This is the accessible text file for GAO report number GAO-14-103 
entitled 'Drinking Water: EPA Has Improved Its Unregulated Contaminant 
Monitoring Program, but Additional Action Is Needed' which was 
released on February 10, 2014. 

This text file was formatted by the U.S. Government Accountability 
Office (GAO) to be accessible to users with visual impairments, as 
part of a longer term project to improve GAO products' accessibility. 
Every attempt has been made to maintain the structural and data 
integrity of the original printed product. Accessibility features, 
such as text descriptions of tables, consecutively numbered footnotes 
placed at the end of the file, and the text of agency comment letters, 
are provided but may not exactly duplicate the presentation or format 
of the printed version. The portable document format (PDF) file is an 
exact electronic replica of the printed version. We welcome your 
feedback. Please E-mail your comments regarding the contents or 
accessibility features of this document to Webmaster@gao.gov. 

This is a work of the U.S. government and is not subject to copyright 
protection in the United States. It may be reproduced and distributed 
in its entirety without further permission from GAO. Because this work 
may contain copyrighted images or other material, permission from the 
copyright holder may be necessary if you wish to reproduce this 
material separately. 

United States Government Accountability Office: 
GAO: 

Report to Congressional Requesters: 

January 2014: 

Drinking Water: 

EPA Has Improved Its Unregulated Contaminant Monitoring Program, but 
Additional Action Is Needed: 

GAO-14-103: 

GAO Highlights: 

Highlights of GAO-14-103, a report to congressional requesters. 

Why GAO Did This Study: 

EPA's UCMR program collects data on unregulated contaminants in the 
nation's drinking water. EPA uses these data and other information to 
make regulatory determinations—-decisions on whether to regulate 
additional drinking water contaminants. It is currently in its third 
data collection cycle, UCMR3. 

GAO was asked to examine the UCMR program. This report examines: (1) 
the extent to which EPA implemented GAO's prior recommendations to 
improve the program and opportunities, if any, to strengthen it 
further; (2) the factors EPA considered when it selected the UCMR3 
contaminants and the limitations, if any, it faced in selecting them; 
and (3) the extent to which UCMR data support regulatory 
determinations. 

GAO reviewed EPA documents, surveyed 48 subject matter experts, 
assessed the UCMR program against statutory requirements and other 
standards, and interviewed EPA officials. 

What GAO Found: 

The Environmental Protection Agency (EPA) has implemented all of the 
recommendations GAO made in its May 2011 report to improve the 
Unregulated Contaminant Monitoring Rule (UCMR) program. In that 
report, GAO recommended that EPA (1) monitor for the full 30 
contaminants allowed by statute, (2) monitor for most or all 
contaminants using a more robust monitoring approach, and (3) select 
sufficiently sensitive minimum reporting levels (MRL) for monitoring 
contaminants. EPA now requires public water systems to monitor for 30 
contaminants in the UCMR3 program, using its most robust monitoring 
approach for a majority of these contaminants, and setting MRLs as low 
as can be reliably measured, according to EPA. The Safe Drinking Water 
Act (SDWA) requires EPA to vary the monitoring frequency based on the 
type of contaminant likely to be found, but EPA used a standard 
monitoring frequency for all contaminants. This may result in 
inaccurate estimates of the occurrence of sporadically occurring 
microbes (e.g., viruses) or pesticides, according to experts GAO 
surveyed and studies it reviewed. In such cases, the monitoring data 
may not provide reliable estimates of contaminant occurrence. 

EPA used 10 factors to select the 30 contaminants for UCMR3, but its 
selection process faced some limitations. Officials told GAO that the 
contaminants did not have to meet all 10 of the selection factors to 
be chosen, but 3 were very important (1) the availability of an 
analytical method to detect contaminants, (2) the reliability of 
health effects information on the contaminants, and (3) the need for 
data to support regulatory determinations for priority contaminants. 
However, EPA is limited by a statutory cap of 30 contaminants every 5 
years, which restricts its ability to collect data on additional 
contaminants that could have been monitored for little additional 
cost. SDWA's legislative history reflected concerns with the ability 
of public water systems to absorb such costs, but many of the 
analytical methods EPA is using for UCMR3 are able to test a single 
sample of drinking water for more than one contaminant at a time. 
However, because of the limit of 30, EPA cannot always take advantage 
of this efficiency and is unable to gain economies of scale using 
monitoring that is already under way. 

EPA uses UCMR data to support regulatory determinations but faces a 
time lag when doing so. EPA has used UCMR data to support 10 out of 12 
regulatory determinations it has made since 2008 and is currently 
using UCMR data to inform the determinations expected in 2015. 
However, a time lag between the statutory deadline for making 
regulatory determinations and when UCMR data are available delays 
determinations on given contaminants until the following cycle. The 2-
year time frame SDWA originally established from the time EPA 
publishes the UCMR list to when it makes regulatory determinations has 
not provided enough time for the agency to incorporate the UCMR data 
into the determinations. The UCMR3 monitoring, data collection, and 
analysis overlap with the time when EPA will be making its regulatory 
determinations for contaminants from its most recent Contaminant 
Candidate List. Consequently, UCMR data are not available to support 
regulatory determinations for contaminants during the cycle in which 
they are monitored; rather, UCMR data typically are not used until the 
next cycle. EPA officials told GAO that most of the UCMR3 data, which 
are being collected from 2013 to 2015, will be used to support the 
regulatory determinations it expects to issue in 2020 instead of 2015. 

What GAO Recommends: 

Congress should consider amending the Safe Drinking Water Act to allow 
EPA to monitor for more than 30 contaminants under certain 
circumstances, and to adjust statutory time frames so UCMR data can 
inform regulatory determinations in the same cycle. GAO, among other 
things, recommends that EPA vary the monitoring frequency based on 
contaminant type. In commenting on a draft of this report, EPA 
generally agreed with GAO's findings, conclusions, and recommendations. 

View [hyperlink, http://www.gao.gov/products/GAO-14-103]. For more 
information, contact J. Alfredo Gómez (202) 512-3841 or gomezj@gao.gov. 

[End of section] 

Contents: 

Letter: 

Background: 

EPA Has Improved the UCMR Program by Implementing Our Recommendations, 
but Its Monitoring Frequency May Yield Some Inaccurate Estimates: 

EPA Considered 10 Factors to Select Contaminants for Monitoring, but 
Its Selection Process Faced Limitations: 

UCMR Data Support Regulatory Determinations, but Time Lag Results in 
Delays and Reliance on Older Data: 

Conclusions: 

Matters for Congressional Consideration: 

Recommendations for Executive Action: 

Agency Comments: 

Appendix I: Objectives, Scope, and Methodology: 

Appendix II: UCMR3 Contaminants: 

Appendix III: Information on Analytical Methods: 

Appendix IV: Round 2 Survey Questionnaire and Results: 

Appendix V: List of Experts Who Completed the Surveys: 

Appendix VI: Comments from the Environmental Protection Agency: 

Appendix VII: GAO Contact and Staff Acknowledgments: 

Related GAO Products: 

Tables: 

Table 1: Public Water Systems Required to Participate in the Third 
Unregulated Contaminant Monitoring Rule: 

Table 2: EPA Programs to Address Unregulated Contaminants in Drinking 
Water: 

Table 3: Number of Public Water Systems Required to Participate in the 
First, Second, and Third Unregulated Contaminant Monitoring Rules 
(UCMR1, UCMR2, and UCMR3): 

Table 4: Monitoring Frequency by System Size and Water Supply Source 
for the Third Unregulated Contaminant Monitoring Rule: 

Table 5: Contaminants of Emerging Concern Detectable with an 
Analytical Method Already in Use: 

Figures: 

Figure 1: EPA Selection Factors for Identifying Contaminants for 
Monitoring under the Third Unregulated Contaminant Monitoring Rule 
(UCMR3): 

Figure 2: Contaminant Selection and Rule Development Process for the 
Third Unregulated Contaminant Monitoring Rule: 

Figure 3: UCMR3 Timeline: 

Figure 4: UCMR Data Do Not Inform Regulatory Determinations in the 
Same Cycle: 

Figure 5: About 10 Years Typically Pass before Regulatory 
Determinations Are Made for Contaminants on the CCL: 

Figure 6: EPA's Process for Analytical Method Development: 

Abbreviations: 

CCL: Contaminant Candidate List: 

EPA: Environmental Protection Agency: 

GC-MS: gas chromatograph-mass spectrometer: 

MRL: minimum reporting level: 

MTBE: methyl tertiary-butyl ether: 

NAWQA: National Water-Quality Assessment Program: 

OMB: Office of Management and Budget: 

PFC: perfluorinated chemical: 

SDWA: Safe Drinking Water Act: 

UCMR: Unregulated Contaminant Monitoring Rule: 

USGS: United States Geological Survey: 

VOC: volatile organic compound: 

[End of section] 

GAO:
United States Government Accountability Office: 
441 G St. N.W. 
Washington, DC 20548: 

January 9, 2014: 

The Honorable Henry A. Waxman: 
Ranking Member: 
Committee on Energy and Commerce: 
House of Representatives: 

The Honorable Edward J. Markey: 
United States Senate: 

Most Americans receive their drinking water from the more than 160,000 
public water systems in the United States, all of which are subject to 
key requirements of the Safe Drinking Water Act (SDWA).[Footnote 1] 
Under SDWA, the Environmental Protection Agency (EPA) is authorized to 
regulate contaminants in public water systems.[Footnote 2] Currently, 
EPA regulates more than 89 drinking water contaminants. In 1996, in 
the midst of concerns about the number and pace of regulations 
required under the act and the need to better focus limited resources 
on contaminants of greatest public health concern, Congress enacted 
significant amendments to SDWA. The 1996 amendments to SDWA also 
directed EPA, among other things, to identify unregulated contaminants 
that present the greatest public health concern, establish a program 
to monitor drinking water for unregulated contaminants, and decide 
whether or not to regulate at least 5 such contaminants every 5 years. 
[Footnote 3] EPA decisions about whether or not to regulate these 
contaminants are called regulatory determinations. Since 1996, EPA has 
made 21 regulatory determinations and has decided to regulate one of 
the contaminants, perchlorate.[Footnote 4] EPA has identified 116 
unregulated contaminants, including perchlorate and others such as 
formaldehyde and Salmonella enterica, that it believes may occur in 
public water systems and may also require regulation. Those 
unregulated contaminants are listed on EPA's current Contaminant 
Candidate List (CCL).[Footnote 5] 

In addition, EPA has established the Unregulated Contaminant 
Monitoring Rule (UCMR) program, to monitor drinking water for 
unregulated contaminants, under the 1996 amendments to SDWA. As part 
of the UCMR, the SDWA amendments require EPA to issue a list every 5 
years identifying up to 30 contaminants to be monitored by public 
water systems. The UCMR program is currently in its third 5-year 
cycle, known as UCMR3. EPA estimates that UCMR3 will cost $87 million 
from 2012 to 2016, of which $21 million will be paid by EPA. EPA will 
pay for small water systems' costs for the UCMR; the remaining $66 
million will be paid by large water systems.[Footnote 6] The UCMR, 
CCL, and regulatory determinations are related to each other. First, 
the CCL identifies unregulated contaminants that may require 
regulation in the future. Through the UCMR program, public water 
systems monitor for some of the CCL contaminants to determine how 
often and at what level the contaminants occur in drinking water. EPA 
then makes regulatory determinations about contaminants on the CCL 
using UCMR data, information about the contaminants' health effects, 
and other information for support. 

In May 2011, we reported that EPA's monitoring program under the UCMR 
did not use EPA's full statutory authority under SDWA, use 
sufficiently sensitive testing methods, or employ sufficiently robust 
monitoring approaches.[Footnote 7] First, EPA did not use its full 
allotment of 30 contaminants as authorized by SDWA in either of the 
first two UCMR cycles (UCMR1 and UCMR2), which limited the agency's 
progress in obtaining data on drinking water contaminants. Second, in 
some cases, EPA required public water systems to use testing methods 
(i.e., analytical methods) that were not sufficiently sensitive to 
identify the presence of contaminants at EPA's health reference level--
the health benchmark EPA uses in assessing whether to regulate 
specific contaminants. Third, EPA used a limited sample size of 300 
public water systems for UCMR1 for one of the program's monitoring 
approaches, which resulted in occurrence data that, according to EPA 
documents, were not sufficient to make national estimates of 
contaminant occurrence. In the 2011 report, we also recommended that 
EPA use its full statutory authority to monitor for 30 unregulated 
contaminants, select minimum reporting levels that are sufficiently 
sensitive to detect the presence of contaminants in public water 
systems at levels of public health concern, and monitor for most or 
all contaminants using the program's most robust monitoring approach. 

In this context, you asked us to examine the UCMR program. 
Specifically, our objectives for this report were to: (1) evaluate the 
extent to which EPA has implemented the recommendations we made in 
2011 to improve the UCMR program and identify any opportunities to 
strengthen it further; (2) identify the factors that EPA considered 
when it selected the contaminants for monitoring under UCMR3 and the 
limitations, if any, that EPA faced in selecting the contaminants; and 
(3) examine the extent to which UCMR data support regulatory 
determinations. 

To identify the steps EPA has taken in response to the recommendations 
we made in 2011, we reviewed EPA documents, including EPA's proposed 
and final rules for the UCMR program, assessed the UCMR sample design 
against statutory requirements and statistical standards, and 
interviewed EPA officials. To identify the factors EPA considers when 
selecting contaminants, we reviewed EPA documents, including EPA's 
proposed and final rules for the UCMR program and SDWA requirements, 
and interviewed EPA officials. To determine how EPA uses the UCMR data 
to support its regulatory determinations, we reviewed EPA and Federal 
Register documents and interviewed EPA officials. In addition, to 
obtain further information about the UCMR program and identify 
opportunities to strengthen it further, we conducted a two-round 
Delphi survey[Footnote 8] of experts selected for their knowledge of 
one or more of the following areas: drinking water, contaminant 
monitoring, analytical methods, toxicology, risk assessment, and 
environmental regulation. In the first round of the survey, we 
selected 14 experts who demonstrated familiarity with EPA's CCL, UCMR, 
and regulatory determination processes. We asked them 10 open-ended 
questions about how various aspects of the UCMR program helped or 
hindered EPA in (1) selecting priority contaminants, (2) monitoring 
for selected contaminants, and (3) using the data to support 
regulatory determinations. We analyzed the responses provided by the 
experts and developed closed-ended questions for the second round of 
the survey. To identify experts for the second round of the survey, we 
developed a preliminary list of experts, which included the 14 experts 
who completed the first round of the survey and others identified 
through interviews with stakeholders and our review of public comments 
provided on EPA's proposed rules. We contracted with the National 
Academy of Sciences to identify additional experts. We selected 48 
experts for the second round of the survey based on their knowledge of 
one or more of the following topic areas: drinking water, contaminant 
monitoring, analytical methods, toxicology, risk assessment, and 
environmental regulation. For the first round of the survey, we 
received responses from 14 of the 14 experts we surveyed and, for the 
second round, we received responses from 40 of 48 experts we surveyed 
(83 percent response rate). The information that we obtained from the 
expert surveys is not generalizable to all experts that have an 
interest in or knowledge of drinking water contaminant monitoring but 
does provide several experts' views on the UCMR program. For more 
information on our objectives, scope, and methodology see appendix I. 
See appendix IV for a copy of the survey questions and responses for 
the second round of the survey and appendix V for a list of the 
experts who completed the surveys. 

We conducted this performance audit from June 2012 to January 2014 in 
accordance with generally accepted government auditing standards. 
Those standards require that we plan and perform the audit to obtain 
sufficient, appropriate evidence to provide a reasonable basis for our 
findings and conclusions based on our audit objectives. We believe 
that the evidence obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives. 

Background: 

This section presents information on (1) regulated and unregulated 
contaminants in drinking water, (2) the UCMR program and SDWA 
requirements, and (3) laboratories and methods used to analyze 
drinking water. 

Regulated and Unregulated Contaminants in Drinking Water: 

Drinking water contaminants include chemical and microbial substances, 
some of which can harm human health. For example, prolonged exposure 
to arsenic in drinking water can cause skin damage or circulatory 
problems and, over time, may increase cancer risks. Contaminants may 
be naturally occurring or introduced through other means. EPA has 
identified three ways that drinking water generally becomes 
contaminated. First, contaminants can find their way into drinking 
water sources from industrial waste releases, agricultural runoff, and 
other pollution sources. Second, contamination can occur during the 
treatment of water by forming what are known as "disinfection by-
products."[Footnote 9] Third, contaminants can "leach" or dissolve out 
of the materials used for treatment, storage, and distribution of 
water. 

