[Federal Register Volume 65, Number 204 (Friday, October 20, 2000)]
[Proposed Rules]
[Pages 63027-63035]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 00-27034]


=======================================================================
-----------------------------------------------------------------------

ENVIRONMENTAL PROTECTION AGENCY

40 CFR Parts 141 and 142

[WH-FRL-6888-8]
RIN 2040-AB75


National Primary Drinking Water Regulations; Arsenic and 
Clarifications to Compliance and New Source Contaminants Monitoring

AGENCY: Environmental Protection Agency (EPA).

ACTION: Notice of data availability.

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

SUMMARY: The Environmental Protection Agency (EPA) proposed regulations 
for arsenic in drinking water on June 22, 2000 (65 FR 38888), and 
comments on that action were due on September 20, 2000. Since that 
time, EPA has received new risk information which the Agency is 
considering during the development of the final regulation. This 
document summarizes the new risk information received and analyzed by 
the Agency. In addition, this document makes available the cost curves 
used to develop the costs published in the proposal. This information 
does not change the overall technical approach for the proposal. EPA is 
requesting comments on EPA's

[[Page 63028]]

use of the new risk analysis and development of cost estimates for the 
final rule and any comments on other parts of the proposal which would 
change because of the information provided today.

DATES: Your comments on this document must be submitted to EPA in 
writing and should be postmarked or received November 20, 2000.

ADDRESSES: Send written comments to the W-99-16 NODA Arsenic Comments 
Clerk, Water Docket (MC-4101); U.S. Environmental Protection Agency; 
1200 Pennsylvania Ave., NW., Washington, DC 20460. Comments may be 
hand-delivered to the Water Docket, U.S. Environmental Protection 
Agency; 401 M Street, SW; East Tower Basement, room EB-57; Washington, 
DC 20460; (202) 260-3027 between 9 a.m. and 3:30 p.m. Eastern Time, 
Monday through Friday. Comments may be submitted electronically, marked 
docket number W-99-16 NODA, to [email protected]. Please refer to the 
information under the headings ``Additional Information for 
Commenters'' and ``Availability of Docket'' in SUPPLEMENTARY 
INFORMATION for detailed information about filing and docket review.

FOR FURTHER INFORMATION CONTACT: For technical inquiries about risk and 
benefits discussed in this notice, contact Dr. John B. Bennett, (202) 
260-0446, email: [email protected], and for technical inquiries 
about treatment and cost discussed in this notice, contact Jeff Kempic, 
(202) 260-9567, email: [email protected]. For general information 
about this notice, contact Irene Dooley, (202) 260-9531, email: 
[email protected].

SUPPLEMENTARY INFORMATION:

Regulated Entities

    A public water system, as defined in 40 CFR 141.2, provides water 
to the public for human consumption through pipes or other constructed 
conveyances, if such system has ``at least fifteen service connections 
or regularly serves an average of at least twenty-five individuals 
daily at least 60 days out of the year.'' A public water system is 
either a community water system (CWS) or a non-community water system 
(NCWS). A community water system, as defined in Sec. 141.2, is ``a 
public water system which serves at least fifteen service connections 
used by year-round residents or regularly serves at least twenty-five 
year-round residents.'' The definition in Sec. 141.2 for a non-
transient, non-community water system [NTNCWS] is ``a public water 
system that is not a [CWS] and that regularly serves at least 25 of the 
same persons over 6 months per year.'' EPA has an inventory totaling 
over 54,000 community water systems and approximately 20,000 non-
transient, non-community water systems nationwide. Entities potentially 
regulated by this action are community water systems and non-transient, 
non-community water systems. The following table provides examples of 
the regulated entities under this rule.

                       Table of Regulated Entities
------------------------------------------------------------------------
                                    Examples of potentially regulated
           Category                             entities
------------------------------------------------------------------------
 Industry....................   Privately owned/operated community water
                                supply systems using ground water or
                                mixed ground water and surface water.
 State, Tribal, and Local       State, Tribal, or local government-owned/
 Government.                    operated water supply systems using
                                ground water or mixed ground water and
                                surface water.
 Federal Government..........   Federally owned/operated community water
                                supply systems using ground water or
                                mixed ground water and surface water.
------------------------------------------------------------------------

    The table is not intended to be exhaustive, but rather provides a 
guide for readers regarding entities likely to be regulated by this 
action. This table lists the types of entities that EPA is now aware 
could potentially be regulated by this action. Other types of entities 
not listed in this table could also be regulated. To determine whether 
your facility is regulated by this action, you should carefully examine 
the applicability criteria in Secs. 141.11 and 141.62 of the rule. If 
you have any questions regarding the applicability of this action to a 
particular entity, consult the general information person listed in the 
FOR FURTHER INFORMATION CONTACT section.

Additional Information for Commenters

    Please submit an original and three copies of your comments and 
enclosures (including references) and identify your submission by the 
docket number W-99-16 NODA. To ensure that EPA can read, understand, 
and therefore properly respond to comments, the Agency would prefer 
that comments cite, where possible, the paragraph(s) or sections in the 
document or supporting documents to which each comment refers. 
Commenters should use a separate paragraph for each issue discussed. If 
you are submitting your comments electronically and mailing hard 
copies, please indicate on your electronic submission that hard copies 
are being sent separately. Electronic comments must be submitted as a 
WordPerfect 5.1, WP6.1 or WP8 file or as an ASCII file avoiding the use 
of special characters. Comments and data will also be accepted on disks 
in WP 5.1, WP6.1 or WP8, or ASCII file format. Electronic comments on 
this document may be filed online at many Federal Depository Libraries. 
Commenters who want EPA to acknowledge receipt of their comments should 
include a self-addressed, stamped envelope. No facsimiles (faxes) will 
be accepted.

