[Federal Register Volume 64, Number 38 (Friday, February 26, 1999)]
[Notices]
[Pages 9560-9599]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 99-4416]



[[Page 9559]]

_______________________________________________________________________

Part II





Environmental Protection Agency





_______________________________________________________________________



Radon in Drinking Water Health Risk Reduction and Cost Analysis; Notice

  Federal Register / Vol. 64, No. 38 / Friday, February 26, 1999 / 
Notices  

[[Page 9560]]



ENVIRONMENTAL PROTECTION AGENCY

[FRL-6304-3]


Radon in Drinking Water Health Risk Reduction and Cost Analysis

AGENCY: Environmental Protection Agency.

ACTION: Notice and request for public comments and announcement of 
stakeholder meeting.

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

SUMMARY: The Safe Drinking Water Act (SDWA), as amended in 1996, 
requires the U.S. Environmental Protection Agency (EPA) to publish a 
health risk reduction and cost analysis (HRRCA) for radon in drinking 
water for public comment. The purpose of this notice is to provide the 
public with the HRRCA for radon and to request comments on the 
document. As required by SDWA, EPA will publish a response to all 
significant comments to the HRRCA in the preamble to the proposed 
National Primary Drinking Water Regulation (NPDWR) for radon, due in 
August, 1999.
    The goal of the HRRCA is to provide a neutral and factual analysis 
of the costs, benefits, and other impacts of controlling radon levels 
in drinking water. The HRRCA is intended to support future decision 
making during development of the radon NPDWR. The HRRCA evaluates radon 
levels in drinking water of 100, 300, 500, 700, 1000, 2000, and 4000 
pCi/L. The HRRCA also presents information on the costs and benefits of 
implementing multimedia mitigation (MMM) programs to reduce the risks 
of radon exposure in indoor air. The SDWA, as amended, provides for 
development of an Alternative Maximum Contaminant Level (AMCL), which 
public systems may comply with if their State has an EPA approved MMM 
program to reduce radon in indoor air. The concept behind the AMCL and 
MMM option is to reduce radon health risks by addressing the larger 
source of exposure (air levels in homes) compared to drinking water. If 
a State chooses to employ a MMM program to reduce radon risk, it would 
implement a State program to reduce indoor air levels and require 
public water systems to control water radon levels to the AMCL. If a 
State does not choose a MMM program option, a public water system may 
propose a MMM program for EPA approval. Today's notice does not include 
any decisions regarding the choice of a Maximum Contaminant Level (MCL) 
for radon in drinking water. Today's notice also announces a 
stakeholder meeting on the HRRCA and framework for the MMM program.

DATES: The Agency must receive comments on the HRRCA on or before April 
12, 1999. EPA will hold a one day public meeting on Tuesday, March 16, 
1999 from 9 a.m. to 5:30 p.m. EST.

ADDRESSES: Send written comments on HRRCA to the Comment Clerk, docket 
number W-98-30, Water Docket (MC4101), USEPA, 401 M St., SW, 
Washington, DC 20460. Please submit an original and three copies of 
your comments and enclosures (including references).
    Commenters who want EPA to acknowledge receipt of their comments 
should enclose a self-addressed, stamped envelope. No facsimiles 
(faxes) will be accepted. Comments may also be submitted electronically 
to [email protected]. Electronic comments must be submitted as an 
ASCII, WP6.1, or WP8 file avoiding the use of special characters and 
any form of encryption. Electronic comments must be identified by the 
docket number W-98-30. Comments and data will also be accepted on disks 
in WP6.1, WP8, or ASCII file format. Electronic comments on this notice 
may be filed online at many Federal Depository Libraries.
    The record for this notice has been established under docket number 
W-98-30, and includes supporting documentation as well as printed, 
paper versions of electronic comments. The full record is available for 
inspection from 9 a.m. to 4 p.m. EST Monday through Friday, excluding 
legal holidays at the Water Docket, Room EB57, USEPA Headquarters, 401 
M St., SW, Washington, DC 20460. For access to docket materials, please 
call 202-260-3027 to schedule an appointment.
    The stakeholder meeting on the HRRCA and multimedia mitigation 
framework will be held at the offices of at RESOLVE, Inc., 1255 23rd 
Street, N.W,. Suite 275, Washington, DC 20037. Check-in will begin at 
8:30 a.m.

FOR FURTHER INFORMATION CONTACT: For general information, please 
contact the EPA Safe Drinking Water Hotline at 1-800-426-4791 or 703-
285-1093 between 9 a.m. and 5:30 p.m. EST. (For information on radon in 
indoor air, contact the National Safety Council's National Radon 
Hotline at 1-800-SOS-RADON.) The HRRCA, including the appendices, can 
also be accessed on the internet at http://www.epa.gov/safewater/
standard/pp/radonpp/html. For specific information and technical 
inquiries, contact Michael Osinski at 202-260-6252 or 
[email protected].
    For general information on meeting logistics, please contact Sheri 
Jobe at RESOLVE, Inc., at 202-965-6382 or Email: [email protected].

SUPPLEMENTARY INFORMATION: The purpose of the March 16, 1999 
stakeholder meeting is to cover the following key issues, including: 
(1) Discussion of the Health Risk Reduction and Cost Analysis published 
in this notice; and (2) present information and discuss issues related 
to status of development of a framework for multimedia mitigation 
programs. This upcoming meeting is the fifth of a series of 
stakeholders meetings on the NPDWR for radon, intended to seek input 
from State and Tribal drinking water and radon programs, the regulated 
community (public water systems), public health and safety 
organizations, environmental and public interest groups, and other 
stakeholders. EPA encourages the full participation of stakeholders 
throughout this process.
    To register for the meeting, please contact Sheri Jobe at RESOLVE, 
Inc., 1255 23rd Street, N.W,. Suite 275, Washington, DC 20037, Phone: 
202-965-6382, Fax: 202-338-1264, Email: [email protected]. Please 
provide your name, affiliation/organization, address, phone, fax and 
email if you would like to be on the mailing list to receive further 
information about the meeting (including agenda and meeting summary). A 
limited number of tele-conference lines will be available. Please 
indicate whether you would like to participate by phone. Those 
registered for the meeting by February 26, 1999 will receive an agenda, 
logistics sheet, and other information prior to the meeting.

    Dated: January 5, 1999.
Dana D. Minerva,
Acting Assistant Administrator, Office of Water, Environmental 
Protection Agency.

Radon in Drinking Water Health Risk Reduction and Cost Analysis

Table of Contents

1. Executive Summary
2. Introduction
    2.1 Background
    2.2 Regulatory History
    2.3 Safe Drinking Water Act Amendments of 1996
    2.4 Specific Requirements for the Health Risk Reduction and Cost 
Analysis
    2.5 Radon Levels Evaluated
    2.6 Document Structure
3. Health Effects From Radon Exposure
    3.1 Radon Occurrence and Exposure Pathways
    3.1.1 Occurrence
    3.1.2 Exposure Pathways
    3.2 Nature of Health Impacts
    3.3 Impacts on Sensitive Subpopulations

[[Page 9561]]

    3.4 Risk Reduction Model for Radon in Drinking Water
    3.5 Risks from Existing Radon Exposures
    3.6 Potential for Risk Reductions Associated with Removal of Co-
Occurring Contaminants
    3.7 Potential for Risk Increases from Other Contaminants 
Associated with Radon Removal
    3.8 Risk for Ever-Smokers and Never-Smokers
4. Benefits of Reduced Radon Exposure
    4.1 Nature of Regulatory Impacts
    4.1.1 Quantifiable Benefits
    4.1.2 Non-Quantifiable Benefits
    4.2 Monetization of Benefits
    4.2.1 Estimation of Fatal and Non-Fatal Cancer Risk Reduction
    4.2.2 Value of Statistical Life for Fatal Cancers Avoided
    4.2.3 Costs of Illness and Lost Time for Non-Fatal Cancers
    4.2.4 Willingness to Pay to Avoid Non-Fatal Cancers
    4.3 Treatment of Monetized Benefits Over Time
5. Costs of Radon Treatment Measures
    5.1 Drinking Water Treatment Technologies and Costs
    5.1.1 Aeration
    5.1.2 Granular Activated Carbon (GAC)
    5.1.3 Storage
    5.1.4 Regionalization
    5.1.5 Radon Removal Efficiencies
    5.1.6 Pre-Treatment to Reduce Iron and Manganese Levels
    5.1.7 Post-Treatment--Disinfection
    5.2 Monitoring Costs
    5.3 Water Treatment Technologies Currently In Use
    5.4 Cost of Technologies as a Function of Flow Rates and Radon 
Removal Efficiency
    5.5 Choice of Treatment Responses
    5.6 Cost Estimation
    5.6.1 Site and System Costs
    5.6.2 Aggregate National Costs
    5.6.3 Costs to Community Water Systems
    5.6.4 Costs to Consumers/Households
    5.6.5 Costs to Non-Transient Non-Community Systems
    5.7 Application of Radon Related Costs to Other Rules
6. Results: Costs and Benefits of Reducing Radon in Drinking Water
    6.1 Overview of Analytical Approach
    6.2 Health Risk Reduction and Monetized Health Benefits
    6.3 Costs of Radon Mitigation
    6.4 Incremental Costs and Benefits of Radon Removal
    6.5 Costs to Community Water Systems
    6.6 Costs and Impacts to Households
    6.7 Summary of Cost and Benefit Analysis
    6.8 Sensitivities and Uncertainties
    6.8.1 Uncertainties in Risk Reduction and Health Benefits 
Calculations
    6.8.2 Uncertainty in Cost and Impact Calculations
7. Implementation Scenarios--Multimedia Mitigation Programs
    7.1 Multimedia Mitigation Programs
    7.2 Implementation Scenarios Evaluated
    7.3 Multimedia Mitigation Cost and Benefit Assumptions
    7.4 Annual Costs and Benefits of Multimedia Mitigation Program 
Implementation
    7.6 Sensitivities and Uncertainties

List of Tables and Figures

Table 3-1. Radon Distributions by Region
Table 3-2. Radon Distribution in Public Water Systems
Table 3-3. Population Exposed Above Various Radon Levels By System 
Size
Table 3-4. Estimated Radon Unit Lifetime Fatal Cancer Risks in 
Community Water Systems
Table 3-5. Radon Treatment Assumptions to Calculate Residual Fatal 
Cancer Risks
Table 3-6. Annual Fatal Cancer Risks for Exposures to Radon from 
Community Water Systems
Table 3-7. Radon Risk Reductions Across Various Effluent Levels and 
Percent Removals
Table 3-8. Radon Risk Reduction from Treatment Compared to DBP Risks
Table 3-9. Annual Lung Cancer Death Risks Estimates from Radon 
Progeny for Ever-Smokers, Never-Smokers, and the General Population
Table 4-1. Proportion of Fatal Cancers by Exposure Pathway and 
Estimated Mortality
Table 4-2. Estimated Medical Care and Lost-Time Costs Per Case for 
Survivors of Lung Cancer
Table 4-3. Estimated Medical Care and Lost-time Costs Per Case for 
Survivors of Stomach Cancer
Table 5-1. Unit Treatment Costs by Removal Efficiency and System 
Size
Table 5-2. Estimated Proportions of Ground Water Systems With Water 
Treatment Technologies Already in Place
Table 5-3. Decision Matrix For Selection of Treatment Technology 
Options
Table 5-4. Number of Sites per Ground Water System by System Size
Table 6-1. Risk Reduction and Residual Cancer Risk from Reducing 
Radon in Drinking Water
Table 6-2. Estimated Monetized Health Benefits from Reducing Radon 
in Drinking Water
Table 6-3. Risk Reduction and Monetized Benefits Estimates For Ever-
Smokers
Table 6-4. Risk Reduction and Monetized Benefits Estimates For 
Never-Smokers
Table 6-5. Estimated Annualized National Costs of Reducing Radon 
Exposures
Table 6-6. Capital and O&M Costs of Mitigating Radon in Drinking 
Water
Table 6-7. Estimates of the Annual Incremental Costs and Benefits of 
Reducing Radon in Drinking Water
Table 6-8. Number of Community Water Systems Exceeding Various Radon 
Levels
Table 6-9. Average Annual Cost Per System
Table 6-10. Annual Costs per Household for Community Water Systems
Table 6-11. Per Household Impact by Community Water System as a 
Percentage of Median Household Income
Table 6-12. Estimated National Annual Costs and Benefits of Reducing 
Radon Exposures--Central Tendency Estimate
Table 6-13. Total Annual Costs and Fatal Cancers Avoided by System 
Size
Table 6-14. Annual Monetized Health Benefits by System Size
Table 7-1. Central Tendency Estimates of Annualized Costs and 
Benefits of Reducing Radon Exposures with 50% of States Selecting 
the MMM/AMCL Option
Table 7-2. Central Tendency Estimates of Annualized Costs and 
Benefits of Reducing Radon Exposures with 100% of States Selecting 
the MMM/AMCL Option
Figure 3-1. General Patterns of Radon Occurrence in Ground Water
Figure 3-2. EPA Map of Radon Zones in Indoor Air
Figure 6-1. Sensitivity Analysis of Water Mitigation Costs
Figure 7-1. Sensitivity Analysis to Changes in MMM Cost Estimates

Abbreviations Used in This Document

AF: Average Flow
AMCL: Alternative Maximum Contaminant Level
AWWA: American Water Works Association
BAT: Best Available Technology
CWS: Community Water System
DA: Diffused-Bubble Aeration
DBP: Disinfection By-Products
DF: Design Flow
GAC: Granular Activated Carbon
EPA: US Environmental Protection Agency
FACA: Federal Advisory Committee Act
HRRCA: Health Risk Reduction and Cost Analysis
MCL: Maximum Contaminant Level
MCLG: Maximum Contaminant Level Goal
MMM: Multimedia Mitigation program
MSBA: Multi-Stage Diffused Bubble Aeration
NAS: National Academy of Sciences
NDWAC: National Drinking Water Advisory Council
NIRS: National Inorganics and Radionuclides Survey
NPDWR: National Primary Drinking Water Regulation
NTNCWS: Non-Transient Non-Community Water System
OGWDW: Office of Ground Water and Drinking Water
O&M: Operation and Maintenance
OMB: Office of Management and Budget
pCi/l: Picocurie Per Liter
POE GAC: Point-of-Entry Granular Activated Carbon
PTA: Packed Tower Aeration
RIA: Regulatory Impact Analysis
SAB: Science Advisory Board
SDWA: Safe Drinking Water Act, as amended in 1986 and 1996
SDWIS: Safe Drinking Water Inventory System
THM: Trihalomethane
VSL: Value of a Statistical Life
WTP: Willingness To Pay

1. Executive Summary

    This document constitutes the Health Risk Reduction and Cost 
Analysis (HRRCA) in support of development of a National Primary 
Drinking Water Regulation (NPDWR) for radon in drinking water, as 
required by Section 1412(b)(13) of the 1996 Amendments to

[[Page 9562]]

the Safe Drinking Water Act (SDWA). The goal of the HRRCA is to provide 
a neutral and fact-based analysis of the costs, benefits, and other 
impacts of controlling radon levels in drinking water to support future 
decision making during development of the radon NPDWR. The document 
addresses the various requirements for the analysis of benefits, costs, 
and other elements specified by Section 1412(b)(13) of the SDWA, as 
amended.
    This is the first time the Environmental Protection Agency (EPA) 
has prepared a HRRCA under the SDWA, as amended. As such, the EPA is 
very interested in seeking comment on the techniques, assumptions, and 
data inputs upon which the analysis is based. The Agency recognizes 
that there may be other methods of conducting the analysis and 
presenting the data required for this HRRCA, and encourages meaningful 
input from all stakeholders during the public comment period. 
Therefore, the specific analysis and findings presented here are 
intended as an initial effort to frame an analysis that can support 
development of the NPDWR. Since the HRRCA is a cost-benefit tool to 
analyze an array of radon levels during development of the NPDWR, many 
of the issues to be addressed in the regulatory development process 
(e.g. the selection of a Maximum Contaminant Level (MCL), Best 
Available Technology (BAT), and monitoring framework) are not analyzed 
here, but will be presented in the proposed rule.
    The HRRCA evaluates radon levels in ground water supplies of 100, 
300, 500, 700, 1000, 2000, and 4000 pCi/l. The HRRCA also presents 
information on the costs and benefits of implementing multimedia 
mitigation (MMM) programs. The scenarios evaluated are described in 
detail in Section 2.5. This executive summary presents a background on 
the radon in drinking water problem, followed by a summary of findings 
arranged according to each provision for HRRCAs as specified by the 
SDWA, as amended.

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, EPA arranged for the 
National Academy of Sciences (NAS) to assess the health risks of radon 
in drinking water. The NAS released the ``Report on the Risks of Radon 
in Drinking Water,''(NAS Report) in September 1998 (NAS 1998B). 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 (US EPA,1994C).
    NAS recently estimated individual lifetime unit fatal cancer risks 
associated with exposure to radon from domestic water use for ingestion 
and inhalation pathways (Table 3-4). The results show that inhalation 
of radon progeny accounts for most (approximately 89 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.
    The NAS Report confirmed that indoor air contamination arising from 
soil gas typically account 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 HRRCA. 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 HRRCA, EPA has consulted with a broad 
range of stakeholders and technical experts. Participants in a series 
of stakeholder meetings held in 1997 and 1998 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 HRRCA 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. As required by the 1996 
Safe Drinking Water Act (SDWA), EPA has withdrawn the proposed NPDWR 
for radon (US EPA 1997B) and will propose a new regulation by August, 
1999. The HRRCA does not include any decisions regarding the choice of 
a Maximum Contaminant Level (MCL) for radon in drinking water.
    The analysis presented in this HRRCA 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 HRRCA 
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 (US EPA 1998L). This new analysis 
incorporates information from the EPA's 1985 National Inorganic and 
Radionuclides Survey (NIRS) of 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

[[Page 9563]]

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 (Figure 3-1). 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 State level, the distribution of radon concentrations in 
indoor air (Figure 3-2) do not always mirror distributions of radon in 
ground water.
    In addition, the 1996 Amendments to the SDWA introduce two new 
elements into the radon in drinking water rule: (1) an Alternative 
Maximum Contaminant Level (AMCL) and (2) multimedia radon mitigation 
(MMM) programs. The SDWA, as amended, provides for development of an 
AMCL, which public water systems may comply with if their State has an 
EPA approved MMM program to reduce radon in indoor air. The NAS Report 
estimated that the AMCL would be about 4,000 pCi/L, based on SDWA 
requirements. The concept behind the AMCL and MMM option is to reduce 
radon health risks by addressing the larger source of exposure (air 
levels in homes) compared to drinking water. If a State chooses to 
employ a MMM program to reduce radon risk, it would implement a State 
program to reduce indoor air levels and require public water systems to 
control radon levels in drinking water to the AMCL. If a State does not 
choose a MMM program option, a public water system may propose a MMM 
program for EPA approval.

Summary of Findings

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 ground water to 
specific target levels were calculated using the cost curves discussed 
in Section 5.4 and the matrix of treatment options presented in Section 
5.5. 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 Safe Drinking Water Suite Model 
(US EPA 1998N). The equations and parameter values relating system size 
to flow rates are presented in Appendix C. The technologies addressed 
in the cost estimation included a number of aeration and granular 
activated carbon (GAC) technologies described in Section 5.1, as well 
as storage, regionalization, and disinfection as a post-treatment. To 
estimate costs, water systems were assumed, with a few exceptions, 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.
    The costs of reducing radon to various levels are summarized in 
Table 6-5, which shows that, as expected, aggregate radon mitigation 
costs increase with decreasing radon levels. The cost ranges presented 
in the table represent plausible upper and lower bounds of 50 percent 
above to 50 percent below the central tendency estimates. 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.