EPA regulates drinking water contaminants under SDWA by issuing 
legally enforceable standards, known as National Primary Drinking 
Water Regulations or "primary standards" that generally limit the 
levels of these contaminants in public water systems. EPA has issued 
primary standards addressing more than 89 contaminants. States 
generally enforce these standards through compliance monitoring. Under 
compliance monitoring, public water systems collect a sample of 
drinking water, analyze the sample for specific contaminants, and 
report the results to states. To analyze the samples, water systems 
and states use analytical methods developed by government agencies, 
universities, and other organizations. EPA evaluates analytical 
methods for drinking water and approves those it determines meet 
agency requirements for monitoring contaminants. Laboratories 
analyzing drinking water compliance samples must be certified by the 
state or EPA. EPA establishes schedules for how often water systems 
must monitor for regulated contaminants. The monitoring frequency 
generally depends on the contaminant, previous monitoring results, the 
type of source water, and the size of the population served by the 
water system. 

Unregulated drinking water contaminants include chemical and microbial 
substances and are not currently subject to primary standards under 
SDWA.[Footnote 10] Exposure to unregulated drinking water contaminants 
may pose a human health risk, particularly to vulnerable 
subpopulations such as pregnant women, infants, the elderly, and those 
with compromised immune systems.[Footnote 11] Unregulated contaminants 
include industrial chemicals, military chemicals, pesticides, 
bacteria, and other microbes and chemicals. They also include 
contaminants of emerging concern--contaminants characterized by a 
perceived, potential or real threat to human health or the environment 
or by a lack of published health standards. Examples include 
pharmaceuticals, such as prescription and over-the-counter drugs, and 
ingredients in personal care products, such as cosmetics and soaps. A 
contaminant may also be considered to be emerging because a new source 
of public exposure has been discovered or a new detection method or 
treatment technology has been developed. 

UCMR Program and SDWA Requirements: 

Under the UCMR program, EPA collects data about the occurrence of 
unregulated contaminants in drinking water. The UCMR program collects 
data on the extent of the occurrence of unregulated contaminants in 
all public water systems serving more than 10,000 people and in a 
nationally representative sample of smaller systems. These data inform 
EPA's decisions about whether to issue primary standards for 
unregulated contaminants. Under the UCMR, EPA is to identify up to 30 
unregulated contaminants that may be present in drinking water and 
establishes a monitoring program to obtain data on the extent of their 
occurrence. These data help EPA determine whether drinking water 
contamination is occurring and estimate its level and frequency. EPA 
established a three-tiered approach for monitoring UCMR contaminants 
that includes (1) assessment monitoring, (2) screening survey, and (3) 
pre-screen testing. EPA's approach is based primarily on the 
availability of analytical methods to monitor the contaminants. 
Assessment monitoring is used to monitor contaminants that can be 
detected by well-established analytical methods that are widely used 
in drinking water laboratories. The screening survey is used to 
monitor contaminants that can be detected by more recently developed 
analytical methods that are used in fewer drinking water laboratories. 
Pre-screen testing is used to monitor contaminants that can be 
detected by very new or specialized analytical methods that are used 
in a small set of drinking water laboratories. 

Approximately 6,000 public water systems are currently participating 
in the UCMR program. Participating public water systems are monitoring 
for contaminants under one or more of the three tiers. EPA varies the 
number of public water systems required to monitor under each tier. 
UCMR monitoring is conducted by all large community and nontransient, 
noncommunity public water systems serving more than 10,000 people. 
[Footnote 12] Monitoring is also conducted by a nationally 
representative sample of other public water systems that EPA selects 
(see table 1). In addition, a sample of community and nontransient, 
noncommunity water systems, and transient water systems are required 
to allow EPA to conduct pre-screen testing under UCMR3.[Footnote 13] 

Table 1: Public Water Systems Required to Participate in the Third 
Unregulated Contaminant Monitoring Rule: 

Monitoring approach: Assessment monitoring using well-established 
analytical methods; 
Type of system: Community water systems and nontransient, noncommunity 
water systems; 
Size of system: Systems serving more than 10,000 people; 
Number of systems: All (about 4,200). 

Size of system: Systems serving 10,000 or fewer people; 
Number of systems: Representative sample of 800 systems. 

Monitoring approach: Screening survey using more recently developed 
analytical methods; 
Type of system: Community water systems and 
nontransient, noncommunity water systems; 
Size of system: Systems 
serving more than 100,000 people; 
Number of systems: All (about 400). 

Size of system: Systems serving 10,001 to 100,000 people; 
Number of systems: Representative sample of 320 systems. 

Size of system: Systems serving 10,000 or fewer people; 
Number of systems: Representative sample of 480 systems. 

Monitoring approach: Pre-screen testing using specialized analytical 
methods; 
Type of system: Nondisinfecting groundwater systems including 
community water systems, nontransient, noncommunity water systems, and 
transient, noncommunity water systems; 
Size of system: Systems serving 1,000 or fewer people; 
Number of systems: Representative sample of 800 systems. 

Source: GAO analysis of EPA documents. 

[End of table] 

The UCMR program is part of a larger framework established by SDWA 
that also includes the CCL and requires EPA to make regulatory 
determinations for at least five of the contaminants on that list 
every 5 years. The CCL is a list of unregulated contaminants that are 
known or anticipated to occur in public water systems and that may 
require regulation. SDWA directs EPA to select contaminants that pose 
the greatest public health concern for consideration for the CCL and 
regulatory determinations.[Footnote 14] A regulatory determination is 
a formal decision on whether EPA should initiate a rulemaking process 
to develop a primary standard for a specific drinking water 
contaminant. The UCMR, CCL, and regulatory determinations framework is 
implemented by three different branches within EPA's Office of Ground 
Water and Drinking Water's Standards and Risk Management Division: the 
Technical Support Center administers the UCMR program; the Targeting 
and Analysis Branch develops the CCL; and the Standards and Risk 
Reduction Branch leads the development of the regulatory 
determinations. 

SDWA requires EPA to make regulatory determinations for at least five 
of the contaminants on the CCL every 5 years. Regulatory 
determinations can be positive (i.e., a primary standard is needed) or 
negative (i.e., no primary standard is needed). Under SDWA, EPA is to 
make a positive regulatory determination and subsequently issue a 
primary standard when (1) the contaminant may have an adverse effect 
on the health of persons; (2) the contaminant is known to occur or 
there is a substantial likelihood that the contaminant will occur in 
public water systems with a frequency and at levels of public health 
concern; and (3) in the sole judgment of the Administrator, regulation 
of such contaminant presents a meaningful opportunity for health risk 
reduction for persons served by public water systems. EPA can make a 
negative regulatory determination when the SDWA conditions are not 
met. SDWA set recurring 5-year deadlines for the CCL, UCMR, and 
regulatory determinations. The initial deadlines were staggered such 
that EPA was required to publish the CCL in the Federal Register first 
(February 1998), the UCMR list 18 months later (August 1999), and the 
regulatory determinations 2 years after that (August 2001). The cycle 
restarts with the next CCL, UCMR, and regulatory determinations due 5 
years after the first set of deadlines. However, EPA has not always 
met the SDWA deadlines. Moreover, EPA has not issued the CCL, UCMR, 
and regulatory determinations separated by the 18-month and 2-year 
time frames that were originally established by statute.[Footnote 15] 
The cycle is currently in its third iteration. EPA refers to the 
programs in the current cycle as CCL3, UCMR3, and the CCL3 regulatory 
determinations (see table 2). 

Table 2: EPA Programs to Address Unregulated Contaminants in Drinking 
Water: 

Program: Purpose; 
Contaminant Candidate List (CCL): Identify unregulated contaminants 
that are known or anticipated to occur in public water systems; 
Unregulated Contaminant Monitoring Rule (UCMR): Monitor unregulated 
contaminants to determine the extent of their occurrence in drinking 
water; 
Regulatory determinations: Decide whether or not to regulate 
contaminants on the CCL using UCMR data, health effects information, 
and other information. 

Program: SDWA requirement; 
Contaminant Candidate List (CCL): Publish list every 5 years; 
Unregulated Contaminant Monitoring Rule (UCMR): Publish a list of up 
to 30 contaminants for monitoring, every 5 years; 
Regulatory determinations: Decide whether or not to regulate at least 5 
CCL contaminants, every 5 years. 

Program: Outputs; 
Contaminant Candidate List (CCL): 1998 (60 contaminants on CCL1); 2005 
(51 contaminants on CCL2); 2009 (116 contaminants on CCL3); 
Unregulated Contaminant Monitoring Rule (UCMR): 1999 (26 contaminants 
on UCMR1); 2007 (25 contaminants on UCMR2); 2012 (30 contaminants on 
UCMR3); 
Regulatory determinations: 2003 (9 negative determinations for CCL1); 
2008 (11 negative determinations for CCL2); 2011 (1 positive 
determination-out of cycle); 2015 (determinations for CCL3 expected). 

Source: GAO analysis of EPA information. 

[End of table] 

The UCMR program's primary purpose is to inform regulatory 
determinations, but EPA can also use the UCMR data to develop health 
advisories for some of the contaminants it decides not to regulate. 
Health advisories provide information on contaminants that can harm 
human health and are known or anticipated to occur in drinking water. 
They provide technical guidance on health effects to assist EPA 
regional offices, states, and other public health officials. EPA has 
issued new or updated health advisories for nine contaminants it 
decided not to regulate.[Footnote 16] The advisories identify 
concentrations of contaminants at which adverse health effects are not 
anticipated to occur over specific exposure durations (1 day, 10 days, 
several years, and a lifetime). Drinking water health advisories are 
not legally enforceable standards and, according to EPA, are subject 
to change as new information becomes available. 

Laboratories and Methods Used to Analyze Drinking Water: 

Private analytical laboratories and some water utilities' laboratories 
conduct the analytical testing for the UCMR contaminants. All 
laboratories conducting analyses for UCMR3 must be approved by EPA. In 
general, laboratories must first successfully complete the 
registration and application process, which includes testing for 
contaminants using EPA-specified analytical methods, before receiving 
EPA's approval.[Footnote 17] Approved laboratories must adhere to the 
quality assurance and quality control procedures and criteria outlined 
in the methods and rule language to maintain their approval status and 
may be subject to on-site laboratory or paper audits. Laboratories 
also post occurrence data and required quality control data 
electronically in EPA's data system, the Safe Drinking Water Accession 
and Review System, within 120 days of the sample collection date. 
[Footnote 18] 

EPA specifies the analytical methods that laboratories participating 
in the UCMR program use to detect contaminants. Some methods can 
detect a single contaminant while others can detect a group of similar 
contaminants. Because the UCMR collects nationwide data, agency 
officials told us EPA's goal is to use analytical methods that yield 
consistently high-quality monitoring data from public water systems 
across the country. EPA officials told us that the agency develops or 
uses analytical methods for the UCMR program that enable it to meet 
this goal. In particular, agency officials told us these methods must 
(1) have sufficient "holding times";[Footnote 19] (2) be able to be 
used successfully by a variety of labs with varying levels of 
expertise; and (3) be used effectively in different types of drinking 
water. EPA officials said it typically takes 1 or 2 years to develop a 
standard analytical method that meets these requirements and up to 3 
years or longer to develop more complex methods. See appendix III for 
more information on analytical methods. 

EPA Has Improved the UCMR Program by Implementing Our Recommendations, 
but Its Monitoring Frequency May Yield Some Inaccurate Estimates: 

EPA has implemented the three recommendations we made in our May 2011 
report to improve the UCMR program.[Footnote 20] The technical changes 
EPA has made to UCMR3 are significant improvements over prior UCMR 
cycles and have increased the program's productivity and statistical 
strength, among other things. However, the program's monitoring 
frequency may result in inaccurate estimates of the occurrence of some 
contaminants. 

EPA Has Implemented All of GAO's 2011 Recommendations to Improve the 
UCMR Program: 

In UCMR3, EPA implemented our recommendations to (1) monitor for 30 
contaminants; (2) monitor for most or all contaminants using the 
program's more robust monitoring approach; and (3) select minimum 
reporting levels (MRL) that are sufficiently sensitive to detect the 
presence of contaminants in public water systems at levels of public 
health concern as follows: [Footnote 21] 

Monitor for 30 contaminants. In our May 2011 report,[Footnote 22] we 
reported that EPA did not use its full statutory authority to monitor 
for 30 contaminants in implementing the first two cycles of the UCMR 
program. Specifically, EPA required that 26 and 25 contaminants be 
monitored under UCMR1 and UCMR2, respectively. As a result, EPA did 
not obtain occurrence data for 9 additional contaminants, which in 
turn, limited the agency's productivity in obtaining critical 
information to support its regulatory determinations. In its response 
to our recommendation that it use its full statutory authority, EPA 
stated it supports a goal of including as many priority contaminants 
as possible in each 5-year monitoring cycle. Accordingly, in May 2012, 
EPA published its final rule for UCMR3 requiring public water systems 
to monitor for 30 contaminants (see appendix II for a list of the 30 
contaminants).[Footnote 23] 

Use more robust monitoring approach. In 2011, we reported that EPA 
officials described the assessment monitoring approach as "the gold 
standard" for obtaining sufficient data on national occurrence of 
drinking water contaminants to support regulatory determinations, but 
that it had used the more limited screening survey approach for a 
majority of the contaminants monitored under UCMR1 and UCMR2, as shown 
in table 3. The screening survey monitoring approach produces 
estimates of national occurrence with greater uncertainty than those 
provided by assessment monitoring because it requires a smaller sample 
of public water systems to monitor. For example, under UCMR1, the 
screening survey required a sample of 300 public water systems to 
monitor. In contrast, approximately 3,800 systems were required to 
conduct assessment monitoring. In EPA's response to our recommendation 
that it conduct monitoring for most or all contaminants using the 
assessment monitoring approach, EPA stated that it supports a goal of 
using assessment monitoring for as many contaminants as is practical 
to obtain more robust occurrence data for and also to achieve higher 
confidence in the national estimates. EPA is requiring assessment 
monitoring for a majority of the contaminants under UCMR3. In 
addition, for UCMR2 and UCMR3, EPA expanded the sample size of the 
screening survey to increase the statistical strength of the 
monitoring results. We analyzed EPA's methodology for selecting the 
sample of systems for the screening survey under UCMR2 and UCMR3 
against standard statistical criteria, and we determined that the 
larger sample size should allow EPA to make national estimates of 
contaminant occurrence with smaller margins of error and higher levels 
of confidence compared with the first UCMR. In addition, 26 of the 40 
experts who participated in our survey thought the screening survey 
monitoring approach for UCMR3 is likely to provide EPA with sufficient 
data to determine whether contaminants occur with a frequency and at 
levels of public health concern. Further, EPA officials told us that 
the UCMR2 screening survey occurrence data are sufficient for 
supporting regulatory determinations, whereas, in its 2008 regulatory 
determinations notice, EPA stated that the occurrence data were 
insufficient to support regulatory determinations for 12 of the 14 
contaminants monitored under the UCMR1 screening survey. 

Table 3: Number of Public Water Systems Required to Participate in the 
First, Second, and Third Unregulated Contaminant Monitoring Rules 
(UCMR1, UCMR2, and UCMR3): 

Number of contaminants to be monitored: 

Monitoring approach: Assessment monitoring; 
UCMR1: 12; 
UCMR2: 10; 
UCMR3: 21[A]. 

Monitoring approach: Screening survey; 
UCMR1: 14; 
UCMR2: 15; 
UCMR3: 7. 

Monitoring approach: Pre-screen testing; 
UCMR1: 0; 
UCMR2: 0; 
UCMR3: 2. 

Monitoring approach: Total number of contaminants; 
UCMR1: 26; 
UCMR2: 25; 
UCMR3: 30. 

Number of public water systems required to be monitored: 

Monitoring approach: Assessment monitoring; 
UCMR1: about 3,800[B]; 
UCMR2: about 4,200[B]; 
UCMR3: about 5,000[B]. 

Monitoring approach: Screening survey; 
UCMR1: 300[C]; 
UCMR2: about 1,200[C]; 
UCMR3: about 1,200[C]. 

Monitoring approach: Pre-screen testing; 
UCMR1: NA; 
UCMR2: NA; 
UCMR3: 800. 

Source: GAO analysis of EPA information. 

[A] UCMR3 assessment monitoring includes total chromium, a regulated 
contaminant being monitored under another authority. 

[B] Assessment monitoring includes all public water systems serving 
over 10,000 people and a representative sample of 800 systems serving 
10,000 or fewer people. 

