............................. 111 114 16 23 14.5
De-Emanation............................. 153 114 17 28 13.7
ASTM D5072-92............................ 1,622 97 2,217 3,541 -2.6
ASTM D5072-92............................ 16,324 95 14,950 44,400 -4.7
ASTM D5072-92............................ 66,324 94 49,190 210,350 -6.0
----------------------------------------------------------------------------------------------------------------
Notes: (1) All results are reported in methods citations found in Table VIII.B.1.

(h) Accuracy and Precision of the Proposed Methods. While SM 7500-
Rn has the best over-all results in precision and accuracy, the de-
emanation method also shows acceptable performance. The ASTM method
shows similar accuracy and bias, but much larger errors in
repeatability (operator precision) and reproducibility (between-lab
precision). Given this inferior demonstration of precision and the
higher concentrations used in the intra-laboratory studies, it may be
argued that this method should not be proposed as a drinking water
method. However, EPA maintains that the method is similar enough in
substance to SM 7500-Rn that it may serve as an alternate method if the
laboratories use the appropriate quality control measures, i.e., ensure
that the relative percent difference between results on duplicate
samples is within the counting uncertainty 95% confidence interval,
where at least 10% of daily samples are duplicates. This procedure is
described in the 4th edition of the Manual for the Certification of
Laboratories Analyzing Drinking Water, Criteria and Procedures Quality
Assurance (EPA 1997). EPA requests comment on including ASTM D5072-92
as an alternate test method.

C. Laboratory Approval and Certification

1. Background
The ultimate effectiveness of the proposed regulations depends upon
the ability of laboratories to reliably analyze contaminants at
relatively low levels. The Drinking Water Laboratory Certification
Program is intended to ensure that approved drinking water laboratories
analyze regulated drinking water contaminants within acceptable limits
of performance. The Certification Program is managed through a
cooperative effort between EPA's Office of Ground Water and Drinking
Water and its Office of Research and Development. The program
stipulates that laboratories analyzing drinking water compliance
samples must be certified by U.S. EPA or the State. The program also
requires that certified laboratories must analyze PT samples, use
approved methods, and States must also require periodic on-site audits.
External checks of performance to evaluate a laboratory's ability
to analyze samples for regulated contaminants within specific limits is
one of the means of judging lab performance and determining whether to
grant certification. Under a PT program, laboratories must successfully
analyze PT samples (contaminant concentrations are unknown to the
laboratory being reviewed) that are prepared by an organization that is
approved by the primacy entity. Successful annual participation in the
PT program is prerequisite for a laboratory to achieve certification
and to remain certified for analyzing drinking water compliance
samples. Achieving acceptable performance in these studies of known
test samples provides some indication that the laboratory is following
proper practices. Unacceptable performance may be indicative of
problems that could affect the reliability of the compliance monitoring
data.
EPA's previous PE sample program and the approaches to determine
laboratory performance requirements are discussed in 63 FR 47097
(September 3, 1998, ``1998 methods update''). In that notice, EPA
amended the regulations to adopt the universal requirement for
laboratories to successfully analyze a PE sample at least once each
year, addressing the fact that the Agency has not specified PE test
frequency requirements in its current drinking water regulations.
Though not specified in the methods update regulation, PE samples may
be provided by EPA, the State, or by a third party with the approval of
the State or EPA. Under the developing PT program, NIST has accredited
a list of PT sample providers, including a radionuclides PT samples
which will apply to radon.
In addition, guidance on minimum quality assurance requirements,
conditions of laboratory inspections, and other elements of laboratory
certification requirements for laboratories conducting compliance
monitoring measurements are detailed in the 4th edition of the Manual
for the Certification of Laboratories Analyzing Drinking Water,
Criteria and Procedures Quality Assurance (EPA 1997), which can be
downloaded via the internet at ``http://www.epa.gov/OGWDW/
labindex.html''.
2. Laboratory Capacity--Practical Availability of the Methods
In order to determine the practical availability of the methods,
EPA considered three major factors. First, the availability of the
major instrumentation was reviewed. Secondly, several laboratories
performing drinking water analyses were contacted to determine their
potential capabilities to perform radon analyses. Lastly, EPA has
reviewed the current status of the privatized Performance Evaluation
studies program and the on-going measure to implement a uniform
program, highlighting the potential impacts on short-term and long-term
laboratory capacity for radon.
3. Laboratory Capacity: Instrumentation
Regarding instrumentation availability, the major instrumentation
required for LSC is the liquid scintillation counter. Automated
counters capable of what that method terms ``automatic spectral
analysis'' are available from at least a dozen suppliers. The de-
emanation Lucas cell apparatus is the same apparatus that has been used
for radium analyses for many years. In light of the wide availability
and the long history of accessibility of the proper instrumentation,
EPA believes that instrument availability should not be an issue for
radon analytical methods.

[[Page 59298]]

4. Laboratory Capacity: Survey of Potential Laboratories
In order to evaluate the availability of laboratory capacity to
perform radon analyses, EPA contacted the drinking water certification
authorities in the States of California, Maryland, and Pennsylvania.
These states were chosen based both on estimated radon occurrence and
the overall status of the programs. Ultimately, EPA collected
information on the availability and relative costs of radon analyses
for drinking water from a total of nine commercial laboratories.
Eight of the nine laboratories that were contacted do perform radon
analyses. All the laboratories were certified in one or more states to
perform radiochemical analyses. When asked what specific methods were
used, the laboratories responded with either the technique (liquid
scintillation counting) or a specific method citation. EPA Method 913
(which later was revised to become SM 7500-Rn) was cited by two of the
laboratories. EPA Method ``EERF Appendix B'' was cited by another
laboratory. The remaining laboratories indicated that they performed
liquid scintillation analyses and could accommodate requests for
methods employing that technique.
When asked about capacity, the laboratories indicated that they
each perform between 100 and 12,000 analyses per year. The latter
figure came from a laboratory that is currently involved in a large
ground water monitoring project in the western United States. The next
largest estimate was 300 samples per year. However, EPA expects that
like any other type of environmental analysis, given a regulatory
``driver'' to perform the analysis, and given the ability of LSC
analysis to be automated, the laboratory capacity will develop in a
timely manner.
EPA's 1992 Annual Report on Radiation Research and Methods
Validation reports the results of a collaborative study on radon
analysis (EPA 1993) and is another useful source of information
regarding potential radon laboratory capacity. This study employed 51
laboratories with the capability to perform liquid scintillation
analyses. This suggests that at that time there already existed a
substantial capacity for these analyses.
Further, the liquid scintillation apparatus is used for other
radiochemical analyses, including tritium. Information from EPA
regarding the performance evaluation program for tritium analyses
suggests that there are approximately 100-200 laboratories with the
necessary equipment. Much of the capacity for tritium analyses could
also be used for radon (EPA 1997). As of September 1997, 136 of 171
participating laboratories achieved acceptable results for tritium.
While the total number of participants and the number achieving
acceptable results vary between studies, the data indicate that there
is a substantial capability for liquid scintillation analysis
nationwide.
5. Laboratory Capacity: Laboratory Certification and Performance
Evaluation Studies
The availability of laboratories is also dependent on laboratory
certification efforts in the individual states with regulatory
authority for their drinking water programs. Until June of 1999, a
major component of many of these certification programs was their
continued participation in the current EPA Water Supply WS performance
evaluation (PE) program, which included radiochemistry PE studies. Due
to resource limitations, EPA has recently privatized EPA's PE programs,
including the Water Supply studies. EPA has addressed this topic in
public stakeholders meetings and in some recent publications, including
Federal Register notices and its June 1997 ``Labcert Bulletin'', which
can be downloaded from the Internet at ``http://www.epa.gov/OGWDW/
labcert3.html''. The decision to privatize the PE studies programs was
announced in the Federal Register on June 12, 1997 (62 FR 32112). This
notice indicated that in the future the National Institute of Standards
and Technology (NIST) would develop standards for private sector PT
sample providers and would evaluate and accredit these providers, while
the actual development and manufacture of PT samples would fall to the
private sector. Further information regarding the respective roles of
EPA and NIST in the privatized PT program can be downloaded from NIST's
homepage at ``http://ts.nist.gov/ts/htdocs/210/210.htm''. EPA believes
that this program will ensure the continued viability of the existing
PT programs, while maintaining government oversight.
This externalized proficiency testing program is in the process of
becoming operational. Under the externalized PT program:
EPA issues standards for the operation of the program,
NIST administers a program to accredit PT sample
providers,
Non-EPA PT sample providers develop and manufacture PT
sample materials and conduct PT studies,
Environmental laboratories purchase PT samples directly
from PT Sample Providers (approved by NIST or the State), and
Certifying authorities certify environmental laboratories
performing sample analyses in support of the various water programs
administered by the States and EPA under the Safe Drinking Water Act.
NIST is in the process of approving a provider for PT samples for
radionuclides, including radon. States also have the option of
approving their own PT sample providers. At this time, it is difficult
to speculate to what degree this externalization of the PT program will
affect short-term and long-term laboratory capacity for radon. EPA
recognizes that initial implementation problems may arise because of
the potential for near-term limited availability of radon PT samples.
EPA also recognizes that insufficient laboratory capacity may lead to a
short-term increase in analytical costs. In the absence of definitive
information regarding the future PT program, EPA solicits public
comment on this matter.
6. Efforts To Ensure a Uniform Proficiency Testing Program: NELAC
The National Environmental Laboratory Accreditation Conference
(NELAC) is also evaluating the issues surrounding privatization of the
SDWA PT program through its proficiency testing committee. NELAC serves
as a voluntary national standards-setting body for environmental
laboratory accreditation, and includes members from both state and
Federal regulatory and non-regulatory programs having environmental
laboratory oversight, certification, or accreditation functions. One of
the goals for the re-designed SDWA PT program is to be consistent with
NELAC's recommendations.
The members of NELAC meet bi-annually to develop consensus
standards through its committee structure. These consensus standards
are adopted by participants for use in their own programs in pursuit of
a uniform national laboratory accreditation program in which
environmental testing laboratories will be able to receive one annual
accreditation that is accepted nationwide. As part of its accreditation
program, NELAC is developing standards for a proficiency testing
program that addresses all fields of testing, including drinking water.
Recent meetings of the Proficiency Testing Committee of NELAC have
reviewed several important issues, including State selection of PT
sample providers and reciprocity between States.

[[Page 59299]]

These issues are described in more detail elsewhere (NELAC 1999a). The
NELAC Proficiency Testing Committee is currently drafting requirements
for radiochemical proficiency testing under SDWA. The June 15, 1999
draft (NELAC 1999b) of its radiochemical proficiency testing
requirements describes radiochemical PT sample designs, acceptance
limits, and other information.
The intent of the NELAC standards setting process is to ensure that
the needs of EPA and state regulatory programs are satisfied in the
context of a uniform national laboratory accreditation program. EPA
recognizes that cooperating with NELAC is an important part of the re-
design of the Proficiency Testing (PT) program for drinking water,
since NELAC provides a means for states, environmental testing
laboratories, and PT study providers to have direct input into the
process. It is hoped that this mutual effort will minimize the
potential disruption in the process of moving from the old EPA PE
program towards the new privatized PT program. EPA shares NELAC's goal
of encouraging uniformity in standards between primacy States regarding
laboratory proficiency testing and accreditation.
7. Laboratory Capacity: Holding Time
The short holding time for radon, 4 days in Method 7500-Rn,
presents concerns relative to the practical availability of laboratory
capacity as well. The 4-day holding time was also the focus of a number
of comments that EPA received in response to the 1991 proposed rule.
Many commenters were concerned that if a local laboratory is not
available, the only alternative will be to send the samples by
overnight delivery to a laboratory elsewhere. However, this situation
is not unique to the analysis of radon. As evidenced during the data
gathering pursuant to the Disinfection By-Products Information
Collection Rule (DBP ICR), several large commercial laboratories
already account for a sizable share of the market for SDWA analyses for
non-radon parameters, including organics, for which the holding times
are often 7 days. Given that a day would be required for shipping the
samples, only three days would remain for the laboratory to perform the
radon analysis (the day on which the sample is collected being ``day
zero''). Some commenters argued that for a large commercial laboratory
serving the water utilities, this short holding time will make it
difficult if not impossible to perform the necessary analyses within
the holding time. However, through common sense scheduling efforts
between the utility and the laboratory, such as not collecting samples
on Thursdays and Fridays, the holding time issue should be able to be
accommodated in light of the ability of the LSC method to be highly
automated.

D. Performance-Based Measurement System (PBMS)

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 is currently
determining how to adopt PBMS in its drinking water program, but has
not yet made final decisions. When PBMS is adopted in 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 then 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
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 to implement PBMS in
programs under the Safe Drinking Water Act, EPA would then provide
specific instruction on the specified performance criteria and how
these criteria would be used by laboratories for radon compliance
monitoring.

E. Proposed Monitoring and Compliance Requirements for Radon

1. Background
The monitoring regulation for radon proposed in 1991 by EPA
required that groundwater systems monitor for radon at each entry point
to the distribution system quarterly for one year initially. Monitoring
could be reduced to one sample annually per entry point to the
distribution system if the average of all first quarterly samples was
below the MCL. States could allow systems to reduce monitoring to once
every three years if the system demonstrated that results of all
previous samples collected were below the MCL. The proposal also
allowed States to grant waivers to groundwater systems to reduce the
frequency of monitoring, up to once every 9 years, if States determined
that radon levels in drinking water were consistently and reliably
below the MCL. Comments made in response to the proposed monitoring
requirements for radon were mainly concerned that the proposed
monitoring requirements including number of samples and the frequency
of monitoring did not adequately take into account the effect of
seasonal variations in radon levels on determining compliance. Other
commenters felt that sampling at the entry point of the distribution
system was not representative of exposure to radon, and they suggested
that sampling for radon should be done at the point of use.
Since the 1991 proposal EPA has obtained additional information
from States, the waterworks industry and academia on the occurrence of
radon, including data on the temporal variability of radon. Utilizing
this additional data, the Agency performed extensive statistical
analyses to predict how temporal, analytical variations and variations
between individual wells may affect exposure to radon. The results of
these analyses are described in detail in the report ``Methods,
Occurrence and Monitoring Document for Radon'' in the docket for this
rule (USEPA 1999g). As a result of the new information available, EPA
was able to refine the requirements for monitoring and address the
concerns expressed by the commenters on the 1991 proposal.
The proposed monitoring requirements for radon are consistent with
the monitoring requirements for regulated drinking water contaminants,
as described in the Standardized Monitoring Framework (SMF) promulgated
by EPA under the Phase II Rule of the National Primary Drinking Water
Regulations (NPDWR) and revised under Phases IIB and V. The goal of the
SMF is to streamline the drinking water monitoring requirements by
standardizing them within contaminant groups and by synchronizing
monitoring schedules across contaminant groups. A summary of monitoring
requirements in this proposal, the SMF and the 1991 proposal are
provided in Table VIII.E.1.

[[Page 59300]]

Table VIII.E.1.--Comparison of Monitoring Requirements
------------------------------------------------------------------------
Monitoring requirements for radon
-------------------------------------------------------------------------
1999 Proposal--MCL/ SMF for IOCs in
1991 Proposal AMCL groundwater
------------------------------------------------------------------------
Initial Monitoring Requirements
------------------------------------------------------------------------
Four consecutive quarters of Four consecutive Four consecutive
monitoring at each entry point quarters of quarters of
for one year. Initial monitoring at monitoring at
monitoring was proposed to have each entry point. each entry point
been completed by January 1, Initial for sampling
1999. monitoring must points initially
begin by three exceeding MCL.
years from date
of publication of
the final rule in
Federal Register
of 4.5 years from
date of
publication of
the final rule in
Federal Register
(depending on
effective date
applicable to the
State).
------------------------------------------------------------------------
Routine Monitoring Requirements
------------------------------------------------------------------------
One sample annually if average One sample One sample at each
from four consecutive quarterly annually if sample point
samples taken initially is less average from four during the
than MCL. consecutive initial 3 year
quarterly samples compliance period
is less than MCL/ for groundwater
AMCL, and at the systems for
discretion of sampling points
State. below MCL.
------------------------------------------------------------------------
1991 Proposal 1999 Proposal--MCL SMF for IOCs in
Groundwater
------------------------------------------------------------------------
Reduced Monitoring Requirements
------------------------------------------------------------------------
State may allow groundwater State may allow State may allow
systems to reduce the frequency CWS using groundwater
of monitoring to once every groundwater to systems to reduce
three years provided that they reduce monitoring monitoring
have monitored quarterly in the frequency to:. frequency to:
initial year and completed Once every three Once every three
annual testing in the second years if average years if samples
and third year of the first from four subsequently
compliance period. Groundwater consecutive detects less than
systems must demonstrate that quarterly samples MCL and
all previous analytical samples is less than \1/ determined by
were less than the MCL. 2\ the MCL/AMCL, State to be
provided no ``reliably and
samples exceed consistently
the MCL/AMCL. and below MCL.''
if the system is
determined by
State to be
``reliably and
consistently
below MCL/AMCL ''.
------------------------------------------------------------------------
Monitoring Requirements for Radon
------------------------------------------------------------------------
1991 Proposal 1999 Proposal--MCL/ SMF for IOCs in
AMCL Groundwater
------------------------------------------------------------------------
Increased Monitoring Requirements
------------------------------------------------------------------------
Systems monitoring annually or Systems monitoring If the MCL is
once per three year compliance annually would be exceeded in a
period exceed the radon MCL in required to single sample,
a single sample would be increase the system
required to revert to quarterly monitoring if the required to begin
monitoring until the average of MCL/AMCL for sampling
4 consecutive samples is less radon is exceeded quarterly until
than the MCL. Groundwater in a single State determines
systems with unconnected wells sample, the that it is
would be required to conduct system would be ``reliably and
increased monitoring only at required to consistently''
those wells exceeding the MCL. revert to below MCL.
The State may require more quarterly
frequent monitoring than monitoring until
specified. the average of 4
Systems may apply to the State consecutive
to conduct more frequent samples is less
monitoring than the minimum than the MCL/AMCL.
monitoring frequencies Systems monitoring
specified. once every three
years would be
required to
monitor annually
if the radon
level is less
than MCL/AMCL but
above \1/2\ MCL/
AMCL in a single
sample. Systems
may revert to
monitoring once
per three years
if the average of
the initial and
three consecutive
annual samples is
lees than \1/2\
MCL/AMCL.
CWS using
groundwater with
un-connected
wells would be
required to
conduct increased
monitoring only
at those well
which are
affected.
------------------------------------------------------------------------

[[Page 59301]]

Monitoring Requirements for Radon
------------------------------------------------------------------------
1991 Proposal 1999 Proposal--MCL SMF for IOCs in
Groundwater
------------------------------------------------------------------------
Confirmation Samples
------------------------------------------------------------------------
Where the results of sampling Systems may Where the results
indicate an exceedence of the collect sampling indicate
maximum contaminant level, the confirmation an exceedence of
State may require that one samples as the maximum
additional sample be collected specified by the contaminant
as soon as possible after the State. The level, the State
initial sample was taken [but average of the may require that
not to exceed two weeks] at the initial sample one additional
same sampling point. The and any sample be
results of the of the initial confirmation collected as soon
sample and the confirmation samples will be as possible after
sample shall be averaged and used to determine the initial
the resulting average shall be compliance. sample was taken
used to determine compliance. [but not to
exceed two weeks]
at the same
sampling point.
The results of
the initial
sample and the
confirmation
sample shall be
averaged and the
resulting average
shall be used to
determine
compliance.
------------------------------------------------------------------------
Grandfathering of Data
------------------------------------------------------------------------
If monitoring data collected If monitoring data States may allow
after January 1, 1985 are collected after previous sampling
generally consistent with the proposal of the data to satisfy
requirements specified in the rule are the initial
regulation, than the State may consistent with sampling
allow the systems to use those the requirements requirements
data to satisfy the monitoring specified in the provided the data
requirements for the initial regulation, then were collected
compliance period. the State may after January 1,
allow the systems 1990.
to use those data
to satisfy the
monitoring
requirements for
the initial
compliance period.
------------------------------------------------------------------------
Monitoring Requirements for Radon
------------------------------------------------------------------------
1991 Proposal 1999 Proposal--MCL SMF for IOCs in
Groundwater
------------------------------------------------------------------------
Waivers
------------------------------------------------------------------------
State may grant waiver to The State may The State may
groundwater systems to reduce grant a grant waiver to
the frequency of monitoring, up monitoring waiver groundwater
to nine years. If State to systems to systems after
determines that radon levels in reduce the conducting
drinking water are ``reliably frequency of vulnerability
and consistently'' below the monitoring to up assessment to
MCL. to one sample reduce the
every nine years frequency of
based on previous monitoring, up to
analytical nine years, if
results, State determines
geological that radon levels
characteristics in drinking water
of source water are ``reliably
aquifer and if a and
State determines consistently''
that radon levels below the MCL.
in drinking water System must have
are ``reliably three previous
and samples.
consistently'' Analytical
below the MCL/ results of all
AMCL. previous samples
Analytical results taken must be
of all previous below MCL.
samples taken
must be below \1/
2\ the MCL/AMCL.
------------------------------------------------------------------------

In developing the proposed compliance monitoring requirements for
radon, EPA considered:
(1) The likely source of contamination in drinking water;
(2) The differences between ground water and surface water systems;
(3) The collection of samples which are representative of consumer
exposure;
(4) Sample collection and analytical methods;
(5) The use of appropriate historical data to identify vulnerable
systems and to specify monitoring requirements for individual systems;
(6) The analytical, temporal and intra-system variance of radon
levels;
(7) The use of appropriate historical data and statistical analysis
to establish reduced monitoring requirements for individual systems;
and
(8) The need to provide flexibility to the States to tailor
monitoring requirements to site-specific conditions by allowing them
to:

--Grant waivers to systems to reduce monitoring frequency, provided
certain conditions are met.
--Require confirmation samples for any sample exceeding the MCL/AMCL.
--Allow the use of previous sampling data to satisfy initial sampling
requirements.
--Increase monitoring frequency.
--Decrease monitoring frequency.
2. Monitoring for Surface Water Systems
CWSs relying exclusively on surface water as their water source
will not be required to sample for radon. Systems that rely in part on
ground water would be considered groundwater systems for purposes of
radon monitoring. Systems that use ground water to supplement surface
water during low-flow periods will be required to monitor for radon.
Ground water under the influence of surface water would be considered
ground water for this regulation.
3. Sampling, Monitoring Schedule and Initial Compliance for CWS Using
Groundwater
EPA is retaining the quarterly monitoring requirement for radon as
proposed initially in the 1991 proposal to account for variations such
as sampling, analytical and temporal variability in radon levels.
Results of analysis of data obtained since 1991, estimating
contributions of individual sources of variability to overall variance
in the radon data sets evaluated, indicated that sampling and
analytical variance contributes less than 1 percent to the overall
variance. Temporal variability within single wells accounts

[[Page 59302]]

for between 13 and 18 percent of the variance in the data sets
evaluated, and a similar proportion (12-17 percent) accounts for
variation in radon levels among wells within systems. (USEPA 1999g)
The Agency performed additional analyses to determine whether the
requirement of initial quarterly monitoring for radon was adequate to
account for seasonal variations in radon levels and to identify non-
compliance with the MCL/AMCL. Results of analysis based on radon levels
modeled for radon distribution for ground water sources (USEPA 1999g)
and systems (USEPA 1998a) in the U.S. show that the average of the
first four quarterly samples provides a good indication of the
probability that the long-term average radon level in a given source
would exceed an MCL or AMCL. Tables VIII.E.2 and VIII.E.3 show the
probability of the long-term average radon level exceeding the MCL and
AMCL at various averages obtained from the first four quarterly samples
from a source.

Table VIII.E.2.--The Relationship Between the First-Year Average Radon
Level and the Probability of the Long-Term Radon Average Radon Levels
Exceeding the MCL
------------------------------------------------------------------------
Then the probability that the
If the average of the first four long-term average radon level
quarterly samples from a source is in that source exceeds 300 pCi/
L is
------------------------------------------------------------------------
Less than 50 pCi/L..................... 0 percent.
Between 50 and 100 pCi/L............... 0.5 percent.
Between 100 and 150 pCi/L.............. 0.4 percent.
Between 150 and 200 pCi/L.............. 7.2 percent.
Between 200 and 300 pCi/L.............. 26.8 percent.
------------------------------------------------------------------------

Table VIII.E.3.--The Relationship Between the First-Year Average Radon
Level and the Probability of the Long-Term Radon Average Radon Levels
Exceeding the AMCL
------------------------------------------------------------------------
Then the probability that the
If the average of the first four long-term average radon level
quarterly samples from a source is in that source exceeds 4000 pCi/
L is
------------------------------------------------------------------------
Less than 2,000 pCi/L.................. Less than 0.1 percent.
Between 2,000 and 2,500 pCi/L.......... 9.9 percent.
Between 2,500 and 3,000 pCi/L.......... 15.1 percent.
Between 3,000 and 4,000 pCi/L.......... 32.9 percent.
------------------------------------------------------------------------

The Agency proposes that systems relying wholly or in part on
ground water will be required to initially sample quarterly for radon
for one year at each well or entry point to the distribution system.
All samples will be required to be of finished water, as it enters the
distribution system after any treatment and storage. If the average of
the four quarterly samples at each well is below the MCL/AMCL,
monitoring may be reduced to once a year at State discretion. Systems
may be required to continue monitoring quarterly in instances where the
average of the quarterly samples at each well is below but close to the
MCL/AMCL. The reason for this is that in such cases, there is a good
chance for the long-term average radon level to exceed the MCL/AMCL.
Systems already on-line must begin initial monitoring for
compliance with the MCL/AMCL by the compliance dates specified in the
rule (i.e., 3 years after the date of promulgation or 4.5 years after
the date of promulgation). Monitoring requirements for new sources will
be determined by the State. The compliance dates are discussed in
detail in Section VII.E, Compliance Dates.
The Agency is retaining the requirement as proposed in 1991 to
sample at the entry point to the distribution system. Sampling at the
entry point allows the system to account for radon decay during storage
and removal during the treatment process. The reason for not allowing
sampling at the point of use is that this approach would not take into
account higher exposure levels that may be encountered at locations
upstream from the sampling site. In addition, sampling at the entry
point will make it easier to identify and isolate possible contaminant
sources within the system. The sample collection sites at each entry
point to the distribution system and the monitoring schedule requiring
sampling for four consecutive quarters proposed herein is consistent
with the SMF. This approach streamlines monitoring since the same
sampling points can be used for the collection of samples for other
source-related contaminants.
EPA specifically requests comments on the following aspects of the
proposed monitoring requirements:
The appropriateness of the proposed initial monitoring
period.
The availability and capabilities of laboratories to
analyze radon samples collected during the initial compliance period.
The Agency recognizes that short-term implementation problems may arise
to meet the initial monitoring deadline because of the potential
limited availability of radon performance evaluation (PE) samples used
to evaluate and certify laboratories.
The appropriateness of the proposed number and frequency
of samples required to monitor for radon.
The designation of sampling locations at the entry point
to the distribution system which is located after any treatment and
storage. Comments are also solicited on the definition of sampling
points that are representative of consumer exposure.
Designating sampling locations and frequencies that permit
simultaneous monitoring for all regulated contaminants, whenever
possible and advantageous. The proposed sampling locations would be
such that the same sampling locations could be used for the collection
of samples for other source-related contaminants such as the volatile
organic chemicals and inorganic chemicals, which would simplify sample
collection efforts.
EPA also solicits comments on whether the monitoring requirements
should include additional monitoring for radon as a source of consumer
exposure from the distribution system. Results of investigations in
Iowa indicate that in some instances, pipe scale deposited in the
distribution system can be a source of exposure to radon. Community
ground water systems could be required to collect an additional sample
from the distribution system during the initial year of monitoring, at
the same time the entry point sample is collected, and continue to
collect samples from the distribution system annually if it is shown
that exceedence of the MCL/AMCL is caused by the release of radon from
deposited scale in the interior of the distribution system. Results
obtained from distribution samples could provide information on the
extent and frequency

[[Page 59303]]

of occurrence of radon originating from distribution systems.
4. Increased/Decreased Monitoring Requirements
Initial compliance with the MCL/AMCL will be determined based on an
average of four quarterly samples taken at individual sampling points
in the initial year of monitoring. Systems with averages exceeding the
MCL/AMCL at any sampling point will be deemed to be out of compliance.
Systems in a non-MMM State exceeding the MCL will have the option to
develop and implement a local MMM program in accordance with the
timeframe discussed in Section VII.E, Compliance Dates without
receiving a MCL violation.
Systems exceeding the MCL/AMCL will be required to monitor
quarterly until the average of four consecutive samples is less than
the MCL/AMCL. Systems will then be allowed to collect one sample
annually if the average from four consecutive quarterly samples is less
than the MCL/AMCL and if the State determines that the system is
reliably and consistently below the MCL/AMCL.
Systems will be allowed to reduce monitoring frequency to once
every three years (one sample per compliance period) per well or
sampling point, if the average from four consecutive quarterly samples
is less than \1/2\ the MCL/AMCL and the State determines that the
system is reliably and consistently below the MCL/AMCL. As shown in
Tables VIII.E.2 and VIII.E.3, EPA believes that there is sufficient
margin of safety to allow for this since there is a small probability
that long term average radon levels will exceed the MCL/AMCL.
Systems monitoring annually that exceed the radon MCL/AMCL in a
single sample will be required to revert to quarterly monitoring until
the average of four consecutive samples is less than the MCL/AMCL.
Community ground water systems with unconnected wells will be required
to conduct increased monitoring only at those wells exceeding the MCL/
AMCL. Compliance will be based on the average of the initial sample and
three consecutive quarterly samples.
Systems monitoring once per compliance period or less frequently
which exceed \1/2\ the MCL/AMCL (but do not exceed the MCL/AMCL) in a
single sample would be required to revert to annual monitoring. Systems
may revert to monitoring once every three years if the average of the
initial and three consecutive annual samples is less than \1/2\ the
MCL/AMCL. Community ground water systems with unconnected wells will be
required to conduct increased monitoring only at those wells exceeding
the MCL/AMCL.
States may grant a monitoring waiver reducing monitoring frequency
to once every nine years (once per compliance cycle) provided the
system demonstrates that it is unlikely that radon levels in drinking
water will occur above the MCL/AMCL. In granting the monitoring waiver,
the State must take into consideration factors such as the geological
area where the water source is located, and previous analytical results
which demonstrate that radon levels do not occur above the MCL/AMCL.
The monitoring waiver will be granted for up to a nine year period.
(Given that all previous samples are less than \1/2\ the MCL/AMCL, then
it is highly unlikely that the long-term average radon levels would
exceed the MCL/AMCL.)
If the analytical results from any sampling point are found to
exceed the MCL/AMCL (in the case of routine monitoring) or \1/2\ the
MCL/AMCL (in the case of reduced monitoring), the State may require the
system to collect a confirmation sample(s). The results of the initial
sample and the confirmation sample(s) shall be averaged and the
resulting average shall be used to determine compliance.
EPA specifically requests comments on the following aspects of the
proposed monitoring requirements :
Allowing systems at State discretion, to reduce monitoring
frequencies as long as the system demonstrates that its radon levels
are maintained below the MCL/AMCL. For example, all community ground
water systems would be required to collect one sample from each entry
point to the distribution system (located after any treatment and
storage) quarterly at first and annually after compliance is
established. MCL/AMCL exceedence would trigger reverting to quarterly
sampling until compliance with the MCL/AMCL is reestablished.
Compliance is reestablished when the average of four consecutive
quarterly samples is below the MCL/AMCL.
Allowing States to reduce monitoring requirements to not
less than once every three years if the average radon levels from four
consecutive quarterly samples is less than \1/2\ the MCL/AMCL, and the
State determines that the radon levels in the drinking water are
reliably and consistently below \1/2\ the MCL/AMCL. A single sample
exceeding \1/2\ the MCL/AMCL would trigger reverting to sampling
annually. Comments are solicited on the criteria allowing the utility
to revert to monitoring once every three years if the average of the
initial and three consecutive annual samples is less than \1/2\ the
MCL/AMCL.
Factors affecting State discretion to grant waivers. In
addition, the Agency solicits comments on the advisability of reducing
the monitoring frequency up to nine years between samples. Comments are
solicited on the requirement that all previous samples (that might be
used to identify systems which are very unlikely to exceed the MCL/
AMCL) must be below \1/2\ the MCL/AMCL in order for a system to qualify
for a waiver.
Allowing States to require the collection of confirmation
samples to verify initial sample results as specified by the State, and
to use the average of the initial sample and the confirmation samples
to determine compliance.
5. Grandfathering of Data
At a State's discretion, sampling data collected since the proposal
could be used to satisfy the initial sampling requirements for radon,
provided that the system has conducted a monitoring program and used
analytical methods that meet proposal requirements. The Agency wants to
provide water suppliers with the opportunity to synchronize their radon
monitoring program with monitoring for other contaminants and to get an
early start on their monitoring program if they wish to do so.
The Agency solicits comments on the advisability of allowing the
use of monitoring data obtained since the proposal to satisfy the
initial monitoring requirements.