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 6-1 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. 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 160 fatal cancers and 9.2 non-fatal 
cancers per year are associated with radon exposures through CWSs. At a 
radon level of 4,000 pCi/l, approximately 2.2 fatal cancers and 0.1 
non-fatal cancers per year are prevented. At the lowest level evaluated 
(100 pCi/l), approximately 115 fatal and 6.6 non-fatal cancers per year 
would be prevented.
    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 
6-5. 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, with low- and 
high-end values of $700,000 (1997$) and $16.3 million (1997$), 
respectively, used for the purposes of sensitivity analysis. These 
figures span the range of VSL estimates from 26 studies reviewed in 
EPA's recent draft guidance on benefits assessment (US EPA 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

[[Page 9564]]

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), with a 
range between $169,000 and $1.05 million (1997$). The combined fatal 
and non-fatal health benefits are summarized in Table 6-2. The annual 
health benefits range from $13 million for a radon level of 4000 pCi/l 
to $673 million at 100 pCi/l. The ranges in the last column of Table 6-
2 illustrate how benefits vary when the upper and lower bound estimates 
of the VSL and WTP measures are used.
    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 peace of mind benefits specific 
to reduction of radon risks that may not be adequately captured in the 
VSL estimate. In addition, treating radon in drinking water with 
aeration oxidizes arsenic into a less soluble form that is easier to 
remove with conventional 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 is a non-quantifiable 
benefit that can be attributed to some components of a MMM program. 
Providing such information might allow households to make informed 
choices about the appropriate level of risk reduction given their 
specific circumstances and concerns. These potential benefits are 
difficult to quantify because of the uncertainty surrounding their 
estimation. However, they are likely to be somewhat less significant 
relative to the monetized benefits estimates.

Incremental Costs and Benefits of Radon Removal

    Table 6-7 summarizes the central tendency and the upper and lower 
bound estimates of the incremental costs and benefits of radon exposure 
reduction. Both the annual incremental costs and benefits increase as 
the radon level decreases from 4000 pCi/l down to 100 pCi/l. 
Incremental costs and benefits are within 10 percent of each other at 
radon levels of 1000, 700, and 500 pCi/l. The table also illustrates 
the wide ranges of potential incremental costs and benefits due to the 
uncertainty inherent in the estimates. There is substantial overlap 
between the incremental costs and benefits at each radon level.

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 (Table 6-10). 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.
    Another significant finding is that, like the per system costs, 
costs per household (which are a function of per system costs) are 
relatively constant across different radon levels within each system 
size category. For example, there is less than one dollar per year 
variation in household costs, regardless of the radon level being 
considered for households served by large public or private systems 
(between $6 and $7 annually), by medium public or private systems 
(between $10 and $11), and by small public or private systems (between 
$19 and $20 annually). Similarly, for very small systems (501-3300 
people), the cost per household is consistently about $34 annually for 
public systems and about $40 annually for private systems, varying 
little with the target radon level. Only for very very small systems is 
there a noticeable variation in household costs across radon levels. 
The range for per household costs for public CWSs serving 25-500 people 
is $87 per year (at 4,000 pCi/l) to $135 per year (at 100 pCi/l). The 
corresponding range for private CWSs is $139 to $238 per year. For 
households served by the smallest public systems (25-100 people) the 
range of cost per household ranges from $292 per year at 4,000 pCi/l to 
$398 per year at 100 pCi/l. For private systems, the range is $364 per 
year to $489 per year, respectively.

Summary of Annual Costs and Benefits

    Table 6-12 reveals that at a radon level of 4000pCi/l (equivalent 
to the AMCL estimated in the NAS Report), annual costs are 
approximately twice the annual monetized benefits. For radon levels of 
1000pCi/l to 300 pCi/l, the central tendency estimates of annual costs 
are above the central tendency estimates of the monetized benefits, 
although they are within 10 percent of each other. However, as shown in 
Tables 6-2 and 6-5, due to the uncertainty in the cost and benefit 
estimates, there is a very broad possible range of potential costs and 
benefits that overlap across all of the radon levels evaluated.

Benefits From the Reduction of Co-Occurring Contaminants

    The occurrence patterns of other 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.

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

[[Page 9565]]

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% of the total fatal cancer risk from radon in 
drinking water, approximately 30% 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 1998A, 1998B). 
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 (US EPA 1999A), will also experience the bulk of the risk 
reduction from radon exposure reduction in drinking water supplies.

Risk Increases From Other Contaminants Associated With Radon Exposure 
Reduction

    As discussed in Section 5.1, the need to install radon treatment 
technologies may require some systems that currently do not disinfect 
to do so. Case studies (US EPA 199D) 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 systems 
either had disinfection already in place or did not add it). In 
practice, the tendency to add disinfection 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 (US EPA 1998I). This rule identified MCLs for the 
major DBPs, 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 is proportional to the concentration of organic precursor 
contaminants, which tend to be much lower in ground water than in 
surface water.
    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 1998B). The report concluded that, based upon median and 
average total trihalomethane (THM) levels taken from EPA's 1981 
Community Water System Survey, a typical ground water CWS would face 
incremental individual lifetime cancer risk due to chlorination 
byproducts of 5 x 10-5. It should be emphasized that this 
risk is based on average and median THM occurrence information that 
does not segregate systems that disinfect from those that do. 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.
    A more meaningful comparison is to look at the trade-off between 
risk reduction from radon treatment in cases where disinfection is 
added with the added risks from DBP formation. This trade-off will 
affect only a minority of systems since a majority of ground water 
systems already have disinfection in place. For the smallest systems 
size category, approximately half of all CWSs already have disinfection 
in place. The proportion of systems having disinfection in place 
increases as the size categories increase, up to >95% for large systems 
(Table 5-2). In addition, although EPA is using the conservative 
costing assumption that all systems adding aeration or GAC would 
disinfect, not all systems adding aeration or GAC would have to add 
post-disinfection or, if disinfecting, may use a disinfection 
technology that does not forms DBPs. For those ground water systems 
adding treatment with disinfection, this trade-off tends to be 
favorable since the combined risk reduction from radon removal and 
microbial risk reduction outweigh 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 3.7. As noted by the NAS Report, these risk 
reductions outweigh the increased risk from DBP exposure for those 
systems that chlorinate as a result of adding radon treatment.
    The ratios between risk reduction from radon removal and the risks 
from THMs at levels equal their MCLs (a conservative assumption) are 
shown in Table 3.8. The data indicate that the risk ratios are 
favorable for treatment with disinfection, ignoring microbial risk 
reduction, even assuming the worst case scenario that ground water 
systems have THM levels at the MCL. It is worth noting that there is 
the possibility that 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.

Other Factors: Uncertainty in Risk, Benefit, and Cost Estimates

    Estimates of health benefits from radon reduction are uncertain. A 
few of the variables affecting the uncertainty in the benefit estimates 
include the distribution of radon in ground water systems, the NAS's 
risk models for ingestion and inhalation risks, and the transfer factor 
used to estimate indoor air radon activity levels. EPA plans to include 
an uncertainty analysis of radon in drinking water risks with the 
proposed rule. Monetary benefit estimates are also strongly 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.
    Estimates of the regulatory costs also have associated uncertainty. 
The major factors affecting this uncertainty include assumptions 
regarding the distribution of radon levels among ground water systems 
and among treatment sites within systems, uncertainties in unit cost 
models, the assumed prevalence of the various compliance decisions, and 
the exclusion of NTNCWSs in the HRRCA's national cost estimates.
    To deal with a lack of information regarding the intra-system 
variability of

[[Page 9566]]

radon levels between treatment sites (source wells), the national cost 
estimates are based on the assumption that all CWSs above a target 
radon level, as estimated by system-level average radon occurrence 
predictions from the occurrence model, will install separate treatment 
systems at each site. Ideally, occurrence information at each treatment 
site will provide a better estimate of national costs, since the wells 
within a water system would exhibit a range of radon occurrence levels, 
some of which may be below the target radon level, others above this 
level. Since it is not obvious whether the system-level approach will 
lead to either a positive or negative bias in the national cost 
estimates, EPA is in the process of performing an analysis of the 
intra-system variability for radon occurrence and will include this 
analysis in support of the upcoming proposed rule.
    There are also significant uncertainties in estimated treatment 
unit costs and in the decision-trees that are used to model national 
level compliance decisions that will by made by the system-size 
stratified universe of drinking water systems in response to a range of 
radon influent levels. It is possible to estimate the uncertainties in 
both the unit costs and the decision-tree by performing sensitivity 
analyses for the factors affecting costs. Regarding unit costs, this 
analysis leads to a spread in costs that adequately resembles the 
``real-world'' as shown by ranges in treatment cost case studies. 
Regarding the uncertainty in the decision-tree, it is unfortunately not 
possible to verify results in this way. However, since there are so few 
technologies to mitigate radon in water, the decision-tree is fairly 
robust.

Other Impacts: Costs and Benefits of Multimedia Mitigation Program 
Implementation Scenarios

    In addition to evaluating the costs and benefits across a range of 
radon levels, two scenarios were evaluated that reduce radon exposure 
through the use of MMM programs. The two scenarios evaluated assume: 
(1) 50 percent of States (all water systems in those States) select MMM 
implementation; and (2) 100 percent of States select MMM. These two 
scenarios are described in detail in Section 7. 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.
    In calculating the cost of MMM programs, the cost per fatal cancer 
case avoided was estimated at $700,000 (1997$). This value was 
originally estimated by EPA in 1992 using 1991 data. The same nominal 
value is used in the HRRCA based on anecdotal evidence from EPA's 
Office of Radiation and Indoor Air (ORIA) that there has been an 
equivalent offset between a decrease in testing and mitigation costs 
since 1991 and the expected increase due to inflation in the years 
1992-1997. This dollar amount reflects that real testing and mitigation 
costs have decreased, while nominal costs have remained approximately 
constant.
    Tables 7-2 and 7-3 illustrate that, as expected, the costs of 
reducing radon exposures decrease with increasing numbers of States 
(i.e. CWSs) selecting the MMM implementation scenario. Also, as would 
be expected, the annual costs of implementing MMM are, on average, 
lower compared to reducing radon exposures in drinking water alone. 
Central tendency estimates of the total annualized benefits exceed the 
annualized costs for both the 50 and 100 percent MMM participation 
scenarios over all radon levels. The cost per fatal cancer case avoided 
is also lower for both the 50 and 100 percent MMM implementation 
scenarios compared to the scenario in which no States elect to develop 
a MMM program. In addition, the cost per fatal cancer case avoided is 
significantly lower for the MMM scenario with 100 percent of the States 
electing the MMM program compared to when 50 percent of the States 
choose the MMM scenario, especially at the lower radon levels. The 
costs and benefits estimates are also broken out into their respective 
MMM and water mitigation components. With the exception of 4000pCi/l 
(the NAS estimated AMCL), annual monetized benefits are significantly 
larger than annual costs for the MMM component of the total costs. For 
the water mitigation component, the annual costs are larger than the 
annual monetized benefits across all radon levels.

2. Introduction

2.1  Background

    This Health Risk Reduction and Cost Analysis (HRRCA) provides the 
Environmental Protection Agency's (EPA) analysis of potential costs and 
benefits of different target levels for radon in drinking water. The 
HRRCA builds on several technical components, including estimates of 
radon occurrence in drinking water supplies, 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 regulation for radon in drinking water. 
As required by the 1996 Safe Drinking Water Act (SDWA), EPA has 
withdrawn the proposed regulation for radon in drinking water (US EPA 
1997B) and will propose a new regulation by August, 1999.
    One of the most important inputs used by EPA in the HRRCA is the 
National Academy of Sciences (NAS) September 1998 report ``Risk 
Assessment of Radon in Drinking Water'' (NAS Report). EPA has used the 
NAS 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 estimated cancer deaths potentially 
prevented by reducing radon levels. The NAS Report is the most 
comprehensive accumulation of scientific data gathered to date on radon 
in drinking water. SDWA required the NAS assessment, which generally 
affirms EPA's earlier scientific conclusions and analyses on the risks 
of exposure to radon and progeny in drinking water.
    The analysis presented in this HRRCA 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 also have been updated. The HRRCA 
follows EPA policies with regard to the methods and assumptions used in 
cost and benefit assessment.
    In updating key analyses and developing the framework for the cost-
benefit analysis presented in the HRRCA, EPA has consulted with a broad 
range of stakeholders and technical experts. Participants in a series 
of stakeholder meetings held in 1997 and 1998 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. EPA convened an expert panel in 
Denver in November of 1997 to review treatment technology costing 
approaches. The panel made a number of

[[Page 9567]]

recommendations for modification to EPA cost estimating protocols that 
have been incorporated into the radon cost estimates. EPA also 
consulted with a subgroup of the National Drinking Water Advisory 
Council (NDWAC) on evaluating the benefits of drinking water 
regulations. The NDWAC was formed in accordance with the Federal 
Advisory Committee Act (FACA) to assist and advise EPA. A variety of 
stakeholders participated in the NDWAC benefits working group, 
including utility company staff, environmentalists, health 
professionals, State water program staff, a local elected official, 
economists, and members of the general public.
    The American Water Works Association (AWWA) convened a ``Radon 
Technical Work Group,'' in 1998 that provided technical input on EPA's 
update of technical analyses (occurrence, analytical methods, and 
treatment technology), and discussed conceptual issues related to 
developing guidelines for multimedia mitigation programs. Members of 
the Radon Technical Work Group included representatives from State 
drinking water and indoor air programs, public water systems, drinking 
water testing laboratories, environmental groups and the U.S. 
Geological Survey. EPA also held a series of conference calls with 
State drinking water and indoor air programs, to discuss issues related 
to developing guidelines for multimedia mitigation programs.

2.2  Regulatory History

    Section 1412 of the Safe Drinking Water Act (SDWA), as amended in 
1986, requires the EPA to publish Maximum Contaminant Level Goals 
(MCLGs) and to promulgate National Primary Drinking Water Regulations 
(NPDWRs) for contaminants that may cause an adverse effect on human 
health and that are known or anticipated to occur in public water 
supplies. In response to this charge, the EPA proposed NPDWRs for 
radionuclides, including radon, in 1991 (US EPA 1991). The proposed 
rule included a maximum contaminant level (MCL) of 300 pCi/l for radon 
in drinking water, applicable to both community water systems and non-
transient non-community water systems. A community water system (CWS) 
is defined as a public water system with at least 15 or more service 
connections or that regularly serves at least 25 year-round residents. 
A non-transient non-community system (NTNCWS) is a public water system 
that is not a CWS and that regularly serves at least 25 of the same 
persons for at least six months per year. Examples of NTNCWSs include 
those that serve schools, offices, and commercial buildings. Under the 
proposed rule, all CWSs and NTNCWSs relying on ground water would have 
been required to monitor radon levels quarterly at each point of entry 
to the distribution system. Compliance monitoring requirements were 
based on the arithmetic average of four quarterly samples. The 1991 
proposed rule required systems with one or more points of entry out of 
compliance to treat influent water to reduce radon levels below the MCL 
or to secure water from another source below the MCL.
    The proposed rule was accompanied by an assessment of regulatory 
costs and economic impacts, as well as an assessment of the risk 
reduction associated with implementation of the MCL. The Agency 
received substantial comments on the proposal and its supporting 
analyses from States, water utilities, and other stakeholder groups. 
Comments from the water industry questioned EPA's estimates of the 
number of systems that would be out of compliance with the proposed 
MCL, as well as the cost of radon mitigation. EPA's Science Advisory 
Board (SAB) provided extensive comments on the risk assessment used by 
the Agency to support the proposed MCL. The SAB recommended that EPA 
expand the analysis of the uncertainty associated with the risk and 
risk reduction estimates. In response to these comments, the assessment 
was revised twice, once in 1993 and again in 1995 (US EPA 1995). Both 
of the revised risk analyses provided detailed quantitative uncertainty 
analysis.

2.3  Safe Drinking Water Act Amendments of 1996

    In the 1996 Amendments to the Safe Drinking Water Act, Congress 
established a new charter for public water systems, States, and EPA to 
protect the safety of drinking water supplies. Among other mandates, 
amended Section 1412(b)(13) directed EPA to withdraw the drinking water 
standards proposed for radon in 1991 and to propose a new MCLG and 
NPDWR for radon by no later than August 6, 1999. As noted above, the 
amendments require NAS to conduct a risk assessment for radon in 
drinking water and an assessment of risk reduction benefits from 
various mitigation measures to reduce radon in indoor air (Section 
1412(b)(13)(B)). In addition, the amendments introduce two new elements 
into the radon in drinking water rule: (1) An Alternative Maximum 
Contaminant Level (AMCL) and (2) multimedia radon mitigation (MMM) 
program.
    If the MCL established for radon in drinking water is more 
stringent than necessary to reduce the contribution to radon in indoor 
air from drinking water to a concentration that is equivalent to the 
national average concentration of radon in outdoor air, EPA is required 
to simultaneously establish an AMCL that would result in a contribution 
of radon from drinking water to radon levels in indoor air equivalent 
to the national average concentration of radon in outdoor air (Section 
1412(b)(13)(F)). If an AMCL is established, EPA is to publish 
guidelines for State programs, including criteria for multimedia 
measures to mitigate radon levels in indoor air, to comply with the 
AMCL.
    States may develop and submit to EPA for approval an MMM program to 
decrease radon levels in indoor air (Section 1412(b)(13)(G)). These 
programs may rely on a variety of mitigation measures, including public 
education, testing, training, technical assistance, remediation grants 
and loan or incentive programs, or other regulatory and non-regulatory 
measures. EPA shall approve a State's program if it is expected to 
achieve equal or greater health risk reduction benefits than would be 
achieved by compliance with the more stringent MCL. If EPA does not 
approve a State program, or a State does not propose a program, public 
water supply systems may propose their own MMM programs to EPA, 
following the same procedures outlined for States. Once the MMM 
programs are established, EPA is required to re-evaluate them no less 
than every five years.

2.4  Specific Requirements for the Health Risk Reduction and Cost 
Analysis

    Section 1412(b)(13)(C) of the 1996 Amendments requires EPA to 
prepare a Health Risk Reduction and Cost Analysis (HRRCA) to be used to 
support the development of the radon NPDWR. SDWA requires the HRRCA be 
published for public comment by February 6, 1999, six months before the 
rule is to be proposed. In the preamble of the proposed rule, EPA must 
include a response to all significant public comments on the HRRCA.
    The HRRCA must also satisfy the requirements established in Section 
1412(b)(3)(C) of the amended SDWA. According to these requirements, EPA 
must analyze each of the following when proposing an NPDWR that 
includes a MCL: (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

[[Page 9568]]

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.
    To the extent possible, this HRRCA follows the new cost-benefit 
framework being developed by the Office of Ground Water and Drinking 
Water (OGWDW) . As provided in the SDWA, as amended, 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. More in-depth 
discussions of input data and assumptions will be provided in a 
companion ``Analytical Support Document'' and an in-depth presentation 
and discussion of the results will appear in a separate ``Cost/Benefit 
Document'' that will accompany the proposed rule. The HRRCA by itself 
does not constitute the complete Regulatory Impact Analysis (RIA), but 
serves as a foundation upon which the RIA can be developed for the 
proposed rule.

2.5  Radon Levels Evaluated

    The HRRCA is intended to present preliminary estimates of the 
potential costs and benefits of various levels of controlling radon in 
drinking water. The HRRCA assumes that all systems drawing water from 
sources above a defined radon level will employ treatment technologies 
to meet the target level or ``regionalize'' to obtain water from 
another source with lower radon levels. This analysis evaluates radon 
levels of 100, 300, 500, 700, 1,000, 2,000, and 4,000 pCi/l. The 
analysis did not include any provisions for exemptions or phased 
compliance and assumed that a simple quarterly monitoring scheme would 
be used to determine the need for mitigation and ongoing compliance.
    The HRRCA also evaluates national costs and benefits of MMM 
implementation scenarios, with States choosing to reduce radon exposure 
in drinking water through an Alternative Maximum Contaminant Level 
(AMCL) and radon risks in indoor air through MMM programs. Based on NAS 
recommendations, the AMCL level that is evaluated is 4,000 pCi/l. Under 
the scenarios that include an AMCL, the HRRCA assumes that a portion of 
the States would adopt an AMCL supplemented with MMM programs to 
address indoor air radon risks. In the absence of information 
concerning the number of States that would choose to implement radon 
risk reduction through the use of AMCL plus multimedia programs, the 
HRRCA assumes that either 50 or 100 percent of the systems in the 
United States would choose to implement MMM programs and comply with 
the AMCL. For the MMM implementation scenarios, a single multimedia 
cost estimate is used, based on the cost-effectiveness of current 
voluntary mitigation efforts. These issues are discussed in more detail 
in Section 7.