Availability of Docket

    The docket for this document has been established under number W-
99-16-II, and includes supporting documentation as well as printed, 
paper versions of electronic comments. The docket is available for 
inspection from 9 a.m. to 4 p.m., Monday through Friday, excluding 
legal holidays, at the Water Docket; EB 57; in the East Tower basement 
of U.S. EPA; 401 M Street, SW; Washington, DC. For access to docket 
materials, please call (202) 260-3027 to schedule an appointment.

Abbreviations Used

%--percent
AIC--Akaike information criterion
CWS--community water system
EB--East Tower Basement
ED01--Effective dose which results in 1% excess lifetime 
risk
EPA--U.S. Environmental Protection Agency
et al.--et alibi, Latin for ``and others''
FR--Federal Register
i.e.--id est, Latin for ``that is''
kg--kilograms, 2.2 pounds
L--Liter, also referred to as lower case ``l'' in older citations
LED01--a 95% lower confidence limit for ED01
MDBP--microbial/disinfection by-product
MCL--maximum contaminant level

[[Page 63029]]

mg--milligrams--one thousandth of a gram, 1 milligram = 1,000 
micrograms
microgram (g)--One-millionth of gram (3.5  x  
10-\8\ oz., 0.000000035 oz.)
g/L--micrograms per liter
MOE01--margin of exposure, ratio of ED01 to MCL
NAS--National Academy of Sciences
NODA--Notice of Data Availability
NRC--National Research Council, operating agency of NAS
NRC--National Research Council, the operating arm of NAS
O&M--operation and maintenance
ppb--Parts per billion. Also, g/L or micrograms per liter
RIA--Regulatory Impact Analysis
U.S.--United States
VSL--Value of a statistical life
WTP--Willingness to pay

How Does This Document Relate to the June 22, 2000 Proposal?

    In the Thursday, June 22, 2000, Federal Register the U.S. 
Environmental Protection Agency (EPA) proposed regulations for arsenic 
and clarifications to compliance and new source contaminants monitoring 
(65 FR 38888). This document applies only to the arsenic part of the 
proposal. Specifically, EPA noted that ``Further work on the risk 
assessment will also be done before the final rule is issued to analyze 
the risks of internal cancers (65 FR 38888 at 38899).'' This document 
discusses new risk information and EPA's subsequent risk analysis.
    On page 39835 of the June 22, 2000, arsenic proposal, EPA noted 
that the unit cost curves are in the November 1999 ``Technologies and 
Costs for the Removal of Arsenic from Drinking Water.'' It has come to 
EPA's attention that the cost curves used to develop the costs that are 
included in the proposal and supporting Regulatory Impact Analysis 
(RIA) are in an earlier version of this document dated April 1999. This 
document announces the availability of the April 1999 version, with 
curves that more accurately reflect the analysis in the preamble and 
the RIA. The overall approach to cost estimation in the proposed rule 
and the proposed Maximum Contaminant Level (MCL) remain unchanged.

What New Risk Data Has EPA Analyzed?

    In the proposal we calculated bladder cancer benefits and risks 
using the bladder cancer risk analysis from the 1999 National Research 
Council (NRC) report, Arsenic in Drinking Water. We also estimated lung 
cancer benefits in a ``What If'' analysis based on a qualitative 
statement about lung cancer deaths from the 1999 NRC report. At that 
time we noted that a peer-reviewed lung cancer risk study would 
probably become available before the final rule came out (65 FR 38888 
at 38944). This Spring, we received a copy of a peer-reviewed article 
by Morales et al. (2000). This article presented additional analyses of 
bladder cancer risks as well as estimates of lung and liver cancer 
risks for the same Taiwanese population analyzed in the NRC report. 
This document makes available for public comment the Morales et al. 
(2000) information and the Agency's analysis of the bladder and lung 
cancer risks from that paper.

What Is in the Article by Morales et al.?

    The article ``Risk of Internal Cancer from Arsenic in Drinking 
Water'' (Morales et al., 2000) presents an assessment of the magnitude 
of risk for cancers of the bladder, liver and lung from exposure to 
arsenic in water, based on data from 42 villages in an arseniasis 
endemic region of Taiwan. The authors calculated excess lifetime risk 
estimates using several Poisson regression models and a multistage-
Weibull model. (Excess lifetime risk is the additional probability of 
disease or death due to the given cause over the course of a lifetime.) 
Risk estimates are expressed as ED01, the concentration at 
which 1% additional lifetime risk of death is incurred; 
LED01, a 95% lower confidence limit for ED01; and 
MOE01(50), the ``margin of exposure,'' or the ratio of 
ED01 to the current MCL of 50 g/L. The authors 
found that risk estimates are sensitive to the choice of model, to 
whether a comparison population is used to define the unexposed disease 
mortality rates, and whether the comparison population is all of Taiwan 
or just an unexposed portion of the population in the study area. The 
authors noted that some of the factors that may affect the magnitude of 
risk could not be evaluated quantitatively: The ecological nature of 
the data, the nutritional status of the study population, and the 
dietary intake of arsenic. Despite all these sources of uncertainty, 
however, the analysis suggests that the current standard of 50 
g/L is associated with a substantial increased risk of cancer 
and thus is not suffciently protective of public health. (The authors 
state that ``the risk associated with a concentration of 50 g/
L is approximately 1 in 300, based on linear extrapolation from the 
point of departure. * * * This is an extremely high value.'')
    The Morales et al. (2000) article uses several statistical models 
to estimate bladder, lung, and liver cancer risk from arsenic exposure. 
It also presents the combined risk of all three cancers. The risk 
assessments are based on a study from Taiwan published by Chen et al. 
(1985), with the data grouped at the village level. These data are also 
used for the bladder cancer risk analysis in the 1999 NRC report. 
Morales et al. (2000) examine issues of dose-response modeling for the 
generalized linear model. The authors identify several Poisson and 
multistage-Weibull models which fit the data about equally well. They 
prefer the Poisson models, in part because the fit of the Weibull 
models is more sensitive to the omission of subsets of individual 
villages. The models are based on mortality data from Taiwan, and model 
results are transferred to the United States (U.S.) without adjustment 
for differences in mortality-to-incidence ratios for the various 
illnesses. The authors adjusted the risk analyses to reflect 
differences in average population weight and in the consumption of 
drinking water between the U.S. and Taiwan (assuming a representative 
person in the U.S. weighs 70 kg and drinks 2 liters of water per day 
vs. a Taiwanese weighing 55 kg and drinking 3.5 liters). Two comparison 
populations, one from all of Taiwan and one from southwestern Taiwan, 
were used in the modeling to estimate background levels of risk.
    The various model results present considerable variability in 
cancer risk estimates for arsenic. The authors propose several reasons 
for the variability, including the large variability of exposure among 
people within each village and use of a comparison population in the 
analysis. The authors also suggest that a variety of factors for which 
data were not available, including the dietary intake of inorganic 
arsenic, could influence or even confound these models. They observe 
that ``* * * this is an ecological study wherein only relatively simple 
exposure and population characteristics could be measured. It will be 
important to consider this and other sources of uncertainty when 
interpreting the results (Morales et al., 2000).'' The authors 
conclude, however, that it seems likely that arsenic is contributing to 
excess cancer mortality in the U.S. based on their evaluation of 
combined risks of bladder, lung, and liver cancer: ``Despite the 
considerable variation in estimated ED01, the results are 
sobering and indicate that current standards are not adequately 
protective against cancer (Morales et al., 2000).''