IX. State Implementation

This section describes the regulations and other procedures and
policies States have to adopt, or have in place, to implement today's
proposed rule. States must continue to meet all other conditions of
primacy in 40 CFR part 142.
Section 1413 of the SDWA establishes requirements that a State must
meet to obtain or maintain primacy enforcement responsibility (primacy)
for its public water systems. These include: (1) Adopting drinking
water regulations that are no less stringent than Federal NPDWRs in
effect under Section 1412(b) of the Act; (2) adopting and implementing
adequate procedures for enforcement; (3) keeping records and making
reports available on activities that EPA requires by regulation; (4)
issuing variances and exemptions (if allowed by the State) under
conditions no less stringent than allowed by Sections 1415 and 1416;
(5) adopting

[[Page 59304]]

and being capable of implementing an adequate plan for the provision of
safe drinking water under emergency situations; and (6) adopting
authority for administrative penalties.
40 CFR part 142 sets out the specific program implementation
requirements for States to obtain primacy for the public water supply
supervision (PWSS) program, as authorized under SDWA 1413 of the Act.
In addition to meeting the basic primacy requirements, States may be
required to adopt special primacy provisions pertaining to a specific
regulation. States are required by 40 CFR 142.12 to include these
regulation-specific provisions in an application for approval of their
program revisions. To maintain primacy for the PWS program and to be
eligible for interim primacy enforcement authority for future
regulations, States must adopt today's rule, when final, along with the
special primacy requirements discussed next. Interim primacy
enforcement authority allows States to implement and enforce drinking
water regulations once State regulations are effective and the State
has submitted a complete and final primacy revision application. Under
interim primacy enforcement authority, States are effectively
considered to have primacy during the period that EPA is reviewing
their primacy revision application.

A. Special State Primacy Requirements

In addition to adopting drinking water regulations at least as
stringent as the regulations described in the previous sections, EPA
requires that States adopt certain additional provisions related to
this regulation, in order to have their drinking water program revision
application approved by EPA. States have two options when implementing
this rule. States may adopt the AMCL and implement a State-wide MMM
program plan or States may adopt the MCL. If a State chooses to adopt
the MCL, CWSs in that State have the option to develop and implement a
State-approved local MMM program plan and comply with the AMCL.
To ensure that the State program includes all the elements
necessary for a complete enforcement program, EPA is proposing that 40
CFR part 142 be amended to require the following in order to obtain
primacy for this rule:
(1) Adoption of the promulgated Radon Rule, and
(2) One of the following, depending on which regulatory option the
State chooses to adopt:
(a) If a State chooses to develop and implement a State-wide MMM
program plan and adopt the AMCL, the primacy application must contain a
copy of the State-wide MMM program plan meeting the four criteria in 40
CFR Part 141 Subpart R and the following: a description of how the
State will make resources available for implementation of the State-
wide MMM program plan, and a description of the extent and nature of
coordination between interagency programs (i.e., indoor radon and
drinking water programs) on development and implementation of the MMM
program plan, including the level of resources that will be made
available for implementation and coordination between interagency
programs (i.e., indoor air and drinking water programs).
(b) If a State chooses to adopt the MCL, the primacy application
must contain a description of how the State will implement a program to
approve local CWS MMM program plans prepared to meet the criteria
outlined in 40 CFR Part 141 Subpart R. In addition, the primacy
application must contain a description of how the State will ensure
local CWS MMM program plans are implemented and the extent and nature
of coordination between interagency programs (i.e., indoor radon and
drinking water programs) on development and implementation of the MMM
program, including the level of resources that will be made available
for implementation and coordination between interagency programs (i.e.,
indoor air and drinking water programs), as well as, a description of
the reporting and record keeping requirements for the CWSs.
States are required to submit their primacy revision application
packages by two years from the date of publication of the final rule in
the Federal Register. For States adopting the AMCL, EPA approval of a
State's primacy revision application is contingent on submission of and
EPA approval of the State's MMM program plan. Therefore, EPA is
proposing to require submission of State-wide MMM program plans as part
of the complete and final primacy revision application. This will
enable EPA to review and approve the complete primacy application in a
timely and efficient manner in order to provide States with as much
time as possible to begin to implement MMM programs. In accordance with
Section 1413(b)(1) of SDWA and 40 CFR 142.12(d)(3), EPA is to review
primacy applications within 90 days. Therefore, although the SDWA
allows 180 days for EPA review and approval of MMM program plans, EPA
expects to review and approve State primacy revision applications for
the AMCL, including the State-wide MMM program plan, within 90 days of
submission to EPA.
EPA is proposing that States notify CWSs of their decision to adopt
the MCL or AMCL at the time they submit their primacy application
package to EPA (24 months after publication of the final rule). If a
State adopts the MCL, CWSs choosing to implement a local CWS MMM
program and comply with the AMCL will be required to have completed
initial monitoring, notify the State of their intention, and begin
developing a plan 4 years after the rule is final. EPA is particularly
concerned that these CWSs have sufficient time to develop MMM program
plans with local input and allow for State approval. Therefore, it is
EPA's expectation that States will be submitting complete and final
primacy revision applications by 24 months from the date of publication
of the final rule in Federal Register. In reviewing any State requests
for extensions of time in submitting primacy revision applications, EPA
will consider whether sufficient time will be provided to CWSs to
develop and get State approval of their local MMM program plans prior
to implementation.

B. State Record Keeping Requirements

Today's rule does not include changes to the existing recordkeeping
provisions required by 40 CFR 142.14. MMM record keeping requirements
will be addressed in each State's primacy revision application
submission to meet the special primacy requirements for radon (40 CFR
142.16).

C. State Reporting Requirements

Currently States must report to EPA information under 40 CFR 142.15
regarding violations, variances and exemptions, enforcement actions and
general operations of State public water supply programs.
In accordance with the Safe Drinking Water Act (SDWA), EPA is to
review State MMM programs at least every five years. For the purposes
of this review, the States with EPA-approved MMM program plans shall
provide written reports to EPA in the second and fourth years between
initial implementation of the MMM program and the first 5-year review
period, and in the second and fourth years of every subsequent 5-year
review period. EPA will review these programs to determine whether they
continue to be expected to achieve risk reduction of indoor radon using
the information provided in the two biennial reports. EPA requests
comment on this approach. These reports are required to include the
following information:

[[Page 59305]]

A quantitative assessment of progress towards meeting the
required goals described in Section VI. A., including the number or
rate of existing homes mitigated and the number or rate of new homes
built radon-resistant since implementation of the States' MMM program:
and
A description of accomplishments and activities that
implement the program strategies outlined in the implementation plan
and in the two required areas of promoting increased testing and
mitigation of existing homes and promoting increased use of radon-
resistant techniques in construction of new homes.
If goals were defined as rates, the State must also
provide an estimate of the number of mitigations and radon-resistant
new homes represented by the reported rate increase for the two-year
period.
If the MMM program plan includes goals for promoting
public awareness of the health effects of indoor radon, testing of
homes by the public; testing and mitigation of existing schools; and
construction of new public schools to be radon-resistant, the report is
also required to include information on results and accomplishments in
these areas.
EPA will use this information in discussions and consultations with
the State during the five-year review to evaluate program progress and
to consider what modifications or adjustments in approach may be
needed. EPA envisions this review process will be one of consultation
and collaboration between EPA and the States to evaluate the success of
the program in achieving the radon risk reduction goals outlined in the
approved programs. If EPA determines that a MMM program in not
achieving progress towards its goals, EPA and the State shall
collaborate to develop modifications and adjustments to the program to
be implemented over the five year period following the review. EPA will
prepare a summary of the outcome of the program evaluation and the
proposed modification and adjustments, if any, to be made by the State.
States that submit a letter to the Administrator by 90 days after
publication of the final rule committing to develop an MMM program
plan, must submit their first 2-year report by 6.5 years from
publication of the final rule. For States not submitting the 90-day
letter, but choosing subsequently to submit an MMM program plan and
adopt the AMCL, the first 2-year report must be submitted to EPA by 5
years from publication of the final rule. States shall make available
to the public each of these two-year reports, as well as the EPA
summaries of the five-year reviews of a State's MMM program, within 90
days of completion of the reports and the review.
In primacy States without a State-wide MMM program, the States
shall provide a report to EPA every five-years on the status and
progress of CWS MMM programs towards meeting their goals. The first of
such reports would be due 5 years after CWSs begin implementing a local
MMM program which is 5.5 years from publication of the final rule.

D. Variances and Exemptions

Section 1415 of the SDWA authorizes the State to issue variances
from NPDWRs (the term ``State'' is used in this preamble to mean the
State agency with primary enforcement responsibility, or ``primacy,''
for the public water supply system program or EPA if the State does not
have primacy). The State may issue a variance under Section 1415(a) if
it determines that a system cannot comply with an MCL due to the
characteristics of its source water, and on condition that the system
install BAT. Under Section 1415(a), EPA must propose and promulgate its
finding identifying the best available technology, treatment
techniques, or other means available for each contaminant, for purposes
of Section 1415 variances, at the same time that it proposes and
promulgates a maximum contaminant level for such contaminant. EPA's
finding of BAT, treatment techniques, or other means for purposes of
issuing variances may vary, depending upon the number of persons served
by the system or for other physical conditions related to engineering
feasibility and costs of complying with MCLs, as considered appropriate
by the EPA. The State may not issue a variance to a system until it
determines among other things that the variance would not pose an
unreasonable risk to health (URTH). EPA has developed draft guidance,
``Guidance in Developing Health Criteria for Determining Unreasonable
Risks to Health'' (USEPA 1990) to assist States in determining when an
unreasonable risk to health exists. EPA expects to issue final guidance
for determining when URTH levels exist later this year. When a State
grants a variance, it must at the same time prescribe a schedule for
(1) compliance with the NPDWR and (2) implementation of such additional
control measures as the State may require.
Under Section 1416(a), the State may exempt a public water system
from any MCL and/or treatment technique requirement if it finds that
(1) due to compelling factors (which may include economic factors), the
system is unable to comply or develop an alternative supply, (2) the
system was in operation on the effective date of the MCL or treatment
technique requirement, or, for a newer system, that no reasonable
alternative source of drinking water is available to that system, (3)
the exemption will not result in an unreasonable risk to health, and
(4) management or restructuring changes cannot be made that would
result in compliance with this rule. Under Section 1416(b), at the same
time it grants an exemption the State is to prescribe a compliance
schedule and a schedule for implementation of any required interim
control measures. The final date for compliance may not exceed three
years after the NPDWR effective date except that the exemption can be
renewed for small systems for limited time periods.
EPA will not list ``small systems variance technologies'', as
provided in Section 1415(e)(3) of the Act, since EPA has determined
that affordable treatment technologies exist for all applicable system
sizes and water quality conditions. As stated in this Section of the
Act, if the Administrator finds that small systems can afford to comply
through treatment, alternate water source, restructuring, or
consolidation, according to the affordability criteria established by
the Administrator, then systems are not eligible for small systems
variances. Small systems will, however, still be able to apply for
``regular'' variances and exemptions, pursuant to Sections 1415 and
1416 of the Act.

E. Withdrawing Approval of a State MMM Program

If EPA determines that a State MMM program is not achieving
progress towards its MMM goals, and the State repeatedly fails to
correct, modify and adjust implementation of its MMM program after
notice by EPA, EPA may withdraw approval of the State's MMM program
plan. The State will be responsible for notifying CWSs of the
Administrator's withdrawal of approval of the State-wide MMM program
plan. The CWSs in the State would then be required to comply with the
MCL within one year from date of notification, or develop a State-
approved CWS MMM program plan. EPA will work with the State to develop
a State process for review and approval of CWS MMM program plans that
meet

[[Page 59306]]

the required criteria and establish a time frame for submittal of
program plans by CWSs that choose to continue complying with the AMCL.
The review process will allow for local public participation in
development and review of the program plan.

X. What Do I Need To Tell My Customers? Public Information
Requirements

A. Public Notification

Sections 1414(c)(1) and (c)(2) of the SDWA, as amended, require
that public water systems notify persons served when violations of
drinking water standards occur. EPA recently proposed to revise the
current public notification regulations to incorporate new statutory
provisions enacted under the 1996 SDWA amendments (64 FR 25963, May 13,
1999). The purpose of public notification is to alert customers in a
timely manner to potential risks from violations of drinking water
standards and the steps they should take to avoid or minimize such
risks.
Today's regulatory action would add violation of the radon NPDWR to
the list of violations requiring public notice under the May 13, 1999,
proposed public notification rule. Today's action would make three
changes to the proposed public notification rule.
First, Appendix A to Subpart Q would be modified to
require a Tier 2 public notice for violations of the MCL and AMCL for
all community water systems. Under the proposed rule, Tier 2 public
notices would be required for violations and situations with potential
to have serious adverse effects on human health. Tier 2 public notices
must be distributed within 30 days after the violation is known, and
must be repeated every three months until the violation is resolved.
Second, Appendix A would also be modified to require a
Tier 3 public notice for all radon monitoring and testing procedure
violations and for violations of the Multimedia Mitigation (MMM)
Program Plan. Tier 3 public notices must be distributed within a year
of the violation and could, at the water system's option, be included
in the annual Consumer Confidence Report (CCR).
Third, Appendix B to Subpart Q would be modified to add
standard health effects language, which public water systems are
required to use in their notices when violations of the AMCL or MMM
occur. EPA proposes that the standard health effects language for these
violations, to be included in CCR annual reports and public notices.
The language for violation of the (A)MCL would be as follows: ``People
who use drinking water containing radon in excess of the (A)MCL for
many years may have an increased risk of getting lung and stomach
cancer.'' The language for violation of the MMM would be as follows:
``Your water system is not complying with requirements to promote the
reduction of lung cancer risks from radon in indoor air, which is a
problem in some homes. Radon is a naturally occurring radioactive gas
which may enter homes from the surrounding soil and may also be present
in drinking water. Because your system is not complying with applicable
requirements, it may be required to install water treatment technology
to meet more stringent standards for radon in drinking water. The best
way to reduce radon risk is to test your home's indoor air and, if
elevated levels are found, hire a qualified contractor to fix the
problem. For more information, call the National Safety Council's Radon
Hotline at 1-800-SOS-RADON.'' The standard health effects language
public water systems are to use in their public notice would be
identical to that used in the annual CCR.
The final public notification rule is expected to be published
around December, 1999, well in advance of the August, 2000, deadline
for the final radon regulation. The final public notification
requirements for radon, therefore, will be published with the final
radon rule. The Agency will republish the tables in Appendices A and B
to Subpart Q of Part 141 with all necessary changes in the final rule.

B. Consumer Confidence Report

Section 1414(d) of the SDWA requires that all community water
systems provide annual water quality reports (or consumer confidence
reports (CCRs)) to their customers. In their reports, systems must
provide, among other things, the levels and sources of all detected
contaminants, the potential health effects of any contaminant found at
levels that violate EPA or State rules, and short educational
statements on contaminants of particular interest.
Today's action updates the standard CCR rule requirements in
subpart O and adds special requirements that reflect the multimedia
approach of this rule. The intent of these provisions is to assist in
clearer communication of the relative risks of radon in indoor air from
soil and from drinking water, and to encourage public participation in
the development of the State or CWS MMM program plans. Systems that
detect radon at a level that violates the A/MCL would have to include
in their report a clear and understandable explanation of the violation
including: the length of the violation, actions taken by the system to
address the violation, and the potential health effects (using the
language proposed today for Appendix C to subpart O: ``People who use
drinking water containing radon in excess of the (A)MCL for many years
may have an increased risk of getting lung and stomach cancer''). This
approach is comparable to that used for other drinking water
contaminants.
In addition, recognizing the novelty of the MMM approach and the
interest that consumers may have in participating in the design of the
MMM program, today's action also proposes that any system that has
ground water as a source must include information in its report in the
years between publication of the final rule and the date by which
States, or systems, will be required to implement an MMM program. This
information would include a brief educational statement on radon risks,
explaining that the principal radon risk comes from radon in indoor
air, rather than drinking water, and for that reason, radon risk
reduction efforts may be focused on indoor air rather than drinking
water. This information will also note that many States and systems are
in the process of creating programs to reduce exposure to radon, and
encourage readers to call the Radon Hotline (800-SOS-RADON) or visit
EPA's radon web site (www.epa.gov/iaq/radon) for more information. A
system would be able to use language provided in the proposed rule by
EPA or could chose to tailor the wording to its specific local
circumstances in consultation with the primacy agency. EPA recognizes
that this creates a slight additional burden on community water system
operators, but believes that the value of strong public support for,
and participation in, the creation of the MMM program outweighs this
burden. EPA also recognizes that this notice may provoke some
confusion, since CCRs would alert consumers to the risks presented by a
contaminant which most systems have never monitored in their water,
although the notice would state that the system would be testing and
would provide customers with the results. EPA is requesting comment on
this proposed notice.
Finally, the Agency will republish the tables in Appendices A, B,
and C to Subpart O of Part 141 with all necessary changes in the final
rule.

[[Page 59307]]

Risk Assessment and Occurrence

XI. What Is EPA's Estimate of the Levels of Radon in Drinking
Water?

A. General Patterns of Radon Occurrence

Radon levels in ground water in the United States are generally
highest in New England and the Appalachian uplands of the Middle
Atlantic and Southeastern States. There are also isolated areas in the
Rocky Mountains, California, Texas, and the upper Midwest where radon
levels in ground water tend to be higher than the United States
average. The lowest ground water radon levels tend to be found in the
Mississippi Valley, lower Midwest, and Plains States. The following map
shows the general patterns of radon occurrence in those States for
which data are available.

BILLING CODE 6560-50-P

[[Page 59308]]

[GRAPHIC] [TIFF OMITTED] TP02NO99.008

BILLING CODE 6560-50-C

[[Page 59309]]

In addition to large-scale regional variation, radon levels in
ground water vary significantly over a smaller area. Local differences
in geology tend to greatly influence the patterns of radon levels
observed at specific locations. (This means, for example, that not all
radon levels in New England are high and not all radon levels in the
Gulf Coast region are low). Over small distances, there is often no
consistent relationship between radon levels in ground water and
uranium or other radionuclide levels in the ground water or in the
parent bedrock (Davis and Watson 1989). Similarly, no significant
geographic correlation has been found between radon levels in
groundwater systems and the levels of other inorganic contaminants.
Radon may be found in groundwater systems where other contaminants (for
example, arsenic) also occur. However, finding a high (or low) level of
radon does not indicate that a high (or low) level of other
contaminants will also be found. Similarly, there is little evidence
that radon occurrence is correlated with the presence of organic
pollutants. In estimating the costs of radon removal, EPA has taken
into account the fact that other contaminants, such as iron and
manganese, may also be present in the water. High levels of iron and
manganese may complicate the process of radon removal and increase the
costs of mitigation.
Radon is released rapidly from surface water. Therefore, radon
levels in supplies that obtain their water from surface sources (lakes
or reservoirs) are very low compared to groundwater levels.
Because of its short half life, there are relatively few man-made
sources of radon exposure in ground water. The most common man-made
sources of radon ground water contamination are phosphate or uranium
mining or milling operations and wastes from thorium or radium
processing. Releases from these sources can result in high ground water
exposures, but generally only to very limited populations; for
instance, to persons using a domestic well in a contaminated aquifer as
a source of potable water (USEPA 1994a).

B. Past Studies of Radon Levels in Drinking Water

A number of studies of radon levels in drinking water were
undertaken in the 1970s and early 1980s. Most of these studies were
limited to small geographic areas, or addressed systems that were not
representative of community systems throughout the U.S. The first
attempt to develop a comprehensive understanding of radon levels in
public water supplies was the National Inorganics and Radionuclides
Survey (NIRS), which was undertaken by the EPA in 1983-1984. As part of
NIRS, radon samples were analyzed from 1,000 community groundwater
systems throughout the United States. The size distribution of systems
sampled was the same as the size distribution of groundwater systems in
U.S., and the geographic distribution was approximately consistent with
the regional distribution of systems. Because of the limited number of
samples, however, the number of radon measurements in some States was
quite small. Table XI.B.1 summarizes the regional patterns of radon in
drinking water supplies as seen in the NIRS database.

Table XI.B.1.--Radon in Community Ground Water Systems by Region (All System Sizes)
----------------------------------------------------------------------------------------------------------------
Geometric
Region Arithmetic mean Geometric mean standard
(pCi/L) (pCi/L) deviation (pCi/L)
----------------------------------------------------------------------------------------------------------------
Appalachian............................................ 1,127 333 4.76
California............................................. 629 333 3.09
Gulf Coast............................................. 263 125 3.38
Great Lakes............................................ 278 151 3.01
New England............................................ 2,933 1,214 3.77
Northwest.............................................. 222 161 2.23
Plains................................................. 213 132 2.65
Rocky Mountains........................................ 607 361 2.77
----------------------------------------------------------------------------------------------------------------
Source: USEPA 1999g.
Note: These distributions are described in two ways. First, the arithmetic means (average values) are given. In
addition, the geometric mean and geometric standard deviation are given. This approach is taken because the
distributions of radon in groundwater systems are not ``normal'' bell-shaped curves. Instead, like many
environmental data sets, it was found that the logarithms of the radon concentrations were normally
distributed (``lognormal distribution.'') The geometric mean corresponds to the center of a bell-shaped
``normal'' distribution when radon concentrations are expressed in logarithms. The geometric standard
deviation is a measure of the spread of the bell-shaped curve, expressed in logarithmic form.

The NIRS has the disadvantage that the samples were all taken from
within the water distribution systems, making estimation of the
naturally occurring influent radon levels difficult. In addition, the
NIRS data provide no information to allow analysis of the variability
of radon levels over time or within individual systems. Thus, while the
NIRS data provide statistically valid estimates of radon levels in the
systems that were sampled, they do not adequately represent radon
levels in some individual States, especially in large systems.
The NIRS data formed the basis for EPA's first estimates of the
levels of radon in community groundwater systems in the United States
(Wade Miller 1990). They formed the basis for estimating the impacts of
EPA's 1991 Proposed Rule. These estimates were updated in 1993, using
improved statistical methods to estimate the distributions of radon in
different size systems (Wade Miller 1993.)

C. EPA's Most Recent Studies of Radon Levels in Ground Water

EPA's current re-evaluation of radon occurrence in ground water
(USEPA 1999g) uses data from a number of additional sources to
supplement the NIRS information and to develop estimates of the
national distribution of radon in community ground water systems of
different sizes. EPA gathered data from 17 States where radon levels
were measured at the wellhead, rather than in the distribution systems.
The Agency then evaluated the differences between the State (wellhead)
data and the NIRS (distribution system) data. These differences were
then used to adjust the NIRS data to make them more representative of
ground water radon levels in the States where no direct

[[Page 59310]]

measurements at the wellhead had been made. EPA solicits any additional
data on radon levels in community water systems, particularly in the
largest size categories.
Table XI.C.1 summarizes EPA's latest estimates of the distributions
of radon levels in ground water supplies of different sizes. It also
provides information on the populations exposed to radon through
community water systems (CWS). In this table, radon levels and
populations are presented for systems serving population ranges from 25
to greater than 100,000 customers. The CWSs are broken down into the
following system size categories:
Very very small systems (25-500 people served), further
subdivided into 25-100 and 101-500 ranges, in response to comments
received on the 1991 proposal;
Very small systems (501-3,300 people);
Small systems (3,301-10,000 people);
Medium systems (10,001-100,000 people); and
Large systems (greater than 100,000 people).

Table XI.C.1.--Radon Distributions in Community Groundwater Systems
----------------------------------------------------------------------------------------------------------------
System Size (Population Served)
-----------------------------------------------------------------------------------
25-100 101-500 501-3,300 3,301-10,000 >10,000 All systems
----------------------------------------------------------------------------------------------------------------
Total Systems............... 14,651 14,896 10,286 2,538 1,536 43,907
Geometric Mean Radon Level, 312 259 122 124 132 232
pCi/L......................
Geometric Standard Deviation 3.0 3.3 3.2 2.3 2.3 3.0
Arithmetic Mean............. 578 528 240 175 187 442
Population Served (Millions) 0.87 3.75 14.1 14.3 55.0 88.1
Radon Level, pCi/L.......... Proportions of Systems Exceeding Radon Levels (percent)
100......................... 84.7 78.7 56.9 60.4 62.9 74.0
300......................... 51.4 45.1 22.1 14.3 16.2 39.0
500......................... 33.6 29.1 11.4 4.6 5.5 24.2
700......................... 23.4 20.3 6.8 1.8 2.3 16.5
1000........................ 14.7 12.9 3.6 0.6 0.8 10.2
2000........................ 4.7 4.4 0.8 0.0 0.1 4.9
4000........................ 1.1 1.1 0.1 0.0 0.0 0.8
----------------------------------------------------------------------------------------------------------------
Sources: USEPA 1999g; Safe Drinking Water Information System (1998).

Systems were broken down in this fashion because EPA's previous
analyses have shown that the distributions of radon levels are
different in different size systems. In the updated occurrence
analysis, insufficient data were available to accurately assess radon
levels in various subcategories of largest systems. Thus, data from the
two largest size categories were pooled to develop exposure estimates.

D. Populations Exposed to Radon in Drinking Water

Based on data from the Safe Drinking Water Information System
(SDWIS), the Agency estimates that approximately 88.1 million people
were served by community ground water systems in the United States in
1998. Using the data in Table XI.C.1, systems serving more than 500
people account for approximately 95 percent of the population served by
community ground water systems, even though they represent only about
33 percent of the total active systems. The largest systems (those
serving greater than 10,000 people) serve approximately 62.5 percent of
the people served by community ground water systems, even though they
account for only 3.5 percent of the total number of systems.
As noted previously, the average radon levels vary across the
system size categories. As shown in Table XI.C.1, the average system
geometric mean radon levels range from approximately 120 pCi/L for the
larger systems to 312 pCi/L for the smallest systems. The average
arithmetic mean values for the various system size categories range
from 175 pCi/L to 578 pCi/L, and the population-weighted arithmetic
mean radon level across all the community ground water supplies is 213
pCi/L (calculations not shown). The bottom panel of Table XI.C.1 shows
the proportions of the systems with average radon levels greater than
selected values.
Table XI.D.1 presents the total populations in homes served by
community ground water systems at different radon levels, broken down
by system size category. These data show that approximately 20 percent
of the total population served by community ground water systems are
served by systems where the average radon levels entering the system
exceed 300 pCi/L and 64 percent of this population are served by
systems with average radon levels above 100 pCi/L. Less than one-tenth
of one percent of the population is served by systems obtaining their
water from sources with radon levels above 4,000 pCi/L.