2.6  Document Structure

    The HRRCA is organized into 7 sections and a number of appendices. 
The appendices, while not included in this Federal Register Notice, are 
available in the docket for review and can be downloaded from the web 
at www.epa.gov/safewater/standard/pp/radonpp/html. Section 3 discusses 
the health effects of exposure to radon. Section 4 describes the 
assumptions and methods for estimating quantifiable benefits and 
assessing non-quantifiable benefits. Section 5 discusses the water 
treatment and MMM methods used to calculate the national costs of the 
various radon levels examined. Section 6 presents the results of the 
cost and benefit analysis of reducing radon levels in drinking water, 
and evaluates economic impacts on households. In addition, the major 
sources of uncertainty associated with the estimates of costs, 
benefits, and economic impacts are identified. Section 7 estimates the 
costs and benefits of two different implementation scenarios in which 
States and water systems elect to develop and implement a MMM program 
and comply with the AMCL. Appendices provide details of the risk 
calculations, cost curves for treatment technologies, methods used to 
calculate system flows, and detailed breakdown summaries of the cost, 
benefit and impact calculations.

3. Health Effects of Radon Exposure

    This Section presents an overview of the major issues and 
assumptions addressed in order to characterize the health impacts and 
potential benefits of reductions in radon exposures. The methods that 
have been used to characterize risk and benefits in the HRRCA are also 
described. The assumptions and methods presented below are used in 
Section 4 to derive detailed estimates of the health reduction benefits 
of different radon levels in ground water supplies.

3.1  Radon Occurrence and Exposure Pathways

    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 (US EPA 1998L). This new analysis 
incorporates information from the EPA 1985 National Inorganic and 
Radionuclides Survey (NIRS) of 1000 community ground water systems 
throughout the United States, along with supplemental data provided by 
the States, water utilities, and academic researchers.
    The new study also addressed a number of issues raised by public 
comments on the previous occurrence analysis. These include 
characterization of regional and temporal variability in radon levels, 
variability in radon levels across different-sized water systems, 
impact of sampling point, and the proper statistical techniques for 
evaluating the data.
3.1.1  Occurrence
    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

[[Page 9569]]

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, while 
some 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.
    Radon itself undergoes radioactive decay and has a radioactive 
half-life of about four days. When radon atoms decay they emit 
radiation in the form of alpha particles, and transform into decay 
products, or progeny, which also decay. Unlike radon gas, these progeny 
easily attach to and can be transported by dust and other particles in 
air. The decay of progeny continues until stable, non-radioactive 
progeny are formed. At each step in the decay process, radiation is 
released. The term radon, as commonly used, refers to radon-222 as well 
as its radioactive decay products.
    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 (Figure 3-1). 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 State level, the distribution of radon concentrations in 
indoor air (Figure 3-2) do not always mirror distributions of radon in 
ground water.
    In addition to large-scale regional variation, radon levels in 
ground water also vary significantly over smaller distance scales. 
Local differences in geology tend to greatly influence the patterns of 
radon levels observed at specific locations (e.g., not all radon levels 
in New England are high; not all radon levels in the Gulf Coast region 
are low). Over small distances, there is often no consistent 
relationship between measured radon levels in ground water and radium 
levels in the ground water or in the parent bedrock (Davis and Watson 
1989). Similarly, no significant national correlation has been found 
between radon levels in individual ground water systems and the levels 
of other inorganic contaminants or conventional geochemical parameters. 
Potential correlations between radon levels and levels of organic 
contaminants in ground water have not been investigated, but there is 
little reason to believe any would be found. Radon's volatility is 
rather high compared to its solubility in water. Thus, radon 
volatilizes rapidly from surface water, and measured radon levels in 
surface water supplies are generally insignificant compared to those 
found in ground water.

Figure 3-1. General Patterns of Radon Occurrence in Groundwater in 
the United States

    Figure 3-1 is not printed in the Federal Register. It is available 
in the Water Docket at the address listed in the ADDRESSES section.

Figure 3-2. EPA Map of Radon Zones in Indoor Air

    Figure 3-2 is not printed in the Federal Register. It is available 
in the Water Docket at the address listed in the ADDRESSES section.
    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 (US EPA 1994B).
    Table 3-1 summarizes the regional patterns of radon in drinking 
water supplies as seen in the NIRS database. This survey of 1,000 
ground water systems, undertaken by EPA in 1985, provides the most 
representative national characterization of radon levels in drinking 
water.
    However, 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.

                          Table 3-1.--Radon Distributions by Region (All System Sizes)
----------------------------------------------------------------------------------------------------------------
                                                                                                     Geometric
                                                                    Arithmetic    Geometric Mean     standard
                             Region                                mean (pCi/l)     \1\ (pCi/l)    deviation \2\
                                                                                                      (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
----------------------------------------------------------------------------------------------------------------
\1\ The geometric mean is the anti-log of the average of the logarithms (log base e) of the observations.
\2\ The geometric standard deviation is the anti-log of the standard deviation of the logarithms (log base e) of
  the observations.
 
Source: US EPA 1998L. The values given are not population-weighted, but reflect averages across systems.

    The NIRS data illustrate the wide regional variations in radon 
levels in ground water. The arithmetic mean and geometric mean radon 
levels are substantially higher in New England and the Appalachian 
region (in this analysis, all the States on the east coast between New 
York and Florida) than in other regions of the United States. The large 
differences between the geometric (anti-log of the average of the 
logarithms (log base e) of the observations) and arithmetic means 
indicate how ``skewed'' (i.e., ``stretched'' in a positive direction; a 
bell-shaped curve with a tail out to the right) the radon distributions 
are. The Agency selected a lognormal model as the best approach to 
evaluating these data.
    EPA's current re-evaluation of radon occurrence in ground water 
uses data from a number of additional sources to supplement the NIRS 
information and to develop estimates of the national

[[Page 9570]]

distribution of radon in ground water systems of different sizes. Data 
from 17 States were used to evaluate 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. Table 3-2 
summarizes EPA's latest characterization of the distributions of radon 
levels in ground water supplies of different sizes and populations 
exposed to radon through CWSs.
    In this table, radon levels and populations are presented for 
systems serving various population ranges from 25 to greater than 
100,000. For purpose of estimating costs and benefits, the CWSs are 
aggregated to be consistent with the following system size categories 
identified in the 1996 SDWA, as amended: very very small systems (25-
500 people), further subdivided into 25-100 and 101-500; 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).
    In the updated occurrence analysis, insufficient data were 
available to accurately assess radon levels in the highest CWSs size 
stratum. Thus, data from the two largest size strata were pooled to 
develop exposure estimates for the risk and benefits assessments.
    The Agency estimates that approximately 89.7 million people are 
served by community ground water systems in the United States based on 
an EPA analysis of SDWIS data in 1998). The data in Table 3-2 show that 
systems serving more than 500 people account for approximately 95 
percent of the population served by ground water systems, even though 
they represent only 40 percent the total active systems (USEPA 1997A). 
The estimated system geometric mean radon levels range from 
approximately 120 pCi/l for the largest systems to 312 pCi/l for the 
smallest systems. Arithmetic mean values for the various 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.

                             Table 3-2.--Radon Distributions in Public Water Systems
----------------------------------------------------------------------------------------------------------------
                                                                 System size (population served)
                                                ----------------------------------------------------------------
                                                    25-100      101-500     501-3,300      3,301-      >10,000
-------------------------------------------------------------------------------------------10,000---------------
Total Systems..................................       14,651       14,896       10,286        2,538        1,536
Geometric Mean Radon Level, pCi/l..............          312          259          122          124          132
Geometric Standard Deviation...................          3.0          3.3          3.2          2.3          2.3
Population Served (Millions)...................         0.87         4.18         14.2         14.5         65.9
----------------------------------------------------------------------------------------------------------------
Radon Level, pCi/l.............................       Proportions of Systems Exceeding Radon Levels (percent)
----------------------------------------------------------------------------------------------------------------
100............................................         84.7         78.7         56.9         60.4         62.9
300............................................         51.4         45.1         22.1         14.3         16.2
500............................................         33.6         29.1         11.4          4.6          5.5
700............................................         23.4         20.3          6.8          1.8          2.3
1000...........................................         14.7         12.9          3.6          0.6          0.8
2000...........................................          4.7          4.4          0.8          0.0          0.1
4000...........................................          1.1          1.1          0.1          0.0          0.0
----------------------------------------------------------------------------------------------------------------

    Table 3-3 presents the total exposed population above each radon 
level by system size category. Approximately 20% of the total 
population for all system sizes are above the radon level of 300 pCi/l 
and 63% are above a radon level of 100 pCi/l.

                                        Table 3-3.--Population Exposed Above Various Radon Levels By System Size
                                                                       [Thousands]
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                           Very very     Very very
                   Radon level (pCi/l)                       small         small      Very small      Small        Medium         Large         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
--------------------------------------------------------------------------------------------------------------------------------------------------------

    Radon exposures also arise from NTNCWSs. The Agency estimates that 
approximately 5.2 million people use water from NTNCWSs (US EPA 1998G). 
An analysis of SDWIS data in 1998 shows there are approximately 19,500 
active NTNCWSs in the United States. Over 96 percent of these systems 
serve fewer than 1,000 people. EPA recently identified useful data on 
radon levels in NTNCWSs from six States. A preliminary analysis of data 
from these States suggested that geometric mean radon levels are 
approximately 60 percent higher in NTNCWSs than in CWSs in the same 
size category.
    There are currently no data which enable the agency to determine 
the extent to which the populations exposed to radon from CWSs and 
NTNCWSs overlap. Some portion of individuals exposed through a CWS at 
home may be exposed to radon from a NTNCWS at school or at work.

[[Page 9571]]

Similarly, the same populations may be exposed to radon from two 
different community systems in the course of their normal daily 
activities. Further, in the case of NTNCWSs, it is possible that the 
same individual could be exposed sequentially throughout their life to 
radon from a series of different systems; at school, then at work, etc.
3.1.2  Exposure Pathways
    People are exposed to radon in drinking water in three ways: from 
ingesting radon dissolved in water; from inhaling radon gas released 
from water during household use; and from inhaling radon progeny 
derived from radon gas released from water.
    Typically, indoor air contamination arising from soil gas accounts 
for the bulk of total individual risk due to radon exposure (NAS 
1998B). Nationally, levels of radon in household air average 
approximately 1.25 pCi/l (US EPA 1992A). Usually, the bulk of the radon 
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 recommends that EPA use the central estimate of a transfer factor 
of 1.0 pCi/l for radon in domestic water contributing 1x10-4 
pCi/l to indoor air. As an example, for a typical ground water CWS with 
a radon level of 250 pCi/l, the increment in indoor air activity would 
be 0.025 pCi/l. This is about 2 percent of the average indoor level, 
which is derived mostly from soils.
    As noted, the bulk of radiation exposure through inhalation comes 
from radon progeny, which tend to bind to airborne particulates. When 
the particles are inhaled, they become deposited in the respiratory 
tract, and further radioactive decay results in a radiation dose to the 
respiratory epithelium. In contrast, when radon gas is inhaled, it is 
absorbed through the lung, and much of this fraction remains in the 
body only a short time before being exhaled.
    Direct ingestion of radon gas in water is the other important 
exposure pathway associated with domestic water use. If water is not 
agitated or heated prior to consumption, the bulk (80 to 100 percent) 
of the radon remains in the water and is consequently ingested with it 
(US EPA 1995). Heating, agitation (for example, by a faucet aerator), 
and prolonged standing cause radon to be released and the proportion 
consumed to be reduced. After a person ingests radon in water, the 
radon passes from the gastrointestinal tract into the blood. The blood 
then circulates the radon to all organs of the body before it is 
eventually exhaled from the lungs. When radon and its progeny decay in 
the body, the surrounding tissues are irradiated by alpha particles. 
However, the dose of radiation resulting from exposure to radon gas by 
ingestion varies from organ to organ. Stomach, followed by the tissues 
of colon, liver, kidney, red marrow, and lung appear to receive the 
greatest doses.
    Exposure patterns to radon vary with different exposure settings. 
Depending on the relative radon levels in water and air, water use 
patterns, and exposure frequency and duration, the relative 
contribution of ingestion and inhalation exposure to total risks will 
vary. In the case of domestic water use, inhalation of radon progeny 
accounts for most of the total individual risk resulting from radon 
exposure (Section 3.2). Inhalation exposure to radon from NTNCWSs is 
expected to be less than for CWSs, however, because buildings served by 
these systems tend to be larger, and ventilation rates higher, than the 
corresponding values for domestic exposures. In addition, exposure at 
these facilities tend to be less frequent and of shorter duration than 
exposure from CWSs. Therefore, overall exposures at NTNCWSs will likely 
be lower.

3.2  Nature of Health Impacts

    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 (NAS 1998A). Ingestion of radon in water is suspected of 
being associated with increased risk of tumors of several internal 
organs, primarily the stomach (NAS 1998B). As discussed previously, NAS 
recently estimated the lifetime unit fatal cancer risks associated with 
exposure to radon from domestic water use for ingestion and inhalation 
pathways. EPA subsequently calculated the unit risk of inhalation of 
radon gas to 0.06 percent of the total risk from radon in drinking 
water, using radiation dosimetry data and risk coefficients provided by 
the NAS (NAS 1998B). The lifetime unit fatal cancer risk is defined as 
the lifetime risk associated with exposures to a unit concentration (1 
pCi/l) of radon in drinking water. The findings are summarized in Table 
3-4.

     Table 3-4.--Estimated Radon Unit Lifetime Fatal Cancer Risks in
                         Community Water Systems
------------------------------------------------------------------------
                                     Cancer unit risk    Proportion of
         Exposure pathway              per pCi/l in        total risk
                                          water            (percent)
------------------------------------------------------------------------
Inhalation of radon progeny1......        5.55 x 10-7                 89
Ingestion of radon1...............        7.00 x 10-8                 11
Inhalation of radon gas2..........       3.50 x 10-10               0.06
                                   -------------------------------------
Total.............................        6.25 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).

    These updated risk estimates indicate that inhalation of radon 
progeny accounts for most (approximately 89 percent) of the individual 
risk associated with domestic water use, with almost all of the 
remainder (11 percent) resulting from ingestion of radon gas. 
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. Ingestion of radon also results in 
slightly increased risk cancer of the colon, liver, and other tissues. 
Inhalation of radon gas is estimated to account for approximately 0.06 
percent of the total risk from household radon exposures, and the major 
target organ is again believed to be the lung. In the following 
sections, methods and parameter values developed by the NAS are applied 
to the estimation of baseline population risks and the levels of risk 
reduction associated with the different radon levels.
    Radon, a noble gas, exhibits no other known toxic effects besides 
carcinogenesis. The 1998 NAS report indicates that there is no 
scientific

[[Page 9572]]

evidence to show that exposure to radon is associated with reproductive 
or genetic toxicity. Therefore, the endpoints characterized in the risk 
assessment for radon exposure are primarily increased risk of lung and 
stomach cancers.
    For the purposes of this Health Risk Reduction and Cost Analysis, 
EPA is using the best estimates of radon inhalation and ingestion risks 
provided by the NAS Report. In order to finalize the Agency's estimate 
of lung cancer deaths arising from indoor air exposure, EPA's Office of 
Radiation and Indoor Air is currently assessing various factors 
integral to the approach for estimating the lung cancer risks of 
inhaling radon progeny in indoor air provided in the NAS 1998 report 
``The Health Effects of Exposure to Radon-BEIR VI'' (BEIR VI Report). 
This assessment will be reviewed by the Agency's SAB and may result in 
some adjustment to the estimated unit risk, and its associated 
uncertainty, for inhalation of radon progeny used in this HRRCA

3.3  Impacts on Sensitive Subpopulations

    Populations that might experience disproportional risk as a result 
of radon exposure fall into two general classes: those who might 
receive higher exposures per unit radon in water supplies and those who 
are more sensitive to the exposures they receive. The former group 
includes persons whose domestic water supplies have high radon levels, 
and whose physiological characteristics or behaviors (high metabolic 
rate, high water consumption, large amounts of time spent indoors) 
result in high exposures per unit of exposure concentration. As noted 
above, a portion of the population could be exposed to radon from more 
than one source. For example, a student or worker might be exposed to 
radon from the CWS in the household setting and also from a NTNCWS (or 
from the same or different CWS) at school or work.
    Different age and gender groups may also experience exposure 
dosimetric differences. These differences in radiation dose per unit 
exposure have been taken into account in the BEIR VI Report addressing 
radon in indoor air (NAS 1998A), the NAS Report addressing radon in 
drinking water (NAS 1998B), and the EPA Federal Guidance Report 13 (US 
EPA 1998F).
    The NAS Report concluded that there is insufficient scientific 
information to permit separate cancer risk estimates for subpopulations 
such as pregnant women, the elderly, children, and seriously ill 
persons. The report did note, however, that according to the NAS risk 
model for the cancer risk from ingested radon, which accounts for 11% 
of the total lifetime fatal cancer risk from radon in drinking water, 
approximately 30% of this fatal lifetime cancer risk is attributed to 
exposure between ages 0 to 10.
    The NAS did identify smokers as the only group that is more 
susceptible to inhalation exposure to radon progeny. Inhalation to 
cigarette smoke and radon progeny result in a greater increased risk 
than if the two exposures act independently to induce lung cancer.

3.4  Risk Reduction Model for Radon in Drinking Water

    Risk and risk reduction were estimated using a Monte Carlo model 
that simulated the initial and post-regulatory distributions of radon 
activity levels and population cancer risks. Each iteration of the 
model selected a size stratum of community water systems. The system 
sizes were stratified according to the following populations served: 
<100; 101-500; 501-3,300; 3,301-10,000; and > 10,000 served. For each 
size category, a lognormal distribution of uncontrolled radon levels 
had been defined based on the updated occurrence analysis (USEPA 
1998L). The model sampled randomly from the radon distribution for the 
selected CWS size category to determine if the radon level was above 
the selected maximum exposure level. The proportion of iterations 
choosing each size stratum were determined by the relative national 
populations served by each size stratum of systems. Thus, over a large 
number of iterations (generally, benefit calculations were carried out 
using 20,000 to 50,000 iterations), the model produced a population-
weighted distribution of radon levels.
    In each iteration of the model, the simulated influent radon 
activity level was compared to the maximum radon levels under 
consideration (100, 300, 500, 700, 1000, 2000, and 4000 pCi/l). When 
the simulated influent radon level was less than the target level, the 
simulated level was passed directly to the risk calculation equations. 
The equations calculated population fatal cancer risks from ingestion 
of radon gas, inhalation of radon gas, and inhalation of radon progeny 
using standard exposure factors and unit risk values derived by the 
NAS.
    When the simulated influent radon level in a given iteration 
exceeded a target radon level, the model reduced the value by a 
proportion equivalent to the performance of selected mitigation 
technologies. The degrees of reduction are presented in Table 3-5:

   Table 3-5.--Radon Treatment Assumptions to Calculate Residual Fatal
                              Cancer Risks
------------------------------------------------------------------------
           If the radon level is              Then the treated level is
------------------------------------------------------------------------
Less than the target level................  None; Influent = Effluent.
Above but less than two times the target    Influent = 0.5  x  Effluent.
 level.
Above two times but less than five times    Influent = 0.2  x  Effluent.
 the target level.
Greater than five times the target level..  Influent = 0.01 Effluent.
------------------------------------------------------------------------

    Using this approach implies that a greater level of control is 
achieved than if all the systems were simply assumed to reduce 
exposures to the maximum exposure level. For example, a system with an 
initial uncontrolled concentration of 400 pCi/l would need to employ a 
mitigation technology with a 50 percent removal efficiency to comply 
with a maximum exposure limit of 300 pCi/l, resulting in a final radon 
level of 200 pCi/l. Limited sensitivity analysis suggests that this 
approach does not provide very much in the way of extra risk reduction. 
The preponderance of population risk reduction is achieved by reducing 
radon levels in the relatively few systems that have initial 
uncontrolled values far above the maximum exposure limits, not by the 
relatively small incremental reductions below the target radon levels.