[C] EPA increased the sample size of the screening survey for UCMR2 
and UCMR3. Specifically, under UCMR1, the screening survey included a 
sample of 300 systems (120 systems serving over 10,000 people and 180 
systems serving 10,000 or fewer people). In contrast, under UCMR2 and 
UCMR3, the screening survey includes all systems serving over 100,000 
people; a representative sample of 320 systems serving 10,001 to 
100,000 people; and a representative sample of 480 systems serving 
10,000 or fewer people. 

[End of table] 

Select sufficiently sensitive MRLs. In 2011, we reported that the 
utility of the UCMR occurrence data was reduced because EPA had not 
used sufficiently sensitive MRLs to detect the presence of 
contaminants at levels of public health concern. Specifically, for 9 
of the 20 contaminants for which EPA made regulatory determinations in 
2003 and 2008, the MRL exceeded EPA's health reference level--the 
health benchmark EPA uses to assess whether the detection of a 
specific contaminant may indicate a potential human health risk. 
[Footnote 24] As a result, occurrence of these 9 contaminants at a 
level higher than the health reference level, but lower than the MRL, 
may not have been detected. In its response to our recommendation that 
EPA use MRLs that are sufficiently sensitive, EPA stated it supports a 
goal of establishing minimum reporting levels that are sufficiently 
sensitive to reliably detect the known and likely occurrence of 
contaminants in public water systems at levels of public health 
concern and provide useful and credible information on the occurrence 
of the contaminants in public drinking water systems.[Footnote 25] EPA 
has since revised its approach for selecting MRLs for monitoring 
selected contaminants in an effort to increase the likelihood that the 
monitoring will be sufficiently sensitive to detect the presence of 
contaminants at levels of public health concern. Specifically, EPA 
officials told us that, for UCMR3, the agency set the MRLs based on 
how low the method could reliably measure specific contaminants. 
[Footnote 26] These levels are based on method sensitivity (i.e., not 
adjusted upward based on available health effects information). In 
some cases, the MRLs for UCMR3 are set well below EPA's current health 
reference levels, according to EPA's final rule for UCMR3. However, 
health effects research on some UCMR3 contaminants is still in 
progress, and the health reference levels are subject to change. EPA 
believes that its new approach will be more effective than its past 
approach because, by monitoring for contaminants at the lowest 
concentrations feasible, UCMR3 is more likely to detect contamination 
at levels of public health concern, even if new health effects data 
result in lower health reference levels in the future, as anticipated 
by its final rule for UCMR3. 

UCMR Monitoring Frequency May Yield Inaccurate Estimates of Some 
Contaminant Occurrence: 

EPA has taken steps to improve the UCMR program, but the program's 
standard monitoring frequency may result in inaccurate estimates of 
the occurrence of certain contaminants. In particular, several experts 
we surveyed and studies on contaminant occurrence indicated that the 
UCMR3 standard monitoring frequency may result in inaccurate estimates 
of the occurrence of sporadically occurring microbes (e.g., viruses) 
and seasonally applied pesticides. SDWA calls for EPA to vary the 
monitoring frequency and schedule for public water systems based on 
three criteria (1) the number of persons served by the system, (2) the 
source of the system's water supply, and (3) the contaminants likely 
to be found. Under UCMR3, EPA varies the monitoring frequency for 
systems based on a system's size and water supply (e.g., surface water 
or groundwater), but it does not vary it based on the type of 
contaminant likely to be found. Specifically, under UCMR3, surface 
water systems and groundwater systems under the direct influence of 
surface water[Footnote 27] are required to conduct quarterly 
monitoring of selected chemical contaminants, while groundwater 
systems are required to monitor selected chemical contaminants twice a 
year. In addition, nondisinfecting groundwater systems that serve 
1,000 or fewer people and are located in sensitive hydrogeological 
areas are required to allow EPA monitoring for two microbial 
contaminants (viruses) twice a year (see table 4). The virus 
monitoring in UCMR3 was designed in consideration of where these 
viruses may be more likely to be found--small groundwater systems that 
do not disinfect. Nonetheless, EPA used the standard monitoring 
frequency for viruses; it did not vary the frequency or schedule based 
on the contaminants likely to be found. Similarly, contaminant 
monitoring under UCMR2 occurred quarterly for surface water systems 
and twice a year for groundwater systems for all contaminants. EPA did 
not vary the monitoring frequency or schedule during the seasons when 
pesticides were applied and the likelihood of occurrence was highest. 

Table 4: Monitoring Frequency by System Size and Water Supply Source 
for the Third Unregulated Contaminant Monitoring Rule: 

Assessment monitoring: 

System size: All systems serving more than 10,000 people and a 
representative sample of 800 systems serving 10,000 or fewer people; 

Water supply source: Surface water or groundwater under the direct 
influence of surface water; 
Monitoring frequency: Quarterly during a 1-year period with sampling 
events 3 months apart. 

Water supply source: Groundwater; 
Monitoring frequency: Twice in a 1-year period with sampling events 5-
7 months apart. 

Screening survey: 

System size: All systems serving more than 100,000 people, a 
representative sample of 320 systems serving 10,001 to 100,000 people, 
and a representative sample of 480 systems serving 10,000 or fewer 
people; 

Water supply source: Surface water or groundwater under the direct 
influence of surface water; 
Monitoring frequency: Quarterly during a 1-year period with sampling 
events 3 months apart. 

Water supply source: Groundwater; 
Monitoring frequency: Twice in a 1-year period with sampling events 5-
7 months apart. 

Pre-screen testing: 

System size: A representative sample of 800 nondisinfecting groundwater 
systems serving 1,000 or fewer people; 
Water supply source: Groundwater; 
Monitoring frequency: Twice in a 1-year period with sampling events 5-
7 months apart. 

Source: GAO analysis of EPA documents. 

[End of table] 

When EPA first established the UCMR program, it explained that the 
standard monitoring frequency for public water systems using surface 
water and groundwater under the direct influence of surface water will 
capture variability in contaminant occurrence because one of each 
public water system's quarterly sampling events must occur during what 
EPA considers to be the most vulnerable period for drinking water 
contamination, i.e., the period of peak contamination.[Footnote 28] 
According to EPA's explanation in the Federal Register accompanying 
the final rule for UCMR1, studies show that the greatest risk of 
contaminant occurrence for much of the country is the late-spring/ 
early-summer runoff period from May to July, particularly for 
contaminants such as pesticides and nitrate.[Footnote 29] EPA designed 
the UCMR3 program to include sampling during this period, as 
designated by EPA or the state. Quarterly monitoring for systems using 
surface water or groundwater under the direct influence of surface 
water occurs during one of three possible time frames (1) January, 
April, July, and October; (2) February, May, August, and November; or 
(3) March, June, September, and December. EPA assigns one of these 
three time frames to each participating public water system using 
surface water or groundwater under the direct influence of surface 
water.[Footnote 30] EPA did not specify a vulnerable period for public 
water systems using groundwater in UCMR2 and UCMR3, but it stated that 
samples will be taken during all months and seasons of the year when 
all groundwater systems are considered together, according to agency 
documents.[Footnote 31] 

However, several experts we surveyed and studies on contaminant 
occurrence indicated that the UCMR standard monitoring frequency may 
not accurately characterize the presence of certain contaminants in 
public water systems. In particular, the monitoring frequency may 
result in inaccurate estimates of the occurrence of sporadically 
occurring microbes (e.g., viruses) and pesticides whose peak 
concentrations typically occur during and after the application 
season, primarily from April through July. For example, EPA's 
monitoring frequency may underestimate occurrence for compounds such 
as seasonally applied pesticides, according to one expert we surveyed. 

As noted above, the UCMR3 standard monitoring frequency may result in 
inaccurate estimates of the occurrence of some microbes and pesticides. 

* Microbes. Under UCMR3, EPA selected a representative sample of 800 
nondisinfecting groundwater systems located in sensitive 
hydrogeological areas[Footnote 32] to monitor for two viruses--
enterovirus and norovirus--using the pre-screen testing monitoring 
approach.[Footnote 33] In the notice for its proposed rule for UCMR3, 
EPA states that the agency limited its selection to those systems 
serving 1,000 or fewer persons because these smaller systems are the 
least likely to disinfect the water and, therefore, the most 
vulnerable to contamination with viruses. Monitoring and analyses are 
organized and paid for by EPA through direct contracts with 
laboratories. Under UCMR3, each of the selected systems will be 
monitored for the viruses twice during a 12-month period from January 
2013 to December 2015. The systems' sampling events will be 5 to 7 
months apart. However, more than half of the experts who completed our 
survey (23 of 40) said the twice-yearly monitoring requirement for 
viruses is not likely to provide EPA sufficient data to determine 
whether the contaminants occur with a frequency and at levels of 
public health concern. In addition, 15 experts who completed our 
survey stated that more frequent monitoring may be warranted to 
capture the occurrence of viruses in public water systems, which many 
characterized as variable. Similarly, a technical presentation that 
EPA used to support its rationale for including the viruses in UCMR3 
shows that virus occurrence in groundwater systems is highly variable 
and that groundwater recharge events, like heavy precipitation, 
increase virus contamination levels.[Footnote 34] The technical 
presentation concludes that more frequent sampling is necessary to 
characterize virus occurrence in public water systems. Moreover, in 
the past, EPA varied the monitoring frequency for a microbe that 
public comments on EPA's proposed rule for UCMR1 characterized as 
sporadically occurring.[Footnote 35] Specifically, under UCMR1, 
surface water systems and groundwater systems monitoring for the 
microbial contaminant, Aeromonas, were required to take samples 
quarterly, with additional samples taken, so that samples generally 
were collected each month during the warmest months of the year (July, 
August, and September) because EPA officials told us that this is when 
they expected to find contaminant occurrence.[Footnote 36] Systems 
monitored during one of three possible time frames (1) January, April, 
July, August, September, and October; (2) February, May, July, August, 
September, and November; or (3) March, June, July, August, September, 
and December. In contrast, UCMR3 does not require any additional 
monitoring for viruses. As a result, the data obtained under UCMR3 may 
not provide reliable estimates of the occurrence of viruses in 
drinking water. 

* Pesticides. Under UCMR2, all public water systems serving 100,000 or 
more people and a representative sample of 800 systems serving fewer 
than 100,000 people were required to monitor for nine pesticides using 
the screening survey monitoring approach. These pesticides included 
three widely used herbicides applied for weed control on crops such as 
corn, cotton, and soybeans, and six degradates.[Footnote 37] 
Contaminant monitoring occurred quarterly for surface water systems 
and biannually for groundwater systems. No additional monitoring was 
conducted during the seasons when pesticides were applied and the 
likelihood of occurrence was highest. Two of the 14 experts who 
participated in the first round of our survey said that the UCMR 
standard monitoring frequency may result in inaccurate estimates of 
the occurrence of such pesticides in public water systems. For 
example, one expert said that, following application, pesticides can 
occur at relatively high concentrations but for short periods of time, 
and that, because the UCMR's timing and frequency of monitoring are 
not based on this consideration, the program could completely miss 
that occurrence. Data from a United States Geological Survey (USGS) 
report also show that concentrations of herbicides are usually highest 
during and after the application season, primarily from April through 
July.[Footnote 38] For example, at four stream sites studied, USGS 
found that the presence of atrazine, the dominant herbicide used in 
the Corn Belt during the study period, typically peaked after 
applications in the spring and gradually declined through the summer 
and winter. According to the USGS report, such seasonal patterns of 
pesticide concentrations are generally consistent each year, as long 
as the pesticides are still in use. Pesticide occurrence in 
groundwater also generally coincides with pesticide use patterns. 
Sampling was conducted throughout the year, including intensive 
sampling during the times of highest pesticide use and runoff. This 
sampling approach enabled USGS to conclude that concentrations of 
pesticides in streams and groundwater were typically below water-
quality benchmarks for human health. Thus, seasonal monitoring of 
pesticides is important to assess whether pesticides occur at 
concentrations that may affect human health. Also, high-frequency 
sampling during seasons when intense pesticide use coincides with 
periods of high runoff may be needed to ensure that the highest 
pesticide concentrations are detected. In contrast, EPA's standard 
monitoring frequency requires most public water systems to monitor for 
contaminants only once per season or once over two seasons. As a 
result, some of the data collected under UCMR2 may not yield accurate 
estimates of pesticide occurrence in public water systems. 

EPA Considered 10 Factors to Select Contaminants for Monitoring, but 
Its Selection Process Faced Limitations: 

EPA used 10 selection factors, as well as a working group, and 
internal and external reviews, among other steps, to select 
contaminants for UCMR3. However, the process was lengthy due primarily 
to the reviews and the time required to develop analytical methods. 
The selection process was also limited by a statutory cap of 30 
contaminants, which restricted EPA's ability to collect data on 
additional contaminants that could have been monitored for very little 
additional cost. The selection process also excluded contaminants 
previously monitored in earlier UCMR cycles, which has some 
limitations. 

UCMR3 Contaminant Selection Considered Multiple Factors, but the 
Process Was Lengthy: 

To select contaminants for UCMR3, EPA considered 10 factors and used a 
prioritization process to identify potential contaminants, but took 
over 30 months, or 2 ½ years, to develop and publish the final list in 
the Federal Register after the CCL was published. Had EPA consistently 
met the initial and recurring statutory deadlines for CCL and UCMR, 
they would have been separated by 18 months. 

Contaminant Selection Factors: 

EPA used 10 factors to select contaminants for UCMR3 (see figure 1). 
Officials told us that a contaminant does not have to meet all of the 
selection factors to be chosen; however, three are very important to 
consider as follows: 

* Analytical method availability. Because analytical methods are the 
means by which laboratories test for contaminants, EPA officials told 
us they could not include a contaminant on the UCMR list without also 
having an analytical method available that is sensitive and precise 
enough to provide the data quality and accuracy that EPA needs for the 
UCMR program. See appendix III for more information on analytical 
methods. 

* Health effects information. Because of the importance of monitoring 
for UCMR contaminants that may pose a public health risk and, 
therefore, may need to be regulated, EPA officials told us they placed 
a high priority on contaminants for which they have reliable health 
effects data. 

* Need for data to support regulatory determinations on priority 
contaminants. Because the goal of the UCMR program is to provide 
occurrence data on unregulated contaminants for informing regulatory 
determinations, the Standards and Risk Reduction Branch's priorities 
for data to support a regulatory determination are heavily considered. 

Figure 1: EPA Selection Factors for Identifying Contaminants for 
Monitoring under the Third Unregulated Contaminant Monitoring Rule 
(UCMR3): 

[Refer to PDF for image: list] 

1. Whether an analytical method was available and sensitive enough to 
detect contaminants at the desired level;[A] 

2. Whether information on the health effects of exposure to a 
contaminant was available and reliable; 

3. Whether EPA's Standards and Risk Reduction Branch identified the 
contaminant as a priority for obtaining occurrence data in order to 
support a regulatory determination; 

4. Whether the contaminant had already been monitored under a prior 
Unregulated Contaminant Monitoring Rule (UCMR); if so, it was 
eliminated from further consideration;[A] 

5. Whether the contaminant is listed on the Contaminant Candidate 
List, which serves as the primary basis for developing the UCMR 
list;[A] 

6. Whether EPA identified it as a contaminant of emerging concern;[A] 

7. Whether industry trends show an increase or decrease in the use of 
certain chemicals that could be contaminants; 

8. Whether stakeholders' input through EPA's public meetings, UCMR3 
working group, and public comments on the proposed UCMR rule identify 
a contaminant as a priority; 

9. Whether a contaminant could be monitored using the same analytical 
method being used for contaminants already chosen for the UCMR, which 
would make monitoring it cost-effective; and; 

10. Whether monitoring data from other organizations, such as USGS and 
states, indicate that the contaminant is likely to be present in 
drinking water. 

Source: GAO analysis of EPA information. 

[A] EPA identified these selection factors in the Federal Register. EPA 
officials identified the other factors during interviews with us. 

[End of figure] 

Contaminant Selection Process: 

According to EPA documents and officials, EPA used a prioritization 
process to select potential UCMR3 contaminants. First, EPA considered 
the contaminants listed on CCL3 and identified contaminants of 
emerging concern[Footnote 39] to establish an initial list of about 
150 contaminants. Contaminants for which analytical methods were not 
available were then eliminated from consideration for UCMR3, however 
EPA officials told us that those contaminants may be monitored under a 
future UCMR if an analytical method can be developed. Contaminants 
that were monitored under UCMR1 or UCMR2 were also eliminated from 
further consideration early in the UCMR3 selection process, as 
reported in the Federal Register and agency records.[Footnote 40] 
These two selection factors reduced the list to fewer than 35 
contaminants, according to EPA documentation. In addition, officials 
told us they met regularly with representatives from the Standards and 
Risk Reduction Branch to ensure the list of UCMR contaminants would 
meet their needs for data to support regulatory determinations. 