Table XI.D.1.--Population Exposed Above Various Radon Levels by Community Ground Water System Size (Thousands)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Very very small Very Small Small Medium Large
Radon level (pCi/L) --------------------------------------------------------------------------------- Total
25-100 101-500 501-3,300 3,301-10K 10K-100K >100K
--------------------------------------------------------------------------------------------------------------------------------------------------------
4,000.................................................... 9.4 46 20 0.2 0.9 0.4 77.2
2,000.................................................... 41 183 119 5.7 21.7 11.0 381
1,000.................................................... 128 541 513 85.5 289 147 1,695
700...................................................... 202 848 962 267 859 436 3,558
500...................................................... 290 1,210 1,620 672 2,070 1,050 6,893
300...................................................... 445 1,880 3,140 2,080 6,060 3,070 16,641
100...................................................... 733 3,290 8,080 8,760 23,400 11,900 56,054
--------------------------------------------------------------------------------------------------------------------------------------------------------

[[Page 59311]]

XII. What Are the Risks of Radon in Drinking Water and Air?

A. Basis for Health Concern

The potential hazard of radon was first identified in the 1940s
when an increased incidence of lung cancer in Bohemian underground
miners was shown to be associated with inhalation of high levels of
radon-222 in the mines. By the 1950s, the hazard was shown to be due
mainly to the short half-life progeny of radon-222. Based on a clear
relationship between radon exposure and risk of lung cancer in a number
of studies in miners, national and international health organizations
have concluded that radon is a human carcinogen. In 1988, the
International Agency for Research on Cancer (IARC 1988) convened a
panel of world experts who agreed unanimously that sufficient evidence
exists to conclude that radon causes cancer in humans and in
experimental animals. The Biological Effects of Ionizing Radiation
(BEIR) Committee (NAS 1988, NAS 1999a), the International Commission on
Radiological Protection (ICRP 1987), and the National Council on
Radiation Protection and Measurement (NCRP 1984) also have reviewed the
available data and agreed that radon exposure causes cancer in humans.
EPA has concurred with these determinations and classified radon in
Group A, meaning that it is considered by EPA to be a human carcinogen
based on sufficient evidence of cancer in humans. After smoking, radon
is the second leading cause of lung cancer deaths in the United States
(NAS 1999a).
Most of the radon that people are exposed to in indoor and outdoor
air comes from soil. However, radon in ground water used for drinking
or other indoor purposes can also be hazardous. When radon in water is
ingested, it is distributed throughout the body. Some of it will decay
and emit radiation while in the body, increasing the risk of cancer in
irradiated organs (although this increased risk is significantly less
than the risk from inhaling radon). Radon dissolved in tap water is
released into indoor air when it is used for showering, washing or
other domestic uses, or when the water is stirred, shaken, or heated
before being ingested. This adds to the airborne radon from other
sources, increasing the risk of lung cancer (USEPA 1991, 1994a; NAS
1999b).

B. Previous EPA Risk Assessment of Radon in Drinking Water

1. EPA's 1991 Proposed Radon Rule
Because initial information on the cancer risks of radon came from
studies of underground miners exposed to very high radon levels, not
much consideration was given to non-occupational radon exposure until
recently. As new miner groups at lower radon exposure levels were added
to the data base, it became evident that radon exposures in indoor air,
outdoor air, and drinking water might be important sources of risk for
the U.S. population. In 1991, as part of developing a regulation for
radionuclides and radon in water as required by the 1986 Safe Drinking
Water Act, EPA drafted the Radon in Drinking Water Criteria Document
(USEPA 1991) to assess the ingestion and inhalation risk associated
with exposure to radon in drinking water. EPA estimated that a person's
risk of fatal cancer from lifetime use of drinking water containing one
picocurie of radon per liter (1 pCi/L) is close to 7 chances in 10
million (7 x 10-^-7). Based on this and other
considerations, EPA proposed a rule for regulating radon levels in
public water systems (56 FR 33050).
2. SAB Concerns Regarding the 1991 Proposed Radon Rule
The Radiation Advisory Committee of EPA's Science Advisory Board
(SAB) reviewed EPA's draft criteria document and proposed rule and
identified a number of issues that had not been adequately addressed,
including: (a) Uncertainties associated with the models, model
parameters, and final risk estimates; (b) high exposure from water at
the point of use (e.g., shower); (c) risks from the disposal of
treatment byproducts; and (d) occupational exposure due to regulation
and removal of radon in drinking water. The SAB recommended that EPA
investigate these issues before finalizing the radon rule. The EPA
considered SAB's recommendations in developing the current proposal.
3. 1994 Report to Congress
In 1992, Congress passed Public Law 102-389 (the Chafee-Lautenberg
Amendment to EPA's Appropriation Bill). This law directs the
Administrator of the EPA to report to Congress on EPA's findings
regarding the risks of human exposure to radon and their associated
uncertainties, the costs for controlling or mitigating that exposure,
and the risks posed by treating water to remove radon.
In response to the SAB's comments and the Chafee-Lautenberg
Amendment, EPA drafted a report entitled Uncertainty Analysis of Risks
Associated with Radon in Drinking Water (USEPA 1993b) and presented it
to the SAB in February 1993. This document evaluated the variability
and uncertainty in each of the factors needed to calculate human cancer
risk from water-borne radon in residences served by community
groundwater systems, and used Monte Carlo simulation techniques to
derive quantitative confidence bounds for the risk estimates for each
of the exposure routes to water-borne radon. In addition, the report
summarized the risk estimates from exposure to radon in indoor and
outdoor air.
Based on the data available at the time, EPA estimated that the
total number of fatal cancers that will occur as a result of exposure
to water-borne radon in homes supplied by community groundwater systems
was 192 per year. EPA noted that the risk from water-borne radon is
small compared to the risk of soil-derived radon in indoor air (13,600
lung cancer cases per year) or in outdoor air (520 lung cancer deaths
per year) (USEPA 1992b, 1993b).
The EPA included the findings of this uncertainty analysis with the
SAB review comments in the Report to the United States Congress on
Radon in Drinking Water: Multimedia Risk and Cost Assessment of Radon
(USEPA 1994a). This report also included an assessment of the risk from
exposure to radon at drinking water treatment facilities. The SAB
reviewed the report prepared by EPA, and commended the EPA's
methodologies employed in the uncertainty analysis and the exposure
assessment of radon at the point of use (e.g. showering). However, the
SAB stated that the estimates of risk from ingested radon may have
additional uncertainties in dose estimation and in the use of primarily
the atomic bomb survivor exposure (gamma emission with low linear
energy transfer) in deriving the organ-specific risk per unit dose for
from radon and progeny (alpha particle emission with high linear energy
transfer). The SAB also questioned EPA's estimates of the number of
community water supplies affected, and the extrapolation of the risk of
lung cancer associated with the high radon exposures of uranium miners
to the low levels of exposure experienced in domestic environments. The
SAB recommended that the Agency use a relative risk orientation as an
important consideration in making risk reduction decisions on all
sources of risks attributable to radon. Based on the

[[Page 59312]]

comments and recommendations of the SAB, EPA revised several of the
distributions used in the Monte Carlo analysis and finalized the
Uncertainty Analysis of Risks Associated with Exposure to Radon in
Drinking Water (USEPA 1995).

C. NAS Risk Assessment of Radon in Drinking Water

1. NAS Health Risk and Risk-Reduction Benefit Assessment Required by
the 1996 Amendments to the Safe Drinking Water Act
The 1996 amendments to the Safe Drinking Water Act required EPA to
arrange with the National Academy of Sciences (NAS) to conduct a risk
assessment of radon in drinking water and an assessment of the health-
risk reduction benefits associated with various measures to reduce
radon concentrations in indoor air. The law also directed EPA to
promulgate an alternative maximum contaminant level (AMCL) if the
proposed MCL is less than the concentration of radon in water
``necessary to reduce the contribution of radon in indoor air from
drinking water to a concentration that is equivalent to the national
average concentration of radon in outdoor air.''
2. Charge to the NAS Committee
In accordance with the requirements of the 1996 amendments to the
SDWA, in February 1997, EPA funded the NAS National Research Council to
establish a multidisciplinary committee of the Board of Radiation
Effects Research. This Committee on Risk Assessment of Exposure to
Radon in Drinking Water (the NAS Radon in Drinking Water committee) was
charged to use the best available data and methods to provide the
following:
(a) The best estimate of the central tendency of the transfer
factor for radon from water to air, along with an appropriate
uncertainty range,
(b) Estimates of unit cancer risk (i.e., the risk from lifetime
exposure to water containing 1 pCi/L) for the inhalation and ingestion
exposure routes, both for the general population and for subpopulations
within the general population (e.g., infants, children, pregnant women,
the elderly, individuals with a history of serious illness) that are
identified as likely to be at greater risk due to exposure to radon in
drinking water than the general population,
(c) Unit cancer risks from inhalation exposure for people in
different smoking categories,
(d) Descriptions of any teratogenic and reproductive effects in men
and women due to exposure to radon in drinking water,
(e) Central estimates for a population-weighted average national
ambient (outdoor) air concentration for radon, with an associated
uncertainty range.
The NAS Radon in Drinking Water committee was also asked to
estimate health risks that might occur as the result of compliance with
a primary drinking water regulation for radon. The committee was to
assess the health risk reduction benefits associated with various
mitigation measures to reduce radon levels in indoor air.
3. Summary of NAS Findings
The NAS completed its charge and issued a report entitled ``Risk
Assessment of Radon in Drinking Water'' (NAS 1999b). The NAS report
provides detailed descriptions of the methods and assumptions employed
by the NAS Radon in Drinking Water committee in completing its
evaluation. The following text provides a summary of the NAS report.
(a) National Average Ambient Radon Concentration. Because radon
levels in outdoor air vary from location to location, the NAS Radon in
Drinking Water committee concluded that available data are not
sufficiently representative to calculate a population-weighted annual
average ambient radon concentration. Based on the data that are
available, the NAS Radon in Drinking Water committee concluded that the
best estimate of an unweighted arithmetic mean radon concentration in
ambient (outdoor) air in the United States is 15 Bq/m³
(equal to 0.41 pCi/L of air), with a confidence range of 14 to 16 Bq/
m³ (0.38-0.43 pCi/L air).
(b) Transfer Factor. The relationship between the concentration of
radon in water and the average indoor air concentration of water-
derived radon is described in terms of the transfer factor (pCi/L in
air per pCi/L in water). Most researchers who have investigated this
variable in residences find that it can be described as a lognormal
distribution of values, most conveniently characterized by the
arithmetic mean (AM) and the standard deviation (Stdev), or by the
geometric mean (GM) and the geometric standard deviation (GSD). The NAS
Radon in Drinking Water committee performed an extensive review of both
measured and calculated values of the transfer factor in residences,
with the results summarized in the following Table XII.1:

Table XII.1.--Measured and Modeled Transfer Factors
----------------------------------------------------------------------------------------------------------------
Approach AM Stdev GM GSD
----------------------------------------------------------------------------------------------------------------
Measured....................... 0.87 x 10^-4 1.2 x 10^-4 0.38 x 10^-4 3.3
Modeled........................ 1.2 x 10^-4 2.4 x 10^-4 0.55 x 10^-4 3.5
----------------------------------------------------------------------------------------------------------------
^a Calculated from, GM and GSD.

The committee concluded that there is reasonable agreement between
the average value of the transfer factor estimated by the two
approaches, and identified 1 in 10,000 (1.0 x 10^-4) as the
best central estimate of the transfer factor for residences, with a
confidence bound of about 0.8 to 1.2 x 10^-4. This central
tendency value is the same as has been used in previous assessments
(USEPA 1993b, 1995).
Based on this transfer factor, the NAS committee concluded that the
AMCL for radon in drinking water would be 150,000 Bq/m³ (
about 4,000 pCi/L). That is, a concentration of 4,000 pCi/L of radon in
water is expected to increase the concentration of radon in indoor air
by an amount equal to that in outdoor air.
(c) Biologic Basis of Risk Estimation. Both the BEIR VI Report (NAS
1999a) and their report on radon in drinking water (NAS 1998b)
represent the most definitive accumulation of scientific data gathered
on radon since the 1988 NAS BEIR IV (NAS 1988). These committees'
support for the use of linear non-threshold relationship for radon
exposure and lung cancer risk came primarily from their review of the
mechanistic information on alpha-particle-induced carcinogenesis,
including studies of the effect of single versus multiple hits to cell
nuclei.
The NAS BEIR VI Committee (NAS 1999a) conducted an extensive review
of information on the cellular and molecular mechanism of radon-induced
cancer in order to help support the assessment of cancer risks from low
levels of radon exposure. In the BEIR VI

[[Page 59313]]

report (NAS 1999a), the NAS concluded that there is good evidence that
a single alpha particle (high-linear energy transfer radiation) can
cause major genomic changes in a cell, including mutation and
transformation that potentially could lead to cancer. Alpha particles,
such as those that are emitted from the radon decay chain, produce
dense trails of ionized molecules when they pass through a cell,
causing cellular damage. Alpha particles passing through the nucleus of
a cell can damage DNA. In their report, the BEIR VI Committee noted
that even if substantial repair of the genomic damage were to occur,
``the passage of a single alpha particle has the potential to cause
irreparable damage in cells that are not killed''. Given the convincing
evidence that most cancers originate from damage to a single cell, the
Committee went on to conclude that ``On the basis of these [molecular
and cellular] mechanistic considerations, and in the absence of
credible evidence to the contrary, the Committee adopted a linear non-
threshold model for the relationship between radon exposure and lung-
cancer risk. The Committee also noted that epidemiological data
relating to low radon exposures in mines also indicate that a single
alpha track through the cell may lead to cancer. Finally, while not
definitive by themselves, the results from residential case-control
studies provide some direct support for the conclusion that
environmental levels of radon pose a risk of lung cancer. However, the
BEIR VI Committee recognized that it could not exclude the possibility
of a threshold relationship between exposure and lung cancer risk at
very low levels of radon exposure.
The NAS Committee on radon in drinking water (NAS 1999b) reiterated
the finding of the BEIR VI Committee's comprehensive review of the
issue, that a ``mechanistic interpretation is consistent with linear
non-threshold relationship between radon exposure and cancer risk''.
The committee noted that the ``quantitative estimation of cancer risk
requires assumptions about the probability of an exposed cell becoming
transformed and the latent period before malignant transformation is
complete. When these values are known for singly hit cells, the results
might lead to reconsideration of the linear no-threshold assumption
used at present.@ EPA recognizes that research in this area is on-going
but is basing its regulatory decisions on the best currently available
science and recommendations of the NAS that support use of a linear
non-threshold relationship. EPA recognizes that research in this area
is on-going but is basing its regulatory decisions on the best
currently available science and recommendations of the NAS that support
use of a linear non-threshold relationship.
(d) Unit Risk from Inhalation Exposure to Radon Progeny. The
calculation of the unit risk from inhalation of radon progeny derived
from water-borne radon depends on four key variables: (1) The transfer
factor that relates the concentration of radon in air to the
concentration in water, (2) the equilibrium factor (the level of radon
progeny present compared to the theoretical maximum amount), (3) the
occupancy factor (the fraction of full time that a person spends at
home) and (4) the risk of lung cancer per unit exposure (the risk
coefficient). The values utilized by NAS for each of these factors are
summarized next.
Transfer Factor
The NAS Radon in Drinking Water committee (NAS 1999b) reviewed
available data and concluded that the best estimate of the transfer
factor is 1.0 x 10^-4 pCi/L air per pCi/L water.
Equilibrium Factor
At radiological equilibrium, 1 pCi/L of radon in air corresponds to
a concentration of 0.010 Working Levels (WL) of radon progeny. One WL
is defined as any combination of radioactive chemicals that result in
an emission of 1.3 x 10⁵ MeV of alpha particle energy. One
WL is approximately the total amount of energy released by the short-
lived progeny in equilibrium with 100 pCi of radon. Under typical
household conditions, processes such as ventilation and plating out of
progeny prevent achievement of equilibrium, and the level of radon
progeny present is normally less than 0.010 WL. The equilibrium factor
(EF) is the ratio of the alpha energy actually present in respirable
air compared to the theoretical maximum at equilibrium. Based on a
review of measured values in residences, USEPA (1993b, 1995) identified
a value of 0.4 as the best estimate of the mean, with a credible range
of 0.35 to 0.45. NAS (1999a, 1999b) reviewed the data and also selected
a value of 0.4 as the most appropriate point estimate of EF.
Occupancy Factor
The occupancy factor (the fraction of time that a person spends at
home) varies with age and occupational status. Studies on the occupancy
factor have been reviewed by EPA (USEPA 1992b, 1993b, 1995), who found
that a value of 0.75 is the appropriate point estimate of the mean with
a credible range of 0.65-0.80. Based on a review of available data,
both the BEIR VI committee (NAS 1999a) and the NAS Radon in Drinking
Water committee (NAS 1999b) identified an occupancy factor of 0.7 as
the best estimate to employ in calculation of the inhalation unit risk
from inhalation of radon progeny.
Risk of Lung Cancer Death per Unit Exposure (Risk Coefficient)
There are extensive data on humans (mainly from studies of
underground miners) establishing that inhalation exposure to radon
progeny causes increased risk of lung cancer (NAS 1999a, 1999b). The
basic approach used by NAS to quantify the risk of fatal cancer
(specifically death from lung cancer) from inhalation of radon progeny
in air was to employ empirical dose-response relationships derived from
studies of humans exposed to radon progeny in the environment. The most
recent quantitative estimate of the risk of lung cancer associated with
inhalation of radon progeny has been conducted by the BEIR VI committee
(NAS 1999a), and this analysis was employed by the NAS Radon in
Drinking Water committee (NAS 1999b). The BEIR VI committee reviewed
all of the most current data from studies of humans exposed to radon,
including cohorts of underground miners and residents exposed to radon
in their home, as well as studies in animals and in isolated cells.
Because of differences in exposure level and duration, studies of
residential radon exposure would normally be preferable to studies of
miners for quantifying risk to residents from radon progeny in indoor
air. However, the BEIR VI committee found that the currently available
epidemiological studies of residents exposed in their homes are not
sufficient to develop reliable quantitative exposure-risk estimates
because (a) the number of subjects is small, (b) the difference between
exposure levels is limited, and (c) cumulative radon exposure estimates
are generally incomplete or uncertain. Therefore, the BEIR VI committee
focused their analysis on studies of radon-exposed underground miners.
The method used by the BEIR VI committee was essentially the same
as used previously by the BEIR IV committee (NAS 1988), except that the
database on radon risk in underground miners is now much more
extensive, including 11 cohorts of underground miners, which, in all,
include about 2,700 lung cancers among 68,000

[[Page 59314]]

miners, representing nearly 1.2 million person-years of observations.
Details of these 11 cohorts are presented in the NAS BEIR VI Report
(NAS 1999a). For historical reasons, the measure of exposure used in
these studies is the Working Level Month (WLM), which is defined as 170
hours of exposure to one Working Level (WL) of radon progeny.
Based on evidence that risk per unit exposure increased with
decreasing exposure rate or with increasing exposure duration (holding
cumulative exposure constant), the BEIR VI committee modified the
previous risk model to include a term to account for this ``inverse
dose rate'' effect. Because the adjustment could be based on either the
concentration of radon progeny or the duration of exposure, there are
two alternative forms of the preferred model--the ``exposure-age-
concentration'' model, and the ``exposure-age-duration'' model. For
brevity, these will generally be referred to here as the
``concentration'' and ``duration'' models.
Mathematically, both models can be represented as:

RR=1+ERR=1+(5-14+15-24
15-24+25+
25+)
(1)

Where:

RR=relative risk of lung cancer in a person due to above-average radon
exposure compared to the average background risk for a similar person
in the general population
ERR=Excess relative risk (the increment in risk due to the above-
average exposure to radon)
=exposure-response parameter (excess relative risk per WLM)
5-14=exposures (WLM) incurred from 5-14 years
prior to the current age
15-24=exposures (WLM) incurred from 15-24 years
prior to the current age
25+=exposures (WLM) incurred 25 or more years
prior to the current age
15-24=time-since-exposure factor for risk from
exposures incurred 15-24 years or more before the attained age
25+=time-since-exposure factor for risk from
exposures incurred 25 or more years or more before the attained age
=effect-modification
factor for attained age
=effect-modification factor for exposure
rate or exposure duration

The BEIR VI committee used a two-stage approach for combining
information from the 11 miner studies to derive parameters for the
concentration and duration risk models. First, estimates of model
parameters were derived for each study cohort, and then population-
weighted averages of the parameters were calculated across studies to
derive an overall estimate that takes variation between and within
cohorts into account. The results of the pooled analysis of all of the
miner data indicated that, for a given level of exposure to radon, the
excess relative risk of lung cancer decreases with increasing time
since exposure, decreases as a function of increased attained age,
increases with increasing duration of exposure, and decreases with
increasing exposure rate (the inverse dose rate effect).
The BEIR VI committee applied the risk models to 1985-89 U.S.
mortality data to estimate individual and population risks from radon
in air. At the individual level, the committee estimated the lifetime
excess relative risk (ERR), which is the percent increase in the
lifetime probability of lung cancer death from indoor radon exposure.
For population risks, the committee estimated attributable risk (AR),
which indicates the proportion of lung-cancer deaths that theoretically
may be reduced by reduction of indoor radon concentrations to outdoor
levels.
Extrapolation From Mines to Homes
Because of a number of potential differences between mines and
homes, exposures to equal levels of radon progeny may not always result
in equal doses to lung cells. The ratio of the dose to lung cells in
the home compared to that in mines is described by the K factor. Based
on the best data available at the time, NAS (1991) had previously
concluded that the dose to target cells in the lung was typically about
30 percent lower for a residential exposure compared to an equal WLM
exposure in mines (i.e., K = 0.7). The BEIR VI committee re-examined
the issue of the relative dosimetry in homes and mines. In light of new
information regarding exposure conditions in home and mine
environments, the committee concluded that, when all factors are taken
into account, the dose per WLM is nearly the same in the two
environments (i.e., a best estimate for the K-factor is about 1) (NAS
1999a). The major factor contributing to the change was a downward
revision in breathing rates for miners. Thus, for calculation of risks
from residential exposures, Equation 1 can be applied directly without
adjustment.
Combined Effect of Smoking and Radon
Because of the strong influence of smoking on the risk from radon,
the BEIR VI committee (NAS 1999a) evaluated risk to ever-smokers and
never-smokers separately. The committee had information on 5 of the
miner cohorts, from which they concluded that the combined effects of
radon and smoking were more than additive but less than multiplicative.
As a best estimate the committee determined that never-smokers should
be assigned a relative risk coefficient () about twice that
for ever-smokers, in each of the two models defined previously. This
means that the attributable risk, or the proportion of all lung cancers
attributable to radon, is about twice as high for never-smokers as
ever-smokers. Nevertheless, because the incidence of lung cancer is
much greater for ever-smokers than never-smokers, the probability of a
radon induced lung cancer is still much higher for ever-smokers. This
higher risk in ever-smokers arises from the synergism between radon and
cigarette smoke in causing lung cancer.
Based on the BEIR VI lifetime relative risk results, the NAS Radon
in Drinking Water committee (NAS 1999b) calculated the lifetime risk
(per Bq/m³ air) for each of the two models using the
following basic equation:

Excess lifetime risk=(Baseline risk)* (LRR-1)
Where LRR=lifetime relative risk
Baseline lung cancer risk values used by the NAS Radon in Drinking
Water committee (NAS 1999b) are summarized in Table XII.2.

Table XII.2.--Baseline Lung Cancer Risk
------------------------------------------------------------------------
Smoking Ever- Never-
Gender prevalence smokers \1\ smokers
------------------------------------------------------------------------
Male............................. 0.58 0.116 0.0091
Female........................... 0.42 0.068 0.0059
------------------------------------------------------------------------
\1\ Ever-smokers were defined as persons who had smoked at least 100
cigarettes in their entire life (CDC 1995).

[[Page 59315]]

The NAS Radon in Drinking Water committee (NAS 1999b) adopted the
average of the results from each of the two models as the best estimate
of lifetime risk from radon progeny.
Results: Inhalation Unit Risk for Water-Borne Radon Progeny
Based on the inputs and approaches summarized in the previous
sections, NAS calculated the inhalation unit risk for radon progeny, by
smoking category, with the results described in Table XII.3:

Table XII.3.--Lifetime Unit Risk
--------------------------------------------------------------------------------------------------------------------------------------------------------
Inhalation risk
Smoking category per Bq/m \3\ in air per pCi/L in water Lifetime (yrs) Annual unit risk coefficient (per
(per pCi/L in water) WLM)
--------------------------------------------------------------------------------------------------------------------------------------------------------
Combined.......................... 1.6 x 10^-4 5.93 x 10^-7 74.9 7.92 x 10^-9 5.49 x 10^-4
Ever Smokers...................... 2.6 x 10^-4 9.63 x 10^-7 73.7 1.31 x 10^-8 9.07 x 10^-4
Never Smokers..................... 0.5 x 10^-4 1.85 x 10^-7 76.1 2.43 x 10^-9 1.68 x 10^-4
--------------------------------------------------------------------------------------------------------------------------------------------------------

The NAS Radon in Drinking Water committee (NAS 1999b) estimated
that the uncertainty around the inhalation risk coefficient for radon
progeny can be characterized by a lognormal distribution with a GSD of
1.2 (based on the duration model) to 1.3 (based on the concentration
model). This corresponds to an uncertainty range for the combined
population of about 3.4 x 10-4 to 8.1 x 10-4 lung cancer deaths per
person per WLM.
Inhalation Risks to Subpopulations, Including Children
The NAS Radon in Drinking Water committee concluded that, except
for the lung-cancer risk to smokers, there is insufficient information
to permit quantitative evaluation of radon risks to susceptible sub-
populations such as infants, children, pregnant women, elderly and
seriously ill persons.
The BEIR VI committee (NAS 1999a) noted that there is only one
study (tin miners in China) that provides data on whether risks from
radon progeny are different for children, adolescents, and adults.
Based on this study, the committee concluded that there was no clear
indication of an effect of age at exposure, and the committee made no
adjustments in the lung cancer risk model for exposures received at
early ages.
(e) Unit Risk for Ingestion Exposure. The calculation of the unit
risk from ingestion of radon in water depends on three key variables:
(1) The amount of radon-containing water ingested, (2) the fraction of
radon lost from the water before ingestion, and (3) the risk to the
tissues per unit of radon absorbed into the body (risk coefficient).
The values utilized by NAS for each of these factors are summarized
next.
Water Ingestion Rate
EPA (USEPA 1993b, 1995) performed a review of available data on the
amount of water ingested by residents. In brief, water ingestion can be
divided into two categories: direct tap water (that which is ingested
as soon as it is taken from the tap) and indirect tap water (water used
in cooking, for making coffee, etc.). Available data indicate nearly
all radon is lost from indirect tap water before ingestion, so only
direct tap water is of concern. Based on available data (Pennington
1983; USEPA 1984; Ershow and Cantor 1989, USEPA 1993b, USEPA 1995)
scientists estimated that the mean of the direct tap water ingestion
rate was 0.65 liters per day (L/day), with a credible range of about
0.57 to 0.74 L/day. Based mainly on this analysis, NAS (1999b)
identified 0.6 L/day as the best estimate of direct tap water intake,
and utilized this value in the calculation of the unit risk from radon
ingestion. This value includes direct tap water ingested at all
locations, and so includes both residential and non-residential
exposures.
The analysis conducted for radon in drinking water uses radon-
specific estimates of water consumption, based on guidance from the NAS
Radon in Drinking Water committee. Based on radon's unique
characteristics, this approach is different from the Agency's approach
to other drinking water contaminants.
In general, in calculating the risk for all other water
contaminants, EPA uses 2 liters per day as the average amount of water
consumed by an individual. For radon, the Agency used 0.6 liters per
day to estimate the risks of radon ingestion. The NAS ingestion risk
number is derived from an average risk/radiation coefficient, an
average drinking water ingestion rate, and an average life expectancy.
NAS chose to use an ingestion rate of 0.6 liter per day, based on an
assumption that only 0.6 liters of the ``direct'' water will retain
radon. Since radon is very readily released during normal household
water use, we assume that radon in water used for indirect purposes
(cooking, making coffee, etc) is released before drinking. Only direct
water (drinking from tap directly) is used to estimate ingestion risk.
The Agency solicits comments on this approach to estimating the
ingestion risk of radon in drinking water, particularly the assumption
of 0.6 liters per day direct consumption.
Fraction of Radon Remaining During Water Transfer From the Tap
Because radon is a gas, it tends to volatilize from water as soon
as the water is discharged from the plumbing system into any open
container or utensil. As would be expected, the fraction of radon
volatilized before consumption depends on time, temperature, surface
area-to-volume ratio, and degree of mixing or aeration. A previous
analysis by EPA (USEPA 1995) identified a value of 0.8 as a reasonable
estimate of the mean fraction remaining before ingestion, with an
estimated credibility interval about the mean of 0.7 to 0.9. Because
data are so sparse, and in order to be conservative, NAS assumed a
point estimate of 1.0 for this factor (NAS 1999b).
Risk per Unit of Radon Absorbed (Risk Coefficient)
The NAS Radon in Drinking Water committee reviewed a number of
publications on the risk from ingestion of radon, and noted that there
was a wide range in the estimates, due mainly to differences and
uncertainties in the way radon is assumed to be absorbed across the
gastrointestinal tract. Therefore, the committee developed new
mathematical models of the diffusion of radon in the stomach and the
behavior of radon dissolved in blood and other tissues to calculate the
radiation dose absorbed by tissues following ingestion of radon
dissolved in water (NAS 1999b).
NAS determined that the stomach wall has the largest exposure (and
hence the largest risk of cancer) following oral exposure to radon in
water, but that

[[Page 59316]]

there is substantial uncertainty on the rate and extent of radon entry
into the wall of the stomach from the stomach contents. The ``base
case'' used by NAS assumed that diffusion of radon from the stomach
contents occurs through a surface mucus layer and a layer of non-
radiosensitive epithelial cells before coming into proximity with the
radiosensitive stem cells. Below this layer, diffusion into capillaries
was assumed to remove radon and reduce the concentration to zero. Based
on this model, the concentration of radon near the stem cells was about
30 percent of that in the stomach contents.
The distribution of absorbed radon to peripheral tissues was
estimated by NAS using a physiologically-based pharmacokinetic (PBPK)
model based on the blood flow model of Leggett and Williams (1995). The
committee's analysis considered that each radioactive decay product
formed from radon decay in the body exhibited its own behavior with
respect to tissues of deposition, retention, and routes of excretion
with the ICRP's age-specific biokinetic models The computational method
used by the NAS Radon in Drinking Water committee to calculate the age-
and gender-averaged cancer death risk from lifetime ingestion of radon
is described in EPA's Federal Guidance Report 13 (USEPA 1998d).
Results: Ingestion Unit Risk
The NAS Radon in Drinking Water committee estimated that an age-
and gender-averaged cancer death risk from lifetime ingestion of radon
dissolved in drinking water at a concentration of 1 Bq/L probably lies
between 3.8 x 10^-7 and 4.4 x 10^-6, with 1.9
x 10^-6 as the best central value. This is equivalent to a
lifetime risk of 7.0 x 10^-8 per pCi/L, with a credible
range of 1.4 x 10^-8 to 1.6 x 10^-7 per pCi/L.
This uncertainty range is based mainly on uncertainty in the estimated
dose to the stomach and in the epidemiologic data used to estimate the
risk (NAS 1999b), and does not include the uncertainty in exposure
factors such as average daily direct tap water ingestion rates or radon
loss before ingestion. The lifetime risk estimate of 7.0 x
10^-8 per pCi/L corresponds to an ingestion risk coefficient
of 4.29 x 10^-12 per pCi ingested.
Ingestion Risk to Children
NAS (1999b) performed an analysis to investigate the relative
contribution of radon ingestion at different ages to the total risk.
This analysis considered the age dependence of: radon consumption,
behavior of radon and its decay products in the body, organ size, and
risk. The results indicated that even though water intake rates are
lower in children than in adults, dose coefficients are higher in
children because of their smaller body size. In addition, the cancer
risk coefficient for ingested radon is greater for children than for
adults. Based on dose and stomach cancer risk models, NAS (1999b)
estimated that about 30% of lifetime ingestion risk was due to
exposures occurring during the first 10 years of life. However, the NAS
found no direct epidemiological evidence to suggest that any sub-
population is at increased risk from ingestion of radon. In addition,
ingestion risk as a whole accounts for only 11% of total risk from
radon exposure from drinking water for the general population, with
inhalation accounting for the remaining 89%. The NAS did not identify
children, or any other groups except smokers, as being at significantly
higher overall risk from exposure to radon in drinking water.
(f) Summary of NAS Lifetime Unit Risk Estimates. Table XII.4
summarizes the lifetime average unit risk estimates derived by the NAS
Radon in Drinking Water committee.