What Models Did EPA Choose To Use for Additional Analysis?

    Ten risk models were presented in Morales et al. (2000). Following 
Dr. Louise Ryan's presentation to the SAB

[[Page 63030]]

Drinking Water Committee (SAB, 2000), and after additional consultation 
with the primary authors (Morales and Ryan), EPA chose Model 1 with no 
comparison population for further analysis. In Model 1 the dose effect 
is assumed to follow a linear function and the age effect is assumed to 
follow a quadratic function.
    EPA believes, after consultation with the authors, that the models 
in Morales et al. (2000) with a comparison population are less reliable 
than those without a comparison population. With no comparison 
population, the arsenic dose-response curve is estimated only from the 
study population. Models with a comparison population include mortality 
data from a similar population (in this case either all of Taiwan or 
part of southwestern Taiwan), whose exposure is assumed to be zero. 
Most of the models with comparison populations resulted in dose-
response curves that were supralinear (higher than a linear dose-
response) at low doses. The curves were forced down at zero dose 
because the comparison population consists of a large number of people 
with low risk and assumed zero exposure. EPA believes, based on 
discussions with the authors, that these models are less reliable, for 
two reasons. First, there is no basis in data on arsenic's carcinogenic 
mode of action to consider a supralinear curve to be biologically 
plausible. The conclusion of the NRC panel (NRC, 1999) was that the 
mode of action data led one to expect dose responses that would be 
either linear or less than linear at low dose. However, the NRC 
indicated that available data are inconclusive and ``* * * do not meet 
EPA's 1996 stated criteria for departure from the default assumption of 
linearity.'' Second, models which include comparison populations assume 
that the exposure of the comparison population is zero, and that the 
study and comparison populations are the same in all important ways 
except for arsenic exposure. Neither of these comparison populations 
assumptions may be correct: NRC (1999) notes that ``the Taiwanese-wide 
data do not clearly represent a population with zero exposure to 
arsenic in drinking water''; and Morales et al. (2000) agree that 
``[t]here is reason to believe that the urban Taiwanese population is 
not a comparable population for the poor rural population used in this 
study.'' Moreover, because of the large amount of data in the 
comparison populations, the model results are relatively sensitive to 
assumptions about this group. For these reasons, EPA believes that the 
models without comparison populations are more reliable than those with 
them.
    Of the models that did not include a comparison population, EPA 
believes Model 1 fits the data best, based on the Akaike information 
criterion (AIC), a standard criterion of model fit, applied to the 
Poisson models. EPA did not consider the multi-stage Weibull model for 
additional analysis, because of its greater sensitivity to the omission 
of individual villages (Morales et al., 2000) and to the grouping of 
responses by village (NRC, 1999), as occurs in the Taiwanese data.
    The Poisson regression model (Model 1), without a comparison 
population, gave results for lifetime excess risk of bladder cancer for 
males from arsenic ingestion (about 1.3 in a 1000 at an arsenic level 
of 50 g/L) which were approximately the same as those risks 
found by the NRC (approximately 1 in a 1000 at an arsenic level of 50 
g/L). Among females, lifetime excess risk of bladder cancer is 
estimated to be 2.0 in 1000 at 50 g/L. We also considered 
estimates using this model for excess risks for lung and liver cancer 
due to arsenic. The lung cancer risk estimates, which were comparable 
to the bladder cancer risk estimates, were of special interest to the 
Agency, as the NRC report did not provide a statistical analysis of 
these risks.
    However, EPA did not further consider the Taiwan liver cancer 
estimates for U.S. liver cancer risks. Angiosarcoma liver cancer 
(cancer in the liver's blood vessels) has been linked to arsenic 
exposure in Germany (Roth, 1957, as reported in Smith et al., 1992), 
Chile (Zaldivar et al., 1981, as reported in Smith et al., 1992), and 
the U.S. (Falk et al., 1981, as reported in Smith et al., 1992). 
However, most liver cancers in Taiwan were hepatocellular (i.e., liver 
cell) carcinomas linked to hepatitis (Chen et al., 1985 & 1986), rather 
than angiosarcoma cancer, and are extremely rare in the U.S.