3.5  Risks From Existing Radon Exposures

    In support of the regulatory development process for the revised 
radon rule, EPA has updated its risk assessment for radon exposures in 
drinking water. Previously, EPA developed estimates of risk from total 
population exposure to radon in drinking water in support of the 
proposed rule for radon in 1991 (US EPA 1991). In response to comments 
from the SAB, EPA updated the risk assessment to include an analysis of 
uncertainty in 1993 (US EPA 1993B). The assessment was further revised 
to include revisions to risk factors and other variable values. The 
latest uncertainty analysis was completed in 1995 (US EPA 1995).
    EPA's revised risk analysis in support of this HRRCA takes into 
account new data on radon distributions and exposed populations 
developed in the updated occurrence analysis, as well as new 
information on dose-response relationships developed by the NAS (NAS 
1998B). For the HRRCA,

[[Page 9573]]

population risks are estimated using single-value ``nominal'' estimates 
of the various exposure factors which determine individual risk, and 
Monte Carlo simulation techniques are used to estimate risks associated 
with the distributions of radon exposures from the various size 
categories of CWSs. The risk equations and parameter values used in the 
revised risk assessment are summarized in Appendix A. EPA is currently 
conducting a comprehensive uncertainty analysis of radon risks using 
two-dimensional Monte Carlo methods to better judge the level of 
uncertainty associated with the radon risk estimates.
    Table 3-6 summarizes the results of EPA's revised baseline risk 
assessment. Because the NAS and EPA-derived dose-response and exposure 
parameters factors discussed above were used in the risk assessment, 
the proportions of risk associated with the various pathways were the 
same as shown in Table 3-4. The total estimated population risks 
associated with the current distribution of radon in CWSs was 160 fatal 
cancers per year, 142 of which were associated with progeny inhalation. 
Approximately 18 fatal cancers per year were associated with ingestion 
of radon. These totals are similar to, but somewhat lower than, EPA's 
1991 and 1993 baseline risk estimates (US EPA 1994C). In comparison, 
there are an estimated 15,400 to 21,800 fatal lung cancers per year due 
to inhalation of indoor air contaminated with radon emanating from soil 
and bedrock (NAS 1998A).
    The risks summarized in Table 3-5 do not include any contribution 
from NTNCWSs, Thus, the potential baseline risks and benefits of a 
radon rule may be somewhat underestimated. The limited available data 
concerning radon levels in NTNCWSs suggest that levels may be 
considerably higher (perhaps by 60 percent, on average) than those in 
CWSs of similar size (US EPA 1998L). However, it appears that the 
average exposure per unit activity in NTNCWSs is likely to be lower 
than that for CWSs. Because of the expected lower inhalation exposures, 
water ingestion rates, and frequencies and durations of exposure, the 
individual fatal cancer risk associated with a NTNCWS is expected to be 
lower compared to a CWS with similar radon levels. EPA is currently 
conducting additional analyses of NTNCWS exposures from radon in an 
attempt to refine the current approximate risk estimates.

            Table 3-6.--Annual Fatal Cancer Risks for Exposures to Radon From Community Water Systems
----------------------------------------------------------------------------------------------------------------
                                                                    Annual unit
                                                                    risk (fatal       Annual
                                                                    cancers per     population     Proportion of
                             Pathway                                person per      risk (fatal    total annual
                                                                  year per pCi/l    cancers per   risk (percent)
                                                                    in water)1        year) 2
----------------------------------------------------------------------------------------------------------------
Inhalation of progeny...........................................     7.44 x 10-9             142              89
Ingestion of radon gas..........................................    9.30 x 10-10            17.8              11
Inhalation of radon gas.........................................     4.7 x 10-12             0.1            0.06
      Total.....................................................     8.37 x 10-9             160            100
----------------------------------------------------------------------------------------------------------------
\1\ Derived using NAS lifetime unit fatal cancer risks.
\2\ Estimated through simulation analysis described in Section 3.4; the risk equations and parameter values used
  in the simulation analysis are summarized in Appendix A.

3.6  Potential for Risk Reductions Associated With Removal of Co-
Occurring Contaminants

    Because radon is a naturally occurring ground water contaminant, 
its occurrence patterns are not highly correlated with those of 
industrial pollutants. Similarly, the Agency's re-evaluation of radon 
occurrence has revealed that the geographic patterns of radon 
occurrence are not significantly correlated with naturally occurring 
inorganic contaminants that may pose health risks. Thus, it is not 
likely that a 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. All of the aeration technologies discussed remove 
volatile organic contaminants, as well as radon, from contaminated 
ground water. Similarly, 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 with which radon 
treatment would also reduce risks from other contaminants, and the 
extent of risk reduction that would be achieved, has not been evaluated 
quantitatively in the HRRCA.

3.7  Potential for Risk Increases From Other Contaminants Associated 
With Radon Removal

    As discussed in Section 5.1, the need to install radon treatment 
technologies may require some systems that currently do not disinfect 
to do so. While case studies (US EPA 1998D) 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 
systems either had disinfection already in place or did not add it), in 
practice the tendency to add disinfection 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 (US EPA 1998I). This rule identified MCLs for the 
major DBPs, with which all CWSs and NTNCWSs will have to 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 is proportional to the concentration of 
organic precursor contaminants, which tend to be much lower in ground 
water than in surface water.
    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 1998B). The

[[Page 9574]]

report concluded that, based upon median and average total 
trihalomethane (THM) levels from EPA's 1981 Community Water System 
Survey, a typical ground water CWS will face an incremental individual 
lifetime cancer risk due to chlorination byproducts of 
5x10-5. It should be emphasized that this risk is based on 
average and median THM occurrence information that does not segregate 
systems that disinfect from those that do. 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.
    A more meaningful comparison is to look at the trade-off between 
risk reduction from radon treatment in cases where disinfection is 
added with the added risks from DBP formation. This trade-off will 
affect only a minority of systems since a majority of ground water 
systems already have disinfection in place. For the smallest systems 
size category, approximately half of all CWSs already have disinfection 
in place. The proportions of systems having disinfection in place 
increases as the size categories increase, up to >95% for large systems 
(Table 5-2). In addition, although EPA is using the conservative 
costing assumption that all systems adding aeration or GAC would 
disinfect, not all systems adding aeration or GAC would have to add 
post-disinfection or, if disinfecting, may use a disinfection 
technology that does not forms DBPs. For those ground water systems 
adding treatment with disinfection, this trade-off tends to be 
favorable since the combined risk reduction from radon removal and 
microbial risk reduction outweigh 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 3.7. As noted by the NAS Report, these risk 
reductions outweigh the increased risk from DBP exposure for those 
systems that chlorinate as a result of adding radon treatment.

              Table 3-7.--Radon Risk Reductions Across Various Effluent Levels and Percent Removals
----------------------------------------------------------------------------------------------------------------
                                                  Risk reduction  Risk reduction  Risk reduction  Risk reduction
                  % Removal \1\                     @ 50 pCi/L      @ 100 pCi/L     @ 200 pCi/L     @ 300 pCi/L
----------------------------------------------------------------------------------------------------------------
60..............................................          \2\ NA              NA         1.9E-04         2.8E-04
80..............................................              NA         2.5E-04         5.0E-04         7.6E-04
90..............................................         2.8E-04         5.7E-04         1.1E-03         1.7E-03
99..............................................         3.1E-03         6.2E-03         1.2E-02        1.9E-02
----------------------------------------------------------------------------------------------------------------
\1\ Influent levels used in risk reduction calculations are determined by the relationship, Effluent Level =
  Influent Level*(1--%Removal/100).
\2\ NA = Not applicable since associated influent level would be outside the range of realistic values.

    Comparing the risk reductions in Table 3.7 to the risks from THMs 
at their MCL values (the maximum risk allowable under the DBP rule), 
the ratios between risk reduction from radon removal and the 
conservative assumption that DBPs are present at their MCL values are 
shown in Table 3.8.

                      Table 3-8.--Radon Risk Reduction from Treatment Compared to DBP Risks
----------------------------------------------------------------------------------------------------------------
                                                   Estimated risk ratios (risk reduction from radon removal/risk
                                                                     from THMs at 0.080 mg/L)
                  % Removal \1\                  ---------------------------------------------------------------
                                                  Ratio @ 50 pCi/   Ratio @ 100     Ratio @ 200     Ratio @ 300
                                                         L             pCi/L           pCi/L           pCi/L
----------------------------------------------------------------------------------------------------------------
60..............................................          \2\ NA              NA             1.6             2.4
80..............................................              NA             2.1             4.2             6.3
90..............................................             2.4             4.7             9.5            14.2
99..............................................            26.0            52.0           104.0          155.9
----------------------------------------------------------------------------------------------------------------
Notes: \1\ Influent levels used in risk reduction calculations are determined by the relationship, Effluent
  Level = Influent Level*(1--%Removal/100).
\2\ NA = Not applicable since associated influent level would be outside the range of realistic values.

    As can be seen in Table 3.8, the risk ratios are favorable for 
treatment with disinfection, ignoring microbial risk reduction, even 
assuming the worst case scenario that ground water systems have THM 
levels at the MCL. There is the possibility that 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.

3.8  Risk for Ever-Smokers and Never-Smokers

    As noted previously, cancer risks from inhalation of radon progeny 
are believed to be greater for current and former smokers than for 
``never smokers''. The NAS defines a ``never smoker'' as someone who 
has smoked less than 100 cigarettes in their lifetime. Therefore, 
``ever smokers'' include current and former smokers. EPA and NAS have 
developed estimates of unit risk values (estimates of cancer risks per 
unit of exposure) for radon progeny for ``ever-smokers'' and ``never-
smokers'' as shown in Table 3-9 (US EPA 1999A). The estimated unit risk 
values for inhalation of radon progeny for ever-smokers (and therefore 
the individual and population risk) is approximately 5.5 times greater 
than that for never smokers.
    Because of estimated higher individual risks for smokers, this 
group accounts for a large proportion of the overall population risk 
associated with radon progeny inhalation. The last two columns of the 
table show that, given the current assumptions about smoking prevalence 
and the relative impact of radon progeny on ever smokers and never 
smokers, about 85 percent of the cancer cases from water exposures to

[[Page 9575]]

progeny will occur in the ever-smoker population.

 Table 3-9.--Annual Lung Cancer Death Risk Estimates From Radon Progeny for Ever-Smokers, Never-Smokers, and the
                                               General Population
----------------------------------------------------------------------------------------------------------------
                                                    Annual unit
                                                    risk (fatal   Average annual      Annual       Proportion of
                                                   cancer cases     individual      population     total annual
                 Smoking status                    per year per    risk per year    risk (fatal     population
                                                     pCi/l in       of exposure     cancers per        risk
                                                      water)                           year)
----------------------------------------------------------------------------------------------------------------
Ever............................................       1.31X10-8        2.8X10-6             120              85
Never...........................................       2.44X10-9        5.1X10-7              22              15
Combined........................................       7.44X10-9        1.6X10-6             142             100
----------------------------------------------------------------------------------------------------------------
Source: EPA analyses derived from NAS (1998) estimates.
 
Note: Ever-smoking prevalence was assumed to be 58 percent in males and 42 percent in females, and these rates
  were assumed to be age independent.

4. Benefits of Reduced Radon Exposures

4.1  Nature of Regulatory Benefits

4.1.1  Quantifiable Benefits
    The benefits of controlling exposures to radon in drinking water 
take the form of avoided cancers resulting from reduced exposures. 
Cancer risks (both fatal and non-fatal cancers per year) are calculated 
using the risk model described in Section 3 for the baseline case 
(current conditions) and each of the radon levels. The health benefits 
of controls are estimated as the baseline risks minus the residual 
risks associated with each radon level. The more stringent the radon 
level, the lower the residual risks, and the higher the benefits.
    The primary measures of regulatory benefits that are used in this 
analysis are the annual numbers of fatal and non-fatal cancers 
prevented by reduced exposures. Due to a lack of knowledge about how to 
account for the latency period for radon-induced cancers, it has been 
assumed that risk reduction begins to accrue immediately after the 
reduction of exposures.
    Exposures to radon and its progeny are associated with increases in 
lung cancer risks. Ingestion of radon in drinking water is suspected of 
being associated primarily with increased risks of tumors of the 
stomach, and with lesser risks to the colon, lung, and other organs. 
The first column of Table 4-1 summarizes the estimates of the 
distribution of cancers by organ system for inhalation and ingestion 
exposures given. For purposes of the risk assessment, inhalation of 
progeny and radon gas are assumed to be associated exclusively with 
lung cancer risk. In the case of radon ingestion, stomach cancer 
accounts for the bulk (approximately 87 percent) of the total risk by 
this pathway. Cancers of several other organ systems account for far 
smaller proportions of the cancer risk from radon ingestion, and are 
not included in this analysis.

               Table 4-1.--Proportion of Fatal Cancers by Exposure Pathway and Estimated Mortality
----------------------------------------------------------------------------------------------------------------
                                                                                   Proportion of
                                                                                   fatal cancers
                                                                                   by organ and      Mortality
              Exposure pathway                          Organ affected               exposure      (percent) \2\
                                                                                      pathway
                                                                                   (percent) \1\
----------------------------------------------------------------------------------------------------------------
Inhalation of progeny, radon gas...........  Lung...............................              89              95
Ingestion of radon gas.....................  Stomach............................             9.5              90
                                             Colon..............................             0.4             550
                                             Liver..............................             0.3              95
                                             Lung...............................             0.2              95
                                             General Tissue.....................             0.5              --
----------------------------------------------------------------------------------------------------------------
\1\  Source: US EPA analysis of dosimetry data and organ-specific risk coefficients (NAS 1998).
\2\ Source: US EPA analysis of National Cancer Institute mortality data.

    The last column of Table 4-1 provides estimates of the mortality 
rate associated with the various types of radon-associated cancers. 
These values are used in this analysis to estimate the proportion of 
fatal and non-fatal cancers by organ system and exposure pathway. Both 
of the cancers that account for the bulk of the risk from radon and 
progeny exposures (lung and stomach) have high mortality rates.
4.1.2  Non-Quantifiable Benefits
    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 include any peace of mind benefits specific to reduction 
of radon exposure that may not be adequately captured in the VSL 
estimate. In addition, treating radon in drinking water with aeration 
oxidizes arsenic into a less soluble form that is easier to remove with 
conventional 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 is a non-quantifiable benefit that 
can be attributed to some components of a MMM program. Providing such 
information might allow households to make informed choices about the 
appropriate level of risk reduction given their specific circumstances 
and concerns. These potential benefits are

[[Page 9576]]

difficult to quantify due to the uncertainty surrounding their 
estimation. However, they are likely to be somewhat less in magnitude 
relative to the monetized benefits estimates.

4.2  Monetization of Benefits

4.2.1  Estimation of Fatal and Non-Fatal Cancer Risk Reduction
    The ``direct'' health benefits of the regulation, as discussed 
above, are the reduced streams of cancer cases associated with reduced 
radon exposures. In this analysis, the data in Table 3-6 were used to 
estimate the numbers of fatal cancers of each organ system associated 
with inhalation and ingestion pathway from the risk model described in 
Section 3.1. (These proportions, by the nature of the risk model that 
is used, stay constant for all radon levels.) Subsequently, the total 
number of cancers of each organ system was estimated. This is necessary 
because the output of the risk model is fatal cancers, and the cost of 
illness and willingness to pay for non-fatal cancers are only applied 
to individuals who survive the disease. The total number of cancers per 
year of exposure, and the number of non-fatal cancers were estimated 
from the fatal cancer numbers using the mortality data in Table 4-1. 
Thus, for example, a benefit of 100 cases of fatal lung cancer avoided 
implies approximately 105 total lung cancers avoided, five of which are 
non-fatal. This calculation omits rounding error, and the total number 
of cases is equal to the fatal cases divided by the mortality rate.
    Fatal and non-fatal population cancer risks under baseline 
conditions were estimated first. Then, the residual cancer risks were 
estimated for each of the radon levels. Consistent with the assumptions 
made in the cost analysis, residual water radon levels were calculated 
using a similar range of technology efficiencies. Radon levels were 
assumed to be reduced below baseline levels by either 50, 80, or 99 
percent, using the least stringent reduction which could comply with 
the radon level under evaluation. Benefits took the form of the 
reductions in the numbers of fatal and non-fatal cancers associated 
with each final level compared to the baseline risks.
4.2.2  Value of Statistical Life for Fatal Cancers Avoided
    As one measure of potential benefits, this analysis assigns the 
monetary value of a statistical life saved to each fatal cancer 
avoided. The estimation of the value of a statistical life involves 
inferring individuals' implicit tradeoffs between small changes in 
mortality risk and monetary compensation (US EPA 1998E). A central 
tendency value of $5.8 million (1997$) is used in the monetary benefits 
calculations, with low- and high-end values of $700,000 (1997$) and 
$16.3 million (1997$), respectively, used for the purposes of 
sensitivity analysis. These figures span the range of value of 
statistical life (VSL) estimates from 26 studies reviewed in EPA's 
recent guidance on benefits assessment (US EPA 1998E) which is 
currently being reviewed by EPA's 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. As noted above, no separate 
medical care or lost-time costs are included in the benefits estimate 
for fatal cancers because it is assumed that these costs are captured 
in the VSL for fatal cancers.
4.2.3  Costs of Illness and Lost Time for Non-Fatal Cancers
    Two important elements in the estimation of the economic impacts of 
reduced cancer risks for non-fatal cancers are the reductions in 
medical care costs and the costs of lost time. The costs of medical 
care represent a net loss of resources to society (not considering the 
economic hardship on the cancer patient and family). The cost of lost 
time represents the value of activities that the individual must 
abandon (e.g., productive employment or leisure) as a result of radon-
induced cancer. Together, these two elements are often referred to as 
the costs of illness (COI).
    Medical care and lost-time costs have been estimated for lung and 
stomach cancers, which are the two most common types of tumors 
associated with radon exposures, and which account for 99 percent of 
the total radon-associated cancers. Table 4-2 summarizes the Agency's 
latest medical care and lost-time cost estimates for lung cancer (US 
EPA 1998B, 1998C). Medical care costs have been estimated from survey 
data for ten years after initial diagnosis. The medical costs in the 
first year correspond to the costs of initial treatment, while medical 
costs in subsequent years correspond to the average medical costs 
associated with monitoring and treatment of recurrences among 
individuals who survive to that year. These out-year costs are weighted 
by the proportion of patients surviving to the given year.
    The lost time due to the radon-induced tumors is assumed to be 
concentrated in the first year after diagnosis. This is why the out-
year estimates for the costs of lost time in Table 2-8 are all zero. 
The dollar costs of lost time given in the table are derived by 
assigning values lost productive (work) and leisure (non-productive) 
hours. The costs given in the top row of Table 4-2 correspond to 776 
lost productive hours and 1,493 lost leisure hours per patient. The 
estimates of lost hours are relatively low for lung cancer primarily 
because the average age at diagnosis is advanced (fewer than 34 percent 
of lung cancer patients are diagnosed before age 65).
    Using a discount rate of seven percent, the estimated discounted 
present value in 1997 dollars of combined medical care and lost-time 
costs for a cancer survivor is approximately $108,000. The estimated 
value varies with different discount rates. Using a discount rate of 
three percent, combined costs are $121,600; at ten percent, combined 
costs are approximately $100,200.
    Table 4-3 summarizes the estimation of medical and lost-time costs 
for survivors of stomach cancer. The combined discounted costs for 
stomach cancer are similar to those for lung cancer, but slightly 
higher. At a seven percent discount rate, combined discounted costs for 
stomach cancer are approximately $114,000 (1997$). At three percent, 
they are about $126,300 (1997$). Discounted at ten percent, the average 
combined cost is $106,400 (1997$).