Table XII.4.--Nas Radon in Drinking Water Committee Estimate of Lifetime Unit Risk Posed by Exposure to Radon in
Drinking Water
----------------------------------------------------------------------------------------------------------------
Gender-averaged lifetime unit risk
Exposure route Smoking status --------------------------------------------------
Risk per Bq/L in water Risk per pCi/L in water
----------------------------------------------------------------------------------------------------------------
Inhalation.......................... Ever................... 2.6 x 10^-5 9.6 x 10^-7
Never.................. 0.50 x 10^-5 1.9 x 10^-7
All.................... 1.6 x 10^-5 5.9 x 10^-7
Ingestion........................... All.................... 0.19 x 10^-5 7.0 x 10^-8
---------------------------------------------------------------------------
Total Risk (inhalation + All.................... 1.8 x 10^-5 6.6 x 10^-7
ingestion).
----------------------------------------------------------------------------------------------------------------

(g) Other Health Effects. The NAS Radon in Drinking Water committee
was asked to review teratogenic and reproductive risks from radon. The
committee concluded there is no scientific evidence of teratogenic and
reproductive risks associated with either inhalation or ingestion of
radon.
(h) Relative Magnitude of the Risk from Radon in Water. The NAS
Radon in Drinking Water committee concluded that radon in water
typically adds only a small increment to the indoor air concentration.
The committee estimated the cancer deaths per year due to radon in
indoor air (total), radon in outdoor air, radon progeny from waterborne
radon, and ingestion of radon in water are 18, 200, 720, 160, and 23,
respectively. However, the committee recognized that radon in water is
the largest source of cancer risk in drinking water compared to other
regulated chemicals in water.

D. Estimated Individual and Population Risks

Based on the findings and recommendations of the NAS Radon in
Drinking Water committee, EPA has performed a re-evaluation of the
risks posed by radon in water (USEPA 1999b). This assessment relied
upon the inhalation and ingestion unit risks derived by NAS (1999b),
and calculated risks to individuals and the population by combining the
unit risks derived by NAS with the latest available data on the
occurrence of radon in public water systems (USEPA 1999g).
In brief, the risk to a person from exposure to radon in water is
calculated by multiplying the concentration of radon in the water (pCi/
L) by the unit risk factor (risk per pCi/L) for the exposure pathway of
concern (ingestion, inhalation). The population risk (the total number
of fatal cancer cases per year in the United States due to radon
ingestion in water) is estimated by multiplying the average annual
individual risk (cases per person per year) by the total number of
people exposed. Data which EPA used to

[[Page 59317]]

calculate individual risks and population risks are summarized next.
Radon Concentration in Community Water Systems
The EPA has recently completed a detailed review and evaluation of
the latest available data on the occurrence of radon in community water
systems (USEPA 1999g; see Section XI). In brief, the concentration of
radon in drinking water from surface water sources is very low, and
exposures from surface water systems can generally be ignored. However,
radon does occur in most groundwater systems, with the concentration
values tending to be highest in areas where groundwater is in contact
with granite. In addition, radon concentrations tend to vary as a
function of the size of the water system, being somewhat higher in
small systems than in large systems (USEPA 1999g). Based on EPA's
analysis, the population-weighted average concentration of radon in
community ground water systems is estimated to be 213 pCi/L, with a
credible range of about 190 to 240 pCi/L (USEPA 1999g).
Total Exposed Population
Based on data available from the Safe Drinking Water Information
System (SDWIS), EPA estimates that 88.1 million people (about one-third
of the population of the United States) are served in their residence
by community water supply systems using ground water (USEPA 1998a).
Based on these data on radon occurrence and size of the exposed
population, EPA calculated the risks from water-borne radon to people
exposed at residences served by community groundwater systems. EPA also
calculated revised quantitative uncertainty analysis of the risk
estimates at residential locations, incorporating NAS estimates of the
uncertainty inherent in the unit risks for each pathway. In addition,
EPA performed screening level estimates of risk to people exposed to
water-borne radon in various types of non-residential setting. EPA's
findings are summarized next.
1. Risk Estimates for Ingestion of Radon in Drinking Water
Table XII.5 presents EPA's estimate of the mean individual risk
(fatal cancer cases per person per year) for the people who ingest
water from community ground water systems. This includes exposures that
occur both in the residence and in non-residential settings (the
workplace, restaurants, etc). The lower and upper bounds around the
best estimate were estimated using Monte Carlo simulation techniques
(USEPA 1999b).

Table XII.5.--Estimated Risk from Radon Ingestion at Residential and Non-residential Locations Served by
Community Water Systems
----------------------------------------------------------------------------------------------------------------
Parameter Lower bound Best estimate Upper bound
----------------------------------------------------------------------------------------------------------------
Mean Annual Individual Risk (cancer 3.2 x 10^-8 2.0 x 10^-7 4.3 x 10^-7
deaths per person per year).
Population Risk (cancer deaths per 3 18 38
year).
----------------------------------------------------------------------------------------------------------------

2. Risk Estimates for Inhalation of Radon Progeny Derived From
Waterborne Radon
(a) Inhalation Exposure to Radon Progeny in the Residential
Environment. Table XII.6 presents the EPA's best estimate of the mean
individual risk and population risk of lung cancer fatality due to
inhalation of radon progeny derived from water-borne radon at
residences served by community groundwater systems. Lower and upper
bounds on the individual and population risk estimates were derived
using Monte Carlo simulation techniques.

Table XII.6.--Estimated Risks from Inhalation of Water-Borne Radon Progeny in Residences Served by Community
Ground Water Supply Systems
----------------------------------------------------------------------------------------------------------------
Parameter Lower bound Best estimate Upper bound
----------------------------------------------------------------------------------------------------------------
Mean Annual Individual Risk (lung 7.9 x 10^-7 1.7 x 10^-6 3.0 x 10^-6
cancer deaths per person per year).
Population Risk (lung cancer deaths 70 148 263
per year).
----------------------------------------------------------------------------------------------------------------

Of the total number of lung cancer deaths due to water-borne radon,
most (about 84 percent) are expected to occur in ever-smokers, with the
remainder (about 16 percent) occurring in never-smokers.
Analysis of Peak Exposures and Risks Due to Showering
Both NAS and EPA have paid special attention to the potential
hazards associated with high exposures to radon that may occur during
showering. High exposure occurs during showering because a large volume
of water is used, release of radon from shower water is nearly
complete, and the radon enters a fairly small room (the shower/
bathroom). However, both NAS (1999b) and USEPA (1993b, 1995) concluded
that the risk to humans from radon released during showering was likely
to be small. This is because the inhalation risk from radon is due
almost entirely to radon progeny and not to radon gas itself, and it
takes time (several hours) for the radon progeny to build up from the
decay of the radon gas released from the water. For example, in a
typical shower scenario (about 10 minutes), the level of progeny builds
up to only 2 to 4 percent of its maximum possible value. Thus,
showering is one of many indoor water uses that contribute to the
occurrence of radon in indoor air, but hazards from inhalation of radon
during showering are not of special concern.
(b) Inhalation Exposure to Radon Progeny in the Non-Residential
Environment. The results summarized to this point relate to exposures
which occur in homes. However, on average, people spend about 30
percent of their time at other locations. Surveys of human activity
patterns reveal that time outdoors or in cars accounts for about 13
percent of the time (USEPA 1996), and about 17 percent of the time, on
average across the entire population (including both workers and non-
workers), is spent in non-residential structures. Such non-residential
buildings are presumably all served with water, so exposure to radon
and radon progeny is expected to occur, at least in buildings served by
groundwater. Because data needed to quantify exposure at non-
residential locations are limited, EPA has performed only a screening

[[Page 59318]]

level evaluation to date. This evaluation may be revised in the future,
depending on the availability of more detailed and appropriate input
data.
As with exposures in the home, the largest source of exposure and
risk from water-borne radon in non-residential buildings is inhalation
of radon progeny. Limited data were found on measured transfer factors
in non-residential buildings, so values were estimated for several
different types of buildings based on available data on water use
rates, building size, and ventilation rate, based on the following
basic equation:

TF = (We)/(V)

Where:

W = Water use (L/person/day)
e = Use-weighted fractional release of radon from water to air
V = Building volume (L/person)
= Ventilation rate (air changes/day)

The resulting transfer factor values varied as a function of
building type, based on limited data, but the average across all
building types was about 1 x 10^-4 (the same as for
residences). Very few data were located for the equilibrium factor in
non-residential buildings, so a value of 0.4 (the same as in a
residence) was assumed (USEPA 1999b).
Based on an estimated average transfer factor of 1 x
10^-4 and assuming an average occupancy factor of 17 percent
at non-residential locations, the estimated lifetime and annual risks
of death from lung cancer due to exposure per unit concentration of
radon (1pCi/L) in water are 1.4 x 10^-7 per pCi/L and 1.9
x 10^-9 per pCi/L, respectively.
Assuming a mean radon concentration in water of 213 pCi/L, these
unit risks correspond to lifetime and annual individual risks of 3.1
x 10^-5 and 4.1 x 10^-7 lung cancer deaths per
person. Assuming the same population size of 88.1 million population
exposed to radon through community ground water supplies, EPA's best
estimate of the number of fatal cancer cases per year resulting from
the inhalation of radon progeny in non-residential environments is 36
lung cancer deaths per year (USEPA 1999b) (from the population of
individuals exposed in non-residential settings served by community
ground water supplies).
(c) Analysis of Risk Associated with Exposure at NTNC Locations. A
subset of the water systems serving non-residential populations are the
non-transient non-community (NTNC) systems. Statistics from SDWIS
indicate there are about 5.2 million individuals exposed at buildings
served by NTNC groundwater systems (USEPA 1999b).
Data on radon exposures at locations served by NTNC systems are
limited. However, data are available for water used and population size
at each of 40 strata of NTNC systems (USEPA 1998a). Assuming (a) the
exposure at NTNC locations is occupational in nature with about 8 hr/
day, 250 days/yr, and 25 years per lifetime for workers and 8 hr/day,
180 days/yr, and 12 years per lifetime for students, (b) the same
transfer factor (1 x 10^-4) and equilibrium factor (0.4)
assumed for other non-residential buildings apply at NTNC locations,
and (c) the concentration of radon in water at NTNC locations is about
60 percent higher than in community water systems (mean concentration =
341 pCi/L) (see Section XI of this preamble), then the estimated
population-weighted average individual annual and lifetime lung cancer
risks are 2.6 x 10^-7 and 2.0 x 10^-5,
respectively.
3. Risk Estimates for Inhaling Radon Gas
NAS (1999b) did not derive a unit risk factor for inhalation of
radon gas, but provided in their report a set of annual effective doses
to tissues (liver, kidney, spleen, red bone marrow, bone surfaces,
other tissues) from continuous exposure to 1Bq/m³ of radon
in air. These doses to internal organs from the decay of radon gas
absorbed across the lung and transported to internal sites were based
on calculations by Jacobi and Eisfeld (1980). Based on these dose
estimates, EPA estimated a unit risk value using an approach similar to
that used by NAS to derive the unit risk for ingestion of radon gas in
water. The organ-specific doses reported by Jacobi and Eisfeld were
multiplied by the lifetime-average organ-specific and gender-specific
risk coefficients (risk of fatal cancer per rad) from Federal Guidance
Report No. 13 (USEPA 1998d). Based on an average transfer factor of 1
x 10^-4, and assuming 70 percent occupancy, the estimated
annual average unit risk is 8.5 x 10^-11 cancer deaths per
pCi/L in water. This corresponds to a lifetime average unit risk of 6.3
x 10^-9 per pCi/L. This unit risk excludes the risk of lung
cancer from inhaled radon gas, since this risk is already included in
the unit risk from radon progeny. Based on the population-weighted
average radon concentration of 213 pCi/L, the lifetime average
individual risk is 1.35 x 10^-6 cancer deaths per person,
and the average annual individual risk is 1.8 x 10^-8
cancer deaths per person per year. Based on an exposed population of
88.1 million people, the annual population risk is about 1.6 cancer
deaths/year. The uncertainty range around this estimate, derived using
Monte Carlo simulation techniques, is about 1.0 to 2.7 cancer deaths
per year (USEPA 1999b).
4. Combined Fatal Cancer Risk
The best estimates of fatal cancer risks to residents from
ingesting radon in water, inhalation of waterborne progeny, and
inhalation of radon gas are presented in Table XII.7. As seen, EPA
estimates that an individual's combined fatal cancer risk from lifetime
residential exposure to drinking water containing 1 pCi/L of radon is
slightly less than 7 chances in 10 million (7 x 10^-7), and
that the population risk is about 168 cancer deaths per year
(uncertainty range = 80 to 288 per year). Of this risk, most (88
percent) is due to inhalation of radon progeny, with 11 percent due to
ingestion of radon gas, and less than 1 percent due to inhalation of
radon gas.

Table XII.7.--Summary of Unit Risk, Individual Risk and Population Risk Estimates for Residential Exposure to
Radon in Community Groundwater Supplies
----------------------------------------------------------------------------------------------------------------
Annual
Lifetime unit risk (fatal Annual individual risk population
Exposure pathway cancer cases per person (fatal cancer cases per risk (fatal
per pCi/L) person per year) cancer cases
per year)
----------------------------------------------------------------------------------------------------------------
Radon Gas Ingestion....................... 7.0 x 10^-8 2.0 x 10^-7 18
Radon Progeny Inhalation.................. 5.9 x 10^-7 1.7 x 10^-6 148

[[Page 59319]]

Radon Gas Inhalation...................... 6.3 x 10^-9 1.8 x 10^-8 1.6
---------------------------------------------------------------------
Total (credible bounds)............... 6.7 x 10^-7 (3.6 x 10^-7 1.9 x 10^-6 (0.9 x 10^-6 168 (80-288)
- 9.7 x 10^-7) - 3.3 x 10^-6)
----------------------------------------------------------------------------------------------------------------

EPA believes that radon in community groundwater water systems also
contributes exposure and risk to people when they are outside the
residence (e.g., at school, work, etc.). Although data are limited, a
screening level estimate suggests that this type of exposure could be
associated with about 36 additional lung cancer deaths per year.
Request for Comment
EPA solicits public comments on its assessment of risk from radon
in drinking water. In particular, EPA requests comment and
recommendations on the best data sources and best approaches to use for
evaluating ingestion and inhalation exposures that occur for members of
the public (including both workers and non-workers) at non-residential
buildings (e.g. restaurants, churches, schools, offices, factories,
etc).

E. Assessment by National Academy of Sciences: Multimedia Approach to
Risk Reduction

The NAS report, ``Risk Assessment of Radon in Drinking Water,''
summarized several assessments of possible approaches relating
reduction of radon in indoor air from soil gas to reduction of radon in
drinking water. The NAS Report provided useful perspectives on
multimedia mitigation issues that EPA used in developing the proposed
criteria and guidance for multimedia mitigation programs. The NAS
Committee focused on how the multimedia approach might be applied at
the community level and defined a series of scenarios, assuming that
multimedia programs would be implemented by public water systems. The
report may provide useful perspectives of interest to public water
systems if their State does not develop an EPA-approved MMM program.
For most of the scenarios, the Committee chose primarily to focus
on how to compare the risks posed by radon in indoor air from soil gas
to the risks from radon in drinking water in a home in a local
community. They assessed the feasibility of different activities based
on costs, radon concentrations, different assumptions about risk
reduction actions that might be taken, and other factors.
Overall, the Committee suggested that reduction of indoor radon can
be an alternative and more effective means of reducing the overall risk
from radon. They went on to conclude that mitigation of airborne radon
to achieve equal or greater radon risk reduction ``makes good sense
from a public health perspective.'' They also noted that non-economic
issues, such as equity concerns, could factor into a community's
decision whether to undertake a multimedia mitigation program.
The Committee also discussed the role of various indoor air
mitigation program strategies, or ``mitigation measures'' as they are
described in SDWA. The Committee concluded that an education and
outreach program is important to the success of indoor radon risk
reduction programs, but would not in and of itself be sufficient to
claim that risk reduction took place. Based on an assessment of several
State indoor radon programs, they found that States with effective
programs had several factors in common in the implementation of their
programs. They concluded that the effectiveness of these State programs
were the result of: (1) Promoting wide-spread testing of homes, (2)
conducting radon awareness campaigns, (3) providing public education on
mitigation, and (4) ensuring the availability of qualified contractors
to test and mitigate homes.
These views are consistent with the examples of indoor radon
activities that Congress set forth in the radon provision in SDWA on
which State Multimedia Mitigation programs may rely. These include
``public education, testing, training, technical assistance,
remediation grants and loans and incentive programs, or other
regulatory or non-regulatory measures.'' These measures also represent
many of the same strategies that are integral to the current national
and State radon programs, as well as those outlined in the 1988 Indoor
Radon Abatement Act, sections 304 to 307 (15 U.S.C. 2664-2667).
EPA recognizes, as does the National Academy of Sciences, that
these activities and strategies are important to achieving public
awareness and action to reduce radon, but that these actions are not in
and of themselves actual risk reduction. Therefore, EPA has determined
that State MMM plans will need to set and track actual risk reduction
goals. However, the criteria and guidance for States to use in
designing MMM program plans provides extensive flexibility in choosing
strategies that reflect the needs of individual States.
The Committee discussed the effectiveness of various indoor radon
control technologies and recommended that active sub-slab
depressurization techniques are most effective for controlling radon in
the mitigation of elevated radon levels in existing buildings and in
the prevention of elevated levels in new buildings. (Active systems
rely on mechanically-driven techniques (powered fans) to create a
pressure gradient between the soil and building interior and thus,
prevent radon entry.) The Committee expressed concern over the adequacy
of the scientific basis for ensuring that such methods can be used
reliably as a consistent outcome of normal design and construction
methods. The Committee also noted the limited amount of data available
to quantify the reduction in indoor radon levels expected when such
techniques were used.
The Committee found that much of the comparative data available on
the impact of the passive radon-resistant new construction features is
confined to the impact of the passive thermal stack on radon levels and
not on the other features of the passive radon-resistant new
construction system, such as eliminating leakage paths, sealing utility
penetrations, and prescribing the extent and quality of aggregate
beneath the

[[Page 59320]]

foundation. The Committee found that the passive stack alone yielded
reductions in radon levels as great as 90%, that reductions in radon
levels of about 40% are more typical, and that the effect of the
passive stack may be considerably less in slab-on-grade houses that in
houses with basements. However, the Committee also stated that the
other features in the passive radon-resistant new construction system
contribute to reducing radon levels. EPA notes that there are
substantial difficulties in gathering good comparative data on these
other features because of the significant variability of radon
potential across building sites, even within a small area. In addition
it is impractical to test the same house with and without radon
resistant features. However, based on the Committee's discussion of the
contributions of these other features to reducing radon levels, it is
reasonable to expect that passive systems as a whole achieve greater
reductions in radon than the passive stack alone.
EPA agrees with the Committee's perspective that active radon-
reduction systems, while slightly more expensive, assure the greatest
risk reduction in not only the mitigation of existing homes, but also
in the construction of new homes. EPA also agrees with the Committee's
perspective that more data on passive new construction systems would
allow for more precise estimation of average expected reductions in
radon levels in new homes from application of passive radon-resistant
new construction techniques. However, EPA believes there is sufficient
data and application experience to have a reasonable assurance that the
passive techniques when used in new homes reduce indoor radon levels by
about 50% on average. Further, these techniques have been adopted by
the home construction industry into national model building codes and
by many State and local jurisdictions into their building codes. EPA
recommends that new homes built with passive radon-resistant new
construction features be tested after occupancy and if elevated levels
still exist, the passive systems be converted to active ones. For these
reasons, EPA believes it is appropriate to consider passive radon-
resistant new construction techniques for new homes as one means of
achieving risk reduction through new construction in multimedia
mitigation programs.

Economics and Impacts Analysis

XIII. What Is the EPA's Estimate of National Economic Impacts and
Benefits?

A. Safe Drinking Water Act (SDWA) Requirements for the HRRCA

Section 1412(b)(13)(C) of the SDWA, as amended, requires EPA to
prepare a Health Risk Reduction and Cost Analysis (HRRCA) to be used to
support the development of the radon NPDWR. EPA was to publish the
HRRCA for public comment and respond to significant comments in this
preamble. EPA published the HRRCA in the Federal Register on February
26, 1999 (64 FR 9559). Responses to significant comments on the HRRCA
are provided in Section XIII.H.
The HRRCA addresses the requirements established in Section
1412(b)(3)(C) of the amended SDWA, namely: (1) Quantifiable and non-
quantifiable health risk reduction benefits for which there is a
factual basis in the rulemaking record to conclude that such benefits
are likely to occur as the result of treatment to comply with each
level; (2) quantifiable and non-quantifiable health risk reduction
benefits for which there is a factual basis in the rulemaking record to
conclude that such benefits are likely to occur from reductions in co-
occurring contaminants that may be attributed solely to compliance with
the MCL, excluding benefits resulting from compliance with other
proposed or promulgated regulations; (3) quantifiable and non-
quantifiable costs for which there is a factual basis in the rulemaking
record to conclude that such costs are likely to occur solely as a
result of compliance with the MCL, including monitoring, treatment, and
other costs, and excluding costs resulting from compliance with other
proposed or promulgated regulations; (4) the incremental costs and
benefits associated with each alternative MCL considered; (5) the
effects of the contaminant on the general population and on groups
within the general population, such as infants, children, pregnant
women, the elderly, individuals with a history of serious illness, or
other subpopulations that are identified as likely to be at greater
risk of adverse health effects due to exposure to contaminants in
drinking water than the general population; (6) any increased health
risk that may occur as the result of compliance, including risks
associated with co-occurring contaminants; and (7) other relevant
factors, including the quality and extent of the information, the
uncertainties in the analysis, and factors with respect to the degree
and nature of the risk.
The HRRCA discusses the costs and benefits associated with a
variety of radon levels. Summary tables and figures are presented that
characterize aggregate costs and benefits, impacts on affected
entities, and tradeoffs between risk reduction and compliance costs.
The HRRCA serves as a foundation for the Regulatory Impact Analysis
(RIA) for this proposed rule.

B. Regulatory Impact Analysis and Revised Health Risk Reduction and
Cost Analysis (HRRCA) for Radon

Under Executive Order 12866, Regulatory Planning and Review, EPA
must estimate the costs and benefits of the proposed radon rule in a
Regulatory Impact Analysis (RIA) and submit the analysis to the Office
of Management and Budget (OMB) in conjunction with the proposed rule.
To comply with the requirements of E.O. 12866, EPA has prepared an RIA,
a copy of which is available in the public docket for this proposed
rulemaking. The revised HRRCA is now included as part of the RIA (USEPA
1999f). This section provides a summary of the information from the RIA
for the proposed radon rule.
1. Background: Radon Health Risks, Occurrence, and Regulatory History
Radon is a naturally occurring volatile gas formed from the normal
radioactive decay of uranium. It is colorless, odorless, tasteless,
chemically inert, and radioactive. Uranium is present in small amounts
in most rocks and soil, where it decays to other products including
radium, then to radon. Some of the radon moves through air or water-
filled pores in the soil to the soil surface and enters the air, and
can enter buildings through cracks and other holes in the foundation.
Some radon remains below the surface and dissolves in ground water
(water that collects and flows under the ground's surface). Due to
their very long half-life (the time required for half of a given amount
of a radionuclide to decay), uranium and radium persist in rock and
soil.
Exposure to radon and its progeny is believed to be associated with
increased risks of several kinds of cancer. When radon or its progeny
are inhaled, lung cancer accounts for most of the total incremental
cancer risk. Ingestion of radon in water is suspected of being
associated with increased risk of tumors of several internal organs,
primarily the stomach. As required by the SDWA, as amended, EPA
arranged for the National Academy of Sciences (NAS) to assess the
health risks of radon in drinking

[[Page 59321]]

water. The NAS released the pre-publication draft of the ``Report on
the Risks of Radon in Drinking Water,'' (NAS Report) in September 1998
and published the Report in July 1999 (NAS 1999b). The analysis in this
RIA uses information from the 1999 NAS Report (see Section XII.C of
this preamble). The NAS Report represents a comprehensive assessment of
scientific data gathered to date on radon in drinking water. The
report, in general, confirms earlier EPA scientific conclusions and
analyses of radon in drinking water.
NAS estimated individual lifetime unit fatal cancer risks
associated with exposure to radon from domestic water use for ingestion
and inhalation pathways (Table XIII.1). The results show that
inhalation of radon progeny accounts for most (approximately 88
percent) of the individual risk associated with domestic water use,
with almost all of the remainder (11 percent) resulting from directly
ingesting radon in drinking water. Inhalation of radon progeny is
associated primarily with increased risk of lung cancer, while
ingestion exposure is associated primarily with elevated risk of
stomach cancer.

Table XIII.1.--Estimated Radon Unit Lifetime Fatal Cancer Risks in
Community Water Systems
------------------------------------------------------------------------
Proportion
Cancer unit of total
Exposure pathway risk per pCi/L risk
in water (percent)
------------------------------------------------------------------------
Inhalation of radon progeny \1\........... 5.9 x 10^-7 88
Ingestion of radon \1\.................... 7.0 x 10^-8 11
Inhalation of radon gas \2\............... 6.3 x 10^-9 1
-----------------------------
Total................................. 6.7 x 10^-7 100
------------------------------------------------------------------------
\1\ Source: NAS 1998B.
\2\ Source: Calculated by EPA from radiation dosimetry data and risk
coefficients provided by NAS (NAS 1998B).