How Will the New Data Affect EPA's Risk Analysis?

    This section describes EPA's risk analysis in the June 22, 2000, 
proposed arsenic rulemaking, then extends the analysis to incorporate 
new information from Morales et al. (2000).
    The June 22, 2000, proposed arsenic rulemaking contained an 
analysis of the excess exposed population risks associated with arsenic 
consumption for bladder cancer. This analysis was based on the 1999 
National Research Council (NRC) report, in which the NRC examined risk 
distributions for male bladder cancer in 42 villages in Taiwan. This 
population was exposed to drinking water with arsenic ranging from 10 
to 934 g/L; arsenic exposure estimates were grouped by 
village. To monetize bladder cancer benefits, EPA calculated the number 
of cases potentially avoided, based on the NRC bladder cancer risk 
analyses, for populations exposed to MCL options of 3 g/L, 5 
g/L, 10 g/L, and 20 g/L. The proposal's 
analytic approach included five components. First, EPA used data from 
the recent EPA water consumption study (US EPA, 2000a). Second, we used 
Monte Carlo simulations to develop a distribution of ``relative 
exposure factors,'' which account for individual variations in risk due 
to water consumption and body weight. Third, arsenic occurrence 
estimates (US EPA, 2000c) were used to identify the population exposed 
to levels above 3 g/L. We assumed drinking water exposure 
reflected treatment to 80% of the MCL level, because water systems tend 
to treat below the MCL level in order to provide a margin of safety. 
Fourth, EPA chose four NRC risk distributions (NRC, 1999, from Tables 
10-11 and 10-12) for the analysis, that used Poisson-model derived risk 
estimates, with and without baseline comparison data. Fifth, EPA used 
Monte Carlo simulations to develop estimates of the risks faced by the 
exposed population, using the relative exposure factors, occurrence, 
and the NRC risk distributions. These components of the analysis are 
described in the proposed rulemaking (US EPA, 2000d, section X.A). EPA 
also monetized the potential benefits of avoided lung cancer, using a 
``What If'' analysis based on statements in the NRC report.
    Table 1 shows the mean and 90th percentile bladder cancer incidence 
risks summarized from Tables X-4A, X-4B, X-2A, and X-2B in the June 22, 
2000, arsenic proposed rulemaking (65 FR 38888), after treatment, for 
the U.S. population currently exposed at or above 3 g/L, 5 
g/L, 10 g/L, and 20 g/L. These risk 
distributions are based on bladder cancer mortality data in Taiwan, in 
a section of Taiwan where arsenic concentrations in the water are very 
high by comparison to those in the U.S. It is also an area of low 
incomes (NRC, 1999, pg. 292) and poor diet (NRC, 1999, pg. 295), and 
the availability and quality of medical care is not of high quality, by 
U.S. standards. In its estimate of bladder cancer risk, the Agency 
assumed that within the Taiwanese study area at the time of the study, 
the risk of contracting bladder cancer was relatively close to the risk 
of dying from bladder cancer (that is, that

[[Page 63031]]

the bladder cancer incidence rate was equal to the bladder cancer 
mortality rate). Survival rates for bladder cancer in the U.S. have 
been improving from 1973 to 1996 (i.e., U.S. bladder cancer mortality 
rates decreased overall 24% to 26%). Recent bladder cancer survival 
rates in developing countries range from 23.5% to 66.1%, and are 
currently 45% for bladder cancer in Taiwan, as discussed in the 
proposed rulemaking (65 FR 38888 at 38942). At most, the Agency 
concluded that bladder cancer incidence could be no more than 2 times 
bladder cancer mortality; and that an 80% mortality rate would be 
plausible. The benefits analysis included estimates using an assumed 
mortality rate ranging from 80% to 100%.

    Table 1.--Bladder Cancer Risks From the June 22, 2000 Proposal: Mean (From Tables X-4A and X-4B) and 90th
Percentile (From Tables X-2A and X-2B) Lifetime Incidence Risks,\1\ for U.S. Populations Exposed at or Above MCL
  Options, After Treatment \2\ (Lower Bounds: Low NRC Risk, CWS Water Consumption; Upper Bounds: High NRC Risk,
                                            Total Water Consumption)
----------------------------------------------------------------------------------------------------------------
                                                                             90th percentile  exposed population
          MCL g/L                Mean exposed  population risk                     risk
----------------------------------------------------------------------------------------------------------------
3...................................  2.1-4.5  x  10-\5\                    4-7  x  10-\5\
5...................................  3.6-7.5  x  10-\5\                    6-12  x  10-\5\
10..................................  5.5-11.4  x  10-\5\                   1-2  x  10-\4\
20..................................  6.9-13.9  x  10-\5\                   1.4-2.8  x  10-\4\
----------------------------------------------------------------------------------------------------------------
\1\ Actual risks could be lower, given the various uncertainties discussed, or higher, as these estimates assume
  a 100% mortality rate.
\2\ The risk analysis assumed exposure at 80% of the MCL level, because water systems tend to treat below the
  MCL level in order to provide a margin of safety.

    The Morales et al. (2000) article provided a new analysis of 
bladder cancer risk. Although the data used were the same as used by 
the NRC to analyze bladder cancer risk in their 1999 publication, 
Morales et al. (2000) consider more dose-response models and evaluate 
how well they fit the Taiwanese data. Therefore the Agency decided to 
examine the implications of the new bladder cancer risk assessment from 
Morales et al. (2000), as well as the lung cancer risk assessment. 
Using the same analytical approach as in the arsenic proposed rule 
(with Monte Carlo simulations combining relative exposure factors, 
occurrence estimates, and risk distributions), the Agency recalculated 
the mean bladder cancer risks for U.S. populations, based on the risk 
estimates from Morales et al. (2000), derived from Model 1 with no 
comparison population. The results are shown in Table 2, along with the 
bladder cancer risks remaining, after treatment, for the 90th 
percentile U.S. population.