          Table 4-2.--Estimated Medical Care and Lost-Time Costs Per Case for Survivors of Lung Cancer
----------------------------------------------------------------------------------------------------------------
                                                            Medical care       Cost of lost       Cost of lost
                                                               costs             leisure        productive time
                  Year after diagnosis                     (undiscounted      (undiscounted      (undiscounted
                                                         1997 dollars) \1\  1997 dollars) \2\  1997 dollars) \2\
----------------------------------------------------------------------------------------------------------------
1......................................................            $34,677             $9,886            $14,393

[[Page 9577]]

 
2......................................................              9,936                  0                  0
3......................................................              9,383                  0                  0
4......................................................              8,969                  0                  0
5......................................................              8,604                  0                  0
6......................................................              8,262                  0                  0
7......................................................              7,934                  0                  0
8......................................................              7,609                  0                  0
9......................................................              7,287                  0                  0
10.....................................................              6,974                  0                  0
Discounted Present Value at 7 Percent..................             85,225              9,390             13,671
Total Discounted Value (1997 dollars)..................           108,287
----------------------------------------------------------------------------------------------------------------
\1\ Medical care cost estimates derived from US EPA 1998B.
\2\ Lost productive and leisure hours estimates from US EPA 1998B; value of productive time estimated at $12.47/
  hr, value of leisure hour estimated at $9.64/hour (from US EPA 1998J).


         Table 4-3.--Estimated Medical Care and Lost-Time Costs Per Case for Survivors of Stomach Cancer
----------------------------------------------------------------------------------------------------------------
                                                                              Cost of lost        Cost of lost
                                                       Medical care costs        leisure        productive time
                 Year after diagnosis                  (Undiscounted 1997  (undiscounted 1997    (undiscounted
                                                          dollars) \1\        dollars) \2\     1997 dollars) \2\
----------------------------------------------------------------------------------------------------------------
1....................................................          $37,507.28          $19,337.84             13,288
2....................................................            9,328.23                0                     0
3....................................................            8,749.24                0                     0
4....................................................            8,265.39                0                     0
5....................................................            7,829.62                0                     0
6....................................................            7,423.51                0                     0
7....................................................            7,035.81                0                     0
8....................................................            6,663.46                0                     0
9....................................................            6,300.32                0                     0
10...................................................            5,946.38                0                     0
Discounted Present Value at 7 Percent................           82,997.35           18,368                12,621
Total Discounted Value (1997 dollars)................         113,987
----------------------------------------------------------------------------------------------------------------
\1\ Medical care cost estimates derived from US EPA 1998C.
\2\ Lost productive and leisure hours estimates from US EPA 1998C; value of productive time estimated at $12.47/
  hr, value of leisure hour estimated at $9.64/hour (from US EPA 1998J).

4.2.4  Willingness to Pay to Avoid Non-Fatal Cancers
    As was the case for fatal cancers, willingness to pay (WTP) 
measures of the values of avoiding serious non-fatal illness have also 
been developed. These WTP measures were developed because the cost of 
illness estimates may be seen as understating total willingness to pay 
to avoid non-fatal cancers. The main reason that the cost of illness 
understates total WTP is the failure to account for many effects of 
disease--it ignores pain and suffering, defensive expenditures, lost 
leisure time, and any potential altruistic benefits (US EPA 1998E). 
Recently, EPA applied one such study to evaluate the benefits of 
avoiding non-fatal cancers in the Regulatory Impact Analysis for the 
Stage I Disinfection By-Products Rule (US EPA 1998M). That study 
estimated a range of WTP to avoid chronic bronchitis ranging from 
168,600 to 1,050,000 with a central tendency (mean) estimate of 536,000 
(Viscusi et al. 1991). In the benefits assessment, EPA uses the central 
tendency measure as a surrogate for the cost of avoiding non-fatal 
cancers and an alternative to the cost of illness measures discussed 
above. The high and low ends of the range are used in sensitivity 
analysis of the monetized benefit estimates.

4.3  Treatment of Monetized Benefits Over Time

    The primary measures of regulatory benefits that are used in this 
analysis are the annual numbers of expected fatal and non-fatal cancers 
prevented by reduced exposures to radon in drinking water. The monetary 
valuation of fatal cancer risks used is a result of a benefits transfer 
exercise from the risk of immediate accidental death to the risk of 
fatal cancer. No adjustments to the benefits calculations have been 
made to reflect the time between the reduction in exposure and the 
diagnosis and illness or possible death from cancer. Also, no 
adjustments have been made for any other factors which might affect the 
valuation. Cancer valuations could be adjusted for how they differ from 
accidental death valuations with respect to timing (latency) and with 
respect to other factors that may affect individuals' willingness-to-
pay for cancer risk reduction, including dread, pain and suffering, the 
degree to which the risk is voluntary or involuntary, and the amount by 
which life spans are shortened. Such adjustments have been under debate 
in the academic literature. In the absence of quantitative evidence on 
the relative impact of each factor, EPA has not adjusted the benefits 
estimates in this HRRCA to account for the factors discussed here. The 
Agency is currently reviewing the various issues raised; at this time 
no Agency policy regarding any such adjustments is in place.

[[Page 9578]]

5. Costs of Radon Treatment Measures

    This section describes how the costs and economic impacts of 
reductions in radon exposures were estimated. The most commonly used 
and cost-effective technologies for mitigating radon are described, 
along with the degree of radon removal that can be achieved. Costs of 
achieving specified radon removal levels for specific flow rates are 
discussed, along with the need for pre-and post-treatment technologies. 
The methods used to estimate treatment costs for single systems and 
aggregate national costs are explained, and the approach for 
translating the costs into economic impacts on affected entities is 
also described.

5.1  Drinking Water Treatment Technologies and Costs

    The two most commonly employed methods for removing radon from 
water supplies are aeration and granular activated carbon (GAC) 
absorption. These treatment approaches can be technically feasible and 
cost-effective over a wide range of removal efficiencies and flow 
rates. In addition to the radon treatment technologies themselves, 
specific pre-or post-treatment technologies may also be required. When 
influent iron and manganese levels are above certain levels, pre-
treatment may be required to remove or sequester these metals and avoid 
fouling the radon removal equipment. Also, aeration and GAC absorption 
may introduce possible infectious particulates into the treated water. 
Thus, disinfection is generally required as a post-treatment when radon 
reduction technologies are installed.
    When only low removal efficiency is required, and sufficient 
capacity is available, simple storage may in some cases be sufficient 
to reduce radon levels in water below specified radon levels. Radon 
levels rapidly decrease through natural radioactive decay, and if 
storage is in contact with air, through volatilization. Therefore, 
storage has also been included in the cost analysis.
    In some cases, water systems will choose to seek other sources of 
water rather than employ expensive treatment technologies. Systems may 
choose a number of strategies, such as shutting down sources with high 
radon levels and pumping more from sources with low levels, or 
converting from ground water to surface water. In the cost analysis, 
however, it has been assumed that such options will not be available to 
most systems, and they will need to obtain water from other systems. 
This option is referred to as ``regionalization'' in the following 
discussions.
    These general families of technologies, along with the specific 
variants used in the cost analysis, are described.
5.1.1  Aeration
    Because of radon's volatility, when water containing radon comes 
into contact with air, the radon rapidly diffuses into the gas phase. 
Several aeration technologies are available. As will be discussed in 
more detail below, the specific technology adopted in response to the 
rule will depend on the system's influent radon level, size, and the 
degree of radon removal that is required. The following common aeration 
technologies have been included in this analysis. Other aeration 
technologies are available (spray aeration, tray aeration, etc.) that 
can potentially be used by water systems to remove radon. These 
technologies have not been included in the analysis either because they 
have technical characteristics that limit their use in public water 
systems, or because their removal efficiencies are lower, and/or their 
unit costs are higher than the three aeration technologies included in 
the analysis.
    Packed Tower Aeration (PTA). During PTA treatment, the water flows 
downward by gravity and air is forced upward through a packing material 
that is designed to promote intimate air-water contact. The untreated 
water is usually distributed on the top of the packing with sprays or 
distribution trays and the air is blown up a column by forced or 
induced draft. This design results in continuous and thorough contact 
of the liquid with , air (US EPA 1998O). In terms of radon removal, PTA 
is the most effective aeration technology. Radon removal efficiencies 
of up to 99.9 percent are technically feasible and not prohibitively 
expensive for most applications. In this analysis, two different PTA 
treatments are used to estimate radon removal cost. The costs are 
dependant on the degree of reduction required to achieve compliance 
with the allowable radon level. The first design is capable of reducing 
radon levels by 80 percent; the second and more costly version reduces 
radon in drinking water by 99 percent.
    Diffused Bubble Aeration (DA). Aeration is accomplished in the 
diffused-air type equipment by injecting bubbles of air into the water 
by means of submerged diffusers or porous plates. The untreated water 
enters the top of the basin and exits from the bottom [having been] 
treated, while the fresh air is blown from the bottom and is exhausted 
from the top (US EPA 1998O). Diffused bubble aeration can achieve radon 
removal efficiencies greater than 90 percent. In this analysis, a DA 
system with a removal efficiency of 80 percent is used as the basis for 
estimating compliance costs.
    Multiple Stage Bubble Aeration (MSBA). MSBA is a variant of DA 
developed for small to medium water supply systems (US EPA 1998O). MSBA 
units consist of shallow, partitioned trays. Water passes through 
multiple stages of bubble aeration of relatively shallow depth. In this 
analysis, an MSBA radon removal efficiency of 80 percent is assumed.
    All of the aeration technologies discussed above are assumed to be 
``central'' treatments in the cost analysis. That is, a single large 
installation is used to treat water from a given source, prior to the 
water entering the distribution system to serve many users. It is also 
technically feasible to apply some of these technologies at the point 
of entry (e.g. just before water from the distribution system enters 
the household where it is to be used). However, most aeration 
technologies are only cost-effective at minimum flows far above that 
corresponding to the water usage rate of a typical household, and thus 
would not likely be selected as the treatment of choice.
    Also, in all of the aeration systems just discussed, the radon 
removed from water is released to ambient (outdoor) air. In this 
analysis, it has been assumed that the air released from aeration 
systems will not itself require treatment, result in appreciable risks 
to public health, or result in increased permitting costs for water 
systems. For the 1991 proposed rule, EPA conducted analyses on radon 
emissions and potential risks associated with radon and its progeny as 
they disperse from a water treatment facility (US EPA 1988, 1989). In 
summary, these analyses concluded that the annual risk of fatal cancer 
from radon and its progeny in off-gas emissions was 2,700 times smaller 
(108 cases/0.04 cases) than the annual risk of fatal cancer from radon 
and its progeny from tap water after all ground water systems were at 
or below the 1991 target level of 300 pCi/L. Using the occurrence 
estimates at that time, the off-gas risk was estimated to be 4800 times 
smaller (192 cases/0.04 cases) than the radon in tap water risk if no 
water mitigation was done (US EPA 1994C). The EPA's SAB reviewed the 
Agency's report and concluded that: (1) while the uncertainty analysis 
could be upgraded to lend greater scientific credibility, the results 
of modeling would not likely change, i.e., the risk posed by release of 
radon through treatment would be less

[[Page 9579]]

than that posed by drinking untreated water; and (2) it is likely that 
the conservative assumptions adopted by EPA in its air emissions 
modeling resulted in overestimates of risk (US EPA 1994C).
5.1.2  Granular Activated Carbon (GAC)
    The second major category of radon removal technology is treatment 
with granular activated carbon. GAC adsorption removes contaminants 
from water by the attraction and accumulation of the contaminant on the 
surface of carbon. The magnitude of the available surface area for 
adsorption to occur is of primary importance, while other chemical and 
electrochemical forces are of secondary significance. Therefore, high 
surface area is an important factor in the adsorption process (US EPA 
1998O). GAC systems are commonly used in water supply systems to remove 
pesticides or other low-volatility organic chemicals that cannot be 
removed by aeration. Radon can also be captured by GAC filtration, but 
the amounts of carbon and the contact times needed to produce a high 
degree of radon removal are generally much greater than those required 
to remove common organic contaminants. For most system sizes and design 
configurations evaluated in this study, aeration can achieve the same 
degree of radon reduction at lower cost than GAC. However, in the cost 
analysis for the radon rule, it has been assumed that a small minority 
of systems will nonetheless choose GAC technology over aeration 
alternatives, due to system-specific needs (e.g., land availability). 
Also, POE GAC (see below) may be cost-effective for systems serving 
only a few households. Depending on the specific design and operating 
characteristics, GAC can remove up to 99.9 percent of influent radon, 
but high removal efficiencies require large amounts of carbon and long 
contact times.
    Two types of GAC systems have been evaluated: Central GAC and Point 
of Entry GAC (POE GAC). Central GAC refers to a design configuration in 
which the activated carbon treatment takes place at a central treatment 
facility, prior to entry into the distribution system. GAC may be 
combined with other treatments and may be used to remove contaminants 
other than radon in large, centralized facilities. In this analysis, 
costs are estimated for central GAC systems with removal rates of 50, 
80, and 99 percent. POE GAC generally refers to small- to medium-sized 
carbon filtration units placed in the water distribution system just 
before use occurs (e.g., before water enters a residence from the 
distribution system.) System maintenance involves periodic replacement 
of the filter units. As noted previously, POE GAC may be the most cost-
effective treatment for very small systems serving few households. 
Costs are estimated for POE GAC with removal rates of 99%.
5.1.3  Storage
    Another technology that may be practical when only a relatively 
slight reductions in radon levels are needed is the storage of water 
for a period of time necessary for radioactive decay and volatilization 
to reduce radon to acceptable levels. Depending on the configuration of 
the vessel, storage for 24 to 48 hours may be sufficient to reduce 
radon levels by 50 percent or more. The mode of removal is a 
combination of radon decay and transfer of the radon from the water to 
the storage tank headspace, which is refreshed through ventilation (US 
EPA, 1998D). It has been assumed that a proportion of the smallest CWSs 
(serving 500 people or fewer) with relatively low influent radon levels 
and sufficient storage capacity may choose storage as the preferred 
radon treatment technology. In estimating costs for the storage option, 
it is assumed that the entire capital and O&M costs of the storage 
system is attributable to the need to reduce radon levels. In fact, the 
majority of CWSs choosing storage are likely to already have at least 
some storage capacity available (ten percent of small systems have 
atmospheric storage in place (US EPA 1997A)). These systems may be able 
to add ventilation and/or other mechanisms to increase air/water 
contact with a small capital investment, which supports the conclusion 
that the present assumption of no storage in place is a conservative 
assumption.
5.1.4  Regionalization
    The last technology whose costs are included in the HRRCA is 
regionalization. In this analysis, regionalization is defined as the 
construction of new mains to the nearest system with water below the 
required radon level. This cost is estimated to be $280,000 per system 
(1997$). The cost of actually purchasing water is not included in 
regionalization costs, for several reasons. In the first case, 
regionalization may involve the actual consolidation of water systems, 
and thus there may be no charge to the system which is 
``regionalized''. In addition, the system which supplies the water to 
the regionalized system will still incur the same (or nearly the same) 
costs for radon treatment as before regionalization and could be 
expected to pass them on to the regionalized system. This assumes that 
the water production cost ($/kgal) for the CWS before it regionalizes 
is equal to the unit price ($/kgal) it will pay to the water system 
from which it purchases water. In reality, this will over-estimate 
costs in some cases and under-estimate in others. Including a water 
purchase price in the cost estimate for regionalization without 
correcting it for the removal of water production costs would lead to 
an over-estimate in the costs of regionalization.
5.1.5  Radon Removal Efficiencies
    The amount of radon that the various technologies can remove from 
water varies according to their specific design and operating 
characteristics. At the most costly extreme, both aeration and GAC 
technologies can remove 99 percent or more of the radon in water. Less 
costly alternative designs remove less radon. In this analysis, one or 
more cost estimates have been developed for the technologies discussed 
above, corresponding to one or more radon removal levels. Approximate 
cost ranges for achieving specified radon reduction efficiencies using 
the various technologies are shown in Table 5-1. These costs are 
estimated based on flow rates for a single installation, which may 
treat water for an entire system or from a single source. For the 
aeration and GAC technologies, costs have also been derived for 
combined radon removal and post-treatment technologies, as discussed 
below. The basis for the derivation of these cost estimates is 
described in more detail in Section 5.4.
    The procedures used to decide what proportion of CWSs will adopt 
the various radon removal technologies is described in more detail in 
Section 5.5. In general, however, the large majority of the systems are 
assumed to select the least-cost technology required to achieve a 
target radon level. Other systems, for reasons of technical 
feasibility, may need to choose more costly treatment technologies.
5.1.6  Pre-Treatment to Reduce Iron and Manganese Levels
    Pre-treatment technologies may also need to be part of radon 
reduction systems. Aeration and GAC technologies can be fouled by high 
concentrations of iron and manganese (Fe/Mn). EPA believes that Fe/Mn 
concentrations greater than 0.3 mg/l would generally require 
pretreatment to protect aeration/GAC systems from fouling. However, 
since this level is near to the secondary MCL, it is believed that 
essentially all systems with iron and manganese levels

[[Page 9580]]

above 0.3 are likely to already be treating to remove or sequester 
these metals. Therefore, costs of adding Fe/Mn treatment to radon 
removal systems are not included in the HRRCA. Preliminary EPA 
estimates suggest that inclusion of Fe/Mn treatment costs will not 
significantly effect overall cost estimates for radon removal. More 
detailed analysis will be presented when the proposed NPDWR is 
published.

BILLING CODE 6560-50-P

[[Page 9581]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.000



BILLING CODE 6560-50-C

[[Page 9582]]

5.1.7  Post-Treatment--Disinfection
    In addition to pre-treatment requirements, the installation of some 
radon reduction technology may also require post-treatment, primarily 
to reduce microbial contamination. Both aeration and GAC treatment may 
introduce potentially infectious particulate contamination, which must 
be addressed before the water can enter the distribution system. The 
treatment of water for other contaminants may also introduce microbial 
contamination. This is one reason why the majority of systems already 
use disinfection technologies. As will be discussed in more detail 
below, a substantial proportion of ground water systems (ranging from 
50 percent in the smallest size category, to about 68 percent of the 
largest systems) already disinfect. Costs of disinfection are only 
attributed to the radon rule only for that proportion of systems not 
already having disinfection systems in place. For systems that do not 
already disinfect, chlorination is assumed to be the treatment of 
choice. Alternative technologies are available, for example UV 
disinfection, but chlorination is widely used in all size classes of 
water supply systems, and the chlorination is considered to provide a 
reasonable basis for estimating disinfection costs.

5.2  Monitoring Costs

    While not strictly speaking a water treatment technology, ground 
water monitoring will play an important role in any strategy to reduce 
radon exposures. Therefore, monitoring costs have been included as a 
cost element in the cost analysis. Although EPA has not yet defined a 
monitoring strategy for the proposed NPDWR, it is clear that systems 
will, first, have to sample influent water to determine the need for 
treatment, and second, continue to monitor after treatment (or after a 
decision is made not to mitigate). For the purpose of developing 
national cost estimates, it has been assumed that all systems will have 
to conduct initial quarterly monitoring of all sources, and continue to 
conduct radon monitoring and analysis indefinitely after the rule is 
implemented. This is a conservative assumption (likely to overstate 
monitoring costs) because in reality a large proportion of systems with 
radon levels below the MCL will probably be allowed to monitor less 
frequently after the initial monitoring period.
    Monitoring costs are simply the unit costs of radon analyses times 
the number of samples analyzed. The number of intake sites per system 
is estimated from SDWIS data, as discussed in Section 5.7. The cost of 
analyzing each sample is estimated to be between $40 and $75, with an 
representative cost of $50 per sample used for the national cost 
estimate (US EPA 1998K).