The NAS Report confirmed that indoor air contamination arising from
soil gas typically accounts for the bulk of total individual risk due
to radon exposure. Usually, most radon gas enters indoor air by
diffusion from soils through basement walls or foundation cracks or
openings. Radon in domestic water generally contributes a small
proportion of the total radon in indoor air.
The NAS Report is one of the most important inputs used by EPA in
the RIA. EPA has used the NAS's assessment of the cancer risks from
radon in drinking water to estimate both the health risks posed by
existing levels of radon in drinking water and also the cancer deaths
prevented by reducing radon levels.
In updating key analyses and developing the framework for the cost-
benefit analysis presented in the RIA, EPA has consulted with a broad
range of stakeholders and technical experts. Participants in a series
of stakeholder meetings held in 1997, 1998, and 1999 included
representatives of public water systems, State drinking water and
indoor air programs, Tribal water utilities and governments,
environmental and public health groups, and other Federal agencies.
The RIA builds on several technical components, including estimates
of radon occurrence in drinking water, analytical methods for detecting
and measuring radon levels, and treatment technologies. Extensive
analyses of these issues were undertaken by the Agency in the course of
previous rulemaking efforts for radon and other radionuclides. Using
data provided by stakeholders, and from published literature, the EPA
has updated these technical analyses to take into account the best
currently available information and to respond to comments on the 1991
proposed NPDWR for radon.
The analysis presented in the RIA uses updated estimates of the
number of active public drinking water systems obtained from EPA's Safe
Drinking Water Information System (SDWIS). Treatment costs for the
removal of radon from drinking water have also been updated. The RIA
follows current EPA policies with regard to the methods and assumptions
used in cost and benefit assessment.
As part of the regulatory development process, EPA has updated and
refined its analysis of radon occurrence patterns in ground water
supplies in the United States (USEPA 1998l). This new analysis
incorporates information from the EPA's 1985 National Inorganic and
Radionuclides Survey (NIRS) of approximately 1000 community ground
water systems throughout the United States, along with supplemental
data provided by the States, water utilities, and academic research.
The new study also addressed a number of issues raised by public
comments in the previous occurrence analysis that accompanied the 1991
proposed NPDWR, including characterization of regional and temporal
variability in radon levels, and the impact of sampling point for
monitoring compliance.
In general, radon levels in ground water in the United States have
been found to be the highest in New England and the Appalachian uplands
of the Middle Atlantic and Southeastern States. There are also isolated
areas in the Rocky Mountains, California, Texas, and the upper Midwest
where radon levels in ground water tend to be higher than the United
States average. The lowest ground water radon levels tend to be found
in the Mississippi Valley, lower Midwest, and Plains States. When
comparing radon levels in ground water to radon levels in indoor air at
the States level, the distributions of radon concentrations in indoor
air do not always mirror distributions of radon in ground water.
2. Consideration of Regulatory Alternatives
(a) Regulatory Approaches. The RIA evaluates MCL options for radon
in ground water supplies of 100, 300, 500, 700, 1000, 2000, and 4000
pCi/L. As Table VII.1 in Section VII of the preamble illustrates, the
costs and benefits increase as the radon level decreases and the
benefit-cost ratios are very similar at each level. The RIA also
presents information on the costs and benefits of implementing
multimedia mitigation (MMM) programs. The scenarios evaluated are
described in detail in Sections 9 and 10 of the RIA (USEPA 1999f).
Based on the analysis shown in the report, the selected regulatory
alternative discussed next has a significant multimedia mitigation
component. For more information on this analysis, please refer to the
RIA.
(b) Selected Regulatory Alternatives. A CWS must monitor for radon
in drinking water in accordance with the regulations, as described in
Section VIII of this preamble, and report their results to the State.
If the State determines that

[[Page 59322]]

the system is in compliance with the MCL of 300 pCi/L, the CWS does not
need to implement a MMM program (in the absence of a State program),
but must continue to monitor as required.
As discussed in Section VI, EPA anticipates that most States will
choose to develop a State-wide MMM program as the most cost-effective
approach to radon risk reduction. In this case, all CWSs within the
State may comply with the AMCL of 4000 pCi/L. Thus, EPA expects the
vast majority of CWSs will be subject only to the AMCL. In those
instances where the State does not adopt this approach, the proposed
regulation provides the following requirements:
(i) Requirements for Small Systems Serving 10,000 People or Less.
The EPA is proposing that small CWSs serving 10,000 people or less must
comply with the AMCL, and implement a MMM program (if there is no state
MMM program). This is the cut-off level specified by Congress in the
1996 Amendments to the Safe Drinking Water Act for small system
flexibility provisions. Because this definition does not correspond to
the definitions of ``small'' for small businesses, governments, and
non-profit organizations previously established under the RFA, EPA
requested comment on an alternative definition of ``small entity'' in
the preamble to the proposed Consumer Confidence Report (CCR)
regulation (63 FR 7620, February 13, 1998). Comments showed that
stakeholders support the proposed alternative definition. EPA also
consulted with the SBA Office of Advocacy on the definition as it
relates to small business analysis. In the preamble to the final CCR
regulation (63 FR 4511, August 19, 1998), EPA stated its intent to
establish this alternative definition for regulatory flexibility
assessments under the RFA for all drinking water regulations and has
thus used it for this radon in drinking water rulemaking. Further
information supporting this certification is available in the public
docket for this rule.
EPA's regulation expectation for small CWSs is the MMM and AMCL
because this approach is a much more cost-effective way to reduce radon
risk than compliance with the MCL. (While EPA believes that the MMM
approach is preferable for small systems in a non-MMM State, they may,
at their discretion, choose the option of meeting the MCL of 300 pCi/L
instead of developing a local MMM program). The CWSs will be required
to submit MMM program plans to their State for approval. (See Sections
VI.A and F for further discussion of this approach).
SDWA Section 1412(b)(13)(E) directs EPA to take into account the
costs and benefits of programs to reduce radon in indoor air when
setting the MCL. In this regard, the Agency expects that implementation
of a MMM program and CWS compliance with 4000 pCi/L will provide
greater risk reduction for indoor radon at costs more proportionate to
the benefits and commensurate with the resources of small CWSs. It is
EPA's intent to minimize economic impacts on a significant number of
small CWSs, while providing increased public health protection by
emphasizing the more cost-effective multimedia approach for radon risk
reduction.
(ii) Requirements for Large Systems Serving More Than 10,000
People. The proposal requires large community water systems, those
serving populations greater than 10,000, to comply with the MCL of 300
pCi/L unless the State develops a State-wide MMM program, or the CWS
develops and implements a MMM program meeting the four regulatory
requirements, in which case large systems may comply with the AMCL of
4,000pCi/L. CWSs developing their own MMM plans will be required to
submit these plans to their State for approval.
(c) Background on the Selection of the MCL and AMCL. For a
description of EPA's process in selecting the MCL and AMCL, see Section
VII.D of today's preamble.

C. Baseline Analysis

Data and assumptions used in establishing baselines for the
comparison of costs and benefits are presented in the next section.
While the rule as proposed does not require 100 percent compliance with
an MCL, an analysis of these full compliance scenarios are required by
the SDWA, as amended, and were an important feature in the development
of the NPDWR for radon.
1. Industry Profile
Radon is found at appreciable levels only in systems that obtain
water from ground water sources. Thus, only ground water systems would
be affected by the proposed rule. The following discussion addresses
various characteristics of community ground water systems that were
used in the assessment of regulatory costs and benefits. Table XIII.2
shows the estimated number of community ground water systems in the
United States. This data originally came from EPA's Safe Drinking Water
Information System (SDWIS) and are summarized in EPA's Drinking Water
Baseline Handbook (USEPA, 1999c). EPA estimates that there were 43,908
community ground water systems active in December 1997 when the SDWIS
data were evaluated. Approximately 96.5 percent of the systems serve
fewer than 10,000 customers, and thus fit EPA's definition of a
``small'' system (see 63 FR 44512 at 44524-44525, August 19, 1998).
Privately-owned systems comprise the bulk of the smaller size
categories, whereas most larger systems are publicly owned.

Table XIII.2.--Number of Community Ground Water Systems in the United States \1\
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
System size category
Primary source/ownership ------------------------------------------------------------------------------------------------------------------------------------
25-100 101-500 501-1,000 1,001-3,301 3,301-10,000 10,001-50,000 50,001-100,000 100,001-1,000,000 >1,000,000 Total
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
Total...................................................... 14,232 15,070 4,739 5,726 2,489 1,282 139 70 2 43,908
Public..................................................... 1,202 4,104 2,574 3,792 1,916 997 113 52 2 14,764
Private.................................................... 12,361 9,776 1,705 1,531 459 243 24 14 0 26,252
Purchased-Public........................................... 114 427 265 272 84 36 1 4 0 1,203
Purchased-Private.......................................... 171 347 101 79 13 3 1 0 0 718
Other...................................................... 384 416 94 52 17 3 0 0 0 971
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Source: USEPA 1999c.

In addition to the number of affected systems, the total number of
sources (wells) is an important determinant of potential radon
mitigation costs. Larger systems tend to have larger numbers of sources
than small ones, and it has been

[[Page 59323]]

conservatively assumed in the mitigation cost analysis that each source
out of compliance with the MCL or AMCL would need to install control
equipment.
Table XIII.3 summarizes the estimated number of wells per ground
water system. Both the number of wells and the variability in the
number of wells increases with the number of customers served. These
characteristics of community ground water sources are included in the
mitigation cost analysis discussed in Section 7 of the RIA (USEPA
1999f).
2. Baseline Assumptions
In addition to the characteristics of the ground water suppliers,
other important ``baseline'' assumptions were made that affect the
estimates of potential costs and benefits of radon mitigation. Two of
the most important assumptions relate to the distribution of radon in
ground water sources and the technologies that are currently in place
for ground water systems to control radon and other pollutants.
As noted in Section 3 of the RIA (USEPA 1999f), EPA has recently
completed an analysis of the occurrence patterns of radon in
groundwater supplies in the United States (USEPA 1999g). This analysis
used the NIRS and other data sources to estimate national distributions
of groundwater radon levels in community systems of various sizes. The
results of that analysis are summarized in Table XIII.4. These
distributions are used to calculate baseline individual and population
risks, and to predict the proportions of systems of various sizes that
will require radon mitigation.

Table XIII.3.--Estimated Average Number of Wells Per Groundwater System \1\
--------------------------------------------------------------------------------------------------------------------------------------------------------
System size category
-----------------------------------------------------------------------------------------------------------------------
25-100 101-500 501-1,000 1001-3,301 3,301-10,000 10,001-50,000 50,001-100,000 100,001-1,000,000
--------------------------------------------------------------------------------------------------------------------------------------------------------
Average Number of Wells 1.5 (0.2) 2.0 (0.2) 2.3 (0.2) 3.1 (0.3) 4.6 (1.1) 9.8 (1.8) 16.1 (2.2) 49.9 (12.7)
(Confidence Interval)..........
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Source: USEPA 1999c.

Table XIII.4.--Distribution of Radon Levels in U.S. Groundwater Sources
----------------------------------------------------------------------------------------------------------------
Population served
Statistic ---------------------------------------------------------------------
25-100 101-500 501-3,300 3,301-10,000 >10,000
----------------------------------------------------------------------------------------------------------------
Geometric Mean, pCi/L..................... 312 259 122 124 132
Geometric Standard Deviation, pCi/L....... 3.04 3.31 3.22 2.29 2.31
Arithmetic Mean........................... 578 528 240 175 187
----------------------------------------------------------------------------------------------------------------

The costs of radon mitigation are affected to some extent by the
treatment technologies that are currently in place to mitigate radon
and other pollutants, and by the existence of pre- and post-treatment
technologies that affect the costs of mitigation. EPA has conducted an
extensive analysis of water treatment technologies currently in use by
groundwater systems. Table XIII.5 shows the proportions of ground water
systems with specific technologies already in place, broken down by
system size (population served). Many ground water systems currently
employ disinfection, aeration, or iron/manganese removal technologies.
This distribution of pre-existing technologies serves as the baseline
against which water treatment costs are measured. For example, costs of
disinfection are attributed to the radon rule only for the estimated
proportion of systems that would have to install disinfection as a
post-treatment because they do not already disinfect. The cost analysis
assumes that any system affected by the rule will continue to employ
pre-existing radon treatment technology and pre- and post-treatment
technologies in their efforts to comply with the rule. Where pre- or
post-treatment technologies are already in place it is assumed that
compliance with the radon rule will not require any upgrade or change
in the pre- or post-treatment technologies. Therefore, no incremental
cost is attributed to pre- or post-treatment technologies. This may
underestimate costs if pre- or post-treatment technologies need to be
changed (e.g., a need for additional chlorination after the
installation of packed tower aeration). The potential magnitude of this
cost underestimation is not known, but is likely to be a very small
fraction of total treatment costs.

Table XIII.5.--Estimated Proportions of Groundwater Systems With Water Treatment Technologies Already in Place (Percent) \1\
--------------------------------------------------------------------------------------------------------------------------------------------------------
System Size (Population Served)
-------------------------------------------------------------------------------------------------------------
Water treatment technologies in place 100,001
25-100 101-500 501-1,000 1,001-3,300 3,301-10,000 10,001-50,000 50,001-100,000 1,000,000
--------------------------------------------------------------------------------------------------------------------------------------------------------
Fe/Mn removal & aeration & disinfection... 0.4 0.2 1.2 0.6 2.9 2.2 3.1 2
Fe/Mn removal & aeration.................. 0 0.1 0.2 0.1 0.4 0.1 0.4 0.1
Fe/Mn removal & disinfection.............. 2.1 5.1 8.3 3 7.8 7.4 9.7 6.8
Fe/Mn removal............................. 1.9 1.5 1.5 1 1.1 0.4 1.1 0.2
Aeration & disinfection only.............. 0.9 3.2 9.8 13.7 20.9 19.7 18.6 19.9
Aeration only............................. 0.8 1 1.8 2.9 2.9 1 2.1 0.6
Disinfection only......................... 49.6 68.2 65 65 56.3 66 58.3 68.3

[[Page 59324]]

None...................................... 44.3 20.7 12.2 13.7 7.7 3.2 6.7 2.1
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\. Source: EPA analysis of data from the Community Water System Survey (CWSS), 1997, and Safe Drinking Water Information System (SDWIS), 1998.

The treatment baseline assumptions shown in Table XIII.5 were used
in the initial analysis for the development of the NPDWR for radon.
These assumptions were used to establish the costs of 100 percent
compliance with an MCL. Another analysis, which portrays the costs of
the rule as recommended in this proposed rulemaking, is provided in the
results section of this summary and also in Section 9 of the RIA.

D. Benefits Analysis

11. Quantifiable and Non-Quantifiable Health Benefits
The quantifiable health benefits of reducing radon exposures in
drinking water are attributable to the reduced incidence of fatal and
non-fatal cancers, primarily of the lung and stomach. Table XIII.6
shows the health risk reductions (number of fatal and non-fatal cancers
avoided) and the residual health risk (number of remaining cancer
cases) at various radon in water levels.

Table XIII.6.--Residual Cancer Risk and Risk Reduction from Reducing Radon in Drinking Water
----------------------------------------------------------------------------------------------------------------
Risk Risk
Residual Residual reduction reduction
fatal cancer non-fatal (fatal (non-fatal
Radon Level (pCi/L in water) risk (cases cancer risk cancers cancers
per year) (cases per avoided per avoided per
year) year)\1\ year)\1\
----------------------------------------------------------------------------------------------------------------
(Baseline)............................................... 168 9.7 0 0
4,0002 \2\............................................... 165 9.5 2.9 0.2
2,000.................................................... 160 9.4 7.3 0.4
1,000.................................................... 150 8.8 17.8 1.1
700...................................................... 141 8.3 26.1 1.5
500...................................................... 130 7.6 37.6 2.2
300...................................................... 106 6.1 62.0 3.6
100...................................................... 46.8 2.8 120 7.0
----------------------------------------------------------------------------------------------------------------
Notes:
\1\ Risk reductions and residual risk estimates are slightly inconsistent due to rounding.
\2\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA provisions of Section 1412(b)(13).

Since preparing the prepublication edition of the NAS Report, the
NAS has reviewed and slightly revised their unit risk estimates. EPA
uses these updated unit risk estimates in calculating the baseline
risks, health risk reductions, and residual risks. Under baseline
assumptions (no control of radon exposure), approximately 168 fatal
cancers and 9.7 non-fatal cancers per year are associated with radon
exposures through CWSs. At a radon level of 4,000 pCi/L, approximately
2.9 fatal cancers and 0.2 non-fatal cancers per year are prevented. At
300 pCi/L, approximately 62.0 fatal cancers and 3.6 non-fatal cancers
are prevented each year.
The Agency has developed monetized estimates of the health benefits
associated with the risk reductions from radon exposures. The SDWA, as
amended, requires that a cost-benefit analysis be conducted for each
NPDWR, and places a high priority on better analysis to support
rulemaking. The Agency is interested in refining its approach to both
the cost and benefit analysis, and in particular recognizes that there
are different approaches to monetizing health benefits. In the past,
the Agency has presented benefits as cost per life saved, as in Table
XIII.7.
The costs of reducing radon to various levels, assuming 100 percent
compliance with an MCL, are summarized in Table XIII.7, which shows
that, as expected, aggregate radon mitigation costs increase with
decreasing radon levels. For CWSs, the costs per system do not vary
substantially across the different radon levels evaluated. This is
because the menu of mitigation technologies for systems with various
influent radon levels remains relatively constant and are not sensitive
to percent removal.

Table XIII.7.--Estimated Annualized National Costs of Reducing Radon Exposures
[$Million, 1997]
----------------------------------------------------------------------------------------------------------------
Central
tendency Total Total cost per
Radon level (pCi/L) estimate of annualized fatal cancer
annualized national costs case avoided
costs \2\ \3\
----------------------------------------------------------------------------------------------------------------
4000 \1\........................................................ 34.5 43.1 14.9
2000............................................................ 61.1 69.7 9.5

[[Page 59325]]

1000............................................................ 121.9 130.5 7.3
700............................................................. 176.8 185.4 7.1
500............................................................. 248.8 257.4 6.8
300............................................................. 399.1 407.6 6.6
100............................................................. 807.6 816.2 6.8
----------------------------------------------------------------------------------------------------------------
\1\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

\2\ Costs include treatment, monitoring, and O&M costs only.
\3\ Costs include treatment, monitoring, O&M, recordkeeping, reporting, and state costs for administration of
water programs.

An alternative approach presented here for consideration as one
measure of potential benefits is the monetary value of a statistical
life (VSL) applied to each fatal cancer avoided. Since this approach is
relatively new to the development of NPDWRs, EPA is interested in
comments on these alternative approaches to valuing benefits, and will
have to weigh the value of these approaches for future use.
Estimating the VSL involves inferring individuals' implicit
tradeoffs between small changes in mortality risk and monetary
compensation. In the HRRCA, a central tendency estimate of $5.8 million
(1997$) is used in the monetary benefits calculations. This figure is
determined from the VSL estimates in 26 studies reviewed in EPA's
recent draft guidance on benefits assessment (USEPA 1998e), which is
currently under review by the Agency's Science Advisory Board (SAB) and
the Office of Management and Budget (OMB).
It is important to recognize the limitations of existing VSL
estimates and to consider whether factors such as differences in the
demographic characteristics of the populations and differences in the
nature of the risks being valued have a significant impact on the value
of mortality risk reduction benefits. Also, medical care or lost-time
costs are not separately included in the benefits estimate for fatal
cancers, since it is assumed that these costs are captured in the VSL
for fatal cancers.
For non-fatal cancers, willingness to pay (WTP) data to avoid
chronic bronchitis is used as a surrogate to estimate the WTP to avoid
non-fatal lung and stomach cancers. The use of such WTP estimates is
supported in the SDWA, as amended, at Section 1412(b)(3)(C)(iii): ``The
Administrator may identify valid approaches for the measurement and
valuation of benefits under this subparagraph, including approaches to
identify consumer willingness to pay for reductions in health risks
from drinking water contaminants.''
A WTP central tendency estimate of $536,000 is used to monetize the
benefits of avoiding non-fatal cancers (Viscusi et al. 1991). The
combined fatal and non-fatal health benefits are summarized in Table
XIII.8. The annual health benefits range from $17.0 million for a radon
level of 4000 pCi/L to $702 million at 100 pCi/L.

Table XIII.8.--Estimated Monetized Health Benefits from Reducing Radon
in Drinking Water
------------------------------------------------------------------------
Monetized
health
benefits,
central
Radon level (pCi/L) tendency
(annualized,
$millions,
1997)\1\
------------------------------------------------------------------------
4,000 \2\............................................... 17.0
2,000................................................... 42.7
1,000................................................... 103
700..................................................... 152
500..................................................... 219
300..................................................... 362
100..................................................... 702
------------------------------------------------------------------------
Notes:
\1\ Includes contributions from fatal and non-fatal cancers, estimated
using central tendency estimates of the VSL of $5.8 million (1997$),
and a WTP to avoid non-fatal cancers of $536,000 (1997$).
\2\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on
SDWA provisions of Section 1412(b)(13).

Reductions in radon exposures might also be associated with non-
quantifiable benefits. EPA has identified several potential non-
quantifiable benefits associated with regulating radon in drinking
water. These benefits may include any customer peace of mind from
knowing drinking water has been treated for radon. In addition, if
chlorination is added to the process of treating radon via aeration,
arsenic pre-oxidation will be facilitated. Neither chlorination nor
aeration will remove arsenic, but chlorination will facilitate
conversion of Arsenic (III) to Arsenic (V). Arsenic (V) is a less
soluble form that can be better removed by arsenic removal
technologies. In terms of reducing radon exposures in indoor air, it
has also been suggested that provision of information to households on
the risks of radon in indoor air and available options to reduce
exposure may be a non-quantifiable benefit that can be attributed to
some components of a MMM program. Providing such information might
allow households to make more informed choices than they would have in
the absence of an MMM program about the need for risk reduction given
their specific circumstances and concerns. In the case of the proposed
radon rule, it is not likely that accounting for these non-quantifiable
benefits would significantly alter the overall assessment.
The benefits calculated for this proposal are assumed to begin to
accrue on the effective date of the rule and are based on a calculation
referred to as the ``value of a statistical life'' (VSL), currently
estimated at $5.8 million. The VSL is an average estimate derived from
a set of 26 studies estimating what people are willing to pay to avoid
the risk of premature mortality. Most of these studies examine
willingness to pay in the context of voluntary acceptance of higher
risks of immediate accidental death in the workplace in exchange for
higher wages. This value is sensitive to differences in population
characteristics and perception of risks being valued.
For the present rulemaking analysis, which evaluates reduction in
premature mortality due to carcinogen exposure, some commenters have
argued that the Agency should consider an assumed time lag or latency
period in these calculations. Latency refers to the difference between
the time of initial exposure to environmental carcinogens and the onset
of any resulting cancer. Use of such an approach might reduce
significantly the present value estimate.

[[Page 59326]]

The BEIR VI model and U.S. vital statistics, on which the estimate of
lung cancers avoided is based, imply a probability distribution of
latency periods between inhalation exposure to radon and increased
probability of cancer death. EPA is interested in receiving comments on
the extent to which the presentation of more detailed information on
the timing of cancer risk reductions would be useful in evaluating the
benefits of the proposed rule.
Latency is one of a number of adjustments or factors that are
related to an evaluation of potential benefits associated with this
rule, how those benefits are calculated, and when those economic
benefits occur. Other factors which may influence the estimate of
economic benefits associated with avoided cancer fatalities include (1)
A possible ``cancer premium'' (i.e., the additional value or sum that
people may be willing to pay to avoid the experiences of dread, pain
and suffering, and diminished quality of life associated with cancer-
related illness and ultimate fatality); (2) the willingness of people
to pay more over time to avoid mortality risk as their income rises;
(3) a possible premium for accepting involuntary risks as opposed to
voluntary assumed risks; (4) the greater risk aversion of the general
population compared to the workers in the wage-risk valuation studies;
(5) ``altruism'' or the willingness of people to pay more to reduce
risk in other sectors of the population; and (6) a consideration of
health status and life years remaining at the time of premature
mortality. Use of certain of these factors may significantly increase
the present value estimate. EPA therefore believes that adjustments
should be considered simultaneously. The Agency also believes that
there is currently neither a clear consensus among economists about how
to simultaneously analyze each of these adjustments nor is there
adequate empirical data to support definitive quantitative estimates
for all potentially significant adjustment factors. As a result, the
primary estimates of economic benefits presented in the analysis of
this rule rely on the unadjusted $5.8 million estimate. However, EPA
solicits comment on whether and how to conduct these potential
adjustments to economic benefits estimates together with any rationale
or supporting data commenters wish to offer. Because of the complexity
of these issues, EPA will ask the Science Advisory Board (SAB) to
conduct a review of these benefits transfer issues associated with
economic valuation of adjustments in mortality risks. In its analysis
of the final rule, EPA will attempt to develop and present an analysis
and estimate of the latency structure and associated benefits transfer
issues outlined previously consistent with the recommendations of the
SAB and subject to resolution of any technical limitations of the data
and models.

E. Cost Analysis

1. Total National Costs of Compliance with MCL Options
Table XIII.9 summarizes the estimates of total national costs of
compliance with the range of potential MCLs considered. The table is
divided into two major groupings; the first grouping displays the
estimated costs to systems and the second grouping displays the
estimated costs to States. State costs, presented in Table XIII.9, were
developed as part of the analyses to comply with the Unfunded Mandates
Reform Act (UMRA) and also the Paperwork Reduction Act (PRA).
Additional information on State costs is provided in Section 8 of the
RIA and also in Section VIII of this preamble.

Table XIII.9.--Summary of Estimated Costs Under the Proposed Radon Rule Assuming 100% Compliance With an MCL of
300 pCi/L
[$ Millions] \1\
----------------------------------------------------------------------------------------------------------------
10 percent
3 percent cost 7 percent cost cost of
of capital of capital capital
----------------------------------------------------------------------------------------------------------------
Costs to Water Systems
----------------------------------------------------------------------------------------------------------------
Total Capital Costs (20 years, undiscounted).............. 2,463 2,463 2,463
----------------------------------------------------------------------------------------------------------------
Annual Costs
----------------------------------------------------------------------------------------------------------------
Annualized Capital.............................................. 165.6 232.5 289.4
Annual O&M...................................................... 152.4 152.4 152.4
-----------------------------------------------
Total Annual Treatment.................................... 318.0 385.0 441.8
-----------------------------------------------
Monitoring Costs................................................ 14.1 14.1 14.1
Recordkeeping and Reporting Costs \2\........................... 6.1 6.1 6.1
-----------------------------------------------
Total Annual Costs to Water Systems \3\................... 338.2 405.1 461.6
----------------------------------------------------------------------------------------------------------------
Costs to States
----------------------------------------------------------------------------------------------------------------
Administration of Water Programs................................ 2.5 2.5 2.5
-----------------------------------------------
Total Annual State Costs.................................. 2.5 2.5 2.5
Total Annual Costs of Compliance \4\...................... 340.6 407.6 464.4
----------------------------------------------------------------------------------------------------------------
1. Assumes no MMM program implementation costs (e.g., all systems comply with 300 pCi/L).
2. Figure represents average annual burden over 20 years.
3. Costs include treatment, monitoring, O&M, recordkeeping, and reporting costs to water systems.
4. Totals have been rounded. Costs include treatment, monitoring, O&M, recordkeeping, reporting, and state
costs for administration of water programs.

[[Page 59327]]

2. Quantifiable and Non-quantifiable Costs
The capital and operating and maintenance (O&M) costs of mitigating
radon in Community Water Systems (CWSs) were estimated for each of the
radon levels evaluated. The costs of reducing radon in community ground
water to specific target levels were calculated using the cost curves
discussed in Section 7.5 and the matrix of treatment options presented
in Section 7.6 of the RIA. For each radon level and system size
stratum, the number of systems that need to reduce radon levels by up
to 50 percent, 80 percent and 99 percent were calculated. Then, the
cost curves for the distributions of technologies dictated by the
treatment matrix were applied to the appropriate proportions of the
systems. Capital and O&M costs were then calculated for each system,
based on typical estimated design and average flow rates. These flow
rates were calculated on spreadsheets using equations from EPA's
Baseline Handbook (USEPA 1999e). The equations and parameter values
relating system size to flow rates are presented in Appendix C of the
RIA. The technologies addressed in the cost estimation included a
number of aeration and granular activated carbon (GAC) technologies
described in Section 7.2 of the RIA, as well as storage,
regionalization, and disinfection as a post-treatment. To estimate
costs, water systems were assumed, with a few exceptions to simulate
site-specific problems, to select the technology that could reduce
radon to the selected target level at the lowest cost. CWSs were also
assumed to treat separately at every source from which water was
obtained and delivered into the distribution system.
EPA has attempted to note potential non-quantifiable benefits when
the Agency believes they might occur, as in the case of peace-of-mind
benefits from radon reduction. The Agency recognizes that there may
also be non-quantifiable disbenefits, such as anxiety on the part of
those near aeration plants or those who find out that their radon
levels are high. It is not possible to determine whether the net
results of such psychological effects would be positive or negative.
The inclusion of non-quantifiable benefits and costs in this analysis
are not likely to alter the overall results of the benefit-cost
analysis for the proposed radon rule.

F. Economic Impact Analysis

A summary analysis of the impacts on small entities is shown in
Section XIV.B of this preamble (Regulatory Flexibility Act). An
analysis of the impacts on State, local, and tribal governments is
shown in Section XIV.C (Unfunded Mandates Reform Act). For information
on how this proposed rulemaking may impact Indian tribal governments,
see Section XIV.I of today's preamble. Information on the types of
information that States will be required to collect, as well as EPA's
estimate of the burden and reporting requirements for this proposed
rulemaking, is shown in Section XIV.D (Paperwork Reduction Act). EPA's
assessment of the impacts that this proposed rulemaking may have on
low-income and minority populations, as well as any potential concerns
regarding children's health, are shown in Section XIV.F (Environmental
Justice) and Section XIV.G (Protection of Children from Environmental
Health Risks and Safety Risks) of today's preamble.

G. Weighing the Benefits and Costs

1. Incremental Costs and Benefits of Radon Removal

Table XIII.10.--Estimates of the Annual Incremental Risk Reduction, Costs, and Benefits of Reducing Radon in Drinking Water Assuming 100% Compliance
With an MCL
[$ Millions 1997]
--------------------------------------------------------------------------------------------------------------------------------------------------------
Radon Level, pCi/L
------------------------------------------------------------------------------------------
4000 \1\ 2000 1000 700 500 300 100
--------------------------------------------------------------------------------------------------------------------------------------------------------
Incremental Risk Reduction, Fatal Cancers Avoided Per Year... 2.9 4.4 10.5 8.4 11.5 24.4 58.4
Incremental Risk Reduction, Non-Fatal Cancers Avoided Per 0.2 0.3 0.6 0.4 0.8 1.3 3.5
Year........................................................
Annual Incremental Monetized Benefits, $ Million Per Year.... 17.0 25.7 61.0 48.7 67.1 142 341
Annual Incremental Radon Mitigation Costs, $ Million Per Year 34.5 26.6 60.8 54.9 72.0 150.3 408.5
\2\.........................................................
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).
\2\ Costs include treatment, monitoring, and O&M costs only.

2. Impacts on Households
The cost impact of reducing radon in drinking water at the
household level was also assessed. As expected, costs per household
increase as system size decreases as shown in Table XIII.11.