 Table 2.--Bladder Cancer: Mean and 90th Percentile Lifetime Incidence Risks,\1\ for U.S. Populations Exposed At
   or Above MCL Options, after Treatment \2\ (Morales Risk, Low Water Consumption for Lower Bound, High Water
                                          Consumption for Upper Bound)
----------------------------------------------------------------------------------------------------------------
              MCL g/L                  Mean exposed  population risk   90th percentile  population risk
----------------------------------------------------------------------------------------------------------------
3...........................................  4.9-6.0  x  10-5                  1-1.2  x  10-4
5...........................................  8.4-10.2  x  10-5                 1.8-2.0  x  10-4
10..........................................  1.2-1.47  x  10-4                 2.6-3.1  x  10-4
20..........................................  1.55-1.89  x  10-4                3.5-4.1  x  10-4
----------------------------------------------------------------------------------------------------------------
\1\ Actual risks could be lower, given the various uncertainties discussed, or higher, as these estimates assume
  a 100% mortality rate.
\2\ The risk analysis assumed exposure at 80% of the MCL level, because water systems tend to treat below the
  MCL level in order to provide a margin of safety.

    The Agency also estimated the mean and 90th percentile lung cancer 
risks for U.S. populations, using the same analytical approach and the 
risk estimates from Morales et al. (2000), derived from Model 1 with no 
comparison population. The results are shown in Table 3.

[[Page 63032]]



 Table 3.--Lung Cancer: Mean Lifetime Incidence Risks,\1\ for U.S. Populations Exposed at or Above MCL Options,
   After Treatment \2\ (Morales Risk, Low Water Consumption for Lower Bound, High Water Consumption for Upper
                                                     Bound)
----------------------------------------------------------------------------------------------------------------
          MCL g/L                Mean exposed  population risk        90th percentile  population risk
----------------------------------------------------------------------------------------------------------------
3..................................  4.9-6.1  x  10-5                        1.0-1.2  x  10-4
5..................................  8.2-10.5  x  10-5                       1.7-2.1  x  10-4
10.................................  1.21-1.46  x  10-4                      2.7-3.1  x  10-4
20.................................  1.52-1.87  x  10-4                      3.4-4.3  x  10-4
----------------------------------------------------------------------------------------------------------------
\1\ Actual risks could be lower, given the various uncertainties discussed, or higher, as these estimates assume
  a 100% mortality rate.
\2\ The risk analysis assumed exposure at 80% of the MCL level, because water systems tend to treat below the
  MCL level in order to provide a margin of safety.

    EPA believes, based upon this most recent risk information, that 
the combined risk of excess cases of lung and bladder cancer 
attributable to arsenic in drinking water could be at least twice that 
of bladder cancer alone. However, EPA will need to conduct additional 
analyses of this risk information, together with additional analyses of 
the various uncertainties associated with the underlying data, and of 
comments submitted in response to the proposed rule, to develop its 
best estimate of the overall risk in support of a final rulemaking.

How Did EPA Analyze the Lower Bound of its Risk Estimates?

    The Agency performed a sensitivity analysis of the lower bound risk 
estimates, considering the effect on risk estimates of exposure to 
arsenic through water used in preparing food in Taiwan. The 1988 EPA 
``Special Report on Ingested Inorganic Arsenic'' contained the 
following discussion:

    For the studied population, rice and sweet potatoes were the 
main staple and might account for as much as 80% of food intake per 
meal. For the purpose of discussion we will assume that a man in the 
study population ate one cup of dry rice and two pounds of potatoes 
per day and that the amount of water required to cook the rice and 
potatoes was about 1 L. Under this assumption, the risk calculated 
before is overestimated by about 30% (1 L/ 3.5 L). This calculation 
considers only the water used for cooking; the arsenic content in 
the rice and potatoes that might have been absorbed from soil 
arsenic is not considered because of the lack of information.

The Taiwanese staple foods were dried sweet potatoes and rice (Wu et 
al., 1989). Both the 1988 EPA report and the 1999 NRC report assumed 
that an average Taiwanese male weighed 55 kg and drank 3.5 liters of 
water daily, and that an average Taiwanese female weighed 50 kg and 
drank 2 L of water daily. Using these assumptions, along with an 
assumption that Taiwanese men and women ate one cup of dry rice and two 
pounds of sweet potatoes a day, the Agency re-estimated risks for 
bladder and lung cancer, using one additional liter water consumption 
for food preparation (i.e., the water absorbed by hydration during 
cooking). The food consumed in Taiwan contains more arsenic than in the 
U.S.: on average, about 50 g/day in Taiwan, versus about 10 
g/day in the U.S. (NRC, 1999, pp. 50-51). Thus our analysis 
may still overstate the risk to the U.S. population, when the total 
consumption of inorganic arsenic (from food preparation and drinking 
water) is considered. Results of the EPA analysis considering water 
used in cooking are shown in Table 4, using the NRC bladder cancer 
risk, the Morales et al. (2000) bladder cancer risk, and the Morales et 
al. (2000) lung cancer risk estimates utilized earlier in this 
Document. Table 5 shows the cancer risks remaining, after treatment to 
80% MCL options, for high percentile U.S. populations, providing a 
sensitivity analysis for the lower bound risk taking into account the 
arsenic intake from water used in cooking dried foods.