5.3  Water Treatment Technologies Currently In Use

    EPA has conducted an extensive analysis of water treatment 
technologies currently in use by ground water supply systems (Table 5-
2). This table 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 Fe/Mn 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.
    Within current EPA cost models, the estimate of the number of sites 
(entry points into the distribution system) is ideally broken down into 
three parts: estimates of the average national occurrence of the 
contaminant in drinking water systems, the intra-system variability of 
the contaminant concentration, and the typical number of sites within 
system size categories. In prior RIAs, EPA modeled all drinking water 
systems requiring treatment as installing centralized treatment, which 
assumes that there is one point of treatment within a system. A more 
accurate estimate of treatment would be to calculate costs according to 
treatment installed at each well site that is predicted to be above the 
target radon level within a water system. This intra-system variability 
analysis accounts for the fact that, in reality, multi-site water 
systems do not necessarily have the same radon level at each site. 
However, because the analysis of intra-system variability for radon 
occurrence is not yet complete, it is not possible to use this approach 
to calculate treatment costs. For future rules, including the proposed 
rule for radon, EPA will calculate national cost estimates based on the 
number of sites rather than by the system as a whole. These estimates 
will more accurately reflect the percentage of the population receiving 
drinking water that has been treated in some way and will result in 
more accurate national compliance cost estimates.
    The cost analysis assumes that any system affected by the rule will 
continue to employ pre-existing radon treatment technology and pre-and 
post-treatments in their efforts to comply with the rule. Where pre-or 
post-treatments are already in place, but radon treatment is currently 
not taking place, it is assumed that compliance with the radon rule 
will not require any upgrade or change in the pre-or post-treatments. 
Therefore, no incremental cost is attributed to pre-or post-treatment 
technologies. This may underestimate costs if pre-or post-treatments 
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 5-2.--Estimated Proportions of Ground Water Systems With Water Treatment Technologies Already in Place
                                                  (Percent) \1\
----------------------------------------------------------------------------------------------------------------
                                                          System size (population served)
 Water treatment technologies in -------------------------------------------------------------------------------
              place                25-100    101-500   501-1K    1K-3.3K  3.3K-10K   10K-50K  50K-100K   100K-1M
----------------------------------------------------------------------------------------------------------------
Fe/Mn Removal & Aeration &             0.4       0.2       1.2       0.6       2.9       2.2       3.1       2.0
 Disinfection...................
Fe/Mn Removal & Aeration........       0.0       0.1       0.2       0.1       0.4       0.1       0.4       0.1
Fe/Mn Removal & Disinfection....       2.1       5.1       8.3       3.0       7.8       7.4       9.7       6.8
Fe/Mn Removal...................       1.9       1.5       1.5       1.0       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       1.0       1.8       2.9       2.9       1.0       2.1       0.6
Disinfection Only...............      49.6      68.2      65.0      65.0      56.3      66.0      58.3      68.3

[[Page 9583]]

 
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.

5.4  Cost of Technologies as a Function of Flow Rates and Radon Removal 
Efficiency

    EPA has developed a set of cost curves that describe the 
relationships between the capital and operating and maintenance costs 
of the various treatment technologies, flow rates, and the degree of 
radon removal that is required (US EPA 1998A, 1998O). Cost curves were 
developed using the most recent available data and standard cost 
estimation methodologies. Separate functions for capital and operation 
and maintenance (O&M) costs have been developed for each technology and 
radon removal rate. For all of the technologies except regionalization, 
both the capital and O&M cost curves are functions of flow rates. 
Capital costs are estimated as a function of the design flow (DF) of 
the technology. The DF for a technology is equal to a technology's 
maximum flow capacity, or the largest amount of water that can be 
processed per unit time. The DF is typically two to three times greater 
than the average amount of water treated by a given system. O&M costs 
are functions of the average flow (AF) through the system. Labor, 
treatment chemicals and materials, periodic structure maintenance, and 
water stewardship expenses are estimated based on daily average flows. 
The cost curves developed by OGWDW for the various radon removal 
technologies are provided in Appendix B.

5.5  Choice of Treatment Responses

    The Agency has developed a set of assumptions regarding the choices 
that CWSs will make in deciding how to mitigate water radon levels to 
meet specific exposure reduction requirements. These assumptions have 
been developed taking into account the expected influent radon levels, 
the degree of radon removal needed to reach specified levels, the types 
of technologies that would be technically feasible and cost-effective 
for systems of a given size, and the distribution of pre-existing 
technologies shown in Table 5-2. Generally, it is assumed that a system 
will choose the least-cost alternative technology to achieve a given 
radon level. For example, to achieve a radon level of 100 pCi/l, all 
systems with average influent levels below 100 would not need to 
mitigate, systems with influent radon levels between 100 and 200 pCi/l 
would need to employ technologies that achieve 50 percent reduction, 
systems with influent levels between 200 and 500 pCi/l would employ 
technologies capable of 80 percent radon removal, and systems with 
influent radon above 500 pCi would employ technologies with removal 
efficiencies of 99 percent. In actuality, removal efficiencies would be 
more variable; e.g., a removal efficiency of 90 percent, rather than 99 
percent, could be employed for radon levels between 500 and 1,000 pCi/
l. However, this cost analysis has been limited to three removal 
efficiencies to simplify the analysis. EPA does not believe that this 
has introduced any significant bias into the assessment.
    Table 5-3 presents the estimated proportions of systems of given 
sizes that are expected to choose specified radon reduction 
technologies for given degrees of radon removal. Most systems in most 
size classes are assumed to choose aeration as the preferred radon 
reduction technology with or without disinfection, depending on the 
proportion of systems in that size stratum already disinfecting. This 
is because some form of aeration is generally the most cost-effective 
option for a given degree of radon reduction. For small systems and low 
required removal efficiencies, multistage fixed-bed (MSBA) and diffused 
bubble aeration (DA) tend to be the most cost-effective. For large 
systems and high removal efficiencies, packed tower aeration (PTA) is 
the only feasible aeration technology.
    Small proportions of the smallest system size categories (less than 
5 percent in all cases) are assumed to choose central GAC with or 
without disinfection. A few percent of the smallest systems are also 
assumed to choose POE GAC. Storage is assumed to be a viable option for 
two percent of small systems where radon reduction of 50 percent or 
less is required, and regionalization is assumed to be feasible for one 
percent of the smallest systems. EPA has assumed in this HRRCA that no 
systems would choose spray aeration or alternative source technologies. 
It is believed that these technologies would be chosen only rarely, and 
their omission has not biased the compliance cost estimates. This issue 
will be addressed in more detail in the proposed NPDWR.

                           Table 5-3.--Decision Matrix for Selection of Treatment Technology Options: Up to 50 Percent Removal
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                 Percent of system size category (population served) choosing treatment technology
               Treatment technology option               -----------------------------------------------------------------------------------------------
                                                             <100       101-500    501-1000    1001-3.3K   3301-10K     10-50K      50-100K    100-1000K
--------------------------------------------------------------------------------------------------------------------------------------------------------
PTA (80)................................................         2.6         7.8        16.8        31.9        60.8        86.9        86.3        96.4
PTA (80) + disinfection.................................         2.4         2.2         3.2         8.1         9.2         3.2        13.7         3.6
MSBA/STA (80)...........................................        13.2        21.8        22.7        15.9         8.7         0.0         0.0         0.0
MSBA/STA (80) + disinfection............................        11.8         6.2         4.3         4.1         1.3         0.0         0.0         0.0
DA (80).................................................        31.7        43.4        42.7        31.9        17.4         9.7         0.0         0.0
DA (80) + disinfection..................................        28.3        12.6         8.3         8.1         2.6         0.4         0.0         0.0
Retrofit Spray..........................................         0.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
GAC (50)................................................         2.6         2.3         0.8         0.0         0.0         0.0         0.0         0.0

[[Page 9584]]

 
GAC (50) + disinfection.................................         2.4         0.7         0.2         0.0         0.0         0.0         0.0         0.0
POE GAC (99)............................................         2.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
Storage (50)............................................         2.0         2.0         1.0         0.0         0.0         0.0         0.0         0.0
Regionalization (99)....................................         1.0         1.0         0.0         0.0         0.0         0.0         0.0         0.0
Alternate source (99)...................................         0.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
All Systems.............................................       100         100         100         100         100         100         100         100
PTA (80)................................................         4.2        10.9        20.2        31.9        60.8        96.5        86.3        96.4
PTA (80) + disinfection.................................         3.8         3.1         3.8         8.1         9.2         3.5        13.7         3.6
MSBA/STA (80)...........................................        14.8        21.0        21.0        15.9         8.7         0.0         0.0         0.0
MSBA/STA (80) + disinfection............................        13.2         6.0         4.0         4.1         1.3         0.0         0.0         0.0
DA (80).................................................        29.6        42.8        42.0        31.9        17.4         0.0         0.0         0.0
DA (80) + disinfection..................................        26.4        12.2         8.0         8.1         2.6         0.0         0.0         0.0
Retrofit Spray..........................................         0.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
GAC (80)................................................         2.6         2.3         0.8         0.0         0.0         0.0         0.0         0.0
GAC (80) + disinfection.................................         2.4         0.7         0.2         0.0         0.0         0.0         0.0         0.0
POE GAC (99)............................................         2.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
Regionalization (99)....................................         1.0         1.0         0.0         0.0         0.0         0.0         0.0         0.0
Alternate source (99)...................................         0.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
All Systems.............................................       100         100         100         100         100         100         100         100
PTA (99)................................................        15.3        26.5        35.3        47.8        69.4        96.5        86.3        96.4
PTA (99) + disinfection.................................        13.7         7.5         6.7        12.2        10.6         3.5        13.7         3.6
MSBA/STA (99)...........................................        34.3        49.1        48.7        31.9        17.4         0.0         0.0         0.0
MSBA/STA (99) + disinfection............................        30.7        13.9         9.3         8.1         2.6         0.0         0.0         0.0
GAC (99)................................................         1.6         1.6         0.0         0.0         0.0         0.0         0.0         0.0
GAC (99) + disinfection.................................         1.4         0.4         0.0         0.0         0.0         0.0         0.0         0.0
POE GAC (99)............................................         2.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
Regionalization (99)....................................         1.0         1.0         0.0         0.0         0.0         0.0         0.0         0.0
Alternate source (99)...................................         0.0         0.0         0.0         0.0         0.0         0.0         0.0         0.0
      Totals............................................       100         100         100         100         100         100         100         100
--------------------------------------------------------------------------------------------------------------------------------------------------------
Notes:
1. Technology abbreviations: PTA = packed tower aeration, MSBA/STA = multi-stage bubble aeration, GAC = granular activated carbon, POE GAC = point of
  entry granular activated carbon. Numbers in parentheses indicate removal efficiencies.
2. Capital costs for small systems include land costs. For large systems, it is assumed that additional land is not required.
3. Sequestration costs are included in PTA and MSBA/STA capital costs.
4. Additional housing costs are included in PTA, MSBA/STA, and GAC capital costs and are weighted under the assumption that 50% of small systems will
  require additional housing, 100% of large systems will require additional housing.
5. Permitting costs are included and are assumed to be 3% of capital costs, with a minimum of $2500.
6. Pump and blower redundancies are included in capital costs.

5.6  Cost Estimation

5.6.1  Site and System Costs
    The costs of reducing radon in ground water to specific radon 
levels was calculated using the cost curves discussed in Section 5.4 
and the matrix of treatment options presented in Section 5.5. For each 
radon level and system size stratum, the number of systems required 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 Safe Drinking Water Suite Model (US EPA 1998N). The 
equations and parameter values relating system size to flow rates are 
presented in Appendix C.
    The distributions of influent radon levels in the various system 
size categories were calculated using the results of EPA's updated 
radon occurrence analysis (exceedance proportions calculated from data 
in US EPA 1998L).
    Capital and O&M costs were estimated separately for each ``site'' 
(a separate water source, usually a well) within systems. Where systems 
obtained water from only one site, costs are calculated by applying the 
entire system flow rate to the appropriate cost curves. Where systems 
consisted of more than one site, the total system flow rate was divided 
by the number of sites, capital and O&M costs were then calculated for 
the resulting flow rate, and the total system cost was obtained by 
multiplying this result by the number of sites in the system. This 
approach provides conservative cost estimates, because it assumes that 
separate treatment systems would be built at each site. This approach 
also obscures some of the effects of variability in system sizes on 
costs, because each system in a given size category is assumed to have 
the same flow rate.
    Table 5-4 summarizes the numbers of sites per system for the 
various size categories of combined public and private community ground 
water systems. The average ranges from 1.1 site per system serving less 
than 100 people to almost nine sites per system serving greater than 
100,000 people. The distributions of the numbers of sites per systems 
are very skewed, with ninetieth-percentile values ranging from 2 to 20 
sites per system for the smallest and largest size categories, 
respectively. A large proportion of the systems serving 10,000 people 
or less obtain water from only one site. Public and private water 
systems differ with regard to system design and average flows. For

[[Page 9585]]

this reason, separate cost estimates have been developed for the public 
and private community ground water systems.

   Table 5-4.--Numbers of Sites per Ground Water System by System Size
------------------------------------------------------------------------
                                                                 90th
                                                  Average     percentile
        System size (population served)          sites per    sites per
                                                   system       system
------------------------------------------------------------------------
25-100........................................          1.1            2
101-500.......................................          1.2            2
501-1,000.....................................          1.4            3
1,001-3,300...................................          1.7            4
3,301-10,000..................................          2.3            4
10,001-50,000.................................          3.9           10
50,000-100,000................................          8.7           20
>100,000......................................          8.8          20
------------------------------------------------------------------------
Source: EPA analysis of CWSS data, 1998.

    In addition to the costs of radon treatment and disinfection, 
monitoring costs were also calculated for each system. As noted 
previously, the average cost of monitoring was estimated to be $50 per 
sample, and it was assumed that each site in a system would need to be 
monitored quarterly. Monitoring costs were added as an ongoing cost 
stream to the O&M costs.
5.6.2  Aggregated National Costs
    The estimated costs of reducing radon levels to meet different 
radon levels were estimated by summing the costs for the individual 
sites and systems in each size category and influent range. Separate 
totals were compiled for capital and O&M costs. Capital costs were 
annualized (over 20 years at a seven per cent discount rate) and added 
to the annual O&M costs to provide single aggregate estimates of 
national costs for each radon level. This approach implicitly assumes 
that treatment devices have useful lives that are identical to the 
period of financing. In reality, the useful life and period of 
financing are not necessarily the same. The aggregate cost estimates 
are presented in Section 6. As will be discussed in more detail below, 
separate cost estimates were developed for implementation options 
involving MMM programs and are presented in Section 7. Summary outputs 
of the spreadsheet models used to estimate costs are provided in 
Appendix D.
5.6.3  Costs to Community Water Supply Systems
    As noted above, costs were estimated separately for public and 
private ground water systems. Costs per system were calculated by 
dividing total costs for a given size category of public or private 
system by the total number of systems needing to mitigate radon. The 
results of these assessments are presented in Section 6.
5.6.4  Costs to Consumers/Households
    Costs to households have also been calculated for public and 
private ground water systems. Costs are calculated by multiplying the 
average annual treatment costs per thousand gallons by the estimated 
average household consumption (83,000 gal/year). This approach assumes 
that all water systems pass incremental costs attributable to the radon 
rule on to system's residential customers and that the residential 
customers will pay the same proportion of costs as other users. Average 
household costs are calculated separately for public and private 
community water systems across various system-size categories. Per 
household costs are then compared to median household income data (US 
EPA 1998H) for the same system-size categories. These impacts are 
discussed in Section 6.
5.6.5  Costs of Radon Treatment by Non-Transient Non-Community Systems
    Very little data are available that will support the development of 
detailed estimates of radon treatment costs for the NTNCWS that could 
be affected by a radon NPDWR. EPA is currently conducting a more 
detailed evaluation of the characteristics of NTNCWSs that will be 
completed in time for the proposed rule.

5.7  Application of Radon Related Costs to Other Rules

    The baseline for the radon rule compliance cost estimates presented 
in this draft HRRCA consists of the pre-existing treatment technology 
distribution shown in Table 5-2. As the radon rule is implemented, 
however, other rules may also require additional systems to install new 
technologies (e.g., disinfection). Thus, attributing all costs of 
increased use of disinfection at systems with high radon levels to the 
radon rule would overstate its cost. At the present time, EPA has not 
quantified the potential degree to which the costs of the radon rule 
may be overstated.

6. Results: Costs and Benefits of Reducing Radon in Drinking Water

    This section presents benefit, cost, and impact estimates for the 
various radon levels. Section 6.1 provides an overview of the 
analytical approach. Sections 6.2 and 6.3 present the monetized benefit 
and cost estimates for the various radon levels evaluated. Section 6.3 
summarizes the economic impacts on the various affected entities. 
Section 6.5 compares the costs and benefits of the radon levels 
evaluated. Section 6.6 presents a brief summary of the major 
uncertainties in the cost, benefit, and impact estimates.
    The presentation of costs and benefits in this Section is based on 
analysis of radon levels of 100, 300, 500, 700, 1,000, 2,000, and 4,000 
pCi/l in CWSs served by ground water.

6.1  Overview of Analytical Approach

    The analysis of benefits quantifies the reduction in health risks/
impacts to the general population and considers the risks to 
potentially sensitive subpopulations (qualitatively). The evaluated 
health benefits of the rule consist of reduced fatal and non-fatal 
cancer risks, and the monetary surrogates for these benefits have been 
estimated, as described in Section 4.0. The national cost estimates 
developed include the capital and O&M costs to reduce radon, along with 
pre- and post-treatment costs where appropriate, as well as monitoring 
costs. Record keeping and reporting costs and implementation costs to 
States and government entities will be addressed in the RIA prepared 
for the proposed rule.
    The costs and benefits of a radon NPDWR will result in economic 
impacts on affected individuals, corporate entities, and government 
entities. In this analysis, the impacts on water systems and households 
have been evaluated. These include: (1) the cost to systems of 
different sizes and ownership types, and (2) changes in water costs to 
households as a proportion of income. Public systems include those 
owned by government entities. Private systems consist of investor-owned 
entities that provide drinking water as their primary line of business. 
Ancillary systems include drinking water systems that are operated 
incidentally to another business. The vast majority of ancillary 
systems are mobile home parks, but some are schools, hospitals, and 
other entities. The economic impacts of the MMM programs on systems or 
households have not been calculated, because there is no information at 
present as to how these programs would be funded or upon whom the costs 
would fall.

6.2  Health Risk Reduction and Monetized Health Benefits

    The probabilistic risk model was used to calculate the cancer risk 
reduction benefits of the various levels. Risk reduction benefits were 
calculated by subtracting the estimated population risk (number of 
fatal cancers per year at a particular radon level) from the

[[Page 9586]]

baseline (pre-regulation) population cancer risk due to radon exposure. 
Estimates of the number of non-fatal cancers avoided were developed as 
described in Section 4.2.1. The results of this analysis are summarized 
in Table 6-1. Under the baseline scenario, the estimated number of 
fatal cancers per year caused by radon exposures in domestic water 
supplies is 160, and the number of non-fatal cancers is 9.2. As radon 
levels decrease, residual risks decrease, and the risk reduction 
benefits increase. Since very few people are exposed at levels above 
2,000 pCi/l, the benefit of controls in this range is relatively small 
(fewer than 7 cancers prevented per year). The health risk reduction 
benefits then increase rapidly as radon levels decrease because 
progressively larger populations are affected as more and more systems 
are required to mitigate exposures.

            Table 6-1.--Residual Cancer Risk and Risk Reduction From Reducing Radon in Drinking Water
----------------------------------------------------------------------------------------------------------------
                                                Residual fatal   Residual non-   Risk reduction   Risk reduction
                                                 cancer risk     fatal cancer    (fatal cancers     (non-fatal
         Radon level (pCi/l in water)             (cases per      risk (cases     avoided per    cancers avoided
                                                    year)          per year)       year) \1\      per year) \1\
----------------------------------------------------------------------------------------------------------------
(Baseline)...................................            160               9.2              0                0
4,000 \2\....................................            158               9.1              2.2              0.1
2,000........................................            153               8.8              6.5              0.4
1,000........................................            143               8.2             16                0.9
700..........................................            135               7.8             25                1.4
500..........................................            124               7.1             36                2.1
300..........................................            101               5.8             58                3.4
100..........................................             44.8             2.6            115                6.6
----------------------------------------------------------------------------------------------------------------
\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).

    At the lowest level (100 pCi/l) analyzed, the residual cancer risk 
(the cancer risk occurring after controls are installed) is 
approximately 45 fatal cancers per year. The risk reduction from this 
radon level is 115 fatalities per year, a reduction of approximately 72 
percent from the baseline of 160 per year. A similar proportional 
reduction in non-fatal cancers is seen with decreasing radon levels.
    The monetary valuation methods discussed in Section 4 were applied 
to these risk reductions, as shown in Table 6-2. The central tendency 
benefits estimates are based on a VSL of $5.8 million (1997$) and a WTP 
to avoid fatal cancers of $536,00 (1997$). The ranges of benefits 
estimated using the upper and lower bound estimates of the VSL and WTP 
to avoid non-fatal cancers are also provided in the table.