Table XIII.11.--Annual Costs per Household for Community Water Systems to Treat to Various Radon Levels \1\
[$, 1997]
----------------------------------------------------------------------------------------------------------------
VVS (25- VVS (101- VS (501- S (3301- M (10,001-
Radon level (pCi/L) 100) 500) 3300) 10K) 100K) L (> 100K)
----------------------------------------------------------------------------------------------------------------
Households Served by PUBLIC Systems Above Radon Level by Population Served
----------------------------------------------------------------------------------------------------------------
4000 \2\........................ 256.5 91.0 22.7 14.3 6.2 4.5
2000............................ 259.0 92.8 23.5 14.9 7.1 5.2
1000............................ 262.5 94.8 24.6 15.4 8.6 6.4
700............................. 264.4 96.0 25.2 15.9 9.6 7.2
500............................. 266.3 97.1 25.9 16.4 10.6 8.1

[[Page 59328]]

300............................. 269.5 99.3 26.9 17.4 12.4 9.5
100............................. 278.8 107.1 29.1 20.1 16.2 12.8
----------------------------------------------------------------------------------------------------------------
Households Served by PRIVATE Systems Above Radon Level by Population Served
----------------------------------------------------------------------------------------------------------------
4000 \2\........................ 372.4 141.1 30.3 22.8 6.6 4.4
2000............................ 375.8 143.7 31.2 23.7 7.5 5.1
1000............................ 380.5 146.3 32.6 24.7 9.1 6.3
700............................. 383.1 147.8 33.4 25.4 10.1 7.1
500............................. 385.6 149.4 34.2 26.2 11.2 7.9
300............................. 389.8 152.2 35.5 27.7 13.1 9.4
100............................. 401.5 162.4 37.9 32.1 17.1 12.6
----------------------------------------------------------------------------------------------------------------
\1\ Reflects total household costs for systems to treat down to these levels. Because EPA expects that most
systems will comply with the AMCL/MCL, most systems will not incur these household costs.
\2\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

Costs to households are higher for households served by smaller
systems than larger systems for two reasons. First, smaller systems
serve far fewer households than larger systems and, consequently, each
household must bear a greater percentage share of the capital and O&M
costs. Second, smaller systems tend to have higher influent radon
concentrations that, on a per-capita or per-household basis, require
more expensive treatment methods (e.g., one that has an 85 percent
removal efficiency rather than 50 percent) to achieve the applicable
radon level.
To further evaluate the impacts of these household costs, the costs
per household were compared to median household income data for each
system-size category. The results of this calculation, presented in
Table XIII.12 for public and private systems, indicate a household's
likely share of average incremental costs in terms of the median
income. Actual costs for individual households will reflect higher or
lower income shares depending on whether they are above or below the
median household income (approximately $30,000 per year) and whether
the water system incurs above average or below average costs for
installing treatment. For all system sizes but very very small private
systems, average household costs as a percentage of median household
income are less than one percent for households served by either public
or private systems. Average impacts exceed one percent only for
households served by very very small private systems, which are
expected to face average impacts of 1.12 percent at the 4,000 pCi/l
level and 1.35 percent at the 300 pCi/l level and for households served
by very very small public systems at the 300 pCi/l level, whose average
costs barely exceed one percent. Similar to the average cost per
household results on which they are based, average household impacts
exhibit little variability across radon levels.

Table XIII.12.--Per Household Impact by Community Groundwater Systems as a Percentage of Median Household Income
[Percent]
--------------------------------------------------------------------------------------------------------------------------------------------------------
Average Impact to Households Served by Public Systems Average Impact to Households Served by Private Systems
Exceeding Radon Levels Exceeding Radon Levels
Radon level, pCi/L ------------------------------------------------------------------------------------------------------------------
VVS (25- VVS (101- VVS (25- VVS (101-
100) 500) VS S M L 100) 500) VS S M L
--------------------------------------------------------------------------------------------------------------------------------------------------------
4000 \1\............................. 0.86 0.30 0.13 0.06 0.03 0.02 1.12 0.35 0.16 0.07 0.04 0.02
2000................................. 0.92 0.36 0.12 0.05 0.02 0.01 1.19 0.42 0.16 0.09 0.02 0.01
1000................................. 0.96 0.38 0.13 0.05 0.02 0.01 1.24 0.44 0.16 0.09 0.03 0.01
700.................................. 0.98 0.38 0.13 0.06 0.03 0.02 1.27 0.45 0.17 0.09 0.03 0.01
500.................................. 1.00 0.39 0.13 0.06 0.03 0.02 1.30 0.45 0.17 0.09 0.03 0.01
300.................................. 1.05 0.40 0.14 0.06 0.03 0.02 1.35 0.47 0.18 0.10 0.04 0.02
100.................................. 1.17 0.44 0.15 0.07 0.05 0.03 1.51 0.51 0.19 0.12 0.05 0.02
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

3. Summary of Annual Costs and Benefits
Table XIII.13 reveals that at a radon level of 4000 pCi/L
(equivalent to the AMCL estimated in the NAS Report), annual costs of
100 percent compliance with an MCL are approximately twice the annual
monetized benefits. For radon levels of 1000 pCi/L to 300 pCi/L, the
central tendency estimates of annual costs are above the central
tendency estimates of the monetized benefits.

[[Page 59329]]

Table XIII.13.--Estimated National Annual Costs and Benefits \1\ of Reducing Radon Exposures Assuming 100%
Compliance with an MCL--Central Tendency Estimate
[$ Millions, 1997]
----------------------------------------------------------------------------------------------------------------
Annualized Total Annual
Radon level (pCi/L) treatment annualized Cost per fatal monetized
costs \2\ costs \3\ cancer avoided benefits
----------------------------------------------------------------------------------------------------------------
4000 \4\....................................... 34.5 43.1 14.9 17.0
2000........................................... 61.1 69.7 9.5 42.7
1000........................................... 121.9 130.5 7.3 103
700............................................ 176.8 185.4 7.1 152
500............................................ 248.8 257.4 6.8 219
300............................................ 399.1 407.6 6.6 362
100............................................ 807.6 816.2 6.8 702
----------------------------------------------------------------------------------------------------------------
Notes:
\1\ Benefits are calculated for stomach and lung cancer assuming that risk reduction begins immediately.
Estimates assume a $5.8 million value of a statistical life and willingness to pay of $536,000 for non-fatal
cancers.
\2\ Costs are annualized over twenty years using a discount rate of seven percent. Costs include treatment,
monitoring, and O&M costs.
\3\ Costs include treatment, monitoring, O&M, recordkeeping, reporting, and state costs for administration of
water programs.
\4\ 4000 pCi/L is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

Because the costs of compliance with an MCL for small systems
outweigh the benefits at each radon level (Table XIII.14), the MMM
option was recommended for small systems to alleviate some of the
financial burden to these systems and the households they serve and to
realize equivalent or greater benefits at much lower costs. The results
of the benefit-cost analyses for MMM implementation scenarios are shown
at the end of this section and also in Section 9 of the RIA.

Table XIII.14.-- Estimated Annual Costs and Benefits for 100% Compliance With an MCL by System Size
[$Millions, 1997]
--------------------------------------------------------------------------------------------------------------------------------------------------------
System size
Radon level (pCi/l) Parameter \1\ -----------------------------------------------------------------------------------
25-100 101-500 501-3300 3301-10,000 10,001-100K >100K
--------------------------------------------------------------------------------------------------------------------------------------------------------
4000................................. Benefits..................... 0.16 0.79 2.7 2.8 7.0 3.6
Costs........................ 7.8 14.3 6.3 2.9 2.7 0.5
2000................................. Benefits..................... 0.41 2.0 6.8 6.9 17.7 9.0
Costs........................ 13.2 22.7 11.6 5.7 6.3 1.6
1000................................. Benefits..................... 1.0 4.8 16.3 16.7 42.6 21.6
Costs........................ 23.1 36.5 24.7 13.4 18.9 5.3
700.................................. Benefits..................... 1.5 7.1 24.1 24.6 62.9 31.9
Costs........................ 30.6 46.5 36.3 21.1 32.8 9.5
500.................................. Benefits..................... 2.1 10.2 34.7 35.4 90.6 45.9
Costs........................ 39.4 57.9 50.8 32.0 53.0 15.6
300.................................. Benefits..................... 3.5 16.9 57.3 58.6 150 75.9
Costs........................ 55.6 79.3 78.8 56.1 99.3 26.9
100.................................. Benefits..................... 7.2 32.7 111 113 290 147
Costs........................ 93.4 134 147 122 238 73.5
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Costs do not include recordkeeping, reporting, or state costs for administration of water programs. Recordkeeping and reporting costs are estimated
at $6.1 million for all system sizes and State administration costs for water programs are estimated at $2.5 million.

Total costs to public and private water systems, by size, were also
evaluated in the RIA. Table XIII.15 presents the total annualized costs
for public and private systems by system size category for all radon
levels evaluated in the RIA. The costs are comparable for public and
private systems across system sizes for all options. This pattern may
be due in large part to the limited number of treatment options assumed
to be available to either public or private systems in mitigating
radon.

Table XIII.15.--Average Annual Cost Per System
[$Thousands, 1997]
--------------------------------------------------------------------------------------------------------------------------------------------------------
Average costs to public systems exceeding radon levels Average costs to private systems exceeding radon levels
-------------------------------------------------------------------------------------------------------------------
Radon Level (pCi/l) VVS (25- VVS (101- VVS (25- VVS (101-
100) 500) VS S M L 100) 500) VS S M L
--------------------------------------------------------------------------------------------------------------------------------------------------------
4000................................ 8.2 12.4 18.5 49.3 82.3 484.9 7.6 10.1 15.6 43.7 72.1 468.5
2000................................ 8.3 12.6 19.1 51.3 94.1 560.7 7.7 10.3 16.2 45.5 82.4 541.8
1000................................ 8.4 12.9 26.6 60.1 115.9 693.4 7.8 10.5 16.8 47.3 100.2 670.2
700................................. 8.5 13.0 27.2 61.9 129.0 758.3 7.9 10.6 17.1 48.7 111.7 752.7
500................................. 8.5 13.2 27.8 63.7 143.2 847.8 7.9 10.7 17.5 50.3 123.9 841.6
300................................. 8.6 13.5 28.8 67.4 167.1 1000.4 8.0 10.9 18.1 53.3 144.7 992.9
100................................. 8.9 14.6 31.0 77.2 219.1 1345.3 8.2 11.6 19.1 61.8 189.6 1333.1
--------------------------------------------------------------------------------------------------------------------------------------------------------

[[Page 59330]]

Annual Per System Cost for those Systems Below Radon Levels: Monitoring Costs Only
--------------------------------------------------------------------------------------------------------------------------------------------------------
All................................. 0.3 0.3 0.4 0.6 1.1 2.6 0.3 0.3 0.4 0.6 1.1 2.6
--------------------------------------------------------------------------------------------------------------------------------------------------------

4. Benefits From the Reduction of Co-Occurring Contaminants
The occurrence patterns of industrial pollutants are difficult to
clearly define at the national level relative to a naturally occurring
contaminant such as radon. Similarly, the Agency's re-evaluation of
radon occurrence has revealed that the geographic patterns of radon
occurrence are not significantly correlated with other naturally
occurring inorganic contaminants that may pose health risks. Thus, it
is not likely that a clear relationship exists between the need to
install radon treatment technologies and treatments to remove other
contaminants. On the other hand, technologies used to reduce radon
levels in drinking water have the potential to reduce concentrations of
other pollutants as well. Aeration technologies will also remove
volatile organic contaminants from contaminated ground water.
Similarly, granular activated carbon (GAC) treatment for radon removal
effectively reduces the concentrations of organic (both volatile and
nonvolatile) chemicals and some inorganic contaminants. Aeration also
tends to oxidize dissolved arsenic (a known carcinogen) to a less
soluble form that is more easily removed from water. The frequency and
extent that radon treatment would also reduce risks from other
contaminants has not been quantitatively evaluated.
5. Impacts on Sensitive Subpopulations
The SDWA, as amended, includes specific provisions in Section
1412(b)(3)(C)(i)(V) to assess the effects of the contaminant on the
general population and on groups within the general population such as
children, pregnant women, the elderly, individuals with a history of
serious illness, or other subpopulations that are identified as likely
to be at greater risk of adverse health effects due to exposure to
contaminants in drinking water than the general population. The NAS
Report concluded that there is insufficient scientific information to
permit separate cancer risk estimates for potential subpopulations such
as pregnant women, the elderly, children, and seriously ill persons.
The NAS Report did note, however, that according to the NAS model for
the cancer risk from ingested radon, which accounts for 11 percent of
the total fatal cancer risk from radon in drinking water, approximately
30 percent of the fatal lifetime cancer risk is attributed to exposure
between ages 0 to 10.
The NAS Report identified smokers as the only group that is more
susceptible to inhalation exposure to radon progeny (NAS 1999b).
Inhalation of cigarette smoke and radon progeny result in a greater
increased risk than if the two exposures act independently to induce
lung cancer. NAS estimates that ``ever smokers'' (more than 100
cigarettes over a lifetime) may be more than five times as sensitive to
radon progeny as ``never smokers'' (less than 100 cigarettes over a
lifetime). Using current smoking prevalence data, EPA's preliminary
estimate for the purposes of the HRRCA is that approximately 85 percent
of the cases of radon-induced cancer will occur among current and
former smokers. This population of current and former smokers, which
consists of 58 percent of the male and 42 percent of the female
population, will also experience the bulk of the risk reduction from
radon exposure reduction in drinking water supplies.
6. Risk Increases From Other Contaminants Associated With Radon
Exposure Reduction
As discussed in Section 7.2 of the RIA, the need to install radon
treatment technologies may require some systems that currently do not
disinfect to do so. Case studies (US EPA 1998j) of twenty-nine small to
medium water systems that installed treatment (24 aeration, 5 GAC) to
remove radon from drinking water revealed only two systems that
reported adding disinfection (both aeration) with radon treatment (the
other systems either had disinfection already in place or did not add
it). In practice, the tendency to add other disinfection with radon
treatment may be much more significant than these case studies
indicate. EPA also realizes that the addition of chlorination for
disinfection may result in risk-risk tradeoffs, since, for example, the
disinfection technology reduces potential for infectious disease risk,
but at the same time can result in increased exposures to disinfection
by-products (DBPs). This risk-risk trade-off is addressed by the
recently promulgated Disinfectants and Disinfection By-Products NPDWR
(63 FR 69390). This rule identified MCLs for the major DBPs, with which
all CWSs and NTNCWSs must comply. These MCLs set a risk ceiling from
DBPs that water systems adding disinfection in conjunction with
treatment for radon removal could face. The formation of DBPs
correlates with the concentration of organic precursor contaminants,
which tend to be much lower in ground water than in surface water. In
support of this statement, the American Water Works Association's
WATERSTATS survey (AWWA 1997) reports that more than 50% of the ground
water systems surveyed have average total organic carbon (TOC) raw
water levels less than 1 mg/L and more than 80% had TOC levels less
than 3 mg/L. On the other hand, WATERSTATS reports that less than 6% of
surface water systems surveyed had raw water TOC levels less than 1 mg/
L and more than 50% had raw water TOC levels greater than 3 mg/L. In
fact, this survey reports that more than 85% of surface water systems
had finished water TOC levels greater than 1 mg/L.
The NAS Report addressed several important potential risk-risk
tradeoffs associated with reducing radon levels in drinking water,
including the trade-off between risk reduction from radon treatment
that includes post-disinfection with the increased potential for DBP
formation (NAS 1999b). The report concluded that, based upon median and
average total trihalomethane (THM) levels taken from a 1981 survey,
ground water systems would face an incremental individual lifetime
cancer risk due to chlorination

[[Page 59331]]

byproducts of 5 x 10^-5. It should be emphasized that this
risk is based on average and median Trihalomethane (THM) occurrence
information that does not segregate systems that disinfect from those
that do. It should also be noted that this survey pre-dates the
promulgation of the Stage I Disinfection Byproducts Rule by almost
twenty years. Further, the NAS Report points out that this average DBP
risk is smaller than the average individual lifetime fatal cancer risk
associated with baseline radon exposures from ground water (untreated
for radon), which is estimated at 1.2 x 10^-4 using a mean
radon concentration of 213 pCi/L.
While this risk comparison is instructive, a more meaningful
relationship for the proposed radon rule would be to compare the trade-
off between radon risk reduction from radon treatment and introduced
DBP risk from disinfection added along with radon treatment. EPA
emphasizes that this risk trade-off is only of concern to the small
minority (<1%) of small ground water systems with radon levels above
the AMCL of 4000 pCi/L and to the small minority of large ground water
systems that are not already disinfecting. Presently, approximately
half of all small community ground water systems already have
disinfection in place, as shown in Table XIII.5. The proportion of
systems having disinfection in place increases as the system's size
increases; >95% of large ground water systems currently disinfect. In
terms of the populations served, 83% of persons served by small
community ground water systems (those serving 10,000 persons or fewer)
already receive disinfected drinking water and 95% of persons served by
large ground water systems already receive disinfected drinking water.
As shown in Tables XIII.16 and XIII.17, even for those ground water
systems adding both radon treatment and disinfection, this risk-risk
trade-off tends to be very favorable, since the risk reduction from
radon removal greatly outweighs the added risk from DBP formation.
An estimate of the risk reduction due to treatment of radon in
water for various removal percentages and finished water concentrations
is provided in Table XIII.16. These risk reductions are much greater
than NAS's estimate of the average lifetime risk from DBP exposure for
ground water systems, by factors ranging from 3.5 for low radon removal
efficiencies (50%) to more than 130 for higher radon removal
efficiencies (>95%).

Table XIII.16.--Radon Risk Reductions Resulting from Water Treatment
------------------------------------------------------------------------
Required Reduced lifetime risk
Radon Influent (Raw Water) level, removel resulting from Water
pCi/L efficiency Treatment for Radon in
(percent) Drinking Water \1\
------------------------------------------------------------------------
500.............................. 52 1.7 x 10 ^-\4\
750.............................. 68 3.4 x 10 ^-\4\
1000............................. 76 5.1 x 10 ^-\4\
2500............................. 90 1.5 x 10 ^-\3\
4000............................. 94 2.5 x 10 ^-\3\
10000............................ 98 6.5 x 10 ^-\3\
------------------------------------------------------------------------
\1\ Assumes that water is treated to 80% of the radon MCL.

Table XIII.17 demonstrates the risk-risk trade-off between the risk
reduction from radon removal and the risks introduced from total
trihalomethanes (TTHM) for two scenarios: (1) the resulting TTHM level
is 0.008 mg/L (10% of the TTHM MCL) and (2) the resulting TTHM level is
0.080 mg/L (the TTHM MCL). The table demonstrates that the risk-risk
trade-off is favorable for treatment with disinfection, even for
situations where radon removal efficiencies are low (50%) and TTHM
levels are present at the MCL. While accounting quantitatively for the
increased risk from DBP exposure for systems adding chlorination in
conjunction with treatment for radon may somewhat decrease the
monetized benefits estimates, disinfection may also produce additional
benefits from the reduced risks of microbial contamination.

Table XIII.17.--Radon Risk Reduction from Treatment Compared to DBP
Risks
------------------------------------------------------------------------
Estimated risk ratios: (lifetime risk
reduction from radon removal \1\ /
lifetime average risk from TTHMs in
chlorinated groundwater)
Radon influent (Raw Water) level --------------------------------------
pCi/L TTHMs
present at TTHMs
(NAS) \2\ 10% of TTHM present at
MCL (0.080 MCL
mg/L) \3\
------------------------------------------------------------------------
500.............................. 4 30 3
750.............................. 7 60 6
1000............................. 10 90 9
2500............................. 30 300 30
4000............................. 50 500 50
10000............................ 130 1200 120
------------------------------------------------------------------------
Notes: \1\ From Table XIII.16.
\2\ From Appendix D in: National Research Council, Risk Assessment of
Radon in Drinking Water, National Academy Press, Washington, DC. 1999.
DBP concentrations are from a 1981 study and therefore pre-date the
Stage 1 DBP NPDWR.
\3\ US EPA Regulatory Impact Analysis for the Stage 1 Disinfectants/
Disinfection Byproducts Rule. Prepared by The Cadmus Group. November
12, 1998. Analysis is based on the 95% upper confidence interval value
from the Integrated Risk Information System (IRIS) lifetime unit risks
for each THM. TTHM is assumed to comprised by 70% chloroform, 21%
bromodichloromethane, 8% dibromochloromethane, and 1% bromoform.
\4\ US EPA. Regulatory Impact Analysis for the Stage 1 Disinfectants/
Disinfection Byproducts Rule. Based on the 95% upper confidence
interval value from the Integrated Risk Information System (IRIS) for
the lifetime unit risk for dibromochloromethane (2.4 x 10 ^-\6\ risk
of cancer case over 70 years of exposure).

[[Page 59332]]

7. Other Factors: Uncertainty in Risk, Benefit, and Cost Estimates
Estimates of health benefits from radon reduction are uncertain.
EPA is including an uncertainty analysis of radon in drinking water
risks in Section XII of the preamble to the proposed radon rule. A
brief discussion on the uncertainty analysis is also shown in Section
10 of the RIA (USEPA 1999f) for radon in drinking water. Monetary
benefit estimates are also affected by the VSL estimate that is used
for fatal cancers. The WTP valuation for non-fatal cancers has less
impact on benefit estimates because it contributes less than 1 percent
to the total benefits estimates, due to the fact that there are few
non-fatal cancers relative to fatal cancers and they receive a much
lower monetary valuation.
8. Costs and Benefits of Multimedia Mitigation Program Implementation
Scenarios
In addition to evaluating the costs and benefits across a range of
radon levels, EPA has evaluated five scenarios that reduce radon
exposure through the use of MMM programs. The implementation
assumptions for each scenario are described in the next section. These
five scenarios are described in detail in Section 9 of the RIA. For the
MMM implementation analysis, systems were assumed to mitigate water to
the 4,000 pCi/L Alternative Maximum Contaminant Level (AMCL), if
necessary, and that equivalent risk reduction between the AMCL and the
radon level under evaluation would be achieved through a MMM program.
Therefore, the actual number of cancer cases avoided is the same for
the MMM implementation scenarios as for the water mitigation only
scenario. A complete discussion on why MMM is expected to achieve equal
or greater risk reduction is shown in Section VI.B of the preamble for
the proposed radon rule.
For the RIA, EPA used a simplified approach to estimating costs of
mitigating indoor air radon risks. A point estimate of the average cost
per life saved under the current voluntary radon mitigation programs
served as the basis for estimating the costs of risk reduction under
the MMM options. The Agency has estimated the average screening and
mitigation cost per fatal lung cancer avoided to be approximately
$700,000, assuming the current distribution of radon in indoor air,
that all homes would be tested for radon in indoor air, and that all
homes at or above EPA's voluntary action level of 4 pCi/L would be
mitigated. This value was originally derived based on data gathered in
1991. The same value has been used in the RIA, without adjustment for
inflation, after discussions with personnel from EPA's Office of
Radiation and Indoor Air indicated that screening and mitigation costs
have not increased since 1991.
9. Implementation Scenarios
EPA evaluated the annual cost of five MMM implementation scenarios
that span the range of participation in MMM programs that might occur
when a radon NPDWR is implemented. Each scenario assumes a different
proportion of States will comply with the AMCL and implement MMM
programs. It has been assumed that ``50 percent of States'' implies 50
percent of systems in the U.S; ``60 percent of States'' implies 60
percent of systems, and so on.

Scenario A: 50 percent of States implement MMM programs.
Scenario B: 60 percent of States implement MMM programs.
Scenario C: 70 percent of States implement MMM programs.
Scenario D: 80 percent of States implement MMM programs.
Scenario E: 95 percent of States implement MMM programs.

States that do not implement MMM programs instead must review and
approve any system-level MMM programs prepared by community water
systems. In these States, regardless of scenario, 90 percent of systems
are assumed to comply with the AMCL and to implement a system-level MMM
program and 10 percent are assumed to comply with the MCL. EPA requests
comment on whether this is an appropriate assumption.
10. Costs and Benefits of MMM Implementation Scenarios
Table XIII.18 shows the total annual system-level and State-level
costs for each MMM scenario, assuming an MCL of 300 pCi/L and AMCL of
4,000 pCi/L. Additional MMM scenario cost and benefit tables for MCL
levels of 100, 500, 700, 1000, 2000, and 4000 pCi/L are shown in
Appendix E of the RIA. System, State, and MMM mitigation costs decrease
from $121.1 million to $60.4 million as the percentage of States
implementing MMM programs increases from 50 to 95 percent. System-level
costs decrease from $104 million to $47 million as the percentage of
States implementing MMM programs increases from 50 to 95 percent. Costs
for actual mitigation of radon in indoor air rise from $3.9 million to
$4.1 million as the percentage of States implementing MMM programs
rises from 50 to 95 percent. Note that these mitigation costs are
relatively flat because all scenarios assume that 95 percent or more of
the risk reduction will be achieved through MMM at either the State or
local level.
Table XIII.19 represents the ratios of benefits to costs of MMM
programs for each scenario, by system size. Only the ratios in the
bottom row of the table include costs to the States. The balance of the
numbers presented here represent local benefits and costs only and as
such, somewhat overstate the net benefits of the scenarios. Benefit-
cost ratios are generally less than one for the smallest system size
category (systems serving less than 500 people), but greater than one
for larger systems under all five scenarios. For larger systems,
benefit-cost ratios range from 2.6 for systems serving 501-3,300 people
under Scenario A to approximately 41.4 for systems serving 10,001 to
100,000 people under Scenario E. Overall benefit-cost ratios are over
one for all five scenarios. This pattern is seen primarily because a
larger proportion of smaller systems have influent radon levels
exceeding 4000 pCi/L. A larger proportion of small systems versus large
systems therefore, incur water mitigation costs to comply with the
AMCL.
Table XIII.20 shows the net benefits (benefits minus costs) of the
various MMM implementation scenarios. As would be expected from the
benefit-cost ratios shown in Table XIII.19, all systems serving more
than 500 people realize net positive benefits under all five scenarios.
By far the largest proportion of net benefits is realized by systems
serving 10,001 to 100,000 people.

[[Page 59333]]

Table XIII.18 (A).--Annual System--Level and State--Level Costs Associated with the Multimedia Mitigation and
AMCL Option
[$ Millions/Year] [MCL=300 pCi/L]
----------------------------------------------------------------------------------------------------------------
Scenario A Scenario B Scenario C Scenario D Scenario E 5%
45% implement 36% implement 27% implement 18% implement implement
system-level system-level system-level system-level system-level
MMM program; MMM program; MMM program; MMM program; MMM program;
5% mitigate 4% mitigate 3% mitigate 2% mitigate 5% mitigate
System size water to 300 water to 300 water to 300 water to 300 water to 300
piC/L MCL; 95% piC/L MCL; 96% piC/L MCL; 97% piC/L MCL; 98% piC/L MCL;
mitigate water mitigate water mitigate water mitigate water 99.5% mitigate
to 4000 piC/L to 4000 piC/L to 4000 piC/L to 4000 piC/L water to 4000
AMCL AMCL AMCL AMCL piC/L AMCL
----------------------------------------------------------------------------------------------------------------
System Costs for Water Mitigation ($ millions/year)
----------------------------------------------------------------------------------------------------------------
25-100.......................... 10.2 9.7 9.3 8.8 8.1
101-500......................... 17.6 16.9 16.3 15.6 14.6
501-3300........................ 9.9 9.2 8.5 7.7 6.7
3301-10,000..................... 5.5 5.0 4.5 3.9 3.1
10,001-100,000.................. 7.5 6.6 5.6 4.6 3.2
>100,000........................ 2.0 1.7 1.4 1.1 0.7
-------------------------------------------------------------------------------
Total CWS Water Mitigation 52.7 49.1 45.4 41.8 36.3
Costs......................
----------------------------------------------------------------------------------------------------------------
Water System Administration Costs ($ millions/year)
----------------------------------------------------------------------------------------------------------------
25-100.......................... 17.0 14.0 11.0 8.0 3.7
101-500......................... 17.4 14.3 11.3 8.2 3.8
501-3300........................ 12.0 9.9 7.8 5.7 2.6
3301-10,000..................... 3.0 2.5 1.9 1.4 0.6
10,001-100,000.................. 1.7 1.4 1.1 0.8 0.4
>100,000........................ 0.1 0.1 0.1 0.0 0.0
-------------------------------------------------------------------------------
Total CWS Administrative 51.2 42.1 33.1 24.1 11.1
Costs......................
===============================================================================
Total CWS Water 104.0 91.2 78.5 65.9 47.4
Mitigation and
Administrative Costs...
----------------------------------------------------------------------------------------------------------------

Table XIII.18 (B).--State MMM Administrative Costs
[$ millions/year]
----------------------------------------------------------------------------------------------------------------
Scenario A 50% Scenario B 60% Scenario C 70% Scenario D 80%
of states of states of states of states Scenario E 95%
implement implement implement implement of states
state-wide MMM state-wide MMM state-wide MMM state-wide MMM implement
programs; 45% program; 35% program; 25% program; 15% state-wide MMM
of CWS of CWS of CWS of CWS program; 5% of
implement implement implement implement CWS implement
system-level system-level system-level system-level system-level
MMM program MMM program MMM program MMM program MMM program
----------------------------------------------------------------------------------------------------------------
State costs associated with State-wide MMM program administration, reviewing system-level MMM programs, and
reviewing system-level water mitigation requirements are not distributable across different system sizes.
----------------------------------------------------------------------------------------------------------------
State Administration Costs for 2.5 2.5 2.5 2.5 2.5
Water Mitigation...............
State Administration Costs for 2.9 3.5 4.1 4.7 5.6
State-Level MMM Mitigation.....
State Administration Costs for 7.8 6.1 4.4 2.6 0.9
System-Level MMM Mitigation....
-------------------------------------------------------------------------------
Total State 13.2 12.1 10.9 9.8 8.9
Administration Costs...
----------------------------------------------------------------------------------------------------------------

Table XIII.18 (C).--MMM Testing and Mitigation Costs
[$ million/year]
----------------------------------------------------------------------------------------------------------------
Scenario A Scenario B Scenario C Scenario D Scenario E
----------------------------------------------------------------------------------------------------------------
CWS MMM Costs................... 1.9 1.5 1.1 0.7 0.2
State MMM Costs................. 2.1 2.5 2.9 3.3 3.9
-------------------------------------------------------------------------------
Total MMM Costs............. 3.91 3.95 3.99 4.03 4.12
===============================================================================

[[Page 59334]]

Total Costs (From Tables 121.1 107.3 93.4 79.7 60.4
XIII.18 A, B, and C)...
----------------------------------------------------------------------------------------------------------------

Table XIII.19.--Ratio of Benefits and Costs by System Size for Each Scenario (MCL=300 pCi/L)
--------------------------------------------------------------------------------------------------------------------------------------------------------
System size Benefits, $M Scenario A Scenario B Scenario C Scenario D Scenario E
--------------------------------------------------------------------------------------------------------------------------------------------------------
25-100................................................. 3.5 0.13 0.14 0.17 0.21 0.30
101-500................................................ 16.9 0.48 0.53 0.61 0.70 0.92
501-3,300.............................................. 58.0 2.59 2.98 3.51 4.27 6.23
3,301-10,000........................................... 59.2 6.87 7.85 9.16 11.0 15.61
10,001-100,000......................................... 147.3 15.82 18.35 21.84 26.96 41.43
>100,000............................................... 76.7 37.16 43.70 53.04 67.44 113.68
------------------------------------------------------------------------------------------------
OVERALL........................................ 361.6 2.98 3.37 3.87 4.54 5.99
--------------------------------------------------------------------------------------------------------------------------------------------------------

Table XIII.20.--Net Benefits by System Size for Each Scenario \1\
--------------------------------------------------------------------------------------------------------------------------------------------------------
System size Benefits, $M Scenario A Scenario B Scenario C Scenario D Scenario E
--------------------------------------------------------------------------------------------------------------------------------------------------------
25-100.................................................. 3.5 (24.3) (20.7) (17.1) (13.5) (8.3)
101-500................................................. 16.9 (18.7) (14.8) (11.0) (7.1) (1.6)
501-3,300............................................... 58.0 35.6 38.6 41.5 44.4 48.7
3,301-10,000............................................ 59.2 50.6 51.7 52.7 53.8 55.4
10,001-100,000.......................................... 147.3 138.0 139.3 140.6 141.8 143.7
>100,000................................................ 76.7 74.6 74.9 75.3 75.6 76.0
-----------------------------------------------------------------------------------------------
OVERALL......................................... 361.6 240.5 254.3 268.2 281.9 301.2
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Parentheses indicate negative numbers.