EPA finalized the UCMR3 list through deliberations with other EPA 
staff and stakeholders and by holding internal and external reviews, 
among other steps (see figure 2). In November 2009, a UCMR3 working 
group led by EPA's Technical Support Center and composed of EPA 
national and regional staff, as well as representatives from states, 
began to select the final 30 contaminants. The EPA Office of Science 
and Technology[Footnote 41] conducted health effects evaluations for 
possible UCMR3 contaminants, according to EPA documents and agency 
officials. The UCMR3 working group, whose members included experts 
from the Office of Science and Technology, reviewed this health 
effects information. For each UCMR3 chemical contaminant, EPA 
developed an information sheet that listed known health effects 
information and the sources of that information, among other things. 
EPA used these contaminant information sheets for internal 
considerations and published them in the EPA docket for UCMR3. After 
the working group discussed the list of contaminants, the Technical 
Support Center submitted the proposed and final UCMR contaminant lists 
and associated rules to EPA management for their review. The EPA 
internal reviews involved different offices within EPA, including the 
Office of the EPA Administrator. EPA also submitted the proposed and 
final UCMR lists and rules to the Office of Management and Budget 
(OMB) for its review. On March 3, 2011, EPA published a proposed list 
of 30 contaminants and the associated rule in the Federal Register. 
EPA published the final UCMR3 contaminant list and associated rule on 
May 2, 2012 (see appendix II for list of UCMR3 contaminants). 

Figure 2: Contaminant Selection and Rule Development Process for the 
Third Unregulated Contaminant Monitoring Rule: 

[Refer to PDF for image: process illustration] 

These steps involve regular communication between staff and management: 

UCMR3 working group develops proposed rule: Nov. 2009-June 2010. 

EPA management review: June-Sept. 2010; (Key decision point involving 
input from senior management). 

OMB review: Sept.-Dec. 2010. 

EPA final review of proposed UCMR3 rule: Jan.-Feb. 2011; (Key decision 
point involving input from senior management). 

Proposed UCMR3 rule published in Federal Register: Mar. 2011. 

These steps involve regular communication between staff and management: 

EPA develops final UCMR3 rule: May-Aug. 2011. 

EPA management review: Aug.-Nov. 2011; (Key decision point involving 
input from senior management). 

OMB review: Nov. 2011-Mar. 2012. 

EPA review and final approval: Mar.-Apr. 2012; (Key decision point 
involving input from senior management). 

Final UCMR3 published in Federal Register: May 2012. 

UCMR = Unregulated Contaminant Monitoring Rule. 

Source: GAO analysis of EPA information. 

[End of figure] 

The UCMR3 selection process led EPA to select many contaminants from 
the CCL, EPA's list of the unregulated drinking water contaminants 
that may require regulation because they are known or anticipated to 
occur in public water systems. Specifically, 22 of the UCMR3 
contaminants were also listed on CCL3. The CCL3 process was developed 
using substantial expert input and recommendations from various 
groups, including the National Academy of Science's National Research 
Council, the National Drinking Water Advisory Council, and EPA's 
Science Advisory Board, according to EPA documents. In addition, 37 of 
the 40 experts who participated in our survey said that CCL3 was a 
reasonable source for EPA to use when it began identifying 
contaminants for UCMR3. UCMR3 also included 7 contaminants of emerging 
concern that were not on CCL3, and total chromium.[Footnote 42] 
Further, 33 of the 40 experts said it was worthwhile for EPA to select 
contaminants for the UCMR that are not on the CCL. For example, one 
expert in our survey told us that including such contaminants is the 
best way to identify emerging contaminants that come to light between 
the publication of the CCL and UCMR list. Overall, 30 of the 40 
experts who participated in our survey reported that the contaminants 
EPA selected for UCMR3 warrant monitoring. 

Contaminant Selection Process Was Lengthy: 

EPA's UCMR3 contaminant selection and rule development process took 
more than 2 ½ years to complete. EPA started the formal UCMR3 
selection process in November 2009 but did not publish the final UCMR3 
list until May 2012. SDWA specifies that EPA publish a new UCMR list 
every 5 years. Whereas EPA's previous list was issued in 2007, it met 
the 5-year time frame required by SDWA. However, the UCMR program is 
no longer staggered 18 months after the CCL, as SDWA intended. Rather, 
UCMR3 was issued 31 months after CCL3. Based on the length of its 
process, EPA would have had to begin selecting contaminants in 
September 2008 to have issued UCMR3 within 18 months of CCL3. Further, 
EPA cannot implement the UCMR program until it publishes the final 
UCMR list and rule in the Federal Register, which delays the onset of 
the data collection period. This delay, in turn, delays the 
availability of the UCMR data for making regulatory determinations 
(see figure 3). 

Figure 3: UCMR3 Timeline: 

[Refer to PDF for image: timeline] 

October 2009: CCL3 final list. 

November 2009: UCMR3 selection begins. 

November 2009 - May 2012: UCMR3 contaminant selection and rulemaking. 

May 2012 (31 months after CCL3): UCMR3 final list. 

January 2013 - December 2015: UCMR3 contaminant monitoring. 

Spring 2015: Regulatory determinations 3. 

CCL = Contaminant Candidate List. 
UCMR = Unregulated Contaminant Monitoring Rule. 

Source: GAO. 

[End of figure] 

The UCMR3 contaminant selection and rulemaking process concluded 31 
months after the CCL was issued primarily because EPA waited to 
initiate its working group until CCL3 was finalized, and the EPA and 
OMB reviews were lengthy. First, EPA's UCMR3 working group did not 
begin its work until November 2009, after the final CCL3 had been 
published in October 2009. EPA officials told us that discussions 
about potential contaminants for UCMR3 and analytical methods that 
were likely to be available to detect them began earlier, but they 
waited until November 2009 to begin the formal selection process, in 
part, because of staffing needs. The staff who developed UCMR3 were 
also coordinating and providing technical support to the public water 
systems that were monitoring for contaminants under UCMR2, which took 
place between January 2008 and December 2010. In addition, CCL staff 
serve on the UCMR working group, and UCMR staff also serve on the CCL 
working group, which EPA officials told us limits their availability. 
Second, EPA internal reviews, OMB reviews, and the revisions to the 
UCMR3 proposed and final rules prompted by the reviewers' comments 
added 16 months in total to the process. However, these reviews are 
not always required for this type of activity, according to agency 
guidance. Specifically, EPA's UCMR is conducted under the agency's 
Action Development Process, a process used for developing rules. The 
agency incorporated EPA management and OMB reviews into their process; 
however, these reviews are not always required for a program like the 
UCMR, according to agency guidance, as stated earlier.[Footnote 43] 

In addition, the EPA analytical method development process is complex, 
technically challenging and lengthy, which agency officials also told 
us presents a challenge for EPA to meet its deadline. EPA officials 
said it typically takes 1 to 2 years to develop a standard method that 
meets the requirements needed for the UCMR program and up to 3 years 
or longer for more complex methods. In addition, EPA generally has not 
used a specific analytical method in more than one UCMR cycle, 
although some of those older methods could be used to monitor 
unregulated contaminants for which EPA does not yet have UCMR 
occurrence data. For example, an analytical method used under UCMR1 
could have detected seven CCL3 contaminants, but this method was not 
included in UCMR3. According to industry organizations and EPA 
officials, EPA did not formally reach out to stakeholders when 
developing analytical methods for UCMR 1, 2, or 3. However, EPA is 
formally reaching out to stakeholders to get their input on the 
development of analytical methods that may be used for UCMR4. In May 
2013, EPA held a public meeting and webinar to provide stakeholders 
with an update on analytical methods that EPA has completed or has 
under development for potential use in the UCMR program and requested 
input from attendees to help address some of the technical challenges 
it faces. 

Statutory Cap of 30 Contaminants Limits EPA's Ability to Obtain 
Additional Data at Little Extra Cost: 

The statutory cap limiting the UCMR program to 30 unregulated 
contaminants every 5 years (i.e., each UCMR cycle) impedes EPA's 
ability to obtain occurrence data on additional contaminants that 
could be monitored for little additional effort or cost. Many of the 
analytical methods that EPA is using for UCMR3 are able to test a 
single sample of drinking water for more than one contaminant at a 
time. Therefore, data on additional contaminants could be obtained 
without additional sampling by the public water systems. EPA is 
allowing laboratories to use eight EPA-developed analytical methods to 
test for UCMR3 contaminants and four methods developed by others. 
These methods can detect a total of 67 unregulated contaminants, 
including 4 listed on CCL3 and 8 contaminants of emerging concern, 
according to our analysis of EPA data. There is broad recognition 
throughout the federal government that agencies need to operate 
programs efficiently and maximize efficiencies when possible.[Footnote 
44] In addition, EPA officials who use the UCMR data to make 
regulatory determinations, while sensitive to the practical 
limitations of monitoring for additional contaminants, said that it 
would be helpful to have additional UCMR data if a method already in 
use for UCMR could detect the occurrence of additional contaminants. 
However, because of the SDWA limit of 30 contaminants, EPA cannot 
always take advantage of this efficiency and is unable to gain 
economies of scale by monitoring for additional contaminants using 
monitoring that is already under way. 

Specifically, our analysis indicates that four CCL3 chemicals--1,1,1,2-
Tetrachloroethane; n-propylbenzene; sec-butylbenzene; and methyl 
tertiary-butyl ether (MTBE)--could have been detected using one of the 
UCMR3 analytical methods. As noted earlier, CCL3 represents EPA's list 
of the unregulated drinking water contaminants that may require 
regulation because they are known or anticipated to occur in public 
water systems. EPA had included n-propylbenzene and sec-butylbenzene 
in the proposed list of UCMR3 contaminants, but it did not include 
them on the final UCMR3. Instead, EPA included chromium-6, as well as 
total chromium, which EPA stated became higher priorities for 
monitoring for the agency. Public comments on the draft UCMR3 list 
also supported monitoring these two contaminants. EPA's Standards and 
Risk Reduction Branch, which develops the regulatory determinations, 
supported the inclusion of the two chromiums, as did a majority of the 
experts who participated in our survey (35 of 40 supported adding 
chromium-6 and 28 of 40 supported adding total chromium). However, by 
including these two higher priority contaminants, EPA removed two 
others in light of the cap of 30 contaminants and could not take 
advantage of economies of scale by using a method already selected for 
UCMR3 to obtain additional data. In addition, MTBE could also have 
been monitored using the same method. However, EPA officials told us 
they decided not to include MTBE, which was monitored under UCMR1, 
because they did not need additional information about it. As noted 
earlier, the priorities of EPA's Standards and Risk Reduction Branch 
are a key selection factor for which contaminants to include on the 
UCMR. In the absence of the limit of 30, EPA would have had an 
opportunity to monitor for these CCL3 chemicals, as well as 1,1,1,2-
Tetrachloroethane, as part of UCMR3 at nominal extra cost. EPA 
officials told us that adding a contaminant to an ongoing analysis 
typically costs about $5 to $10. 

Similarly, our analysis indicates that eight contaminants of emerging 
concern known as perfluorinated chemicals, or PFCs, could also have 
been detected by the same analytical method that EPA is using to 
monitor six other PFCs (see table 5).[Footnote 45] In 2009, EPA 
published an action plan to address some of these chemicals, which the 
agency describes as persistent in the environment and in 
people.[Footnote 46] This action plan also states that it can 
reasonably be anticipated that continued exposure to PFCs could harm 
human health. Due to industry's increasing use of various PFCs, EPA 
included six of the chemicals identified in this plan in UCMR3, 
according to EPA officials. However, according to our analysis, five 
additional PFCs from the plan are among the eight that could have been 
detected using the same analytical method but are not on UCMR3. EPA 
selected the six PFCs it thought most important, but the statutory 
limit of 30 contaminants curtailed EPA's opportunity to gather 
occurrence data on eight additional contaminants of emerging concern 
at little additional cost or effort. EPA could have included the 
additional PFCs in UCMR3 but doing so would have reduced the number of 
contaminants selected from CCL3, EPA's list of the unregulated 
drinking water contaminants that may require regulation because they 
are known or anticipated to occur in public water systems. 
Nonetheless, the agency will not have UCMR data on whether or not 
those eight contaminants are present in the nation's drinking water 
until the next UCMR cycle at the very earliest; UCMR4 data collection 
is expected to be completed in 2021. However, EPA has not decided 
which contaminants will be monitored under UCMR4, and so it remains to 
be seen if they are included. 

Table 5: Contaminants of Emerging Concern Detectable with an 
Analytical Method Already in Use: 

Contaminants of emerging concern: 

N-Ethyl perfluorooctanesulfonamido acetic acid; 
N-Methyl perfluorooctanesulfonamido acetic acid; 
Perfluorodecanoic acid; 
Perfluorododecanoic acid; 
Perfluorohexanoic acid; 
Perfluorotetradecanoic acid; 
Perfluorotridecanoic acid; 
Perfluoroundecanoic acid. 

Source: GAO analysis of EPA data. 

Note: All of the contaminants of emerging concern listed in this table 
are perfluorinated chemicals, or PFCs. PFCs are man-made chemicals 
that are used in a wide variety of industrial and commercial products. 
EPA has reported that PFCs have significant adverse effects on 
laboratory animals and wildlife and anticipates that continued 
exposure to PFCs could result in adverse effects on humans. 

[End of table] 

The legislative history of the SDWA Amendments of 1996 generally noted 
concerns with the financial limitations of public water systems and 
their abilities to absorb costs for activities such as monitoring. 
While the number of contaminants being monitored is a factor, costs 
related to analytical methods, such as laboratory equipment and labor, 
are key cost drivers for UCMR monitoring, according to EPA officials. 
EPA officials told us that they carefully consider the cost on public 
water systems when selecting analytical methods for the UCMR program. 
In addition, EPA pays for some of the small public water systems' 
monitoring and must stay within its budget for those costs.[Footnote 
47] EPA documentation indicates that UCMR3 will generally cost public 
water systems $50 to $470 per water sample, per analytical method. EPA 
officials told us that the cost to detect additional contaminants for 
an analytical method already in use was nominal--typically about $5 to 
$10 per water sample, per contaminant. Thus, the cost of adding 
contaminants to an analytical method that is already being used would 
likely pose only a modest cost increase for public water systems. 

UCMR3 Excludes Previously Monitored Contaminants: 

One of EPA's key selection factors for UCMR3 involved generally 
eliminating contaminants that were monitored under a prior UCMR 
round,[Footnote 48] but by eliminating those contaminants, EPA may 
encounter data limitations because (1) the sampling methodology was 
not sufficient, (2) the health information has been updated, or (3) 
the data have become outdated. 

Sampling methodology. One of the sampling methodologies EPA used in a 
prior UCMR did not provide sufficient occurrence data to make 
regulatory determinations for some contaminants, according to EPA's 
regulatory determinations for the CCL2 contaminants.[Footnote 49] 
Specifically, UCMR1 used a limited sample size of 300 public water 
systems for the 14 contaminants it monitored using the screening 
survey approach. This resulted in occurrence data for some 
contaminants that, according to EPA documents, were not sufficient to 
make a national estimate. In its July 2008 Regulatory Determination 
notice, EPA found that a total of 12 contaminants monitored under 
UCMR1 had insufficient occurrence data for making national estimates. 
Three of those contaminants are also listed on CCL3--the pesticides 
disulfoton, diuron, and terbufos. By applying the UCMR selection 
factor to eliminate contaminants monitored under previous UCMR rounds, 
EPA did not consider remonitoring these contaminants in UCMR3, even 
though EPA previously found the existing data were not sufficient to 
make national estimates of contaminant occurrence. However, 31 of the 
40 experts who participated in our survey said it would be worthwhile 
to monitor UCMR contaminants again if they had insufficient data due 
to the sampling size. 

According to EPA's notice in the Federal Register for UCMR1, EPA's 
original intent was to use the screening survey as a first indicator 
of potential contamination and then to remonitor the contaminants 
again using assessment monitoring to obtain nationally representative 
results if the contaminants occurred frequently enough in drinking 
water to warrant doing so.[Footnote 50] However, EPA officials told us 
they later determined that this approach was not feasible because of 
the time it required and did not follow through with its plan to 
remonitor the UCMR1 contaminants. For the UCMR2 and UCMR3 cycles, EPA 
revised its approach and increased the screening survey sample size 
from 300 to 800. Specifically, under UCMR1, the screening survey 
included a sample of 300 public water systems.[Footnote 51] Under 
UCMR2 and UCMR3, the screening survey included a sample of 800 public 
water systems and all of the systems serving more than 100,000 people, 
for a total of approximately 1,200 systems.[Footnote 52] This increase 
allowed EPA to make nationally representative estimates based on the 
occurrence data. 