 Table 4.--Sensitivity Analysis of Mean Lower Bound Incidence Risk Estimates,1, 2 Risks Adjusted for Water Used
                                     in Cooking (CWS Water Consumption Data)
----------------------------------------------------------------------------------------------------------------
        MCL (g/L)               Bladder (NRC)           Bladder  (Morales)          Lung  (Morales)
----------------------------------------------------------------------------------------------------------------
3................................  1.7  x  10-5               3.5  x  10-5              3.6  x  10-5
5................................  2.9  x  10-5               5.7  x  10-5              5.7  x  10-5
10...............................  4.1  x  10-5               8.4  x  10-5              8.4  x  10-5
20...............................  5.1  x  10-5               1.01  x  10-5             1.06  x  10-5
----------------------------------------------------------------------------------------------------------------
\1\ Risks are adjusted under assumption that Taiwanese males and females consume one additional liter of water
  in rehydrating dried rice and sweet potatoes.
\2\ The bladder cancer risks presented in this table provide ``best'' estimates. Actual risks could be lower,
  given the various uncertainties discussed, or higher, as these estimates assume a 100% mortality rate.


[[Page 63033]]


 Table 5.--Sensitivity Analysis of 90th Percentile Lower Bound Incidence Risk Estimates,1, 2 Risks Adjusted for
                               Water Used in Cooking (CWS Water Consumption Data)
----------------------------------------------------------------------------------------------------------------
        MCL (g/L)               Bladder (NRC)           Bladder  (Morales)          Lung  (Morales)
----------------------------------------------------------------------------------------------------------------
3................................  3.5  x  10-5               7.5  x  10-5              7.2  x  10-5
5................................  5.9  x  10-5               1.2  x  10-4              1.2  x  10-4
10...............................  9.0  x  10-5               1.8  x  10-4              1.8  x  10-4
20...............................  1.1  x  10-4               2.3  x  10-4              2.4  x  10-4
----------------------------------------------------------------------------------------------------------------
\1\ Risks are adjusted under assumption that Taiwanese males and females consume one additional liter of water
  in rehydrating dried rice and sweet potatoes.
\2\ The bladder cancer risks presented in this table provide ``best'' estimates. Actual risks could be lower,
  given the various uncertainties discussed, or higher, as these estimates assume a 100% mortality rate.

How Will EPA Evaluate Benefits in the Final Rule?

    The benefits of a regulatory option depend primarily on the number 
of cases of an illness avoided due to the reduction in risk resulting 
from the implementation of the option. For the arsenic proposed rule 
and following established Agency practices, EPA estimated the number of 
cases of bladder cancer avoided using mean exposed population incidence 
risk estimates at various MCL levels (these mean exposed population 
incidence risks are shown in Table 1). We converted lifetime risk 
estimates to annual risk factors, and applied these to the exposed 
population to determine the number of cases avoided (both fatal and 
non-fatal). We adjusted the upper bound bladder cancer number of cases 
estimates by assuming an 80% mortality rate in Taiwan, which is a 
plausible mortality rate for the area of Taiwan during the Chen study. 
The lower bound estimates assumed a 100% mortality rate from bladder 
cancer in Taiwan. For the benefits assessment, EPA used U.S. mortality 
information to divide the number of cases into fatal and non-fatal 
cases avoided. Benefits are assumed to begin to accrue on the effective 
date of the arsenic rule (65 FR 38888 at 38946).
    The avoided cases of fatal bladder cancer are valued by what is 
known as the ``value of a statistical life'' (VSL), currently estimated 
at $6.1 million (in 1999 dollars).\1\ VSL does not refer to the value 
of an identifiable life, but instead to the value of small reductions 
in mortality risks in a population. We used the central tendency 
estimate of $604,000 (1999 dollars) \2\ of the willingness to pay (WTP) 
to avoid a case of chronic bronchitis to monetize the benefits of 
avoiding non-fatal bladder cancers (Viscusi et al., 1991). WTP data for 
avoiding chronic bronchitis has been used before by EPA (the microbial/
disinfection by-product (MDBP) rulemaking) as a surrogate for the WTP 
to avoid non-fatal bladder cancer. EPA summed the monetized benefits 
for fatal and non-fatal bladder cancer cases avoided to obtain total 
monetized benefits for avoided bladder cancer cases (shown in Tables X-
7 and XI-1 of the proposed rule preamble, in 1999 dollars).
---------------------------------------------------------------------------

    \1\ The June 20, 2000, proposal (65 FR 38888) cited the central 
tendency estimate of the VSL as $5.8 million in 1997 $ in the 
preamble text. However, the analyses presented in the proposal's 
tables reflect 1999 $ values, as noted.
    \2\ The June 20, 2000, proposal (65 FR 38888) cited the central 
tendency estimate of the WTP as $536,000 in 1997 $ in the preamble 
text. However, the analyses presented in the proposal's tables 
reflect 1999 $ values, as noted.
---------------------------------------------------------------------------

    In the arsenic proposed rule, EPA also estimated the number of lung 
cancer cases avoided, for the various options considered, using a 
``What If'' analysis, and monetized these cases using the same process 
that was used to monetize the benefits of avoided bladder cancer cases. 
The ``What If'' analysis examined possible benefits from avoided lung 
cancer cases if the number of those cases in the U.S. which were fatal 
in outcome was 2-5 times the number of fatal bladder cancer cases (the 
implicit risk for lung cancer ranged from about half to about twice 
that of the risk for bladder cancer).
    EPA plans to use the benefits evaluation process described in this 
section for the final rule, using the data and analysis of the bladder 
and lung cancer risks described in this document instead of the ``What 
If'' lung cancer analysis included in the proposal. These more 
definitive benefits estimates will be derived from the new risk 
calculations that will accompany the final rule (based upon further 
consideration of additive risk analyses) and other pertinent 
information. Background information on the economic concepts that 
provide the foundation for benefits valuation, and the methods that are 
typically used by economists to monetize the value of risk reductions, 
such as wage-risk, cost of illness, and contingent valuation studies 
are provided in the arsenic RIA.