H. Response to Significant Public Comments on the February 1999 HRRCA

To provide the public with opportunities to comment on the Health
Risk Reduction and Cost Analysis (HRRCA) for radon in drinking water,
the Agency published the HRRCA in the Federal Register on February 26,
1999 (64 FR 9559). The HRRCA was published six months in advance of
this proposal and illustrated preliminary cost and benefit estimates
for various MCL options under consideration for the proposed rule. The
comment period on the HRRCA ended on April 12, 1999, and EPA received
approximately 26 written comments from a variety of stakeholders,
including the American Water Works Association, the National Rural
Water Association, the National Association of Water Companies, the
Association of Metropolitan Water Agencies, State departments of
environmental protection, State health departments, State water
utilities and local water utilities.
Significant comments on the HRRCA addressed the topics of radon
occurrence, exposure pathways, sensitive sub-populations and the risks
to smokers, risks from existing radon exposures, risks associated with
co-occurring contaminants, risk increases associated with radon
removal, the benefits of reduced radon exposures, the costs of radon
treatment measures, the cost and benefit results, and the Multimedia
Mitigation (MMM) program. The following discussion outlines the
significant comments received on the HRRCA and the Agency's response to
these comments.
1. Radon Occurrence
Several commenters had concerns related to EPA's analysis of radon
occurrence. Two commenters felt that the radon levels in Table 3.1 of
the HRRCA were too low and not representative of radon occurrence in
their regions. A California water utility indicated that due to
limitations of the NIRS, EPA should conduct a new national radon
survey, with special emphasis on determining radon levels in the
largest systems, before promulgating the rule. Two commenters from
Massachusetts expressed concerns about radon occurrence. One suggested
that additional analysis of radon variability in individual wells was
required, and another indicated that the effects of storage and
residence time on radon levels in supply systems needed to be taken
into account. One commenter indicated that EPA should more strongly
consider that most risk reductions predicted in the HRRCA come from
reductions in radon levels in the small proportions of systems with
initial very high radon levels.
EPA Response 1-1
As part of the regulatory development process, EPA updated and
refined its analysis of radon occurrence patterns in ground water
supplies in the United States. This new analysis incorporated
information from the EPA 1995 National Inorganic and Radionuclides
Survey (NIRS) of 1000 community ground water systems throughout the
United States, along with supplemental data provided by States, water
utilities, and academic researchers. EPA's current re-evaluation used
data from 17 States to determine the differences between radon levels
in ground water and radon levels in distribution systems in the same
regions. The results of these comparisons were used to estimate
national distributions of radon occurrence in ground water. EPA
believes that the existing NIRS data, along with the Agency's updates
to this data, currently provide the most comprehensive national-level
analysis of radon occurrence patterns in ground water supplies. This
analysis is not intended for the estimation of radon occurrence at the
state-level.

[[Page 59335]]

Variability within the NIRS radon occurrence data was analyzed for
several important contributing factors: within-well (temporal)
variability, sampling and analytical (methods) variability, intra-
system variability (variability between wells within a single system),
and inter-system variability (variability between wells in different
systems). Several important conclusions were drawn from this analysis.
First and foremost is the conclusion that the NIRS data do capture the
major sources of radon occurrence variability and thus can be used
directly, without any additional correction for temporal or sampling
and analytical variability, to provide reasonable national estimates of
radon levels and variability levels in ground water drinking supplies.
In addition, EPA analyzed the additional data sets provided from
stakeholders (described previously) in conjunction with the NIRS radon
data to estimate the magnitudes of the variability sources. Based on
all of these analyses, EPA has concluded that the variability between
systems dominates the over-all variability (it comprises approximately
70 percent of the over-all variability). Temporal variability (13-18
percent), sampling and analytical variability (less than 1 percent),
and intra-system variability (12-17 percent) are relatively minor by
comparison. These results are discussed in detail elsewhere (USEPA
1999b).

Note: These estimates of variability sources apply to national-
level radon occurrence estimates: individual regions may have
systems that show variability sources that deviate significantly
from these values.

This analysis of variability was incorporated into EPA's estimates of
nation-wide radon occurrence and was used in its estimates of the
effects of uncertainty in occurrence information on total national
costs of compliance.
In response to the comment that ``most risk reductions predicted in
the HRRCA come from reductions in radon levels in the small proportions
of systems with initial very high radon levels'', EPA agrees that a
system with high radon levels would benefit more from water mitigation
than a system with much lower initial radon levels, but the vast
majority of the national water mitigation benefits come from systems
that are above the MCL, but not that high above it (e.g., 80 percent
removal required for the system to be at the MCL). This is true since
radon is approximately log-normally distributed (i.e., a much higher
percentage of water systems can be expected to have relatively low
radon levels than relatively high radon levels) and hence most systems
fall into this category. For this reason, the summation of these
smaller per system benefits enjoyed by the large number of systems
nearer the MCL greatly outweigh summation of the larger per system
benefits enjoyed by the minority of systems with very high radon
levels. This is demonstrated in Table 6-2 of the HRRCA (``Estimated
Monetized Benefits from Reducing Radon in Drinking Water''), in which
the central tendency estimate of monetized benefits associated with an
MCL of 500 pCi/L is 212 million dollars and the benefits associated
with an MCL of 100 pCi/L is 673 million dollars. This means that, in
the latter case, 461 million dollars of the benefits come just from the
systems with radon levels between 100 and 500 pCi/L (80 percent removal
required), while the remaining benefits (212 million dollars) come from
the systems with radon levels from 500 pCi/L up to the highest radon
levels.
Five commenters indicated that the estimates of the numbers of
entry points per system used in the HRRCA were incorrect, in that large
systems had far more entry points than the numbers given in Table 5.4
of the HRRCA. Several of these commenters cited data from the Community
Water System Survey (CWSS), showing higher numbers of wells per system
in each system size category than were used for cost calculations in
the HRRCA.
EPA Response 1-2
The relevant distribution for costing out non-centralized treatment
is the number of entry points, not the number of wells. A given entry
point (the point at which treatment is applied) may be fed by several
wells, and hence there is a discrepancy in numbers between the HRRCA,
which reported a distribution of entry points, and Table 1-5 of the
Community Water System Survey (CWSS), which reported the average number
of wells per system. These numbers are related, but not directly
comparable. In general, the average number of entry points for a class
of ground water systems would be expected to be smaller than the
average number of wells. In the HRRCA, the distribution of entry points
per system was estimated from a statistical analysis (``bootstrap
analysis'') of the well and entry point data from the CWSS. This
statistically-calculated distribution was then used to estimate the
percentage of systems within a system size category having a given
number of entry points. However, as part of its uncertainty analysis,
EPA has used the 95% confidence upper bound of the site distribution in
the national cost estimates supporting this proposal. The average
number of entry points per system is roughly 10% higher using this
upper bound analysis. In addition, to test the effects of varying this
distribution on the national costs of compliance, the per system costs,
and the per household costs, EPA conducted an uncertainty analysis
(Monte Carlo analysis including sensitivity) on the distribution by
simultaneously varying both the percentages of systems estimated to
have a particular number of sites and the estimated number of sites.
The results of this analysis are reported both in this notice and in
the Regulatory Impact Analysis. It should be noted that the treatment
unit costs and total number of systems dominated the cost uncertainty
and that the entry point distribution was a relatively minor
contributor to the overall cost uncertainty.
2. Exposure Pathways
A number of issues related to radon exposure pathways were raised.
Several commenters indicated that the risks associated with the build-
up of radon in carbon filters needed to be addressed in HRRCA. Concerns
were also expressed about general population exposures to radon in air
released from aeration facilities and exposures to workers at water
utilities. Another commenter said that EPA should discuss the
persistence of radon in the body after ingestion.
EPA Response 2-1
The risks from radon build-up in carbon filters and radon off-gas
emissions are discussed in some detail in this notice, including an
evaluation of risks, a discussion of references, and responses from a
survey of air permitting boards about the permitting of radon off-gas.
EPA Response 2-2
The persistence of radon in the body following ingestion has been
investigated and the results have been presented in the Criteria
Document for Radon (USEPA 1999b). In brief, radon ingested in water is
well-absorbed from the stomach and small intestine into the bloodstream
and transported throughout the body. Radon is rapidly (within
approximately one hour) excreted from the body via the lungs, so only
about 1 percent of ingested radon undergoes radioactive decay while in
the body. The risks from the retained radon and its decay products in
various organs are calculated by NAS and adopted by EPA in the proposed
rule.

[[Page 59336]]

3. Nature of Health Impacts
No comments were made concerning the general nature of adverse
effects associated with radon exposure. Comments concerning specific
aspects of health impact evaluation are summarized in the following
sections.
(a) Sensitive subpopulations, risks to smokers, non-smokers.
Comments on these sections are addressed together because the majority
of the comments had to do with the characterization of smokers as a
sensitive population. Several commenters noted that most risk reduction
from reducing radon exposure occurs among smokers, and took the
position that EPA should not include risk reductions to smokers in its
benefits assessment, because smoking can be viewed as a voluntary risk.
One commenter suggested that the smokers' willingness to pay for
cigarettes also indicated a willingness to face the risk of smoking.
EPA Response 3-1
The term, ``groups within the general population'' is addressed,
but not comprehensively defined, in the 1996 amendments to the Safe
Drinking Water Act (SDWA, Sec. '1412(b)(3)(C)). The definition of
sensitive subpopulations is an issue for discussion and debate, and EPA
is interested in input from stakeholders. The National Academy of
Sciences (NAS) Radon in Drinking Water Committee, as part of their
assessment of the risks of radon in drinking water, has considered
whether groups within the general population, including smokers, may be
at increased risk. The NAS Committee has indicated, in their Risk
Assessment of Radon in Drinking Water report, that smokers are the only
group within the general population that is more susceptible to
inhalation exposure to radon progeny, but did not specifically identify
smokers as a sensitive subpopulation.
In this proposal, EPA is basing its risk management decision on
risks to the general population. The general population includes
smokers as well as former smokers. The risk assessments for radon in
air and water are based on an average member of the population, which
includes smokers, former smokers, and non-smokers. A more complete
discussion on the risks of radon in drinking water and air is presented
in the NAS's risk assessment report and in Section XII of this
preamble.
(b) Risk reduction model, risks from existing radon exposures.
Commenters raised only one concern associated with the risk model used
to estimate radon reduction benefits. Three commenters suggested that
EPA should consider adopting a threshold-based model for radon
carcinogenesis, and that EPA's current (non-threshold) approach
overestimates radon risks. In support, the commenters cited a recently
published paper (Miller et al, 1999) as providing evidence that a
single alpha particle ``hit'' typical in low-level radon may not be
sufficient to cause cell transformation leading to cancer.
EPA Response 3-2
There are a number of papers that have recently examined the
effects of a single alpha particle on a cell nucleus of mammalian cells
in culture. The authors of this study concluded that cells were more
likely to be transformed to cancer causing cells if there were multiple
alpha particle hits to their nuclei. However, another study, Hei et al.
(1997), using a similar methodology, found direct evidence that a
single ``particle traversing a cell nucleus will have a high
probability of resulting in a mutation'' and concluded that their work
highlighted the need for radiation protection at low doses. Moreover,
follow-up microbeam experiments described by Miller et al. at the 1999
International Congress of Radiation Research demonstrated that one
alpha particle track through the nucleus was indeed sufficient to
induce transformation under some experimental conditions.
Epidemiological data relating to low radon exposures in mines also
indicate that a single alpha track through the cell may lead to cancer.
Finally, while not definitive by themselves, the results from
residential case-control studies provide some direct support for the
conclusion that environmental levels of radon pose a risk of lung
cancer. EPA has based its current risk estimates for radon in drinking
water on the findings of the National Academy of Sciences. Rather than
focus on the results of any one study, the NAS committees based their
conclusions on the totality of data on radon--a weight-of-evidence
approach.
Both the BEIR VI Report (NAS 1999a) and their report on radon in
drinking water (NAS 1998b) represent the most definitive accumulation
of scientific data gathered on radon since the 1988 NAS BEIR IV (NAS
1988). These committees' support for the use of linear-non-threshold
relationship for radon exposure and lung cancer risk came primarily
from their review of the mechanistic information on alpha-particle-
induced carcinogenesis, including studies of the effect of single
versus multiple hits to cell nuclei.
In the BEIR VI report (NAS 1999a), the NAS concluded that there is
good evidence that a single alpha particle (high-linear energy transfer
radiation) can cause major genomic changes in a cell, including
mutation and transformation that potentially could lead to cancer. They
noted that even if substantial repair of the genomic damage were to
occur , ``the passage of a single alpha particle has the potential to
cause irreparable damage in cells that are not killed.'' Given the
convincing evidence that most cancers originate from damage to a single
cell, the committee went on to conclude that ``on the basis of these
[molecular and cellular] mechanistic considerations, and in the absence
of credible evidence to the contrary, the committee adopted a linear-
nonthreshold model for the relationship between radon exposure and
lung-cancer risk. However, the BEIR VI committee recognized that it
could not exclude the possibility of a threshold relationship between
exposure and lung cancer risk at very low levels of radon exposure.''
The NAS committee on radon in drinking water (NAS 1999b) reiterated the
finding of the BEIR VI committee's comprehensive review of the issue,
that a ``mechanistic interpretation is consistent with linear, non-
threshold relationship between radon exposure and cancer risk''. The
committee noted that the ``quantitative estimation of cancer risk
requires assumptions about the probability of an exposed cell becoming
transformed and the latent period before malignant transformation is
complete. When these values are known for singly hit cells, the results
might lead to reconsideration of the linear no-threshold assumption
used at present.'' EPA recognizes that research in this area is on-
going but is basing its regulatory decisions on the best currently
available science and recommendations of the NAS that support use of a
linear non-threshold relationship.
(c)Risk and risk reduction associated with co-occurring
contaminants. Several commenters addressed the issue of risks
associated with co-occurring contaminants. Other commenters indicated a
need to include risks and risk reductions from co-occurring
contaminants.
EPA Response 3-3
The contaminants that may co-occur with radon that are of main
concern are those that can cause fouling of aeration units (or
otherwise impede treatment) and those that are otherwise affected by
the aeration process in such a way as to increase risks. Measures and
costs to avoid aeration fouling are discussed in

[[Page 59337]]

this notice and in the references cited. Arsenic co-occurrence may be
relevant since some systems may have to treat for both, but the
treatment processes are not incompatible. In fact, the only side-effect
of the aeration process that may impact the removal of arsenic would be
the potential oxidation of some fraction of less easily removed As(IV)
form to the more easily removed As(VI) form. There would be no
additional costs due to this effect, and in fact, there may be cost
savings involved. The potential for increased risks due to potential
disinfectant by-product formation after disinfection, is discussed
next.
(d) Risk increases associated with radon removal. Five commenters
said that EPA should include quantitative estimates of the risk
increases associated with increased exposure to disinfection byproducts
(DBPs) in the risk and cost-benefit analyses of the HRRCA. One
commenter said that risks should be apportioned appropriately between
the proposed radon rule and the Groundwater rule. Another commenter
maintained that, contrary to the assertion in the HRRCA, there would be
no reduction in microbial risks due to the increased disinfection
associated with the radon rule because most groundwater sources
currently present no microbial risks.
EPA Response 3-4
EPA would like to highlight that the AMCL/MMM option is the
preferred option for all drinking water systems, which would result in
very few water treatment systems adding disinfection. EPA expects the
radon rule to result in a minority of ground water systems choosing the
MCL option, and of those, many will be larger systems. Since very few
small systems are expected to choose the MCL option , very few systems
are above the AMCL of 4000 pCi/L, and most large ground water systems
already disinfect their water, few systems are expected to add
disinfection in response to the radon rule, i.e., increased risk due to
disinfection by-product formation should not be a significant issue.
However, EPA does evaluate this risk-risk trade-off in this notice for
that minority of systems that will be expected to add disinfection with
treatment for radon. For that minority of systems, the trade-off
between decreased risks from radon and increased risks from
disinfection-by-products is favorable.
4. Benefits of Reduced Radon Exposure
The majority of the comments related to the estimation of benefits
focused on the methods used to monetize reductions in cancer risks.
There were also a few comments on non-quantifiable benefits, and on
several other topics. The previous comments pertaining to risk
reductions to smokers and that benefits from these risk reductions
should be excluded from the HRRCA apply here as well.
(a) Nature of regulatory benefits. There were few comments on this
section, most of which pertained to non-quantifiable benefits. One
commenter indicated that the peace-of-mind non-quantifiable benefit
from radon reduction would be offset by the anxiety of those living
near aeration plants. Another noted that peace-of-mind benefits were
not easy to quantify for non-threshold pollutants like radon and, in
fact, that the regulation of radon might actually increase anxiety by
drawing attention to the risks associated with radon exposures.
Commenters also noted that claiming arsenic reduction as a benefit from
aeration is questionable because there is no demonstrated correlation
between the levels of radon and arsenic in groundwater systems.
EPA Response 4-1
By definition, non-quantifiable benefits cannot be measured and
have not been measured in the HRRCA analysis. Thus, comparisons of
types of such benefits are not very meaningful. EPA attempts to note
these potential benefits when the Agency believes they might occur, as
in the case of peace-of-mind benefits from radon reduction. There may
also be non-quantifiable costs that may offset any non-quantifiable
benefits. These include anxiety on the part of residents near treatment
plants and customers who may not have previously been aware of radon in
their water. As noted elsewhere in this preamble, EPA believes it
unlikely that accounting for these non-quantifiable benefits and costs
quantitatively would significantly alter the overall assessment.
(b) Monetization of benefits. Comments related to risk reduction
have been discussed in previous responses, so are not discussed further
here. Commenters addressed all three approaches to monetizing benefits:
the value of statistical life; the costs of illness; and willingness-
to-pay. A number of commenters suggested the use of Quality-Adjusted
Life Years (QALY) as an alternative approach to the valuation of health
benefits. One commenter indicated that the use of QALYs was a good way
to avoid having to monetize health outcomes. Two commenters indicated
that QALYs had the advantage of being able to take into account the
delayed onset of cancer, as well as reduced incidence. One organization
suggested QALYs as a superior method for combining the benefits from
fatal and non-fatal illness over different time periods; which would be
particularly useful in the case of smokers, whose cancers are likely to
be delayed, but not necessarily prevented, by reductions in radon
exposure.
EPA Response 4-2
The use of QALYs has been extensively discussed within EPA and also
before the Environmental Economics Advisory Committee of EPA's Science
Advisory Board. At this time, current Agency policy is to use Value of
Statistical Life (VSL) estimates for the monetization of risk reduction
benefits. EPA believes QALY calculations to be experimental and not
well established for the types of analyses performed by the Agency.
(c) Value of statistical life (VSL). Several commenters questioned
the use of, or the value selected for, the value of statistical life as
a measure of benefits. Other commenters indicated that the large range
of uncertainty associated with the estimates of risk reduction called
the VSL (and the willingness-to-pay) methods into question, and
indicated that EPA needed to better justify the central-tendency VSL
value selected for use in the HRRCA. They maintained that the VSL
approach would only be appropriate if the VSL estimates were derived
from ``similar scenarios'' to those being evaluated in the HRRCA.
Another commenter suggested that using the VSL was inappropriate in
that the VSL dollars did not represent (as do compliance costs) actual
resource losses to society that could be spent on other programs (e.g.
pollution reduction). Thus, the comparison of compliance costs to VSL
costs is not valid. They strongly recommend the use of compliance cost
per life saved as an appropriate measure for judging radon control
options. One commenter indicated that the use of the VSL approach
resulted in greatly over-estimated benefits of radon exposure
reduction, particularly because the VSL for smokers is the same as for
non-smokers and does not account for the age at which mortality is
avoided. Another questioned the validity of the mean VSL value used in
the HRRCA, and indicated that VSL estimates should only come from the
peer-reviewed scientific literature or from Agency documents that had
been subject to public comment.

[[Page 59338]]

EPA Response 4-3
The VSL value, currently recommended by Agency guidance, is derived
from a statistical distribution of the values found in twenty-six VSL
studies, which were chosen as the best such studies available from a
larger body of studies. This examination of studies was undertaken by
EPA's Office of Air and Radiation in the course of its Clean Air Act
retrospective analysis. EPA believes the VSL estimate ($5.8 million,
1997 dollars) to be the best estimate at this time, and is recommending
that this value be used by the various program offices within the
Agency. This estimate may, however, be updated in the future as
additional information becomes available to assist the Agency in
refining its VSL estimate. The VSL estimate is consistent with current
Agency economic analysis guidance, which was recently peer reviewed by
EPA's Science Advisory Board.
d. Costs of illness (COI). Two commenters suggested that EPA should
further review the literature on the costs of illness and develop
better cost measures for the illnesses addressed in the HRRCA.
EPA Response 4-4
EPA believes that the COI data is the most complete analysis of
this type currently underway. The cost of illness (COI) data shown in
the HRRCA were presented as a comparison to Willingness to Pay (WTP) to
avoid chronic bronchitis. The Agency did not use the COI data to
estimate risk reduction valuations for non-fatal cancers because these
estimates can be seen as underestimating the total WTP to avoid non-
fatal cancers. COI may understate total WTP because of its failure to
account for many effects of disease such as pain and suffering,
defensive expenditures, lost leisure time, and any potential altruistic
benefits. It is important to note that the proportion of benefits
attributable to non-fatal cancer cases accounts for less than one
percent of the total benefits in the HRRCA.
(e) Willingness-to-pay. Several commenters questioned EPA's use of
the willingness-to-pay (WTP) approach for monetizing non-fatal cancer
risk reductions. Another suggested that a WTP value for victims of non-
fatal cancers should have been used, instead of the WTP estimates for
chronic bronchitis. It was also suggested that WTP measures would vary
within the general population, and that use of a constant value was
inappropriate.
EPA Response 4-5
EPA believes that the WTP estimates to avoid chronic bronchitis are
the best available surrogate for WTP estimates to avoid non-fatal
cancers. WTP estimates were used in the HRRCA as opposed to COI to
value non-fatal cancer cases. EPA believes that COI may understate
total WTP because of its failure to account for many effects of disease
such as pain and suffering, defensive expenditures, lost leisure time,
and any potential altruistic benefits. It is important to note that the
proportion of benefits attributable to non-fatal cancer cases accounts
for less than one percent of the total benefits in the HRRCA.
(f) Treatment of benefits over time. Many commenters objected to
EPA's assumption that cancer risk reduction, and hence benefits, would
begin to accrue immediately upon the reduction of radon exposures. In
addition, they felt that the failure to discount health benefits
resulted in an overestimation of the benefits. One commenter suggested
that a ``gradual phase-in'' of risk reduction should be incorporated
into the HRRCA benefits calculation. It was also suggested that an
alternative to immediate benefits accrual be used, and that the effects
of the immediate benefits accrual assumption be discussed in detail
with regard to the uncertainties it introduces into the benefits
estimates. One commenter identified the assumption of immediate
benefits as a major source of benefits overestimation. Another comment
asked that EPA provide better justification for assuming immediate
benefits accrual, and suggests instead that a linear phase-in of risk
reduction over 70 years would be more appropriate. Three commenters
also indicate that the failure to take latency of risk reduction into
account and to discount benefits appropriately, greatly biases the
benefits estimates in the upward direction. One commenter indicated
that the failure to discount benefits resulted in a five- to ten-fold
over-estimation.
EPA Response 4-6
These comments address the issue of latency, the difference between
the time of initial exposure to environmental carcinogens and the onset
of any resulting cancer. Qualitative language has been added to the
preamble regarding adjustments, including latency, that could be made
to benefits calculations. This qualitative discussion notes that
latency is one of a number of adjustments related to an evaluation of
potential benefits associated with this rule. EPA believes that such
adjustments should be considered simultaneously. For further
discussion, see section XIII.D of the preamble.
5. Costs of Radon Treatment Measures
(a) Drinking water treatment technologies and costs. All of the
commenters had concerns related to EPA's assumptions and analyses of
costs of radon treatment measures. In fact, one commenter suggested
that the entire section was oversimplified by EPA. Most of the
commenters, however, provided more specific comments which are outlined
next.
EPA Response 5-1
Most, if not all, commenters assumed that EPA would propose that
the risks from radon would be best addressed by drinking water systems
attempting to meet the MCL. Under this scenario, many small systems
would be in situations where they faced very difficult treatment
issues, often with technically difficult and/or expensive solutions.
However, EPA is suggesting that the risks from radon are best addressed
by the combined use of the AMCL with a multi-media mitigation (MMM)
program. Since the proposal also includes a regulatory expectation of
adoption of the AMCL by small systems, EPA believes that many of the
comments received are less applicable to this proposal than if the MCL
were the preferred route of compliance.
(b) Aeration. Several commenters expressed concerns related to
aeration costs. One major concern was EPA's failure to address worker
safety issues, and the associated cost of occupational safety programs,
at treatment plants. A reference to earlier studies of increased risk
to neighbors is provided, but details are not included to evaluate
these studies. Concern was expressed that costs for permitting and
control of radon emissions from treatment plants were not included, and
that the public might react strongly to the presence of a local
treatment plant even if analysis showed the risk to be minimal. Three
commenters noted that the HRRCA failed to consider quantifiable
corrosion control costs associated with aeration. Installation of
aeration for radon removal may also affect lead/copper levels in the
water distribution system, resulting in additional treatment
modifications and costs. Many systems will have to develop a different
corrosion control strategy to comply with the lead and copper rule due
to the radon regulation.
EPA Response 5-2
Worker safety issues for aeration treatment of radon in drinking
water are discussed in today's notice (Section

[[Page 59339]]

VIII.A.3) and are discussed in more detail in other sources (USEPA
1994b, USEPA 1998h). Radon exposure to workers in drinking water
treatment plants has been discussed in the literature (e.g., Fisher et
al. 1996, Reichelt 1996). In fact, these discussions usually apply to
situations where radon is NOT the contaminant being purposely removed,
since there is currently no regulatory driver to do so. When ground
water is exposed to air during treatment for any contaminant, radon may
be released and may accumulate in the treatment facility. The National
Research Council (NAS 1999b) suggests that the air in all groundwater
facilities treating for any contaminant should be monitored for radon
and that ventilation should be investigated as a means of reducing
worker exposure. In support of this position, EPA would further
strongly suggest that systems that attempt to meet the MCL (i.e., that
are in States that do not adopt the AMCL or otherwise choose to meet
the MCL) by installing aeration treatment should take the appropriate
measures to monitor and ventilate the treatment facilities. For those
small systems that choose GAC treatment, other precautions should be
taken to monitor and control gamma exposure. GAC treatment issues are
discussed later in this notice and are discussed in detail elsewhere
(USEPA 1994b, AWWARF 1998 and 1999).
EPA has suggested that occupational exposures be limited to 100
mRem/year, a level well below the upper limit of 5000 mRem/year
approved in by the President in 1987 (``Radiation Exposure Guidance to
Federal Agencies for Occupational Exposure'', as cited in USEPA 1994b).
Based on limited data, it appears that 100 mRem/year is a maintainable
objective within water treatment plants treating for radon or other
contaminants. Exposure level monitoring and mitigation through a
combination of air monitoring and ventilation has been demonstrated to
be feasible and relatively inexpensive (e.g., Reichelt 1996).
Regarding the effects on water corrosivity and the impacts of costs
of corrosion control measures, this notice presents much more detail on
EPA's assumptions. Corrosion control measures are included in national
cost estimates and are discussed in this notice. Case study information
on corrosion control costs associated with aeration are included in the
Radon Technologies and Costs document (USEPA 1999h).
(c) GAC. Two commenters noted that the option for use of granular
activated carbon (GAC) did not address potential problems with
radioactivity buildup in the carbon. In consideration of treatment
methods the two commenters saw no mention of the cost of disposal of
GAC used for radon removal. If not replaced in time it will become a
low level radioactive waste because of Lead 210 and will become
difficult to dispose of. Other issues that need to be addressed
include: will the unit require special shielding; may the charcoal bed
be required to have a radioactive materials license from the State; and
how may radioactive carbon be disposed of?
EPA Response 5-3
Special considerations regarding GAC operations, maintenance, and
ultimate GAC unit disposal are discussed in some detail in Section
VIII.A of this notice, including discussions of the radiation hazards
involved and steps that can be taken to ameliorate these hazards. GAC
disposal costs are included in the operations and maintenance costs in
the model used for cost estimates. Comparisons of modeled GAC capital
and operations & maintenance cost estimates to actual costs reported in
case studies are included in Section VIII of this notice. EPA would
like to strongly emphasize that carbon bed lifetimes (carbon bed
replacement rates) should be designed to preclude situations where
disposal becomes prohibitively expensive or technically infeasible.
Recently, the American Water Works Association Research Foundation
has published a study on the use of GAC for radon removal, which
includes discussions of the issues described previously, that concludes
that GAC is a tenable treatment strategy for small systems when used
properly under the appropriate circumstances (AWWARF 1998a). AWWARF
also reviewed the proper use of GAC for radon removal in its recent
review of general radon removal strategies (AWWARF 1998b). When the
final radon rule is promulgated, a guidance manual will be published
describing technical issues and solutions for small systems installing
treatment.
One commenter suggested that the costs for GAC seemed to be too
high. The figures used in the analysis could be two orders of magnitude
above the costs actually seen by the systems.
EPA Response 5-4
EPA agrees that its GAC cost estimates seem to be very high, as
compared to case studies (USEPA 1999h, AWWARF 1998b). EPA agrees with
others (e.g., AWWARF 1998a and b) that GAC will probably be cost-
effective for very small systems or in a point-of-entry mode. This
issue is addressed in the preamble (Section VIII.A) and GAC will be
included as a small systems compliance technology.
(d) Regionalization. Two commenters questioned a cost of $280,000
as the single cost for regionalization. Assuming $100/foot for an
interconnection, these costs would equate to an interconnection of 2800
feet which seems low. Systems are usually separated by more than one-
half mile. A range of costs may need to be considered rather than a
single number. Smaller systems will have smaller costs, while large
systems will have larger costs. Thus, the charge for regionalization
should vary by systems size. Also, EPA should clarify whether or not
regionalization charges include yearly operation and maintenance costs.
EPA Response 5-5
EPA agrees that the costs of regionalization would be expected to
change with water system size, but, as indicated in the assumptions
outlined in the February 26, 1999 HRRCA, EPA assumed that only very
small systems (those serving fewer than 500) would resort to
regionalization in response to the radon rule. Given that the proposed
rule involves a multi-media approach that greatly encourages small
systems to choose the AMCL of 4000 pCi/L in conjunction with a multi-
media mitigation program, EPA expects that very few systems would
choose regionalization as an option. EPA believes that the assumption
that 1 out of 100 small systems that choose the MCL option would
regionalize is conservative and would only be exercised if
regionalization were cost-competitive with other options, except under
very unusual circumstances. Since the estimate of $250,000 is much more
expensive than any other option modeled for those size categories, this
assumption supports the situation where small systems may be expected
to entertain this option, i.e., where regionalization does not involve
piping water over great distances. This figure is based on a simple
estimate using the cost of installed cast iron pipe at $44 per linear
foot (an average cost for several pipe relevant pipe diameters) from
the 1998 Means Plumbing Cost Data and applying 20 percent for fittings,
excavation, and other expenses to arrive at an estimate of $53 per
linear foot, or $280,000 per linear mile. Purchased water costs ($/
kgal) were assumed to equal the pre-regionalization costs of production
($/kgal), merely as a modeling convenience. In some cases, purchased
water costs may be higher, in