Health effects information. According to EPA officials, health effects 
information for unregulated contaminants is often uncertain and in 
flux because research is often being conducted at the same time UCMR 
monitoring is occurring. Those officials said the new research 
sometimes indicates that health effects occur at lower levels of 
exposure than previously thought. In recognition of this, EPA set the 
monitoring levels, or MRLs, for UCMR3 based on how low the analytical 
method could reliably measure specific contaminants, as reported 
earlier. However, EPA previously set the MRLs for UCMR1 and UCMR2 with 
consideration of the levels of public health concern that were known 
at the time. In May 2011, we reported that EPA set the MRLs for some 
of the UCMR1 and UCMR2 contaminants before it had developed the health 
reference levels for those contaminants. EPA subsequently determined 
that those MRLs exceeded the health reference levels. As a result, 
occurrence of those contaminants at a level higher than the health 
reference level but lower than the MRL might not have been detected, 
and UCMR1 and UCMR2 occurrence data might not accurately capture the 
risk these contaminants pose to public health. Similarly, if the 
health effects information about the UCMR1 and UCMR2 contaminants 
changes in the future, EPA may not have the data it needs to support 
regulatory determinations. Thirty-two of the 40 experts who 
participated in our survey said it would be worthwhile to monitor 
again for UCMR1 and UCMR2 contaminants if new health effects 
information becomes available. 

Age of data. The UCMR data estimated contaminant occurrence at the 
time that they were collected. Due to changing industrial and 
agricultural chemical use and environmental conditions, UCMR data may 
not present an accurate picture of current contaminant occurrence 
after several years. In particular, the UCMR1 data, which were 
collected from 2001 to 2003, are now over 10 years old. Nineteen of 
the 40 experts who participated in our survey said that UCMR data this 
old are not likely to be a reliable indicator of the frequency and 
level of drinking water contamination, and 18 said they did not have 
enough information to judge how long the UCMR data are likely to be a 
reliable indicator. EPA officials and experts told us the shelf life 
of the UCMR data generally depends on the type of contaminant and 
whether its use is likely to have changed over time. According to the 
officials, concentrations of naturally occurring contaminants remain 
relatively constant in drinking water over time but concentrations of 
pesticides and industrial chemicals are subject to change. While some 
of the UCMR1 data may still be a reliable indicator of contaminant 
occurrence, several of the UCMR1 contaminants are pesticides and 
industrial chemicals and the extent of their use has changed since the 
UCMR data were collected. For example, use of the pesticides 
disulfoton and molinate effectively ended following EPA actions in 
2009 and 2004, respectively.[Footnote 53] Use of the pesticides 
terbufos and diuron also decreased significantly from 2003 to 2009, 
according to USGS data. In addition, use of two industrial chemicals, 
MTBE and nitrobenzene, decreased over time, according to EPA data. 
MTBE, a former gasoline additive, has not been used in significant 
quantities since 2005. Nitrobenzene use has fluctuated over time but 
decreased sharply from 1998 to 2005.[Footnote 54] Because the UCMR1 
data estimate contamination at a point in time prior to these changes, 
they likely overestimate current levels of contaminant occurrence. 

EPA might choose not to make regulatory determinations on some 
contaminants if it concludes that the available occurrence data do not 
provide sufficient support, as it did for several contaminants in 
2008. As we reported in May 2011, EPA did not make regulatory 
determinations for some contaminants monitored under a prior UCMR 
round because the occurrence data were not sufficient to do so. 
[Footnote 55] SDWA requires EPA to use the best available data, 
[Footnote 56] but by not remonitoring contaminants when appropriate, 
such as when the data collected in an earlier cycle have become 
outdated, EPA may not have sufficient data to support its regulatory 
determinations. SDWA does not prohibit EPA from remonitoring UCMR 
contaminants; therefore, the agency has the option to do so, for 
example when it believes that remonitoring would provide better 
support for its regulatory determinations. EPA officials told us that 
when developing future UCMRs, they will consider whether adequate 
occurrence data for a contaminant had already been collected under a 
prior UCMR, rather than generally eliminating all previously monitored 
contaminants as part of its contaminants selection process. They also 
said that if the officials who make the regulatory determinations 
asked for a contaminant to be remonitored, they would consider doing 
so. 

UCMR Data Support Regulatory Determinations, but Time Lag Results in 
Delays and Reliance on Older Data: 

EPA uses the UCMR data to support regulatory determinations but faces 
a time lag when doing so. Agency officials told us UCMR data are EPA's 
primary source of data on the occurrence of unregulated contaminants 
in drinking water and are a key input for regulatory determinations. 
However, by the time the complete set of UCMR data is available, the 
recurring 5-year deadline SDWA established for making at least 5 
regulatory determinations has passed, leading EPA, in some cases, to 
rely on older data that may no longer accurately reflect current 
contaminant occurrence in drinking water.[Footnote 57] 

UCMR Data Are EPA's Primary Occurrence Data Source and a Key Input for 
Regulatory Determinations: 

EPA uses the UCMR data, along with information about the contaminants' 
health effects and other information, to make regulatory 
determinations--that is, decisions on whether or not new primary 
drinking water standards are needed for particular unregulated 
contaminants. Regulatory determinations can be positive (i.e., a new 
primary standard is needed) or negative (i.e., no new primary standard 
is needed).[Footnote 58] While SDWA requires EPA to make regulatory 
determinations for at least five CCL contaminants every 5 years, EPA 
officials told us their goal is to make as many regulatory 
determinations as possible, given the available data.[Footnote 59] EPA 
used UCMR1 data to support 9 of the 11 regulatory determinations it 
made in 2008 for the CCL2 contaminants and another regulatory 
determination in 2011. It is currently using UCMR1 and UCMR2 data to 
inform its regulatory determinations for the CCL3 contaminants, which 
EPA expects to issue in 2015. The complete UCMR3 data set will not be 
available until after those regulatory determinations have been 
issued. However, if the timing allows, EPA officials told us they may 
have an opportunity to use a partial UCMR3 data set to corroborate 
other data they have on contaminant occurrence when making the final 
regulatory determinations. 

EPA officials told us that, when they are available, UCMR data are 
EPA's primary data source on unregulated contaminants in drinking 
water because they allow EPA to make national estimates of drinking 
water contamination, as discussed earlier in this report. EPA also has 
data on unregulated contaminants from other sources, such as USGS and 
some states, but those data do not provide nationally representative 
estimates of drinking water contamination. For example, the USGS data 
estimate contamination in source water (i.e., untreated water from 
rivers, lakes, and groundwater aquifers). Since source water can 
become public drinking water, the USGS data can indicate potential 
drinking water contamination. However, these data do not provide 
conclusive information on whether contamination is occurring in public 
water systems. EPA officials told us they use the other data to 
provide additional support for regulatory determinations, but they 
generally do not provide sufficient support on their own. In contrast, 
the UCMR data are generally sufficient for estimating nationwide 
occurrence of drinking water contamination, with the exceptions noted 
earlier in this report. In addition, 35 of the 40 experts who 
participated in our survey said that the UCMR program has been an 
effective way to collect sufficient data to determine whether 
contamination is occurring frequently and at levels of public health 
concern in drinking water. Several experts provided additional 
comments on the overall effectiveness of the UCMR program. For 
example, one expert said that considering the great need for data and 
EPA's limited resources, the UCMR program has been one of the more 
effective tools in the agency's rule development process. Another 
expert said that, given statutory requirements, resources, and data 
availability, EPA has done an admirable job with the UCMR program. 
Finally, one expert said that the UCMR program has been an effective 
way to remove contaminants from the radar screen when they did not 
warrant regulation. 

Time Lag Delays Regulatory Determinations by a Full 5-Year Cycle and 
Results in the Use of Older Data: 

A time lag between the SDWA deadline for regulatory determinations and 
the UCMR data's availability typically delays determinations on given 
contaminants by a full 5-year cycle. SDWA sets deadlines for 
publishing the CCL and UCMR contaminant list and making regulatory 
determinations, but it does not set a deadline for monitoring the UCMR 
contaminants. SDWA established recurring 5-year deadlines for making 
regulatory determinations and for publishing the UCMR list. EPA has 
established a 3-year monitoring period for the UCMR, but this overlaps 
with the time frame during which EPA makes its regulatory 
determinations. EPA also uses some older data that may no longer 
accurately reflect current contaminant occurrence in drinking water 
when it is making regulatory determinations.[Footnote 60] 

Time Lag Typically Delays Regulatory Determinations for Contaminants 
by a Full 5-Year Cycle: 

A time lag between the statutory deadline for regulatory 
determinations and the UCMR data's availability typically delays 
regulatory determinations on given contaminants until the following 
cycle. SDWA directs EPA to publish a new CCL, UCMR, and set of 
regulatory determinations every 5 years, as described earlier in this 
report. The initial timeline staggered the deadlines such that EPA was 
required to publish the UCMR list 18 months after the CCL and issue 
the regulatory determinations 2 years after that. However, in 
practice, the 2-year window SDWA set from the time EPA publishes the 
UCMR list to when it makes regulatory determinations has not provided 
enough time for the agency to monitor the contaminants and incorporate 
the UCMR data into the determinations. Consequently, UCMR data are not 
available for EPA to make regulatory determinations for contaminants 
during the cycle in which they are monitored; rather, UCMR data for 
contaminants in one cycle typically are not used until the next cycle 
or later (see figure 4). 

Figure 4: UCMR Data Do Not Inform Regulatory Determinations in the 
Same Cycle: 

[Refer to PDF for image: illustration] 

Select contaminants for Contaminant Candidate List (CCL): 

1. Contaminant Candidate List 3 (2009); 
Refer to #2 and #4. 

2. Contaminant Candidate List 4 (2014); 
Refer to #5 and #7. 

Select up to 30 contaminants for Unregulated Contaminant Monitoring 
Rule (UCMR): 

3. Unregulated Contaminant Monitoring Rule 2 (2007); 
Refer to #6. 

4. Unregulated Contaminant Monitoring Rule 3 (2012); 
Refer to #7. 

5. Unregulated Contaminant Monitoring Rule 4 (2016). 

Make regulatory determinations on CCL contaminants using UCMR and 
other data: 

6. Regulatory determinations 3 (2015). 

7. Regulatory determinations 4 (2020). 

Source: GAO analysis of EPA data. 

[End of figure] 

EPA officials told us that the timing of the UCMR program is a key 
challenge, particularly in relation to the timing of the CCL and 
regulatory determination deadlines. Since the UCMR program began, the 
full UCMR data set has not been available until a year or more after 
EPA made its regulatory determinations for the same 5-year cycle. The 
time lag is influenced by the recurring 5-year deadlines for the CCL, 
UCMR, and regulatory determinations and increased when EPA missed the 
deadlines for CCL2 and UCMR2. As a result, UCMR2 data were not 
available until nearly 3 years after EPA made the regulatory 
determinations for the CCL2 contaminants. When EPA missed the UCMR2 
deadline, subsequent deadlines were moved to a later date, consistent 
with the 5-year cycle established by SDWA.[Footnote 61] Accordingly, 
the UCMR3 data will lag behind the regulatory determinations for the 
CCL3 contaminants by over a year. Based on EPA's current schedules for 
the three programs, the UCMR4 data will likewise lag behind the 
regulatory determinations for the CCL4 contaminants. EPA's schedule 
indicates that the agency plans to make regulatory determinations for 
the CCL4 contaminants in spring 2020 and complete UCMR4 monitoring in 
December 2020, with the full UCMR4 data set available in 2021. Without 
congressional action to make the time frames more realistic, the time 
lag and resultant delays are expected to continue in subsequent cycles. 

Three-Year Monitoring Period Contributes to Time Lag: 

EPA has established a 3-year monitoring period for the UCMR, which 
contributes to the time lag between the SDWA deadline for regulatory 
determinations and the UCMR data's availability. Before the monitoring 
period begins, EPA spends about 1 year preparing for the monitoring 
that public water systems will conduct. During this time, EPA 
finalizes the monitoring schedules for systems that will be 
participating in the program and approves the laboratories that will 
conduct the analyses, among other things. Each public water system 
monitors for contaminants during a 12-month window throughout the 3-
year period that follows. EPA officials told us they spend up to 1 
year finalizing the data collection and analysis, which includes about 
6 months for public water systems and laboratories to report, review, 
and approve their data.[Footnote 62] The UCMR3 monitoring, data 
collection, and analysis overlap with the time frame during which EPA 
will be making its regulatory determinations for contaminants on CCL3. 
UCMR3 monitoring is occurring from January 2013 through December 2015. 
Regulatory determinations for the CCL3 contaminants are expected in 
spring 2015, according to agency officials.[Footnote 63] Officials who 
make the regulatory determinations told us that it is helpful to have 
the UCMR data about 2 years before publishing the final regulatory 
determinations. They also told us they start using the data once about 
70 percent of the data set is available, approximately 2 years into 
the 3-year monitoring period. For example, EPA began using the UCMR1 
data set to inform its internal deliberations on the regulatory 
determinations for CCL2 after about 70 percent of the data were 
available, according to agency officials. The officials said that by 
this time in the data collection period they saw very little variation 
in contaminant occurrence (i.e., the data were relatively stable). 

EPA officials told us they selected the 3-year monitoring period 
mainly for logistical reasons impacting the agency and the public 
water systems. They told us that the 3-year monitoring period spreads 
the agency's work out over a more manageable period of time. EPA staff 
provide technical support to the public water systems conducting the 
monitoring, and many also have other responsibilities, according to 
EPA officials. The 3-year monitoring period also provides public water 
systems with more flexibility to decide when they will start 
monitoring. EPA officials told us that, after the agency publishes the 
final UCMR in the Federal Register, they assign each system a month 
and year to start monitoring. However, some systems request a later 
start date to incorporate the monitoring into their budgets and 
operations, according to agency officials. EPA officials also said 
laboratory capacity is sufficient under the 3-year period and has 
approved 53 laboratories to participate in UCMR3. EPA officials had 
told us that about 15 laboratories could probably handle the 
analytical work for UCMR3, therefore, laboratory capacity may be 
sufficient under a shorter monitoring period. Officials also told us 
that the 3-year monitoring period is not a given. For example, 
officials told us that a 2-year monitoring period, while challenging, 
may be possible for future cycles, but they would have to evaluate the 
shorter time frame in detail to determine its feasibility. 

UCMR program officials also told us that monitoring over 3 years is 
intended to allow them to account for the effects that extreme weather 
in a given season could have on the data. They told us they did not 
know whether a shorter monitoring period, such as a 2-year period, 
would also allow them to do this because the extent of seasonal 
variability in the UCMR program is unknown. However, they did not 
expect that a shorter monitoring period would address variability to 
the same extent that the current 3-year monitoring period would 
because the longer period of time provides greater opportunity to 
detect it. Nonetheless, as reported earlier, EPA officials who make 
the regulatory determinations observed very few changes in occurrence 
trends after about 70 percent of the UCMR data had been collected, 
about 2 years into the 3-year monitoring period. When asked, experts 
who participated in our survey had mixed views on whether a 2-year 
monitoring period would be likely to provide EPA with sufficient 
information on seasonal fluctuations.[Footnote 64] Specifically, 19 of 
40 respondents thought that it was likely to, 11 thought it was not 
likely to, and 9 said they did not have enough information to judge. 
The UCMR3 data set may provide EPA with information that it can use to 
assess the extent of seasonal variability, according to agency 
officials. Because UCMR3 is monitoring for contaminants at low MRLs, 
it will likely have a higher rate of detections than prior UCMRs and, 
therefore, more data points that EPA can use to detect seasonal 
variability that it would have missed in the past. EPA could use this 
information to help it understand the extent of seasonal variability 
and the feasibility of a shorter monitoring period. 