EPA Benefits Summary and Conclusions

    Morales et al. (2000) assess the risks of lung and bladder cancer 
associated with arsenic consumption in water, based on data from 
Taiwan, using several statistical models. Although the data used were 
the same as used by the NRC (1999). Morales et al. consider more dose-
response models, providing a more exhaustive treatment of model fit. 
They also discuss additional factors, for which data were not 
available, which might influence or confound the analysis. Dose-
response risk estimate for both bladder and lung cancer, derived from 
the best-fitting model, were analyzed further by the Agency. The Agency 
calculated new risk estimates for the U.S. exposed population, for the 
various MICL options under consideration. The resulting risk estimates 
for bladder cancer are higher than those examined in detail in the 
proposal, and the new lung cancer risks are approximately equal to the 
new bladder cancer risks. As noted earlier, EPA believes that the 
combined risk of excess cases from lung and bladder cancer could be at 
least twice that of bladder cancer alone and will be refining its 
overall risk estimate in support of the final rule based on a number of 
factors, with a particular focus on the additive risks of lung and 
bladder cancer. Monetized benefits from avoided cases overall are 
expected to fall within the ranges presented in the June 22 Proposed 
Rule, because of the implicit assumptions of lung cancer risk in the 
``What If'' analysis. However, the lung cancer monetized benefits would 
be more certain, and removed from the ``What If'' categorization. In 
addition, the Agency performed a lower bound sensitivity analysis of 
risk estimates given a variation in the assumption about water used for 
cooking in Taiwan.

What Technologies and Costs Document Is Being Made Available?

    In the June 22, 2000, Federal Register, the EPA presented national 
cost

[[Page 63034]]

estimates of the proposed arsenic rule (65 FR 38888). In several tables 
\3\ in the preamble EPA presented annualized national cost estimates 
for four MCL options (3, 5, 10 and 20 L). The methodology and 
data used to develop these estimates are described in the Regulatory 
Impact Analysis (RIA) (EPA 2000b). This document is making available 
for public comment additional information on the costs of treatment 
technologies EPA: ``Technologies and Costs for the Removal of Arsenic 
From Drinking Water,'' April 1999, which has been placed in the docket, 
and will be made available on EPA's website. EPA used this April 1999 
document to develop the national estimates presented in the proposed 
rule.
---------------------------------------------------------------------------

    \3\ Table VIII-3. Annual Costs of Treatment Trains (Per 
Household); Table IX-11. National Annual Treatment Costs; Table IX-
12. Total Annual Costs Per Household; Table IX-13. Incremental 
National Annual Costs; Table IX-14. Incremental Annual Costs Per 
Household; Table X-7. Estimated Costs and Benefits From Reducing 
Arsenic in Drinking Water; Table XI-1. Estimated Costs and Benefits 
From Reducing Arsenic in Drinking Water; Table XIII-3. Estimated 
Costs and Benefits From Reducing Arsenic in Drinking Water; Table 
XIII-4. Estimated Annualized National Costs of Reducing Arsenic 
Exposures; Table XIII-5. Estimated Annual Costs Per Household and 
(Number of Households Affected); Table XIII-6. Summary of the Total 
Annual National Costs of Compliance with the Proposed Arsenic Rule 
Across MCL Options; Table XIII-7. Estimates of the Annual 
Incremental Risk Reduction, Benefits, and Costs of Reducing Arsenic 
in Drinking Water; Table XIV-2. Average Annual Cost per CWS by 
Ownership; Table XIV-3. Average Compliance Costs per Household for 
CWSs Exceeding MCLs; and Table XIV-4. Average Compliance Costs per 
Household for CWSs Exceeding MCLs as a Percent of Median Household 
Income.
---------------------------------------------------------------------------

    The RIA describes the model (SafeWaterXL) that was used by EPA to 
estimate national costs. The model uses data on arsenic occurrence, 
compliance decision trees, unit treatment technology train costs and 
other relevant data to generate national cost estimates. All of these 
inputs are described in the RIA. The treatment trains that were used in 
the national cost estimation are given in Exhibit 6-1 of the RIA. The 
RIA provides information on treatment technology costs by system size 
in Exhibit 6-2. The exhibit has cost estimates on treatment capital, 
treatment operation and maintenance (O&M), waste disposal capital, and 
waste disposal O&M costs for each treatment train.
    Today's document is advising the public about the availability in 
the docket of ``Technologies and Costs for the Removal of Arsenic from 
Drinking Water,'' April 1999, which provides the unit cost curves 
(regressions) that were used to generate Exhibit 6-1 of the RIA. The 
April 1999 technology and cost document contains curves for several 
removal efficiencies, including the ones corresponding to the removal 
efficiencies identified in Exhibit 6-1 of the RIA.
    The unit cost treatment curves for each technology can be found in 
the April 1999 technology and cost document. The unit cost waste 
disposal curves can be derived from Table 4-1, ``Summary of Residuals 
Characteristics,'' in the technology and cost document. Those 
interested in reproducing the waste disposal curves should consult the 
``Small Water System Byproducts Treatment and Disposal Cost'' (EPA 
1993a) document and the ``Water System Byproducts Treatment and 
Disposal Cost'' (EPA 1993b) document. The former is for small water 
systems, and the latter is for larger ones. An electronic copy of the 
treatment technology and waste disposal equations used in the 
development of the RIA can also be found in the docket.

Why Does the Docket Have a Copy of a Newer Version of the 
Technologies & Costs for Comment?