 Table 6-2.--Estimated Monetized Health Benefits From Reducing Radon in
                             Drinking Water
------------------------------------------------------------------------
                                                Monetized
                                                 health       Range of
                                                benefits,     monetized
                                                 central       health
             Radon Level (pCi/l)                tendency      benefits
                                              (annualized,  (annualized,
                                               $millions,    $millions,
                                                1997) \1\     1997) \2\
------------------------------------------------------------------------
4,000 \3\...................................            13          2-35
2,000.......................................            38         5-106
1,000.......................................            96        12-268
700.........................................           145        18-403
500.........................................           212        26-591
300.........................................           343        43-955
100.........................................           673      84-1875
------------------------------------------------------------------------
\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\ Estimates the range of VSL between $0.7 and $16.3 million (1997$),
  and a range of WTP to avoid non-fatal cancers between $169,000 (1997$)
  and $1.05 million (1997$).
\3\ 4,000 pCi/l is equivalent to the AMCL estimated by the NAS based on
  SDWA provisions of Section 1412(b)(13).

    Using central tendency estimates for each of the monetary 
equivalents, the baseline health costs of fatal and non-fatal cancers 
associated with household radon exposures from CWSs are estimated to be 
$933 million per year. Central tendency estimates of monetized benefits 
range from $13 million per year for a level of 4,000 pCi/l up to $673 
million for the most stringent level of 100 pCi/l. When different 
values for the VSL are used, the benefits estimates change 
significantly. Using a lower bound VSL of $0.7 million, the benefits 
estimates are reduced approximately 9-fold compared to the central 
tendency estimates. Using an upper bound VSL of 16.3 million increases 
the benefits estimates by approximately 3-fold relative to the central 
tendency estimate. Variations in the estimated WTP to avoid non-fatal 
cancers affect benefit total estimates only slightly (i.e., less than 1 
percent), since non-fatal cancers represent a very small proportion of 
estimated radon cancer cases.
    A more detailed breakout of the risk reduction, monetized benefits 
estimates, and the total cost per fatal cancer case avoided for ever-
smokers and never-smokers is provided in Tables 6-3 and 6-4.

                  Table 6-3.--Risk Reduction and Monetized Benefits Estimates for Ever-Smokers1
----------------------------------------------------------------------------------------------------------------
                                                                     Radon level, pCi/l
                                           ---------------------------------------------------------------------
                                              40003     2000      1000       700       500       300       100
----------------------------------------------------------------------------------------------------------------
Fatal Cancers Avoided Per Year............       1.7       5.2      13.2      19.9      29.2      47.1      92.5
Non-Fatal Cancers Avoided Per Year........       0.1       0.3       0.8       1.1       1.7       2.7       5.2
Annual Monetized Health Benefits                10.2      30.6      77.1     115.8     170.0     274.7     539.3
 ($Millions, 1997)--Central Tendency......

[[Page 9587]]

 
Annual Incremental Health Benefits              10.2      20.4      46.5      38.7      54.2     104.7     264.6
 ($Millions/year)--Central Tendency.......
Annual Cost Per Fatal Cancer Avoided             7.0       4.4       3.7       3.7       3.7       4.0      4.3
 ($Millions, 1997) 2......................
----------------------------------------------------------------------------------------------------------------
\1\ Risk reductions for ever- and never-smokers were estimated using the NAS unit risk estimates summarized in
  Table 3-4, an ever-smoking prevalence of 58% males and 42% females, a central VSL estimate of $5.8 million
  (1997$), and central WTP estimate to avoid non-fatal cancer of $536,000 (1997$).
\2\ Total cost estimates come from Table 6-5. The cost per fatal cancer case avoided is calculated by dividing
  the estimates of fatal cancers avoided per year by the annualized mitigation costs for each population. For
  purposes of this analysis, it was assumed that the mitigation costs (for both water and MMM programs) would be
  allocated equally to smoking and non-smoking populations.
\3\ 4000 pCi/l is equivalent to the AMCL estimated by the NAS based on the SDWA provisions of Section
  1412(b)(13).


                  Table 6-4.--Risk Reduction and Monetized Benefits Estimates for Never-Smokers
----------------------------------------------------------------------------------------------------------------
                                                                  Radon Level, pCi/l
                                    ----------------------------------------------------------------------------
                                       4000 *      2000       1000       700        500        300        100
----------------------------------------------------------------------------------------------------------------
Fatal Cancers Avoided Per Year.....       0.4        1.3        3.2        4.8        7.0       11.4       22.3
Non-Fatal Cancers Avoided Per Year.       0.03       0.09       0.22       0.33       0.48       0.78       1.54
Annual Monetized Health Benefits          2.4        7.4       18.6       27.9       41.0       66.3      130.2
 ($Millions, 1997)--Central
 Tendency..........................
Annual Incremental Health Benefits        2.4        5         11.2        9.3       13.1       25.3       63.9
 ($Millions/year)--Central Tendency
Annual Cost Per Fatal Cancer             29.2       18.3       15.3       15.4       15.5       16.4       17.8
 Avoided ($Millions, 1997).........
----------------------------------------------------------------------------------------------------------------
*4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

6.3  Costs of Radon Mitigation

    This section describes the incremental costs associated with each 
of the radon levels. Discussion of the cost results includes: the total 
nationally aggregated cost to all water systems that must comply with 
the target radon levels. These include capital and O&M costs; the 
average annualized cost per system exceeding the applicable radon 
level; the average annualized costs per system and incremental costs 
per household, broken out by public and private water system; and costs 
and impacts to households under each radon level. All costs are 
incremental costs stated in 1997 dollars. Capital costs were annualized 
using a seven percent discount rate and a 20-year amortization period.
6.3.1  Aggregate Costs of Water Treatment
    The total annual nationally aggregated cost varies significantly by 
the specific radon level. Total national cost estimates for CWSs are 
presented in Table 6-5. As demonstrated by the exhibit, water 
mitigation costs increase substantially from the highest radon level 
analyzed ($24 million at 4000 pCi/l) to the lowest level analyzed ($795 
million at 100 pCi/l).

                   Table 6-5.--Estimated Annualized National Costs of Reducing Radon Exposures
                                                [$Million, 1997]
----------------------------------------------------------------------------------------------------------------
                                                                      Central
                                                                     tendency        Range of     Cost per fatal
                       Radon level (pCi/l)                          estimate of     annualized      cancer case
                                                                    annualized    costs (+/-50%)      avoided
                                                                       costs
----------------------------------------------------------------------------------------------------------------
4000*...........................................................              24      12-36                 11.3
2000............................................................              46      23-70                  7.1
1000............................................................              98     49-146                  5.9
700.............................................................             148     75-223                  6.0
500.............................................................             218    109-327                  6.0
300.............................................................             373    187-560                  6.4
100.............................................................             795   398-1193                  6.9
----------------------------------------------------------------------------------------------------------------
*4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

    The costs borne by water systems are made up of annualized capital, 
O&M, and monitoring costs. The contributions of these cost elements are 
broken out in Table 6-6. As the radon level increases (i.e., is made 
less stringent), the proportion of costs due to monitoring increases 
relative to capital and O&M costs.

[[Page 9588]]



                     Table 6-6.--Capital and O&M Costs of Mitigating Radon in Drinking Water
                                                [$Million, 1997]
----------------------------------------------------------------------------------------------------------------
                                                                                      Annual
             Radon levels (pCi/l)                Annual capital  Annual O&M cost    monitoring      Total costs
                                                      cost                             costs
----------------------------------------------------------------------------------------------------------------
4000 *........................................              8.0              5.2            11.4              25
2000..........................................             19.8             15.3            11.4              46
1000..........................................             48.9             37.4            11.4              98
700...........................................             77.9             58.5            11.4             148
500...........................................            119               87.7            11.4             218
300...........................................            210              124              11.4             373
100...........................................            460.             324              11.4            795
----------------------------------------------------------------------------------------------------------------
* 4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

6.4  Incremental Costs and Benefits of Radon Removal

    Table 6-7 summarizes the central tendency and the upper and lower 
bound estimates of the incremental costs and benefits of radon exposure 
reduction. Both the annual incremental costs and benefits increase as 
the radon level is incrementally decreased from 2000 pCi/l down to 100 
pCi/l. The exhibit also illustrates the wide ranges of potential 
incremental costs and benefits due to the uncertainty inherent in the 
estimates. Incremental costs and benefits are within 10 percent of each 
other at radon levels of 1000, 700, and 500 pCi/l. There is substantial 
overlap between the incremental costs and benefits at each radon level.

                         Table 6-7.--Estimates of the Annual Incremental Costs and Benefits of Reducing Radon in Drinking Water
                                                                    [$Millions, 1997]
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                                                   Radon Level, pCi/l
                                                              ------------------------------------------------------------------------------------------
                                                                  4000 *       2,000        1,000         700          500          300          100
--------------------------------------------------------------------------------------------------------------------------------------------------------
Annual Incremental Cost......................................           24           46           52           50           70          156          422
Range of Annual Incremental Costs............................        12-36        11-34        26-76        26-77       34-104       78-233      211-633
Annual Incremental Monetized Benefits........................           13           25           58           48           67          130          329
Range of Incremental Monetized Benefits......................         2-35         3-71        7-162        6-135        8-188       17-364       41-920
Incremental Cost Per Fatal Cancer Case Avoided...............         11.3          5.0          5.2          6.1          6.1          7.0          7.5
--------------------------------------------------------------------------------------------------------------------------------------------------------
* 4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

6.5  Costs to Community Water Systems

    This section examines the regulatory costs that will be incurred by 
individual CWSs at the various radon levels analyzed. Systems above the 
target radon level will incur monitoring costs and treatment costs. 
Systems below the target radon level will incur only monitoring costs.
    The number of CWSs exceeding the applicable radon level increases 
considerably with each decrease in the radon level analyzed as shown 
Table 6-8. The table also shows that the vast majority (90 percent or 
more) of affected systems, regardless of radon level, are very, very 
small (serving 25-500 people) or very small (serving 501-3,300 people).

               Table 6-8.--Number of Community Ground Water Systems Exceeding Various Radon Levels
----------------------------------------------------------------------------------------------------------------
                                     VVSVS
   Exposure level (pCi/l)    --------------------  VS (501-     S (3,301-       M (10,000-        L       Total
                              (25-100)    (101-     3,000)       10,000)         100,000)      (>100K)
------------------------------------------500)------------------------------------------------------------------
4000 \1\....................       364       759         60             5               1            0     1,190
2000........................       949      1448        205            19               8            0     2,630
1000........................      2149      2613        668            75              44            2     5,552
700.........................      3090      3459      1,153           151              94            5     7,951
500.........................      4201      4434      1,796           287             177            9    10,904
300.........................      6302      6233      3,059           657             387           19    16,657
100.........................    10,922    10,349      6,077         1,707             995           48    30,098
----------------------------------------------------------------------------------------------------------------
\1\ 4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).
 
Source: (USEPA 19989L).

    For CWSs that have radon in excess of a given level within each 
size category, the average cost per system to reach the target level 
varies little as the radon levels decrease. This is shown in Table 6-9, 
which presents the average annualized cost per public and private CWS 
by system size category. This pattern is due in large part to the 
limited number of treatment options assumed to be available to systems 
that may (in aggregate) be encountering a relatively wide range of 
radon levels. In some cases (e.g., for very very small systems), the 
average cost per system for a given

[[Page 9589]]

system size increases as the radon level decreases. In other cases, the 
average cost per system remains virtually constant as the radon level 
decreases. These inconsistent patterns are due to two competing 
effects: (1) The average cost will tend to increase because some 
systems must select a more costly treatment option; yet (2) the average 
cost will also tend to decrease with the inclusion of previously 
unaffected systems (those with lower radon levels) that are most likely 
to use lower-cost treatments. The cases where average costs decrease 
with decreasing radon levels are due to the latter effect.
    These results show that changing the radon level affects the number 
of CWSs that must treat for radon, but generally does not significantly 
alter the cost per system for those systems above the target level. 
Moreover, while large systems bear the greatest burden in terms of cost 
per system, there are relatively few large systems with radon levels 
above the exposure scenarios analyzed. The cost per system for CWSs 
with a radon concentration below a target radon level will be the same 
because monitoring costs are dependent on system size and not on 
concentration. Monitoring costs range from less than $250 for the very 
very small systems to almost $2,000 for large systems, again due to the 
larger number of sites requiring monitoring.

BILLING CODE 6560-50-P
[GRAPHIC] [TIFF OMITTED] TN26FE99.001

BILLING CODE 6560-50-C

6.6  Costs and Impacts to Households

    This section reports incremental household costs and impacts 
associated with each radon level, assuming that costs incurred by 
systems above the target radon levels are passed on to the systems' 
customers (i.e., households). Costs per household reflects only 
monitoring and treatment costs to CWSs above the target level. In 
addition, households served by CWSs falling under the target radon 
level also will incur monitoring costs, but no treatment costs. Costs 
for these CWSs are relatively low, however, and are not evaluated at 
the household level. As with per system costs, the results are 
presented separately for public and for private CWSs. This is important 
in considering impacts on households not only because the costs per 
system are different for public versus private systems, but also 
because the smallest private systems tend to serve fewer households 
than do the smallest public systems. Therefore, the average household 
served by a private system must bear a greater percentage of the CWS's 
cost than does the average household served by a public CWS. This is 
particularly important where capital costs make up a large portion of 
total radon mitigation costs.
    The annual cost per household is presented in Table 6-10 for 
households served by public and private CWSs. As expected, costs per 
household increase as system size decreases. Costs per household is 
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 CWS's 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

[[Page 9590]]

50 percent) to achieve the target radon level.
    Another significant finding regarding annual cost per household is 
that, like the per-system costs, household costs (which are a function 
of per system costs) are relatively constant across different radon 
levels within each system size category. For example, there is less 
than $1 dollar per year variation in cost per household, regardless of 
the radon level being considered for households served by large public 
or private systems (between $6 and $7 per year), by medium public or 
private systems (between $10 and $11 per year, and by small public or 
private systems (between $19 and $20 per year). Similarly, for very 
small systems, the costs per household is consistently about $34 per 
year for public systems and consistently about $40 per year for private 
systems, varying little across radon level. Only for very very small 
systems is there a modest variation in household costs. The range for 
per household costs for public systems serving 25-500 people is $87 per 
year (at 4000 pCi/l) to $135 per year (at 100 pCi/l). The corresponding 
range for private systems is $139 to $238 per year. For households 
served by the smallest public system (25-100 people), the range of cost 
per household ranges from $292 per year at 4000 pCi/l to $398 per year 
at 100 pCi/l. For private systems, the range is $364 to $489 per year, 
respectively. Costs per household for very very small systems differ 
more than do household costs for other system size categories because 
very very small systems serve only between 25 and 500 people and, 
consequently, serve fewer households. Therefore, even though per system 
costs show little difference for any system size category, all system 
size categories (other than for very very small systems) spread the 
small difference out among many more households such that the 
difference is indistinguishable.

BILLING CODE 6560-50-P
[GRAPHIC] [TIFF OMITTED] TN26FE99.002

BILLING CODE 6560-50-C

    To further evaluate the impacts of these household costs on the 
households that must bear them, the costs per household were compared 
to median household income data for households in each system-size 
category. The result of this calculation indicates a household's likely 
share of incremental costs in terms of its household income. The 
analysis considers only households served by CWSs with influent radon 
levels that are above the target radon level. Households served by CWSs 
with lower radon levels may incur incremental costs due to new 
monitoring requirements, but these costs are not significant at the 
household level.
    Results are presented in Table 6-11 for public and private CWSs, 
respectively. For all system sizes but one (very very small private 
systems), household costs as a percentage of median household income 
are less than one percent. Impacts exceed one percent only for 
households served by very very small private systems, which are 
expected to face impacts of just under 1.1 percent. Similar to the cost 
per household results on which they are based, household impacts 
exhibit little variability across radon levels.

BILLING CODE 6560-50-P

[[Page 9591]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.003


BILLING CODE 6560-50-C

6.7  Summary of Costs and Benefits

    Table 6-12 summarizes the central tendency estimates of annual 
monetized benefits and annualized costs of the various regulatory 
alternatives. The central tendency national cost estimates are greater 
than the monetized benefits estimates for all radon levels evaluated, 
although they are within 10 percent at levels of 1000, 700, 500, and 
300 pCi/l. Mitigation costs increase more rapidly than the monetized 
benefits as radon levels decrease. However, it is important to 
recognize that due to the uncertainty in the costs and benefits 
estimates, there is a very broad possible range of potential costs and 
benefits that overlap across all of the radon levels evaluated.

Table 6-12.--Estimated National Annual Costs and Benefits of Reducing Radon Exposures--Central Tendency Estimate
                                                [$Millions, 1997]
----------------------------------------------------------------------------------------------------------------
                                                                                                      Annual
                       Radon level (pCi/l)                          Annualized    Cost per fatal     monetized
                                                                       costs      cancer avoided     benefits
----------------------------------------------------------------------------------------------------------------
4000 \3\........................................................              25            11.3              13
2000............................................................              46             7.1              38
1000............................................................              98             5.9              96
700.............................................................             148             6.0             145
500.............................................................             218             6.0             212
300.............................................................             373             6.4             343
100.............................................................             795             6.9             673
----------------------------------------------------------------------------------------------------------------
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.
3. 4000 pCi/l is equivalent to the AMCL estimated by the NAS based on SDWA requirements of Section 1412(b)(13).

    The total annualized cost per fatal cancer case avoided is $11.3 
million at a radon level of 4,000 pCi/l, drops to around $6.0 million 
for radon levels in the range of 1,000 to 500 pCi/l, and increase again 
back to $6.9 million per life saved at the lowest level of 100 pCi/l.