EPA Uses Occurrence Data from Previous UCMR Cycles When Current Data 
Are Not Available: 

EPA has adapted to the time lag by using UCMR data from previous 
cycles to make regulatory determinations. For example, EPA is using 
UCMR1 and UCMR2 data to make regulatory determinations for the CCL3 
contaminants.[Footnote 65] Similarly, EPA used UCMR1 data to make 
regulatory determinations for the CCL2 contaminants. Because EPA has 
carried over many contaminants from one CCL to the next, officials 
told us this is a workable approach.[Footnote 66] However, more than 
half of the experts (25 of 40) who participated in our survey did not 
think using data from a prior UCMR cycle was an effective strategy for 
making regulatory determinations. Many of those experts were concerned 
because the time lag delays regulatory determinations on given 
contaminants and leads EPA to use older data that may not reflect 
current contaminant occurrence in some cases. The time lag delays 
regulatory determinations on contaminants by at least one 5-year 
cycle, and sometimes longer. EPA officials told us that most of the 
UCMR3 data will be used to support regulatory determinations for the 
CCL4 contaminants, which are expected to be issued in 2020. Those 
UCMR3 contaminants that are carried over to CCL4 will wait about 10 
years for regulatory determinations unless EPA uses data from another 
source or makes an out-of-cycle regulatory determination (see figure 
5).[Footnote 67] 

Figure 5: About 10 Years Typically Pass before Regulatory 
Determinations Are Made for Contaminants on the CCL: 

[Refer to PDF for image: timeline] 

October 2009: CCL3 lists contaminant “A.” 

October 2014: CCL4 re-lists contaminant “A.” 

Spring 2015: Regulatory determinations 3; contaminant “A” not 
considered[A]. 

December 2016: UCMR3 data for contaminant “A.” 

October 2019: CCL5 re-lists contaminant “A.” 

Spring 2020: Regulatory determinations 4; contaminant “A”
considered. 

Source: GAO analysis of EPA data. 

CCL = Contaminant Candidate List. 
UCMR = Unregulated Contaminant Monitoring Rule. 

[A] Contaminant "A" not considered because UCMR data are not yet 
available. 

[End of figure] 

Regulatory determinations on given contaminants can also take longer 
to make. For example, 11 of the contaminants EPA included on a short 
list of potential regulatory determinations for CCL3[Footnote 68] were 
first listed on CCL1 in March 1998.[Footnote 69] By the time EPA makes 
its regulatory determinations, about 17 years will have passed since 
they were identified as contaminants known or anticipated to occur in 
public water systems. In addition, there are six contaminants that EPA 
carried over from CCL1 to CCL3 that it has not yet included in a 
regulatory determinations cycle. UCMR data are not yet available for 
any of them, although two--methyl bromide and 1,1-dichloroethane--are 
currently being monitored under UCMR3.[Footnote 70] Methyl bromide is 
a pesticide that was phased out of use in 2005; 1,1,-dichloroethane is 
an industrial chemical. Due to the time lag, EPA's first use of the 
UCMR3 data for these CCL1 contaminants will likely be when it makes 
the regulatory determinations for CCL4.[Footnote 71] This is about 22 
years after they were first identified on the CCL as unregulated 
contaminants that are known or anticipated to occur in public water 
systems and may require regulation. 

EPA is also using some older UCMR data that may not accurately reflect 
current contaminant occurrence in drinking water to make regulatory 
determinations, according to agency documents. Specifically, 6 of the 
32 contaminants EPA included on its short list of potential regulatory 
determinations for CCL3 were monitored from 2001 to 2003 under 
UCMR1.[Footnote 72] EPA officials told us that one of the reasons the 
UCMR1 data were not used for the regulatory determinations it issued 
in 2008 was because health effects information on some of the 
contaminants was not available.[Footnote 73] EPA's July 2008 
regulatory determinations indicate that there were health effects 
information gaps for 3 of the 6 contaminants.[Footnote 74] As noted 
earlier in the report, the UCMR1 data are now over 10 years old and 
may not accurately reflect current levels of drinking water 
contamination in some cases. In particular, the 6 UCMR1 contaminants 
on EPA's short list are industrial chemicals and pesticides, and the 
extent of their use has changed since the UCMR data were collected, as 
described earlier. Four of the 6 contaminants are pesticides: 
disulfoton, molinate, terbufos, and diuron. The other 2 contaminants 
are industrial chemicals: MTBE and nitrobenzene. Further, in its July 
2008 Regulatory Determination notice, EPA stated that it had 
insufficient occurrence data for making national estimates of 
disulfoton, diuron, and terbufos contamination. Occurrence data had 
been collected on those contaminants under UCMR1 but, as discussed 
earlier in this report, the sampling methodology used was not 
sufficient. EPA has not collected new occurrence data on the 
contaminants since then. As a result, the older data may not provide 
an accurate basis for EPA to use when making regulatory determinations. 

Conclusions: 

The overall goal of SDWA is to ensure that public drinking water is 
safe. The UCMR program is an important component of EPA's 
implementation of the act. As EPA's primary data source for assessing 
the nation's potential exposure to unregulated contaminants and the 
levels of that exposure, the UCMR plays a key role in EPA's regulatory 
development process. As a result, the accuracy and timeliness of the 
data it provides are important to support informed decision making. 
The technical changes EPA has made to UCMR3 in response to our 2011 
recommendations are significant improvements over prior UCMR cycles. 
First, by more fully utilizing its statutory authority and monitoring 
for 30 contaminants under UCMR3, EPA has increased the productivity of 
the UCMR program. As a result, more data will be available to inform 
regulatory determinations in the future. Second, by increasing the 
sample size of the public water systems monitoring under the screening 
survey approach, EPA has also increased the statistical strength of 
its monitoring results. The larger sample size should allow EPA to 
make better estimates of exposure to drinking water contaminants. 
Third, by monitoring for contaminants at lower detection levels, or 
MRLs, the UCMR3 data are more likely to remain useful even in light of 
new information about the contaminants' health effects in the future. 

However, some statutory provisions pose limitations for obtaining 
additional and more timely data. First, because SDWA caps the number 
of UCMR contaminants in each 5-year cycle at 30, the agency is not 
able to take advantage of economies of scale by collecting occurrence 
data on additional contaminants. As a result, EPA does not have the 
opportunity to collect data on additional unregulated contaminants at 
little additional effort and cost to the agency or the public water 
systems. The cost to detect additional contaminants with analytical 
methods already being used by the UCMR program can be nominal. Second, 
the 2-year window SDWA initially established from the time EPA 
publishes the UCMR list to when it publishes the regulatory 
determinations does not provide adequate time for the UCMR data to 
support regulatory determinations in the same cycle. As a practical 
matter, the time frames SDWA initially established are not adequate to 
provide enough time for UCMR's data collection period and EPA's 
subsequent use of the data. As a result, regulatory determinations on 
given contaminants are typically delayed by a full 5-year cycle, 
meaning that about 10 years pass from when EPA identifies them on the 
CCL as potential problems and decides whether they need to be 
regulated. Two contaminants--methyl bromide and 1,1-dichloroethane--
will likely wait 22 years after their initial listing for regulatory 
determinations. Without congressional action to make the statutory 
time frames more realistic, the time lag and resultant delays are 
expected to continue in subsequent cycles. 

The UCMR program faces other limitations that may cause EPA to base 
its regulatory determinations on inaccurate or outdated information. 
In particular, EPA's standard monitoring frequency may not be 
accurately estimating the occurrence of certain contaminants in the 
public's drinking water, particularly those that occur sporadically 
like viruses or seasonally like pesticides. By not varying the 
monitoring frequency for contaminants when appropriate, for example by 
increasing the frequency during the months when EPA expects viruses to 
be present, the UCMR data may not provide reliable estimates of 
contaminant occurrence. Similarly, by not remonitoring contaminants 
from prior UCMR cycles when appropriate, EPA may not have sufficient 
data to support its regulatory determinations. EPA continues to use 
data from past UCMR cycles to inform its regulatory determinations, 
but those data sets have potential limitations. For example, EPA 
previously determined that some of the data collected under UCMR1 were 
not sufficient to make national estimates of contaminant occurrence. 
Because EPA has not remonitored those contaminants, the agency still 
does not have sufficient UCMR data about them. Similarly, some of the 
UCMR1 data may be outdated due to changing industrial and agricultural 
chemical use and environmental conditions. As a result, the existing 
UCMR data for certain contaminants may no longer accurately reflect 
current drinking water conditions. We are encouraged by recent 
statements from EPA officials indicating that they will consider 
whether adequate occurrence data for a contaminant had already been 
collected under a prior UCMR when selecting contaminants in the 
future. Finally, the timing of the UCMR program significantly reduces 
the availability of the data to support regulatory determinations. 
Specifically, the time lines for the contaminant selection process and 
monitoring period do not get the UCMR data into the hands of EPA 
officials in time to support regulatory determinations in the same 
cycle. The UCMR data are EPA's preferred source of information on 
contaminant occurrence because the program has been specifically 
designed to yield nationally representative estimates of drinking 
water contamination. However, if the UCMR data are not available, EPA 
either must use data from another source, or delay making regulatory 
determinations on specific contaminants. For most of the UCMR3 
contaminants, over 10 years are likely to pass between the time that 
EPA first listed contaminants on CCL3 in 2009 and the time that 
regulatory determinations are made for those contaminants, expected in 
2020. The SDWA deadlines, EPA's lengthy contaminant selection process, 
and its 3-year monitoring period all contribute to the time lag. As 
stated earlier, to change the statutory deadlines and adjust the time 
frames requires congressional action, and shortening the selection 
process and monitoring period requires EPA action. UCMR program 
officials coordinate closely with the other two EPA offices that make 
regulatory determinations and develop the CCL and are sensitive to the 
impact that timeliness has on each of these processes. We recognize 
that EPA officials must consider a variety of data quality objectives, 
logistical needs, and the burden on public water systems in managing 
the UCMR program, regulatory determinations, and CCL. But, since EPA 
officials consider the UCMR data to be their primary source of 
occurrence data for the regulatory determinations, it is important 
that these data also be timely. However, without a coordinated effort 
to address the time lag, most contaminants on the CCL will continue to 
wait 10 years or longer for a regulatory determination. 

Matters for Congressional Consideration: 

* To take advantage of opportunities to collect UCMR data on 
additional unregulated contaminants, Congress should consider amending 
SDWA to give EPA the flexibility to select more than 30 contaminants 
for monitoring under the UCMR program if high-priority contaminants, 
such as those on the CCL or contaminants of emerging concern, can be 
included at minimal cost, with minimal additional burden on public 
water systems, and while using analytical methods that EPA is already 
employing. 

* To optimize the ability of the UCMR data to support regulatory 
determinations, Congress should consider adjusting the statutory time 
frames for the UCMR and regulatory determinations cycles so that EPA 
can use the UCMR data to support regulatory determinations in the same 
cycle. 

Recommendations for Executive Action: 

We recommend that the Administrator of the Environmental Protection 
Agency direct the Assistant Administrator of the Office of Water to 
take the following three actions: 

* To help ensure that the UCMR data accurately reflect contaminant 
occurrence, vary the monitoring frequency for public water systems 
when the standard UCMR monitoring frequency is not expected to 
accurately detect the presence of contaminants, such as sporadically 
occurring viruses or pesticides that fluctuate seasonally. 

* To help address potential limitations with the UCMR data, when 
considering if adequate data had been collected under a prior UCMR 
during the contaminant selection process, take into account whether: 

- the sampling methodology EPA used in a prior UCMR for a particular 
contaminant provided sufficient occurrence data to make a regulatory 
determination, 

- the contaminant's health information is in flux or has been updated 
and now indicates that adverse health effects occur at a lower level 
than that at which UCMR data were originally collected, or: 

- the data collected in an earlier cycle have become outdated and no 
longer present an accurate picture of contaminant occurrence, for 
example, because industrial or agricultural chemical use has changed. 

* To optimize the ability of the UCMR program to support regulatory 
determinations, convene an internal working group of officials 
responsible for the UCMR, regulatory determinations, and CCL to fully 
examine opportunities to improve the timeliness of the UCMR program. 
Among other things, the working group should consider: 

- whether a shorter process for selecting contaminants would yield a 
high-quality list, or whether the existing process can be started 
sooner, and: 

- whether monitoring for contaminants over a shorter period, instead 
of the current 3-year period, is feasible, given data quality, 
logistical needs, and the burden on public water systems. 

Agency Comments: 

We provided EPA with a draft of this report for its review and 
comment. In written comments, reproduced in appendix VI, EPA generally 
agreed with our findings and conclusions and is committed to 
implementing our recommendations when it develops the next UCMR. 

EPA also provided technical comments, which we incorporated as 
appropriate. 

As agreed with your offices, unless you publicly announce the contents 
of this report earlier, we plan no further distribution until 30 days 
from the report date. At that time, we will send copies of this report 
to the Administrator of EPA and the appropriate congressional 
committees. In addition, this report will be available at no charge on 
the GAO website at [hyperlink, http://www.gao.gov]. 

If you or your staff members have any questions about this report, 
please contact me at (202) 512-3841or gomezj@gao.gov. Contact points 
for our Offices of Congressional Relations and Public Affairs may be 
found on the last page of this report. GAO staff who made key 
contributions to the report are listed in appendix VII. 

Signed by: 

J. Alfredo Gómez: 
Director, Natural Resources and Environment: 

[End of section] 

Appendix I: Objectives, Scope, and Methodology: 

This report addressed the following objectives (1) evaluate the extent 
to which the Environmental Protection Agency (EPA) implemented the 
recommendations GAO made in 2011 to improve the Unregulated 
Contaminant Monitoring Rule (UCMR) program and any opportunities that 
may exist to strengthen it further; (2) identify the factors EPA 
considered when it selected the contaminants for monitoring under 
UCMR3 and the limitations, if any, that EPA faced in selecting the 
contaminants; and (3) examine the extent to which UCMR data support 
EPA's regulatory determinations. 

To obtain information for each of our objectives on issues related to 
monitoring unregulated drinking water contaminants, we surveyed a 
broad range of experts in two rounds using the Delphi method. The 
Delphi method consists of surveying experts in two or more rounds to 
obtain group agreement or consensus on key topics. For this report, 
the first round of the survey was comprised of open-ended questions 
that asked experts to provide their views on the strengths and 
weaknesses of EPA's UCMR program, as well as suggestions for improving 
the program. The second round included primarily closed-ended 
questions developed from our content analysis of the first round 
survey responses. To identify experts for the first round of the 
survey, we reviewed public comments provided on EPA's proposed rules 
for its Contaminant Candidate Lists (CCL), UCMR, and preliminary 
regulatory determinations. We also reviewed public nominations of 
drinking water contaminants submitted to EPA and EPA documentation on 
two expert panels it convened to provide input and review of chemical 
and microbial contaminants for CCL3. We identified 14 experts from 
academia, an analytical laboratory, environmental advocacy 
organizations, professional associations, water utilities, and federal 
and state government agencies. These experts were selected based on 
their demonstrated familiarity with EPA's CCL, UCMR, and regulatory 
determination processes. The first round of the survey--which ran from 
December 12, 2012 to January 4, 2013--asked these experts to respond 
to 10 open-ended questions. These questions asked about various 
aspects of the UCMR program that helped or hindered EPA in (1) 
selecting priority contaminants, (2) monitoring for selected 
contaminants, and (3) using the data to support regulatory 
determinations. We also asked the experts whether programmatic or 
statutory changes were needed to address any challenges identified. We 
received responses from 14 of the 14 experts we surveyed in the first 
round of the survey. 

We performed a content analysis of the responses from the first round 
of the survey to identify the most important issues raised by the 
experts. We used the results of the content analysis to develop 
questions for the second round of the survey. Because this was not a 
sample survey, it had no sampling errors. However, the practical 
difficulties of conducting any survey can introduce nonsampling 
errors, such as difficulties interpreting a particular question, which 
can introduce unwanted variability into the survey results. We took 
steps to minimize nonsampling errors by pretesting the questionnaire 
with five experts. This helped ensure that the questions were clear 
and unbiased, and that the questionnaire did not place an undue burden 
on respondents. An independent reviewer within GAO also reviewed a 
draft of the questionnaire prior to its administration. We made 
appropriate revisions to the content and format of the second round 
survey questionnaire based on the pretests and independent review. 

To identify experts for the second round of the survey, we developed a 
preliminary list of experts, which included the 14 experts who 
completed the first round of the survey and others identified through 
interviews with stakeholders and our review of public comments 
provided on EPA's proposed rules. We contracted with the National 
Academy of Sciences to identify additional experts. We selected 48 
experts for the second round of the survey based on their knowledge of 
one or more of the following topic areas: drinking water, contaminant 
monitoring, analytical methods, toxicology, risk assessment, and 
environmental regulation. Similar to the first round of the survey, 
the selected experts were from academia, analytical laboratories, 
environmental advocacy organizations, industry, water utilities, and 
federal and state government agencies. The second round of the survey 
was administered on the Internet from May 10, 2013 to June 7, 2013. 
Experts were sent an e-mail invitation to complete the survey on a GAO 
web server using a unique username and password. To increase the 
response rate, we followed up with e-mails and personal phone calls to 
the experts to encourage participation in our survey. The second round 
of the survey had an 83 percent response rate. We received responses 
from 40 of 48 experts, including 12 who had also completed round one 
of the survey. The information that we obtained from the expert survey 
is not generalizable to all experts that have an interest in or 
knowledge of drinking water contaminant monitoring issues. The full 
survey and responses are available in appendix IV. The survey 
participants are listed in appendix V. 