[[Page 59340]]

some cases lower. Although EPA does not have many case studies to
support this assumption, it does have information on a Wisconsin case
study in which a small water system (serving 375 persons) regionalized
to connect to a near-by city water supply in 1995, partly in response
to a radium violation. The capital costs for this regionalization case
study was $225,000. There were no reported operations costs associated
with the purchased water. EPA makes no claims that this case study is
typical, but rather that this is the best assumption that it could make
based on the available information. Since this is a minor part of the
over-all national costs and since a more extensive modeling of the
costs of regionalization would necessitate a much more detailed
modeling of the additional benefits of regionalization (which were not
included), this assumption is maintained in the Regulatory Impact
Assessment for this proposed rule.
One commenter also questioned the feasibility of regionalization
for many systems. There are very few locations where this is possible
and just hooking up to a larger supplier is not practical. Many have
systems that are not acceptable to a larger supplier and many larger
suppliers won't accept the liability involved in taking over the small
system.
EPA Response 5-6
Since most small systems are expected to adopt the AMCL/MMM option,
EPA's regionalization assumption (1 percent of the minority of small
systems that choose the MCL option) is consistent with this commenter's
concern. Nevertheless, administrative regionalization is often
feasible, in particular when this does not require new physical
connections, and may be an important element of the long term
compliance strategy for a number of systems.
(e) Pre-treatment to reduce iron/manganese levels. The majority of
the commenters disagreed with EPA's assumptions on the removal of Fe/
Mn. It was assumed that essentially all systems with high Fe/Mn levels
are likely to already be treating to remove or sequester these metals.
Therefore, costs of adding Fe/Mn treatment to radon removal were not
included in the February, 1999 HRRCA (64 FR 9560). Commenters suggested
that this is a poor cost assumption, in that there are many systems
above the secondary MCL for Fe/Mn that do not treat. Of those that
sequester, commenters suggested that existing treatment is ineffective
once Fe/Mn has been oxidized. Therefore, filtration as well as
disinfection would be required for that type of system at a significant
additional cost that needs to be considered when reviewing the HRRCA.
If Fe/Mn is present in the source water, removal treatment will be
necessary to prevent fouling of the radon removal system. Disposal for
the Fe/Mn residuals also presents a special problem with its associated
costs. One commenter noted that by not including the costs of Fe/Mn
removal, EPA is making a poor assumption and may be underestimating
costs.
EPA Response 5-7
EPA recognized that not quantifying the costs associated with the
control of dissolved iron and manganese (Fe/Mn) was potentially a poor
assumption, and indicated that this assumption would be revisited for
the Regulatory Impact Analysis supporting this proposed rule. However,
EPA also indicated that national costs and average per system costs
would probably not be significantly affected in addressing this issue.
While EPA's current modeling results support this conclusion, EPA has
included the costs of adding chemical stabilizers (which minimize Fe/Mn
precipitation and also provide for corrosion control in some cases) by
25 percent of small systems that treat and 15 percent of large systems
that treat. A more detailed discussion on the inclusion of Fe/Mn
treatment costs is provided in Section VIII of the preamble.
To further support its position on Fe/Mn control, EPA has also (1)
analyzed case studies of systems aerating, which include Fe/Mn control
measures for a small minority of the systems, (2) performed an analysis
of the co-occurrence of radon with Fe/Mn in ground water, and (3)
performed an uncertainty analysis on costs, which includes a simulation
of more expensive control measures for Fe/Mn. All of these results are
also discussed in Section VIII of the preamble.
(f) Post treatment-disinfection. Many commenters stated that EPA's
assumption that the majority of groundwater systems already disinfect
is false. Some commenters felt this is inconsistent with the Ground
Water Rule estimates. Commenters suggested that analyses supporting the
proposed groundwater rule estimate that only 50 percent of CWSs and
only 25 percent of NTNCWSs disinfect, while Table 5-2 of the HRRCA
suggests that the majority of water systems using groundwater already
disinfect and that 20 percent of all water systems serving 3,300 or
greater have aeration or disinfection in place.
EPA Response 5-8
The cited analyses supporting the Ground Water Rule (GWR) were
conducted using occurrence estimates at the level of individual entry
points at water systems. The February 1999 Radon HRRCA was conducted
using occurrence estimates at the level of water systems. The GWR and
radon analyses use the same data source for estimating their respective
disinfection-in-place baselines, the 1997 Community Water System Survey
(USEPA 1997a), the only source of information of this type that is
based on a survey that was designed to be statistically representative
of community water systems at the national level. The GWR used a
disinfection-in-place baseline for entry points and the radon HRRCA
used a disinfection-in-place baseline for water systems.
The most desirable level of analysis is at the entry point, but the
only nationally representative data source for radon, the National
Inorganics and Radionuclides Survey, was conducted at the water system
level (samples were taken at the tap), which provides no information
about radon occurrence at individual entry points within water systems.
Radon intrasystem (within system) occurrence variability studies were
not available for the analyses supporting the February 1999 radon
HRRCA. In the interim between publishing the radon HRRCA and today's
proposal, EPA has conducted radon intrasystem variability studies
(based on studies other than NIRS) and has used the results of this
study to estimate radon occurrence at the entry point level. The
current Regulatory Impact Analysis supporting the Radon rule was
conducted at the entry point level, consistent with the Ground Water
Rule.
EPA Response 5-9
The additional costs to which this commenter is referring, namely
the costs of storage for contact time, are included in the costs of the
clearwell, which are included in the costs of the aeration process. In
the scenarios in which disinfection is assumed, EPA does NOT assume
that the systems have a clearwell in place and does include the costs
of adding a clearwell for collection of water after aeration and for
five minutes of disinfection contact time, which EPA believes to be
sufficient for 4-log viral de-activation.
(g) Monitoring costs. One commenter expressed concerns regarding
EPA's calculation of monitoring costs. The commenter suggested that EPA
grossly underestimated the number of wells per

[[Page 59341]]

different water system size in Table 5.4 of the HRRCA (64 FR 9585),
page 9585 and in Appendix D of the HRRCA. As a result, monitoring costs
need to be recalculated by EPA.
EPA Response 5-10
See EPA Response 1-2 for EPA's approach to determining the number
of wells per system.
(h) Choice of treatment responses. As noted previously in Section
G, one commenter questioned whether chlorination would always be the
disinfection technology of choice, as well as EPA's assumption that
existing chlorination practices would not have to be augmented if
aeration were installed. Other commenters on cost issues questioned the
feasibility and practicability of some technologies on cost grounds.
EPA Response 5-11
EPA assumed that chlorination would be the ``typical'' disinfection
technology chosen to model the ``average treatment costs'' (or
``central tendency costs''). There is no way to know beforehand exactly
how the universe of water systems will behave in response to a given
situation, so EPA believes that the best way to model national
compliance costs is to estimate these central tendency costs, then to
use statistical tools to capture the fact that ``real world costs''
will spread around the central tendency costs, rather than being
equivalent to them. By estimating the central tendency costs and using
statistical uncertainty to capture ``real world'' variability
(including variability in disinfection costs), EPA believes that this
modeling technique allows for the fact that real systems will behave in
a variety of ways, including things like choosing different
disinfection technologies.
(i) Site and system costs. A number of issues were raised
concerning site and system cost estimates. Several commenters suggested
that the HRRCA severely underestimated the number of sites per system,
citing the difference between the CWSS data and HRRCA assumptions.
Several commenters noted that the numbers of sources per system in
Table 5-4 of the HRRCA for systems serving 10,001--50,000 were too low.
One commenter maintained that the number of sources per system could
have a significant impact on national treatment costs.
EPA Response 5-12
EPA agrees that the distribution of the number of sites per system
was underestimated and has revised its estimate to be consistent with
the CWSS. However, it should be noted that while the distribution of
the sites per system actually does have an impact on national treatment
costs, this impact is significantly mitigated by the fact that the flow
per well being treated decreases proportionally as the estimated number
of wells per system increases.
(j) Aggregated national costs. Several commenters agreed that the
national average costs masked significant impacts on small systems.
When small systems are considered, the financial impact is large; in
some cases, water bills could double or triple. Providing individual
system costs is critical so that utilities can explain to their
customers the specific costs and benefits for that specific system.
EPA Response 5-13
EPA estimates household impacts for small systems that install
treatment (per household costs) by estimating the costs that small
systems would face (per system costs), then spreading these costs over
the customer base (population served). As demonstrated in the HRRCA,
household costs for small systems are expected to be many times higher
for very small systems than for larger systems. In listing small
systems compliance technologies for radon, EPA estimated the impacts on
small systems by estimating the per system costs and the per household
costs and comparing them to affordability criteria, as described in
this notice and in the references cited. However, it should also be
noted that the vast majority of small systems are expected to comply
with the AMCL/MMM option, rather than the MCL option. Under these
circumstances, less than 1 percent of small systems would have to take
measures to reduce radon levels in their drinking water.
(k) Costs to CWSs. Small systems will bear a significant percentage
of the costs for implementing a radon MCL, but will only accrue a small
proportion of the benefits. At the 300 pCi/L, the two categories of
smallest systems combined would receive 5.6 percent of the benefits at
this level, but would pay 42 percent of the total costs. Several
commenters indicated that the benefit-cost ratio for small systems was
thus highly unfavorable.
EPA Response 5-14
EPA recognizes that small systems experience similar benefits per
customer as large systems, but, due to economies of scale (higher
treatment costs per gallon treated), experience much higher costs per
customer compared to large systems. This, of course, leads to higher
costs at the same level of benefits. However, EPA has also recognized
that radon is a multi-media problem in which most of the risk is
presented from sources other than drinking water and has addressed this
fact by designating the AMCL/MMM option as the preferred option for
small systems. This will greatly lower the per customer costs faced by
small systems and may lead to greater total benefits that accrue to
small systems.
(l) Costs to consumers/households. One commenter thought that the
household consumption presented in the HRRCA (83,000 gal/year) is too
low. This is an understatement because treatment would be required for
all water produced, not just water consumed by households.
EPA Response 5-15
EPA does not assume that per system costs are based only on
residential water use and so does not miscalculate water prices in the
way described by the commenter. To determine the price of water, EPA
calculates per system costs based on both residential and non-
residential consumers (which is the main reason EPA calculates costs
for privately-owned and publically-owned separately, i.e., because they
have different ratios of residential to non-residential consumption).
These per system costs determine the costs per gallon treated (not per
gallon consumed) to determine the water price. The water price may then
be used in conjunction with the household consumption to estimate the
water bills faced by households, since they do pay by the gallon
consumed (and not by the gallon treated).
(m) Application of radon related costs to other rules. Several
commenters addressed the need to include the cumulative impact of
regulations in the RIA. The incremental costs of the regulations for
radon, arsenic, and groundwater systems could substantially change the
affordability analysis for small systems. Thus, treatment decisions
need to be made with an understanding of all the requirements that must
be met so that treatment systems can be designed to meet all
requirements. One commenter suggested a multi-rule cost and benefit
analysis to capture the true costs incurred by these systems.
EPA Response 5-16
The cumulative effects of rules are captured in EPA's
``affordability criteria'', which are described in the publicly
available 1998 EPA document, ``National-Level Affordability Criteria
Under the 1996 Amendments to the Safe

[[Page 59342]]

Drinking Water Act'' (USEPA 1998e). These small system affordability
criteria take into account how much consumers are currently paying for
typical water bills. Since the upcoming regulations will affect these
amounts, the cumulative effect of the costs of the rules will be
explicitly considered in the affordability determinations for small
systems as new rules are issued. EPA recognizes that its method of
basing affordability determinations on average costs does not address
the situation of systems that have significantly above average costs
because they must treat for a number of contaminants simultaneously.
EPA believes this approach is consistent with the requirements of SDWA
for identifying affordable small system technologies and notes that
other SDWA mechanisms may be used to address situations where systems
incur considerably higher costs.
6. Cost and Benefit Results
The main concern of many of the comments regarding this section
suggested that the costs of controlling radon in drinking water far
outweighed possible benefits, especially for small systems. Controlling
indoor air radon was identified as a better use of regulatory and
economic resources by several commenters. Commenters also had concerns
regarding how national total costs, benefits, and economic impacts were
calculated, and regarding the uncertainties in costs and benefits
estimates.
(a) Overview of analytical approach. Many commenters indicated that
the cost-benefit analysis was skewed toward overestimating benefits,
and/or omitted important cost elements. One concern shared by many of
these commenters was that the cost-benefit calculations were biased
because mitigation costs, but not health benefits, were discounted. A
commenter also indicated that too many assumptions had been used to
derive cost and benefit estimates.
EPA Response 6-1
The radon cost benefit analysis was performed according to EPA
guidelines, in an attempt to fairly portray both costs and benefits,
and not leave out important categories of either costs or benefits.
Annual mitigation costs are compared to annual benefits for the
cost benefit comparisons. Annual mitigation costs consist of annualized
capital costs plus yearly operating costs. Annualized costs are
computed under the assumption that capital expenditure are made up
front, with borrowed funds, and the payments are then annualized over a
period of twenty years. Changes in the rate of interest used in the
annualization process will change the annual cost, just like a mortgage
will change with different rates of interest. Adding yearly operating
costs for one year to annualized capital costs for one year gives the
total annual cost for the year. The issue of discounting of benefits is
discussed in Section XIII.D.
In any modeling process, assumptions must be made. To model costs
and benefits, assumptions about those costs and benefits must be made.
The number of assumptions needed depends on the complexity of the
problem addressed, and the time and information available to address
it. We would be interested in information that might inform our
modeling, particularly addressing improvements that could be made to
specific assumptions.
(b) MCL decision-making criteria. A commenter requested that EPA
define explicit decision-making criteria for setting MCL levels, to
assure that the net benefit to society is positive.
Another commenter indicated that, because drinking water radon
accounts for a small portion of total risks, EPA should consider the
relative costs and benefits of mitigation on a case-by-case basis at
individual systems before making regulatory decisions. A commenter
suggested that if the latency of cancer risk reduction and benefits
were discounted properly, the national cost-benefit ratios for radon
mitigation would be between 5:1 and 9:1. They stated that EPA should
not promulgate a rule with net negative benefits, especially in light
of the large economic impacts on small systems.
A commenter indicated that the cost-benefit ratios in Table 6-13 of
the HRRCA imply that regulation of radon in ground water is not
justified. They point out that systems serving 25-3,300 people incur at
least 56 percent of the costs and generate at most 21 percent of the
total benefits at all MCLs. They say that justifying radon control in
drinking water by adding in the benefits of MMM programs is not
justified. Another commenter also maintained that the small, localized
benefits of controlling radon exposures do not come near to justifying
the costs of mitigation.
One commenter said that the decision to set an MCL must take into
account the level of uncertainty in cost and benefit estimates. Another
commenter suggested that the Agency undertake a quantitative
uncertainty analysis of the cost and benefit estimates. Two commenters
said that the closeness of the cost and benefit estimates should be
considered in setting a regulatory level; if uncertainty is large, a
less stringent MCL would be justified.
EPA Response 6-2
EPA has included a detailed discussion on its decision-making
criteria for setting the MCL for radon in drinking water in the
preamble for the proposed rulemaking (see Section VII.D).
(c) National costs of radon mitigation. Two commenters indicated
that the national cost estimates obscured the high costs that would be
borne by individual systems. One commenter indicated that radon
variability in individual wells increases the uncertainty in the cost
estimates. Another commenter said that cost estimates should include
the costs of more frequent lead and copper exceedences brought about by
increased aeration. Other comments on specific cost elements were
summarized in Section 5. One commenter requested that EPA regionally
disaggregate cost and benefit estimates because of structural and
operational differences among water systems. Another commenter
suggested that EPA should conduct a more comprehensive analysis of
costs and benefits, including cost elements not currently addressed,
such as waste management.
EPA Response 6-3
The national costs include an uncertainty analysis which captures
the regional spread in treatment costs. In addition, EPA has estimated
total national costs by assuming that most systems will face ``typical
costs'', but that some will face ``high side'' and some ``low side''
treatment costs. These ``high side'' and ``low side'' cost differences
are largely based on regional considerations, like the costs of land,
structure, and permitting.
(d) Incremental costs and benefits. One commenter indicated that
the incremental costs and benefits of the various MCL options should be
presented in the HRRCA. They question the affordability of radon
mitigation for small systems.
EPA Response 6-4
EPA has provided an analysis of the incremental costs and benefits
of each MCL option in the HRRCA. See Table 6-7, Estimates of the Annual
Incremental Costs and Benefits of Reducing Radon in Drinking Water, in
the February 1999 HRRCA.
(e) Costs to community water systems. One commenter said that a
more accurate picture of costs and impacts (inclusive of State and
local costs) would be needed to make a reasonable

[[Page 59343]]

risk management decision. Another commenter suggested that EPA should
consider the cumulative costs of all drinking water regulations on
drinking water systems.
EPA Response 6-5
See EPA Response 5-14 for EPA's approach to determining the costs
to CWSs. Administrative costs to States were not included in the
February 1999 HRRCA, but have been added in the RIA for the proposed
rule.
(f) Costs and impacts on households. One commenter asked that EPA
explain how it determined what was an ``acceptable'' percentage of
household income that would go to radon mitigation. Another commenter
indicated that household costs should be compared to benefits at the
local, rather than national, level, because benefits and costs are
realized locally. A commenter indicated the median household incomes
for households served by different system sizes are not shown; they
also suggested that household costs as a percentage of income were
underestimated in Table 6-11 of the HRRCA. One commenter said that
expressing household impacts as a proportion of annual income
trivializes it and that costs could more meaningfully be compared to
other types of household expenses (i.e., food, rent). Several
commenters also noted the significant impact the costs could have on
customer water bills for small systems.
EPA Response 6-6
See EPA Response 5-15 for EPA's approach to determining the costs
to households.
(g) Summary of costs and benefits. Comments from one organization
regarding the cost-benefit comparison for radon mitigation were typical
of those received from other sources. They cited the NRC/NAS report as
indicating that only two percent of population risk came from drinking
water and questioned whether the high costs of the rule could justify
the small benefits obtained. They said that the cost-benefit comparison
did not justify regulating radon in ground water, especially in small
systems, where costs were highest and benefits lowest. Another
commenter also pointed out that it would be more cost-effective to
regulate radon in indoor air than in drinking water and further
maintained that spending resources to mitigate radon in water could
actually result in reduced public health protection. They point out
that the cost-benefit ratios for the smallest systems range from 20:1
to 50:1, and suggest that these ratios, rather than the greater
aggregate costs to large systems, should be persuasive in regulatory
decision making. Other commenters suggested the high cost-benefit
ratios did not justify the regulation of small systems.
EPA Response 6-7
The 1996 Safe Drinking Water Act Amendments require EPA to propose
a regulation for radon in drinking water by August 1999. The options
for small systems, proposed for public comment in this rulemaking,
represents EPA's efforts to address stakeholder comments concerning
small systems.
7. Multimedia Mitigation Programs
(a) Multimedia programs. Two commenters indicated that setting the
AMCL at 4,000 pCi/L was justifiable. They suggested that EPA should
utilize on MMM approach as the primary tool for reducing radon risks,
and not use the SDWA to force the States to develop MMM programs.
Several commenters noted that the MCL EPA selects should be
justifiable on cost-benefit grounds, with the MMM program serving as a
supplemental program to allow States to achieve greater risk reduction
at less cost. Another commenter suggested the multimedia approach
allowed under the 1996 amendments to the SDWA should not be used with
regard to radon-222 in water.
EPA Response 7-1
The requirement for implementation of an EPA-approved MMM program
in conjunction with State adoption of the AMCL is consistent with the
statutory framework outlined by Congress in the SDWA provision on
radon. As proposed, States may choose either to adopt the MCL or the
AMCL and an MMM program. EPA recommends that small systems comply with
an AMCL of 4,000 pCi/L and implement a MMM program. See section VII.D
for background on the selection of the MCL and AMCL.
Two commenters believe the radon regulation may result in
litigation against water utilities, local, and State governments if
systems comply with the AMCL rather than the MCL. As a result, some
water utilities could choose to comply with the more stringent MCL
rather than face potential litigation for meeting a ``less stringent
standard,'' regardless of the increased public health protection.
According to one commenter, problems will arise when both the AMCL and
the MCL are required to appear on the annual Consumer Confidence
Report. The public will view the AMCL as an attempt by the water
industry to get around the MCL. This will leave the water utility
vulnerable to toxic tort lawsuits. Because of these problems, the
concept of an MMM program/AMCL is not as attractive as it once
appeared.
EPA Response 7-2
EPA is aware of this concern and the risk communication challenges
of two regulatory limits for radon in drinking water. However, the SDWA
framework requires EPA to set an alternative maximum contaminant limit
for radon if the proposed MCL is more stringent than the level of radon
in outdoor air. It is important to recognize that in State primacy
applications for oversight and enforcement of the drinking water
program, States choosing the MMM approach will be adopting 4,000 pCi/L
as their MCL. In addition, as part of the proposed rule, EPA will be
amending the Consumer Confidence Reporting Rule to reflect the proposed
regulation for radon. Under Sec. 141.153 of the proposed radon rule, a
system operating under an approved multimedia mitigation program and
subject to an Alternative MCL (AMCL) for radon must report the AMCL
instead of the MCL whenever reporting on the MCL is required.
Another commenter questioned the need for regulating radon in water
below 3,000 pCi/L, and maintained that there is no conceivable reason
to regulate it at 100 pCi/L, with or without an MMM program.
EPA Response 7-3
See EPA Response 6-2 for EPA's decision criteria for setting an
MCL.
(b) Implementation scenarios evaluated. One commenter feels that a
``desk top review'' of States likely to adopt an MMM program would give
more useful estimates of MMM acceptance than the HRRCA assumptions of
zero, 50 percent, and 100 percent adoption of MMM programs. This
commenter felt that for an MMM program to be productive, two things are
necessary: (1) relatively high radon concentration in water and (2)
relatively high radon in indoor air.
EPA Response 7-4
For the purposes of the HRRCA, EPA made these assumptions as a
straight forward approach for assessing overall cost implications of
MMM. States are not required to make their determinations on whether to
adopt the MMM approach until after the rule is final in August 2000.
Therefore, EPA did not have this information available when developing
the HRRCA, nor does EPA have this information at this time. However,
discussions with many State

[[Page 59344]]

drinking water and radon program staff suggest that many States are
seriously considering the MMM approach.
EPA expects that MMM programs will be able to achieve indoor radon
risk reduction even in areas of low radon potential. It is important to
keep in mind that the only way to know if a house has elevated indoor
radon levels is to test it. Many homes in low radon potential areas
have been found with levels well above EPA's action level of 4 pCi/L,
often next door to houses with very low levels. EPA estimates that
about 6 million homes in the U.S. of the 83 million homes that should
test are at or above 4 pCi/L. To date only about 11 million homes have
been tested. In addition, EPA is not requiring State MMM program plans
to precisely quantify equivalency in risk reduction between radon in
drinking water and radon in indoor air.
(c) Multimedia mitigation cost and benefit assumptions. Two
commenters indicated that, even if it is not known how the MMM programs
will be funded, the costs of administering such programs should be
included in the HRRCA. Several commenters expressed concerns regarding
the estimated cost of $700,000 per fatal cancer averted. One commenter
felt that using this value is far too optimistic, indicating that the
cost of radon risk reduction under State-mandated MMM programs will
significantly exceed present costs under the voluntary system. To get
the greatest risk reductions at the lowest costs, MMM program should
focus on the houses with the highest radon concentrations. Another
commenter recommended that EPA develop an MMM program that is better
than the existing voluntary programs and further reduces the cost per
fatal cancer avoided. The commenter also requested that EPA supply
background information supporting use of this single MMM program cost
estimate.
EPA Response 7-5
EPA is required under the UMRA to assess the costs to States of
implementing and administering both the MCL and the MMM/AMCL. EPA has
addressed these costs in the preamble of the rule.
EPA believes that the criteria for EPA approval of State MMM
program plans will augment and build on existing State indoor radon
programs and will result in an increased level of risk reduction.
As part of developing the 1992 ``A Citizen's Guide to Radon,'' EPA
analyzed the risk reductions and costs of various radon testing and
mitigation options (USEPA 1992b). Based on these analyses, a point
estimate of the average cost per life saved of the current national
voluntary radon program was used as the basis for the cost estimate of
risk reduction for the MMM option. EPA had previously estimated that
the average cost per fatal lung cancer avoided from testing all
existing homes in the U.S. and mitigation of all those homes at or
above EPA's voluntary action level of 4 pCi./L is approximately
$700,000. This value was originally estimated by EPA in 1991. Since
that time there has been an equivalent offset between a decrease in
testing and mitigation costs since 1992 and the expected increase due
to inflation in the years 1992-1997.
One commenter stated that experiences in Massachusetts showed that
the costs of incorporating passive radon resistant construction
techniques is about the same as current prices for marginal quality
(active) radon mitigation in existing buildings, and disputed the HRRCA
statement that passive techniques are much less expensive. The
commenter supported the NAS findings that the effectiveness of these
techniques in normal construction practice is uncertain.
EPA Response 7-6
Builders have reported costs as low as $100 to install radon
resistant new construction features which is significantly less than
the $350--$500 that was derived in EPA's cost-effectiveness analysis of
the radon model standards. The cost of materials alone for the passive
system will always be less than the cost for an active system which
includes the cost of a fan. In many areas, the majority of the features
for radon-resistant new construction are already required by code or
are common building practice, such as an aggregate layer, ``poly''
sheeting, and sealing and other weatherization techniques. The only
additional cost is associated with the vent stack consisting of PVC
pipe and fittings. In those areas where gravel is not commonly used,
builders can use a drain tile loop or other alternative less costly
than gravel to facilitate communication under the slab. EPA estimates
that the cost to mitigate an existing home ranges from $800 to $2,500
with an average cost of $1,200.
(d) Annual costs and benefits of MMM program implementation.
Several concerns were raised regarding the costs and benefits
associated with MMM program implementation. One commenter suggested
that the MMM program description in the HRRCA provides essentially no
guidance on the point from which additional risk reduction due to MMM
will be measured.
EPA Response 7-7
The HRRCA was not intended to include a discussion and description
of the criteria for EPA approval of State MMM programs. Rather,
proposed criteria are presented in this proposed rule. EPA's proposed
criteria do not entail a determination by the State of the level of
indoor radon risk reduction that has already occurred (``baseline'') as
the basis for determining how much more risk reduction needs to take
place. Rather States, with public participation, are required to set
goals that reflect State and local needs and concerns.
Another commenter states that EPA has underestimated the benefits
of an MMM program. The HRRCA registers only the benefits gained in
relation to water being treated to the MCL. However, according to EPA's
figures, MMM benefits are expected to be much higher than those
achieved by mitigating water alone.
EPA Response 7-8
EPA anticipates that MMM programs will result in sufficient risk
reduction to achieve ``equal or greater'' risk reduction. A complete
discussion on why MMM is expected to achieve equal or greater risk
reduction is shown in Section VI.B of today's preamble. For the
purposes of the HRRCA analyses, EPA made the conservative assumption
that the level of risk reduction would at least be ``equal'' to that
achieved by universal compliance with the MCL.
8. Other Key Comments
(a) Omission of non-transient non-community water systems
(NTNCWSs). Eleven commenters criticized EPA's failure to include
NTNCWSs in the HRRCA. Three commenters indicate that failure to include
NTNCWSs grossly underestimates costs of radon mitigation. Another
commenter also suggests that NTNCWSs should be included in the HRRCA,
to provide a better picture of both costs and benefits. Two commenters
would also like NTNCWSs included because impacts on these systems are
likely to be high. Other commenters maintain that excluding NTNCWSs
skews benefit-cost analyses in favor of regulation. Another commenter
indicates that NTNCWSs, because of the type of wells and aquifers that
they draw from, will be most affected by a radon rule.
EPA Response 8-1
Partly as a result of concerns raised by commenters, and partly as
a result of its own preliminary analysis of exposure and risk, EPA is
not proposing that NTNCWSs be covered by this rule. A more complete
discussion of this issue

[[Continued on page 59345]]

Notices For
2009
2008
2007
2006
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994

[[pp. 59295-59344]] National Primary Drinking Water Regulations; Radon-222

Local Navigation