BILLING CODE 6560-50-P

[[Page 9592]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.004



[[Page 9593]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.005


BILLING CODE 6560-50-C

[[Page 9594]]

6.8  Sensitivities and Uncertainties

6.8.1  Uncertainties in Risk Reduction and Health Benefits Calculations
    The estimates of risk and risk reduction are derived based on 
models which incorporate a number of parameters whose values are both 
uncertain and highly variable. Thus, the estimates of health risks and 
risk reduction are uncertain. In addition, to the extent that age-
specific smoking prevalence rates change, the risk from radon in 
drinking water will change.
    The cost of fatal cancers tend to dominate the monetized benefits 
estimates. Approximately 94 percent of the cancers associated with 
radon exposure and prevented by exposure reduction are fatal cancers of 
the lung and stomach. In addition, the estimated value of statistical 
life ($0.7 to 16.3 million dollars, with a central tendency estimate of 
$5.8 million, 1997$) is much greater than the estimated willingness-to-
pay to avoid non-fatal cancers ($169,000 to $1.05 million, with a 
central tendency estimate of $536,000, 1997$). If the COI measures are 
used, non-fatal cancers account for an even smaller proportion of the 
total monetized costs of cancers, since the medical care and lost-times 
costs for lung and stomach cancer are on the order of $108,000 and 
$114,000, respectively (1997$).
    Unless the VSL is assumed to be near the lower end of its range, 
the assumptions made regarding the monetary value of non-fatal cancers 
are not a major source of uncertainty in the estimates of total 
monetary benefits. For most reasonable combinations of values, the VSL 
is the major contributor to the overall uncertainty in monetized values 
of health benefits. As shown in Table 
6-2, the upper and lower estimates of the monetary benefits for a given 
radon level vary by a factor of approximately 23, corresponding to the 
ratios of the lower- and upper-bound estimates of the VSL.
6.8.2  Uncertainty in Cost and Impact Calculations
    The results of the cost and impact analysis are subject to a 
variety of qualifications. As discussed in Section 5, the analysis is 
subject to a variety of uncertainties in the models and assumptions 
made in developing cost estimates. One important assumption is that for 
all CWSs for which the estimated average radon level exceeds a given 
level, treatment will be necessary at all sites. This is a very 
important assumption, because if systems in reality have only a portion 
of sites above the target level, then mitigation costs could be much 
lower. EPA is currently evaluating intra-system variability in radon 
levels, and will address this issue in more detail in the proposal.
    In addition, CWSs are assumed to select from only a relatively 
small number of treatment methods, and to do so in known, constant, 
proportions. In actuality, systems could select technologies that best 
fit their needs and optimize operating conditions to reduce costs. The 
analysis also relies on various cost-related input data that are both 
uncertain and variable. Some of these variables are entered as 
constants, others as deterministic functions. For example: treatment 
technology cost functions are based on EPA cost curves derived for 
generic systems; households are assumed to use a uniform quantity of 
83,000 gallons/year of drinking water, regardless of geographical 
location, system size, or other factors; MMM program costs are assumed 
to cost $700,000 per fatal cancer case avoided, regardless of the 
specific types or efficiencies of activities undertaken by the 
mitigation programs. One factor that may contribute significantly to 
the overall uncertainty in cost estimates is the set of the nonlinear 
equations (Appendix C) used to convert population served data to 
estimates of average and design flow rates for ground water systems. 
Relatively small errors in the specification of this model could result 
in disproportionately large impacts on the cost estimates. Similarly, 
the cost curves for some of the technologies are highly nonlinear 
function of flow, adding another level of uncertainty to the cost 
estimates.
    Because of the complexity of the various cost models, EPA has not 
conducted a detailed analysis of the uncertainty associated with the 
various models and parameter values. Limited uncertainty analyses have 
been performed, however, to estimate the impact of a few major 
assumptions and models on the overall estimates of mitigation costs. 
First, EPA has analyzed the impacts of errors of plus or minus 50 
percent in the cost curves for the various radon treatment 
technologies. The results of this analysis are shown in Figure 6-1. 
Since water mitigation costs make up the bulk of the total costs of 
meeting radon levels in the absence of MMM programs, the effect of 
these changes is generally to increase or decrease the costs of 
achieving the various levels by slightly less than 50 percent. It can 
be seen from these results that the assumptions regarding costs can 
affect the relationship between costs and monetized benefits. A 
relatively small systematic change in water mitigation costs could 
result in benefit estimates that either exceed, or are less than, a 
wide range of radon levels.
    In addition to assuming across-the board changes in radon 
mitigation costs, EPA also examined the extreme situation in which none 
of the water systems would adopt GAC treatment. Since the GAC 
technologies are the most expensive treatments evaluated, the costs of 
meeting the various radon levels are reduced if GAC is eliminated and 
systems are assumed to employ aeration instead (Figure 6-1). Since, 
however, so few systems are assumed to elect GAC in the first place 
(five percent or less of the smallest systems) the cost decrease of 
eliminating GAC is quite small.

BILLING CODE 6560-50-P

[[Page 9595]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.006


BILLING CODE 6560-50-C

7. Implementation Scenarios--Multimedia Mitigation Programs Option

    This Section presents a preliminary analysis of the likely costs 
and benefits under two different implementation scenarios in which 
States choose to develop and implement multimedia mitigation (MMM) 
programs to comply with the radon NPDWR.

7.1  Multimedia Mitigation Programs

    The SDWA, as amended, provides for development of an Alternative 
Maximum Contaminant Level (AMCL), which public water systems may comply 
with if their State has an EPA approved MMM program to reduce radon in 
indoor air. The idea behind the AMCL and MMM option is to reduce radon 
health risks by addressing the larger source of exposure (air levels in 
homes) compared to drinking water. If a State chooses to employ a MMM 
program to reduce radon risk, it would implement a State program to 
reduce indoor air levels and require public water systems to control 
water radon levels to the AMCL, which is anticipated to be set at 4000 
pCi/l based on NAS's re-evaluation of the radon water to air transfer 
factor. If a State does not choose a MMM program option, a public water 
system may propose a MMM program for EPA approval.
    The Agency is currently developing guidelines for MMM programs, 
which will be published for public comment along with the proposed 
NPDWR for radon in August 1999. For the purpose of this analysis, the 
MMM implementation scenarios are assumed to generate the same degree of 
risk reduction as achieved by mitigating water alone. For example, a 
MMM scenario which includes the AMCL of 4,000 pCi/l and a target water 
level of 100 pCi/l is assumed to generate the same degree of risk 
reduction as the 100 pCi/l level alone. Thus, the HRRCA estimates the 
health risk reduction benefits of MMM implementation options to be the 
same as the benefit that would be achieved reducing radon in drinking 
water supplies alone.

7.2  Implementation Scenarios Evaluated

    EPA has evaluated the annual costs and benefits of two MMM 
implementation assuming (1) all States (and all water systems) would 
adopt MMM programs and comply with the AMCL, and (2) half of the States 
(and half of the water systems) adopt the MMM/AMCL option. These 
scenarios were analyzed in the absence of specific data on States' 
intentions to develop MMM programs. The two scenarios, along with the 
case where the MMM option is not selected by any States or water 
systems (presented in Section 6), span the range of participation in 
MMM programs that might occur when a radon NPDWR is implemented. At 
this point, however, it is not possible to estimate the actual degree 
of State participation. The economic impacts of the MMM programs at the 
system or household level have not been calculated, because there is no 
information at present as to how these programs would be funded or upon 
who the costs would fall.
    The presentation of costs and benefits is based on analysis of 
radon levels of 100, 300, 500, 700, 1,000, 2,000, and 4,000 pCi/l in 
public domestic water supplies, supplemented by States (50 or 100 
percent participation) implementing MMM programs and complying with an 
AMCL of 4,000 pCi/l.
    For the scenario evaluated in which one-half of the States 
(estimated to include 50 percent of all CWSs) were assumed to implement 
a MMM program and comply with an AMCL of 4000 pCi/l option, while the 
other half mitigated

[[Page 9596]]

radon in water to the target radon levels without MMM programs. In the 
other scenario, all of the States (and 100 percent of the CWSs) were 
assumed to adopt MMM programs and comply with the AMCL.

7.3  Multimedia Mitigation Cost and Benefit Assumptions

    For the HRRCA, a simplified approach to estimating the costs of 
mitigating indoor air radon risks was used. Based on analyses conducted 
by EPA (US EPA 1992B, 1994C) a point estimate of the average cost per 
life saved of the current national voluntary radon mitigation program 
was used as the basis for the cost estimate of risk reduction for the 
MMM option. In the previous analysis, the Agency estimated that the 
average cost per fatal lung cancer avoided from testing all existing 
homes in the United States and mitigating all those homes at or above 
EPA's voluntary action level of 4 pCi/l is approximately $700,000 (US 
EPA 1992B). This value was originally estimated by EPA in 1991. The 
same nominal value is used in the HRRCA based on to anecdotal evidence 
from EPA's Office of Radiation and Indoor Air that 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. This dollar amount reflects that real testing and mitigation 
costs have decreased, while nominal costs have remained relatively 
constant. The estimated cost per fatal cancer case avoided by building 
new homes radon-resistant is far lower (Marcinowski 1993). For the 
purposes of this analysis, only the cost per fatal cancer case avoided 
from mitigation of existing homes is used.
    To estimate the national cost of the MMM program's air mitigation 
component, MMM costs were estimated by multiplying the cost per fatal 
cancer case avoided by the number of fatal cases avoided in going from 
a water radon level equal to the AMCL (4,000 pCi/l) to a water level 
equal to various radon levels analyzed in the HRRCA. The number of 
fatal cancer cases avoided was estimated using the risk reduction model 
described in Section 3.

7.4  Annual Costs and Benefits of Multimedia Mitigation Program 
Implementation

    The total annual cost of the radon levels analyzed varies 
significantly depending on assumptions regarding the number of States 
implementing MMM programs. This variation can be seen in Tables 7-1 and 
7-2. Under an assumption that 50 percent of States choose to implement 
MMM programs, the cost of the rule varies from about $38 million per 
year to achieve a radon level in water of 2,000 pCi/l to about $450 
million per year to achieve an level of 100 pCi/l. Assuming that 100 
percent of States implement MMM programs, the cost of the rule varies 
from about $29 million per year to achieve an radon level of 2,000 pCi/
l to about $106 million per year to achieve an level of 100 pCi/l.
    The monetized benefits of both MMM implementation scenarios exceed 
the estimated mitigation costs across all radon levels. When the 50 
percent MMM participation scenario is evaluated, the mitigation costs 
at 2,000 pCi/l are just less than the estimated benefits ($38 million 
versus $39.6 million, respectively). In the case of 100 percent 
multimedia participation, mitigation costs begin at about 65 percent of 
the benefits at a radon level of 2,000 pCi/l, and decrease rapidly so 
that at 100 pCi/l the monetized benefits of radon reduction exceed the 
mitigation costs by almost 7-fold.
    Assuming 50 percent MMM participation, the total cost per fatal 
cancer case avoided is $5.8 million at a radon level of 2,000 pCi/l, 
dropping to around $3.7 million at a level of 500 pCi/l, and increasing 
slightly to about $3.9 at 100, pCi/l (Table 7-1). As expected, the cost 
per fatal cancer case avoided is lowest for the 100 percent MMM 
participation option, ranging from from $4.5 at a radon level of 2,000 
pCi/l to about $900,000 at a level of 100 pCi/l.
    For the 50 percent MMM participation, the incremental cost per 
fatal cancer case avoided decreases from 2000 pCi/l to 500 pCi/l ($8.7 
million to $3.4 million, respectively), then increases to $4.1 million 
at 100 pCi/l. In the case of the 100 percent MMM participation, the 
incremental cost per life saved starts at about $4.3 million for the 
maximum target levels of 2,000 pCi/l, and then drops sharply to about 
700,000 per life saved for the other radon.

 Table 7-1.--Central Tendency Estimates of Annualized Costs and Benefits of Reducing Radon Exposures With 50% of
                                      States Selecting the MMM/AMCL Option
                                                [$million, 1997]
----------------------------------------------------------------------------------------------------------------
                                     Water mitigation component             Multimedia mitigation component
                             -----------------------------------------------------------------------------------
                                                             Cost per                                   Cost per
     Radon level (pCi/l)                            Fatal      fatal                          Fatal      fatal
                               Annual    Annual     cancer    cancer     Annual    Annual     cancer     cancer
                                costs   benefits    cases      case      costs    benefits    cases       case
                                 \2\               avoided    avoided                        avoided    avoided
----------------------------------------------------------------------------------------------------------------
Baseline....................         0         0        0    ........        0           0        0         0
4000........................        25        13        2.2      11.3        0           0        0         0
2000........................        35        25        4.3       8.2        2.3        13        2.2       1.1
1000........................        61        54        9.0       6.6        5.8        42        7.1       0.81
700.........................        86        78       13         6.4        8.6        66       11         0.77
500.........................       121       112       19         6.3       12.7        99       17         0.74
300.........................       199       177       30         6.6       20         164       28         0.73
100.........................       410       341       58         7.0       40         328       56         0.71
----------------------------------------------------------------------------------------------------------------
\1\ Equivalent to the cost of complying with an AMCL of 4000 pCi/l.


[[Page 9597]]


  Table 7-2.--Central Tendency Estimates of Annualized Costs and Benefits of Reducing Radon Exposures With 100% of States Selecting the MMM/AMCL Option
                                                                    [$million, 1997]
--------------------------------------------------------------------------------------------------------------------------------------------------------
                                                                          Water mitigation component               Multimedia mitigation component
                                                                   -------------------------------------------------------------------------------------
                                                                                                  Cost per
                        Radon level (pCi/l)                                               Fatal     fatal                           Fatal      Cost per
                                                                     Annual    Annual    cancer    cancer     Annual     Annual     cancer      fatal
                                                                    costs\1\  benefits    cases     case      costs     benefits    cases    cancer case
                                                                                         avoided   avoided                         avoided     avoided
--------------------------------------------------------------------------------------------------------------------------------------------------------
Baseline..........................................................         0         0       0.0  ........        0.0        0.0        0.0        0.0
4000..............................................................        25        13       2.2      11.3        0.0        0.0        0.0        0.0
2000..............................................................        25        13       2.2      11.3        4.6       25          4.4        1.1
1000..............................................................        25        13       2.2      11.3       12         83         14          0.81
700...............................................................        25        13       2.2      11.3       17        131         23          0.77
500...............................................................        25        13       2.2      11.3       25        198         34          0.74
    300...........................................................        25        13       2.2      11.3       41        328         56          0.73
100...............................................................        25        13       2.2      11.3       80        654        112          0.71
--------------------------------------------------------------------------------------------------------------------------------------------------------
\1\ Equivalent to the cost of complying with an AMCL of 4000 pCi/l.

7.6  Sensitivities and Uncertainties

    EPA conducted a sensitivity analysis associated with potential 
uncertainty in the cost-effectiveness of MMM programs. Since the value 
used is a point estimate ($700,000 per life saved), and since the 
ability to employ MMM programs results in substantial decreases in 
estimated costs, it might be expected that changes in the cost-
effectiveness value would affect the cost estimates for these options 
substantially. Figure 7-1 summarizes the impact of different estimates 
of the cost of MMM programs on the total cost of radon mitigation. 
Costs are graphed for the 50 percent and 100 percent participation 
options for radon level. Costs were estimated for a high-end case 
(assuming a MMM cost 50 percent above the central tendency value), a 
low-end case (50 percent below the central tendency), and for a central 
tendency case that assumes the current $700,000 per life saved as the 
MMM cost.
    The relative impacts of changing MMM costs on the total costs of 
reducing radon exposure can also be seen in Figure 7-1. The figure 
illustrates that the central tendency estimate of monetized benefits is 
e well above the estimated costs for all ranges except for the high-end 
estimate of the 50 percent MMM participation scenario. This is due to 
the greater impact of water mitigation costs relative to the MMM cost 
component to total costs compared to the 100 MMM scenario, where the 
MMM component contributes the largest share to total costs.

BILLING CODE 6560-50-P

[[Page 9598]]

[GRAPHIC] [TIFF OMITTED] TN26FE99.007


BILLING CODE 6560-50-C

References

Davis, RMS and JE Watson Jr. ``The Influence of Radium Concentration 
in Surrounding Rock on Radon Concentration in Ground Water,'' 
University of North Carolina, Chapel Hill: March 13, 1989.
Longtin, J.P. ``Occurrence of Radon, Radium, and Uranium in 
Groundwater.'' Journal of the American Water Works Association. 
July, 1987.
Marcinowski, F. and S. Napolitano. ``Reducing the Risks from 
Radon.'' Air and Waste, Vol. 43, 955-962, 1993.
NAS. 1998A. ``Health Effects of Exposure to Radon--BEIR VI (Pre-
Publication Copy),'' National Academy Press: Washington, DC.
NAS. 1998B. ``Risk Assessment of Radon in Drinking Water (Pre-
Publication Copy),'' National Academy Press: Washington, DC, 
September 15.
US EPA. 1999A. ``Point Estimate of Radon Unit Risks,'' Memo to Mike 
Osinski from Nancy Chiu, Office of Science and Technology, January 
22.
US EPA. 1999B. ``Radon in Drinking Water Health Risk Reduction and 
Cost Analysis: Appendices,'' Office of Ground Water and Drinking 
Water, February.
US EPA. 1998A. ``Cost Evaluation of Small System Compliance Options: 
Point-of-Use and Point-of-Entry Treatment Units,'' Office of Ground 
Water and Drinking Water.
US EPA. 1998B. ``Cost of Lung Cancer (Draft),'' Office of Ground 
Water and Drinking Water, October.
US EPA. 1998C. ``Cost of Stomach Cancer (Draft),'' Office of Ground 
Water and Drinking Water, October.
US EPA. 1998D. ``Evaluation of Full-Scale Treatment Technologies at 
Small Drinking Water Systems: Summary of Available Cost and 
Performance Data,'' Office of Ground Water and Drinking Water, 
December.
US EPA. 1998E. ``Guidelines for Preparing Economic Analyses--Review 
Draft,'' Office of Policy, November.
US EPA. 1998F. ``Health Risks from Low-Level Environmental Exposure 
to Radionuclides,'' Office of Radiation and Indoor Air, Federal 
Guidance Report No. 13, Part I--Interim Version, EPA 401/R-97-014.
US EPA. 1998G. ``Model Systems Report (Draft),'' Office of Ground 
Water and Drinking Water, March.
US EPA. 1998H. ``National-Level Affordability Criteria Under the 
1996 Amendments to the Safe Drinking Water Act,'' Office of Ground 
Water and Drinking Water, August 19.
US EPA. 1998I. ``National Primary Drinking Water Regulations: 
Disinfectants and Disinfection Byproducts; Final Rule,'' 63 FR No. 
241, 69390-69476, December 16.
US EPA. 1998J. ``Potential Benefits of the Ground Water Rule--Draft 
Final Report,'' Office of Ground Water and Drinking Water, February.
US EPA. 1998K. ``Radon Cost Estimate,'' Memo to Bill Labiosa from H. 
McCarty, Office of Ground Water and Drinking Water, December 4.
US EPA. 1998L. ``Re-Evaluation of Radon Occurrence in Ground Water 
Supplies in the United States--External Review Draft,'' Office of 
Ground Water and Drinking Water, September 30.
US EPA. 1998M. ``Regulatory Impact Analysis of the Stage I 
Disinfectants/Disinfection By-Products Rule,'' Office of Ground 
Water and Drinking Water, November 12.
US EPA. 1998N. Safe Drinking Water Suite Model. Inputs from Version 
3.4 of the Cost Library and Version 4.0 of the What If Module.
US EPA. 1998O. ``Technologies and Costs for the Removal of Radon 
From Drinking Water,'' Office of Ground Water and Drinking Water, 
September.

[[Page 9599]]

US EPA 1997A. ``Community Water System Survey. Volume II: Detailed 
Survey Result Table and Methodology Report.'' Office of Ground Water 
and Drinking Water. EPA 815-R-97-001b, January.
US EPA. 1997B. ``Withdrawal of the Proposed NPDWR for Radon-222,'' 
62 FR No. 151, 42221-42222, August 6.
US EPA. 1995. ``Uncertainty Analysis of Risks Associated with Radon 
Exposures in Drinking Water,'' Office of Science and Technology, 
Office of Policy, March.
US EPA. 1994A. ``A Citizen's Guide to Radon (Second Edition): The 
Guide To Protecting Yourself and Your Family from Radon,'' Office of 
Radiation and Indoor Air. EPA-402-K92-001, September.
US EPA. 1994B. ``Final Draft for the Drinking Water Criteria 
Document on Radon,'' Office of Water, September 30.
US EPA. 1994C. ``Report to the United States Congress on Radon in 
Drinking Water, Multimedia Risk and Cost Assessment of Radon,'' 
Office of Water. EPA-811-R-94-001.
US EPA. 1993A. ``EPA's Map of Radon Zones: National Summary,'' 
Office of Air and Radiation, EPA 402-R-93-071, September.
US EPA. 1993B. ``Uncertainty Analysis of Risk Associated with 
Exposures to Radon in Drinking Water,'' Office of Science and 
Technology, April 30.
US EPA. 1992A. ``National Residential Radon Survey: Summary 
Report,'' Office of Air and Radiation, EPA 402-R-92-011, October.
US EPA. 1992B. ``Technical Support Document for the 1992 Citizen's 
Guide to Radon,'' Office of Air and Radiation. EPA 400-R-92-011.
US EPA. 1991. ``National Primary Drinking Water Regulations: 
Radionuclides: Notice of Proposed Rulemaking,'' 56 FR No. 138, 
33050-33127, July 18.
US EPA. 1989. ``Analysis of Potential Radon Emissions from Water 
Treatment Plants Using the MINEDOSE Code,'' Memo to Greg Helms from 
Marc Parotta, Office of Ground Water and Drinking Water.
US EPA. 1988. ``Preliminary Risk Assessment for Radon Emissions from 
Drinking Water Treatment Facilities,'' Memo to Stephen Clark from 
Warren Peters and Chris Nelson, Office of Water.
Viscusi, WK, WA Magat, and J. Huber. ``Pricing Environmental Health 
Risks: Survey Assessments of Risk-Risk and Risk-Dollar Trade-Offs 
for Chronic Bronchitis.'' Journal of Environmental Economics and 
Management, 21:32-51, 1991.

[FR Doc. 99-4416 Filed 2-25-99; 3:08 pm]
BILLING CODE 6560-50-P