To evaluate the extent to which EPA implemented GAO's May 2011 
recommendations to improve the UCMR program, we reviewed EPA's 
proposed and final rules for the UCMR program and interviewed the 
relevant EPA officials who are responsible for designing and 
implementing the UCMR program to identify specific actions the agency 
had taken. To identify any opportunities that may exist to strengthen 
the UCMR program further, we interviewed stakeholders to obtain their 
perspectives on the UCMR program and related issues. In addition, we 
assessed EPA's sampling methodology against standard statistical 
criteria to determine whether the resulting sample sizes are likely to 
provide EPA with sufficient data for making national estimates of 
contaminant occurrence and exposure within an acceptable margin of 
error. 

To identify the factors EPA considered when it selected the 
contaminants for monitoring under UCMR3, we reviewed EPA documentation 
describing the steps the agency took to select the contaminants, 
including UCMR3 working group meeting minutes and EPA's proposed and 
final rules for the UCMR program and the CCL, which serves as EPA's 
primary basis for selecting UCMR contaminants. We also interviewed EPA 
officials who are responsible for designing and implementing the UCMR 
program, including selecting the UCMR contaminants; as well as those 
responsible for developing the CCL and making regulatory 
determinations. To identify the limitations, if any, EPA faced in 
selecting the contaminants, we interviewed stakeholders to obtain 
their perspectives on the UCMR contaminant selection process and 
related issues. In addition, we assessed the UCMR contaminant 
selection process against statutory requirements. 

To examine the extent to which UCMR data support EPA's regulatory 
determinations, we reviewed EPA and Federal Register documents about 
the use of the data for making the CCL2 and CCL3 regulatory 
determinations. In addition, we interviewed EPA officials who are 
responsible for designing and implementing the UCMR program and for 
making regulatory determinations. We also considered the time frames 
for implementing the UCMR program and using UCMR data for making 
regulatory determinations and assessed them against statutory 
requirements and GAO criteria for program timeliness. There is broad 
recognition throughout the federal government that, in addition to 
being effective in producing desired outcomes and other important 
dimensions of performance, agencies need to operate programs and 
provide information about them in a timely and efficient manner. For 
example, the GPRA Modernization Act calls on agencies to set goals and 
measure their progress toward meeting them. Timeliness is an important 
dimension of program performance. In addition, EPA initiated a new 
drinking water strategy in 2010 that recognizes the need to improve 
the timeliness of its efforts to address unregulated contaminants. 

We conducted this performance audit from June 2012 to January 2014 in 
accordance with generally accepted government auditing standards. 
Those standards require that we plan and perform the audit to obtain 
sufficient, appropriate evidence to provide a reasonable basis for our 
findings and conclusions based on our audit objectives. We believe 
that the evidence obtained provides a reasonable basis for our 
findings and conclusions based on our audit objectives. 

[End of section] 

Appendix II: UCMR3 Contaminants: 

Assessment monitoring: 

1,4-dioxane; 
molybdenum; 
cobalt; 
1,2,3-trichloropropane; 
1,3-butadiene; 
chloromethane (methyl chloride); 
1,1-dichloroethane; 
bromochloromethane (Halon 1011); 
bromomethane (methyl 
bromide); 
chlorodiflouromethane (HCFC-22); 
vanadium; 
strontium; 
chromium 6 (hexavalent chromium); 
total chromium[A]; 
chlorate; 
perflourooctanesulfonic acid; 
perflourooctanoic acid; 
perflourononanoic acid; 
perflourohexanesulfonic acid; 
perflouroheptanoic acid; 
perflourobutanesulfonic acid. 

Screening survey: 

17-â-estradiol; 
17-á-ethynylestradiol (ethinyl 
estradiol); 
estrone; 
4-androstrene-3,17-dione; 
estriol; 
equilin; 
testosterone. 

Pre-screen testing: 

enteroviruses; 
noroviruses. 

Source: UCMR3 final rule. 

[A] Total chromium is a regulated contaminant. EPA used a different 
statutory authority, section 1445(a)(1)(A) of the Safe Water Drinking 
Act, to require monitoring for total chromium. 

[End of table] 

[End of section] 

Appendix III: Information on Analytical Methods: 

Analytical methods are laboratory tests that measure the level of 
chemical and microbial contamination in a sample of drinking water. 
Analytical methods specify the volume of the sample to be tested, 
laboratory equipment and procedure used to test the sample, and 
quality control procedures to ensure the test is run correctly. 
Analytical methods vary based on the type of contaminant; different 
equipment and techniques are required for analyzing chemicals versus 
microbes. For example, laboratories use specialized scientific 
equipment such as gas chromatograph-mass spectrometers (GC-MS) to 
detect the presence of some chemicals in drinking water. 

Because the Unregulated Contaminant Monitoring Rule (UCMR) collects 
nationwide data, the Environmental Protection Agency's (EPA) goal is 
to use analytical methods that offer consistent application across the 
country. To get the data quality and accuracy that EPA needs, the 
agency uses analytical methods for UCMR that, among other requirements 
(1) have sufficient "holding times" for shipment of water samples from 
water utilities to laboratories; (2) can be used by a variety of 
laboratories with varying levels of expertise; and (3) can be used 
equally effectively under different drinking water conditions and 
types.[Footnote 75] EPA often develops analytical methods for the UCMR 
program that meet these specific requirements. When EPA develops an 
analytical method, there are many key steps it follows. EPA begins 
with optimizing instrumentation and determining the best calibration, 
which involves setting up the equipment to measure a contaminant at 
the desired level. Next, EPA staff optimize the solid phase 
extraction--a process that involves separating compounds that are 
dissolved or suspended in water from other compounds in the water 
sample according to their physical and chemical properties in order to 
isolate the contaminant for analysis. Then, EPA staff determine which 
preservatives to use that will allow the chemical to be preserved in 
the water sample. They also determine any interferences-if there are 
other chemicals or substances in the water sample that may have 
similar physical and chemical properties and thus interfere with the 
chemical to be analyzed--and how to address such interferences, if 
any. Then, EPA staff determine the holding times for the contaminant 
and the water sample; as noted earlier, EPA must have sufficient 
"holding times" for shipment of water samples from water utilities to 
laboratories. These holding times must not have any negative impact in 
quality on the contaminant to be analyzed in the water sample. After 
following all these steps, if the analytical method meets EPA's data 
quality objectives, then EPA writes the method so that laboratories 
across the country can use and follow the EPA instructions to measure 
the level of contamination in drinking water. EPA then sends the 
written analytical method to several laboratories to ensure the method 
works as written; this process is known as a multilab verification 
process. If the method passes this verification process, then EPA 
publishes and uses the analytical method in a future UCMR. On the 
other hand, if the analytical method does not meet the data quality 
objectives that EPA has set, EPA revises the technical approach in 
order for the earlier steps in the analysis to lead to a method that 
meets the data quality objectives. See figure 6. 

Figure 6: EPA's Process for Analytical Method Development: 

[Refer to PDF for image: process illustration] 

1. Optimize instrumentation. 

2. Optimize calibration. 

3. Optimize extraction if extraction required. 

4. Determine preservatives. 

5. Determine interferences. 

6. Determine aqueous and extract holding times: 
Does method meet data quality objectives? 
In no: go to step 7; 
If yes: go to step 8. 

7. Revise technical approach: 
If EPA determines it is necessary to revise the technical approach, 
the process returns to one of the previous steps, depending on 
specific circumstances. 

8. Write method. 

9. Multilab verification. 

10. Peer review. 

11. Use in future UCMR. 

UCMR = Unregulated Contaminant Monitoring Rule. 

Source: GAO analysis of EPA information. 

[End of figure] 

Developing an analytical method can be a complex, technically 
challenging, and time-consuming process. For example, EPA officials 
cited challenges for developing methods for some CCL3 contaminants 
such as the high water solubility of some contaminants, as well as 
some contaminants' instability in water. In such cases, EPA cannot 
keep these contaminants stable in the water samples for the minimum 
holding times that are required between the collection of samples by 
public water systems and sending the samples to the laboratories for 
analysis. For these contaminants, EPA is still working on techniques 
that will allow the agency to develop an analytical method that works 
effectively under all the conditions noted above that UCMR analytical 
methods must meet. EPA officials said it can take 1 to 2 years to 
develop a standard method that meets their data quality requirements 
and 2 to 3 years for more complex methods. 

Most often, the UCMR program uses EPA analytical methods, but it also 
allowed some methods developed by others as alternatives for UCMR3. 
EPA is also using compliance methods in the UCMR3 program. Compliance 
methods are analytical methods that laboratories are required to use 
when monitoring for regulated drinking water contaminants. For 
example, the analytical method for volatile organic compounds that EPA 
is using for UCMR3 is a compliance method.[Footnote 76] In addition, 
EPA involves stakeholders in its development of analytical methods. In 
May 2013, EPA held a public meeting and webinar to provide 
stakeholders with an update on analytical methods completed or under 
development for potential use in the UCMR program. Several external 
stakeholders participated in the meeting, both in person and via 
webinar. Some of them provided suggestions to address challenges EPA 
was facing with certain methods. 

[End of section] 

Appendix IV: Round 2 Survey Questionnaire and Results: 

The questions we asked experts in Round 2 of our survey are shown 
below. Our survey was composed of closed-and open-ended questions. In 
this appendix, we include all the survey questions and aggregate 
results of responses to the closed-ended questions; we do not provide 
information on responses provided to the open-ended questions. For a 
more detailed discussion of our survey methodology, see appendix I. For 
a list of the experts who completed our survey, see appendix V. 

Survey on EPA's Unregulated Contaminant Monitoring Program: 

U.S. Government Accountability Office: 

EPA uses the Unregulated Contaminant Monitoring Rule (UCMR) program to 
collect data for some contaminants that are suspected to be present in 
drinking water, but do not have health-based standards set under the 
Safe Drinking Water Act. Under the Safe Drinking Water Act, EPA is 
responsible for identifying and regulating drinking water contaminants 
that may pose a public health risk. Specifically, the act directs EPA 
to: 

1. Publish a list every 5 years of contaminants that are a) not 
currently regulated b) known or anticipated to occur in public water 
systems and c) may require regulation. EPA has implemented this 
provision as the Contaminant Candidate List. 

2. Establish a program to monitor unregulated contaminants and publish 
a list every 5 years of up to 30 contaminants to be monitored by public 
water systems. EPA has implemented this provision as the Unregulated 
Contaminant Monitoring Rule (UCMR) program. 

3. Determine whether or not to regulate at least 5 contaminants 
included on the list of unregulated contaminants every 5 years. EPA has 
implemented this provision as the regulatory determination process. The 
act specifies that EPA shall regulate a drinking water contaminant 
under certain conditions. To see these conditions, click here. [In the 
online-version of the survey, this was a live link to the Safe Drinking 
Water Act requirements for making a regulatory determination.] 

Some of the contaminants on the Contaminant Candidate List are 
monitored under the UCMR. EPA uses the monitoring data, along with 
information about the contaminants' health effects, to determine 
whether or not to regulate the contaminants. EPA is currently in its 
third round of the UCMR program (UCMR-3). 

EPA is monitoring 30 contaminants under UCMR-3. Click here to review 
the list. [In the online-version of the survey this was a live link. 
See appendix II of this report for a list of the 30 contaminants.] 

Section 1: UCMR Program Requirements and Considerations: 

The Safe Drinking Water Act requires that EPA publish a list of 
unregulated contaminants that pose the greatest public health concern, 
are known or anticipated to occur in public water systems, and may 
require regulation. This is known as the Contaminant Candidate List 
(CCL). 

1. EPA uses the Unregulated Contaminant Monitoring Rule (UCMR) program 
to collect monitoring data for informing its regulatory determinations 
about some of the contaminants on the CCL. The Safe Drinking Water Act 
requires EPA to select contaminants that pose the greatest public 
health concern for its CCL and for regulatory determinations. 

When selecting contaminants for monitoring under the UCMR, should EPA 
select contaminants that potentially pose the greatest public health 
concern? 

Q1: 

1--Definitely yes; 
Frequency: 24; 
Cumulative Frequency: 24. 

2--Probably yes; 
Frequency: 13; 
Cumulative Frequency: 37. 

3--Probably no; 
Frequency: 1; 
Cumulative Frequency: 38. 

5--Not enough information to judge; 
Frequency: 2; 
Cumulative Frequency: 40. 

2. EPA identified the contaminants for its third UCMR list (UCMR-3) 
largely from contaminants identified in its third Contaminant Candidate 
List (CCL-3). To create CCL-3, EPA identified about 7,500 contaminants 
and then analyzed the contaminants' potential to occur in public water 
systems and to be a public health concern. Based on those criteria, EPA 
selected 116 contaminants for CCL-3 which it determined pose the 
greatest public health concern. 

Was CCL-3 a reasonable source for EPA to use when it began identifying 
contaminants for the UCMR-3? 

Q2: 

1--Definitely yes; 
Frequency: 17; 
Cumulative Frequency: 17. 

2--Probably yes; 
Frequency: 20; 
Cumulative Frequency: 37. 

3--Probably no; 
Frequency: 2; 
Cumulative Frequency: 39. 

5--Not enough information to judge; 
Frequency: 1; 
Cumulative Frequency: 40. 

Please explain your answer to question 2: 

3. The UCMR and CCL are created through separate processes. EPA 
finalized CCL-3 in 2009, and finalized UCMR-3 in 2012. UCMR-3 includes 
22 contaminants from CCL-3 and 8 other contaminants. 

If EPA combined the CCL and UCMR contaminant selection processes, would 
the resulting UCMR list likely be of similar or better quality when 
compared to the UCMR list created under the current separate processes? 

Q3: 

1--Definitely yes; 
Frequency: 5; 
Cumulative Frequency: 5. 

2--Probably yes; 
Frequency: 15; 
Cumulative Frequency: 20. 

3--Probably no; 
Frequency: 10; 
Cumulative Frequency: 30. 

5--Not enough information to judge; 
Frequency: 10; 
Cumulative Frequency: 40. 

[End of table] 

Please explain your answer to question 3: 

4. The Safe Drinking Water Act requires EPA to issue a new list of 
contaminants for monitoring under the UCMR every 5 years. Is 5 years 
enough time to select contaminants and implement a new round of the 
UCMR program? 

Q4: Frequency; Cumulative Frequency. 

1--Definitely yes; 
Frequency: 10; 
Cumulative Frequency: 10. 

2--Probably yes; 
Frequency: 18; 
Cumulative Frequency: 28. 

3--Probably no; 
Frequency: 10; 
Cumulative Frequency: 38. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 40. 

[End of table] 

5. For UCMR-3, EPA selected 8 contaminants that were not on CCL-3. Some 
of these 8 contaminants became priorities between the publication of 
the CCL-3 and UCMR-3 lists. 

Is it worthwhile for EPA to select contaminants for the UCMR when they 
are not on the CCL? 

Q5: 

1--Definitely yes; 
Frequency: 20; 
Cumulative Frequency: 20. 

2--Probably yes; 
Frequency: 13; 
Cumulative Frequency: 33. 

3--Probably no; 
Frequency: 7; 
Cumulative Frequency: 40. 

6. The UCMR list of contaminants is based largely on the CCL, and UCMR 
supports regulatory determinations. 

Do the 30 contaminants EPA selected warrant being included in UCMR-3? 
Click here to review the list. [In the online-version of the survey 
this was a live link. See appendix II of this report for a list of the 
30 contaminants.] 

Q6: 

1--Definitely yes; 
Frequency: 11; 
Cumulative Frequency: 11. 

2--Probably yes; 
Frequency: 19; 
Cumulative Frequency: 30. 

3--Probably no; 
Frequency: 3; 
Cumulative Frequency: 33. 

4--Definitely no; 
Frequency: 3; 
Cumulative Frequency: 36. 

5--Not enough information to judge; 
Frequency: 1; 
Cumulative Frequency: 37. 

6--No answer; 
Frequency: 3; 
Cumulative Frequency: 40. 

7. EPA added chromium-6 and total chromium to the UCMR-3 list based 
largely on the strength of public comments supporting these additions. 