    The EPA has continued to refine and update cost estimates of the 
treatment technologies discussed in the proposed rule. In addition, EPA 
is following the development of emerging technologies that would be 
relevant for arsenic removal. An update of the April 1999 document was 
inadvertently included in the docket: EPA, ``Technologies and Costs for 
Removal of Arsenic from Drinking Water,'' November 1999. This was not, 
however, the version used to develop the RIA. The RIA costs were 
developed using the earlier April 1999 version, which is being provided 
with this document. This data and information is being made available 
to those interested in reproducing our national cost estimates.
    The differences between the cost curves in the April and November 
drafts are attributable to the different design criteria assumptions 
made when running the unit cost models. Three unit cost models were 
used: ``Very Small Systems'' (for systems between 0.015 to 0.100 mgd), 
``Water model'' (for systems between 0.27 and 1.00 mgd), and ``W/W 
Cost'' (for systems between 10 to 200 mgd). The design criteria 
assumptions are described prior to the presentation of the cost curves 
in each document. For example, the design criteria assumptions for 
coagulation assisted microfiltration are listed on page 3-47 of the 
November 1999 document and on page 3-60 of the April 1999 document. EPA 
will continue to refine the cost curves and other cost of compliance 
information and data based on comments submitted on the proposal.

How Will EPA Use the November 1999 Cost Document?

    EPA will carefully consider all comments on the proposed rule and 
will develop new national cost estimates for the final rule, along with 
a new supporting treatment technology and cost document, which would 
update both the April 1999 and November 1999 versions of the treatment 
technology and cost document. The new version that will be developed 
will include cost estimates for emerging technologies, and where 
necessary, updates to the treatment technology cost curves already 
developed. EPA may also develop an updated decision tree to refine and 
improve the cost estimates, based on comments received on the proposal. 
Changes in these inputs to EPA's models for determining the cost of 
compliance and any changes to the national cost estimates generated by 
the model will be presented in the final rule.

References

Chen, C. J., Y. C. Chuang, T. M. Lin and H. Y. Wu. 1985. Malignant 
neoplasms among residents of a Blackfoot disease endemic area in 
Taiwan: High arsenic well water and cancers. ``Cancer Research'' 
45:5895-5899.
Chen, C. J., Y. C. Chuang, T. M. Lin and H. Y. Wu. 1986. A 
retrospective study on malignant neoplasms of bladder, lung and 
liver in Blackfoot disease endemic area in Taiwan. ``British Journal 
of Cancer'' 53:399-405.
Morales, K.H., L. Ryan, K.G. Brown, T-L Kuo, M-M Wu, and C.J. Chen. 
2000. Risk of Internal Cancers from Arsenic in Drinking Water. 
``Environmental Health Perspectives'' 108:655-661.
National Research Council. 1999. ``Arsenic in Drinking Water.'' 
Washington, DC. National Academy Press.
Smith, A.H., C. Hopenhayn-Rich, M.N. Bates, H.M. Goeden, I. Hertz-
Picciotto, H.M. Duggan, R. Wood, M.J. Kosnett, and M.T. Smith. 1992. 
Cancer risks from arsenic in drinking water. ``Environmental Health 
Perspectives'' 97:259-267.
US EPA. 1988. ``Special Report on Ingested Inorganic Arsenic: Skin 
Cancer; Nutritional Essentiality.'' Risk Assessment Forum. EPA/625/
3-87/013. 124 pp. July 1988.
US EPA. 1993a. Small Water System Byproducts Treatment and Disposal 
Cost Document. April 1993.
US EPA. 1993b. Water System Byproducts Treatment and Disposal Cost 
Document. April 1993.
US EPA. 1999a. Technologies and Costs for Removal of Arsenic From 
Drinking Water. EPA-815-R-00-012. Prepared by International 
Consultants, Inc. and Malcolm Pirnie, Inc. under contract to EPA 
OGWDW. November 1999.
US EPA. 1999b. Technologies and Costs for the Removal of Arsenic 
From Drinking Water. Prepared by International

[[Page 63035]]

Consultants, Inc. and Malcolm Pirnie, Inc. under contract 68-C-C6-
0039 with EPA OGWDW. April 1999.
US EPA. 2000a. ``Estimated Per Capita Water Ingestion in the United 
States: Based on Data Collected by the United States Department of 
Agriculture's (USDA) 1994-1996 Continuing Survey of Food Intakes by 
Individuals.'' Office of Water, Office of Standards and Technology. 
EPA-822-00-008. April 2000.
US EPA. 2000b. Proposed Arsenic in Drinking Water Rule: Regulatory 
Impact Analysis (RIA). Prepared by Abt Associates, Inc. under 
contract to EPA OGWDW. EPA 815-R-013. June 2000.
US EPA. 2000c. ``Arsenic Occurrence in Public Drinking Water 
Supplies.'' Public Comment Draft. Office of Water, Washington DC. 
EPA 815-D-00-001. May 2000.
US EPA. 2000d. National Primary Drinking Water Regulations; Arsenic 
and Clarifications to Compliance and New Source Contaminants 
Monitoring; Proposed Rule. Federal Register. Vol. 65, No. 121, p. 
38888. June 22, 2000.
Viscusi, W.K., W.A. Magat, and J. Huber. 1991. 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.
Wu, M.M., T.L. Kuo, Y.H. Hwant, and C.J. Chen. 1989. Dose-response 
relation between arsenic concentration in well water and mortality 
from cancers and vascular diseases. American Journal of 
Epidemiology. 130:1123-1132.

    Authority: 42 U.S.C. 300f, 300g-1, 300g-2, 300g-3, 300g-4, 300g-
5, 300g-6, 300j-4, 300j-9, and 300j-11.

    Dated: October 13, 2000.
J. Charles Fox,
Assistant Administrator for Water.
[FR Doc. 00-27034 Filed 10-19-00; 8:45 am]
BILLING CODE 6560-50-P