7A. Did chromium-6 warrant being included on the UCMR-3 list? 

Q7A: 

1--Definitely yes; 
Frequency: 19; 
Cumulative Frequency: 19. 

2--Probably yes; 
Frequency: 16; 
Cumulative Frequency: 35. 

3--Probably no; 
Frequency: 3; 
Cumulative Frequency: 38. 

5--Not enough information to judge; 
Frequency: 2; 
Cumulative Frequency: 40. 

7B. Did total chromium warrant being included on the UCMR-3 list? 

Q7B: 

1--Definitely yes; 
Frequency: 11; 
Cumulative Frequency: 11. 

2--Probably yes; 
Frequency: 17; 
Cumulative Frequency: 28. 

3--Probably no; 
Frequency: 8; 
Cumulative Frequency: 36. 

5--Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 40. 

Section 2: Factors EPA Considers when Selecting Contaminants for UCMR: 

8. One factor EPA considers when selecting contaminants for UCMR is 
what health effects information is available for the universe of 
unregulated contaminants. 

In general, was enough health effects information available when EPA 
began selecting contaminants in 2009 to identify contaminants that 
potentially pose the greatest public health concern? 

Q8: 

1--Definitely yes; 
Frequency: 2; 
Cumulative Frequency: 2. 

2--Probably yes; 
Frequency: 14; 
Cumulative Frequency: 16. 

3--Probably no; 
Frequency: 15; 
Cumulative Frequency: 31. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 33. 

5--Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 37. 

6--No answer; 
Frequency: 3; 
Cumulative Frequency: 40. 

9. EPA also needs health effects information when determining whether 
or not to regulate a contaminant monitored under UCMR. However, health 
effects information is often being developed for contaminants at the 
same time that the UCMR program is collecting data about contaminant 
occurrence. 

Given this situation, is it worthwhile for EPA to select contaminants 
for UCMR when the health effects information needed to make regulatory 
determinations is not yet available? 

Q9: 

1--Definitely yes; 
Frequency: 8; 
Cumulative Frequency: 8. 

2--Probably yes; 
Frequency: 21; 
Cumulative Frequency: 29. 

3--Probably no; 
Frequency: 8; 
Cumulative Frequency: 37. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 39. 

6--No answer; 
Frequency: 3; 
Cumulative Frequency: 40. 

10. Another factor EPA considers when selecting contaminants for UCMR 
is whether or not the contaminant was included in a prior round of 
UCMR. EPA typically does not select contaminants that were previously 
included. While this policy reduces the burden on public water systems, 
there have been situations in which monitoring again could enable EPA 
to collect better data. The following questions ask about such 
situations. 

10A. In the first round of the UCMR program, the sampling methodology 
that EPA used to select water systems had limitations, and therefore 
the program did not provide sufficient data for making decisions about 
whether or not to regulate some of the contaminants. 

Given the tradeoff between burden on water systems and data quality, 
would it be worthwhile for EPA to consider selecting some of those 
contaminants again for monitoring under UCMR? 

Q10A: 

1--Definitely yes; 
Frequency: 15; 
Cumulative Frequency: 15. 

2--Probably yes; 
Frequency: 16; 
Cumulative Frequency: 31. 

3--Probably no; 
Frequency: 7; 
Cumulative Frequency: 38. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 39. 

5--Not enough information to judge; 
Frequency: 1; 
Cumulative Frequency: 40. 

10B. In prior rounds of the UCMR program, some of the analytical 
methods used were not sensitive enough to detect contaminants at EPA's 
health effects benchmark. More sensitive analytical methods have since 
become available for some of these contaminants. 

Given the tradeoff between burden on water systems and data quality, 
would it be worthwhile for EPA to consider selecting some of those 
contaminants again for monitoring under UCMR? 

Q10B: 

1--Definitely yes; 
Frequency: 18; 
Cumulative Frequency: 18. 

2--Probably yes; 
Frequency: 16; 
Cumulative Frequency: 34. 

3--Probably no; 
Frequency: 3; 
Cumulative Frequency: 37. 

5--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 40. 

10C. New or updated health effects information is now available for 
some contaminants monitored in prior rounds of the UCMR. The new 
information indicates that health effects occur at lower concentrations 
than those previously monitored for. 

Given the tradeoff between burden on water systems and data quality, 
would it be worthwhile for EPA to consider selecting some of those 
contaminants again for monitoring under UCMR? 

Q10C: 

1--Definitely yes; 
Frequency: 19; 
Cumulative Frequency: 19. 

2--Probably yes; 
Frequency: 13; 
Cumulative Frequency: 32. 

3--Probably no; 
Frequency: 4; 
Cumulative Frequency: 36. 

5--Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 40. 

11. Another factor EPA considers when selecting contaminants for UCMR 
is whether analytical methods will be available to detect these 
contaminants in drinking water. EPA typically develops analytical 
methods for the contaminants it plans to monitor. For UCMR-3, EPA also 
considered analytical methods that had been developed by Standard 
Methods and the American Society for Testing and Materials. 

For future rounds of UCMR, will it be worthwhile for EPA to reach out 
to organizations that develop such analytical methods before it selects 
contaminants? 

Q11: 

1--Definitely yes; 
Frequency: 32; 
Cumulative Frequency: 32. 

2--Probably yes; 
Frequency: 8; 
Cumulative Frequency: 40. 

12. Another factor EPA considers when selecting contaminants for UCMR 
is whether the contaminants are similar to others it will be 
monitoring. EPA selected 6 contaminants for UCMR-3 primarily because 
they could be detected by an analytical method that will be used for 
another UCMR-3 contaminant. These 6 were not on the CCL-3, EPA's 
primary source for selecting UCMR-3 contaminants. 

Is it worthwhile for EPA to select contaminants for the UCMR primarily 
because they can be detected by an analytical method that will be used 
for another UCMR contaminant? 

Q12: 

1--Definitely yes; 
Frequency: 10; 
Cumulative Frequency: 10. 

2--Probably yes; 
Frequency: 17; 
Cumulative Frequency: 27. 

3--Probably no; 
Frequency: 12; 
Cumulative Frequency: 39. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 40. 

13. Another factor EPA considers when selecting contaminants is 
stakeholder input. For UCMR-3, EPA convened a State working group, held 
a public stakeholder meeting, and collected public comments on the 
proposed rule published in the Federal Register. 

Did these means of communication give stakeholders enough opportunity 
to provide input on the contaminants that EPA selected for UCMR-3? 

Q13: 

1--Definitely yes; 
Frequency: 10; 
Cumulative Frequency: 10. 

2--Probably yes; 
Frequency: 23; 
Cumulative Frequency: 33. 

3--Probably no; 
Frequency: 5; 
Cumulative Frequency: 38. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 39. 

5--Not enough information to judge; 
Frequency: 1; 
Cumulative Frequency: 40. 

[End of table] 

14. When EPA develops the CCL it is reviewed by EPA's Science Advisory 
Board. In addition, EPA solicited additional expert input from the 
National Drinking Water Advisory Board and from other outside experts 
when it selected contaminants for the CCL-3. 

Should EPA use similar means to obtain additional expert input when it 
selects contaminants for the UCMR? 

Q14: 

1--Definitely yes; 
Frequency: 24; 
Cumulative Frequency: 24. 

2--Probably yes; 
Frequency: 15; 
Cumulative Frequency: 39. 

3--Probably no; 
Frequency: 1; 
Cumulative Frequency: 40. 

15. In addition to EPA, the US Geological Survey, some states, and 
other groups generate contaminant occurrence data. EPA reviews some of 
this other data when selecting the UCMR contaminants. 

Is it worthwhile for EPA to consider this other data on contaminant 
occurrence when it selects contaminants for UCMR? 

Q15: 

1--Definitely yes; 
Frequency: 32; 
Cumulative Frequency: 32. 

2--Probably yes; 
Frequency: 7; 
Cumulative Frequency: 39. 

3--Probably no; 
Frequency: 1; 
Cumulative Frequency: 40. 

Section 3: Collection of Contaminant Occurrence Data through the UCMR 
Program: 

16. EPA varies the frequency for monitoring based primarily on the 
source water type. Under UCMR-3, surface water systems must monitor for 
4 consecutive quarters with sampling events 3 months apart. 

Is the requirement that surface water systems monitor for 4 consecutive 
quarters likely to provide EPA sufficient data to determine whether 
contaminants occur with a frequency and at levels of public health 
concern? 

Q16: 

1--Definitely yes; 
Frequency: 1; 
Cumulative Frequency: 1. 

2--Probably yes; 
Frequency: 26; 
Cumulative Frequency: 27. 

3--Probably no; 
Frequency: 7; 
Cumulative Frequency: 34. 

4--Definitely no; 
Frequency: 5; 
Cumulative Frequency: 39. 

5--Not enough information to judge; 
Frequency: 1; 
Cumulative Frequency: 40.
Please explain your answer to question 16: 

17. Under UCMR-3, ground water systems must monitor twice in a 12-month 
period with sampling events 5-7 months apart. 

Is the requirement that ground water systems monitor twice in a 12-
month period likely to provide EPA sufficient data to determine whether 
contaminants occur with a frequency and at levels of public health 
concern? 

Q17: 

1--Definitely yes; 
Frequency: 3; 
Cumulative Frequency: 3. 

2--Probably yes; 
Frequency: 25; 
Cumulative Frequency: 28. 

3--Probably no; 
Frequency: 10; 
Cumulative Frequency: 38. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 40. 

Please explain your answer to question 17: 

18. Under UCMR-3, non-disinfecting ground water systems located in 
sensitive hydrogeological areas (i.e., karst or fractured bedrock) must 
monitor twice in a 12-month period with sampling events 5-7 months 
apart for 2 viruses. 

Is the requirement that these non-disinfecting ground water systems 
monitor twice in a 12-month period likely to provide EPA sufficient 
data to determine whether the contaminants occur with a frequency and 
at levels of public health concern? 

Q18: 

2-Probably yes; 
Frequency: 9; 
Cumulative Frequency: 9. 

3-Probably no; 
Frequency: 16; 
Cumulative Frequency: 25. 

4-Definitely no; 
Frequency: 7; 
Cumulative Frequency: 32. 

5-Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 36. 

6-No answer; 
Frequency: 4; 
Cumulative Frequency: 40. 

Please explain your answer to question 18: 

19. UCMR-3 is a national study with samples taken over 3 years (January 
2013 through December 2015) and across all months and seasons to 
capture seasonal fluctuations in contaminant occurrence. 

19A. For the 30 contaminants being monitored under UCMR-3, is this 3-
year monitoring timeframe likely to provide EPA sufficient information 
on seasonal fluctuations in contaminant occurrence? Click here to 
review the list of 30 contaminants. [In the online-version of the 
survey this was a live link. See appendix II of this report for a list 
of the 30 contaminants.] 

Q19A: 

1--Definitely yes; 
Frequency: 5; 
Cumulative Frequency: 5. 

2--Probably yes; 
Frequency: 27; 
Cumulative Frequency: 32. 

3--Probably no; 
Frequency: 4; 
Cumulative Frequency: 36. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 37. 

5--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 40. 

19B. For the 30 contaminants being monitored under UCMR-3, would a 2-
year monitoring timeframe instead of the 3-year timeframe be likely to 
provide EPA sufficient information on seasonal fluctuations in 
contaminant occurrence? 

Q19B: 

1--Definitely yes; 
Frequency: 3; 
Cumulative Frequency: 3. 

2--Probably yes; 
Frequency: 16; 
Cumulative Frequency: 19. 

3--Probably no; 
Frequency: 9; 
Cumulative Frequency: 28. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 30. 

5--Not enough information to judge; 
Frequency: 9; 
Cumulative Frequency: 39. 

6--No answer; 
Frequency: 10; 
Cumulative Frequency: 40. 

20. EPA established a three-tiered strategy for monitoring UCMR 
contaminants based primarily on the availability of analytical methods. 
UCMR's Tier I monitoring requires all water systems serving more than 
10,000 people and a random sample of 800 water systems serving 10,000 
or fewer people to monitor those contaminants that can be detected by 
analytical methods that are widely used in drinking water laboratories. 

Is Tier I monitoring likely to provide EPA sufficient data to determine 
whether contaminants occur with a frequency and at levels of public 
health concern? 

Q20: 

1--Definitely yes; 
Frequency: 4; 
Cumulative Frequency: 4. 

2--Probably yes; 
Frequency: 27; 
Cumulative Frequency: 31. 

3--Probably no; 
Frequency: 3; 
Cumulative Frequency: 34. 

5--Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 38. 

6--No answer; 
Frequency: 2; 
Cumulative Frequency: 40. 

21. UCMR's Tier II monitoring requires all water systems serving more 
than 100,000 people, 320 randomly selected water systems serving 10,001 
to 100,000 people, and 480 randomly selected water systems serving 
10,000 or fewer people to monitor those contaminants that can be 
detected by newly developed analytical methods that are used in fewer 
drinking water laboratories. 

Is Tier II monitoring likely to provide EPA sufficient data to 
determine whether contaminants occur with a frequency and at levels of 
public health concern? 

Q21: 

1--Definitely yes; 
Frequency: 1; 
Cumulative Frequency: 1. 

2--Probably yes; 
Frequency: 25; 
Cumulative Frequency: 26. 

3--Probably no; 
Frequency: 8; 
Cumulative Frequency: 34. 

5--Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 38. 

6--No answer; 
Frequency: 2; 
Cumulative Frequency: 40. 

22. UCMR's Tier III monitoring requires a random sample of 800 non-
disinfecting ground water systems located in sensitive hydrogeological 
areas (i.e., karst or fractured bedrock) to monitor those contaminants 
that can be detected by very new or specialized analytical methods that 
are used in a small set of drinking water laboratories. 

Is Tier III monitoring likely to provide EPA sufficient data to 
determine whether contaminants occur with a frequency and at levels of 
public health concern? 

Q22: 

1--Definitely yes; 
Frequency: 1; 
Cumulative Frequency: 1. 

2--Probably yes; 
Frequency: 17; 
Cumulative Frequency: 18. 

3--Probably no; 
Frequency: 11; 
Cumulative Frequency: 29. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 31. 

5--Not enough information to judge; 
Frequency: 8; 
Cumulative Frequency: 39. 

6--No answer; 
Frequency: 1; 
Cumulative Frequency: 40. 

23. Overall, is the three-tiered monitoring strategy described in 
questions 20-22, which varies the number and size of water systems 
based on the availability of analytical methods, likely to provide EPA 
sufficient data to determine whether contaminants occur with a 
frequency and at levels of public health concern? 

Q23: 

2--Probably yes; 
Frequency: 26; 
Cumulative Frequency: 26. 

3--Probably no; 
Frequency: 5; 
Cumulative Frequency: 31. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 32. 

5--Not enough information to judge; 
Frequency: 6; 
Cumulative Frequency: 38. 

6--No answer; 
Frequency: 2; 
Cumulative Frequency: 40. 

Please explain your answer to question 23: 

24. Under UCMR-3, systems must monitor for 5 pathogen indicators. This 
monitoring will be conducted in conjunction with virus monitoring of 
enterovirus and norovirus. 

The pathogen indicators are: Total coliforms, E. coli, Enterococci, 
Bacteriophage, and Aerobic spores. 

Is this type of monitoring likely to help EPA better understand the co-
occurrence of pathogen indicators and viruses? 

Q24: 

1--Definitely yes; 
Frequency: 9; 
Cumulative Frequency: 9. 

2--Probably yes; 
Frequency: 21; 
Cumulative Frequency: 30. 

3--Probably no; 
Frequency: 3; 
Cumulative Frequency: 33. 

4--Definitely no; 
Frequency: 1; 
Cumulative Frequency: 34. 

5--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 37. 

6--No answer; 
Frequency: 3; 
Cumulative Frequency: 40. 

25. Under UCMR-3, systems must monitor for total chromium in 
conjunction with chromium-6. 

25A. Is this type of monitoring likely to provide EPA with valuable 
information on the co-occurrence of total chromium and chromium-6? 

Q25A: 

1--Definitely yes; 
Frequency: 14; 
Cumulative Frequency: 14. 

2--Probably yes; 
Frequency: 20; 
Cumulative Frequency: 34. 

3--Probably no; 
Frequency: 2; 
Cumulative Frequency: 36. 

5--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 39. 

6--No answer; 
Frequency: 1; 
Cumulative Frequency: 40. 

[End of table] 

25B. Is this type of monitoring likely to provide EPA with valuable 
information on the usefulness of monitoring for total chromium as an 
indicator for chromium-6? 

Q25B: 

1--Definitely yes; 
Frequency: 7; 
Cumulative Frequency: 7. 

2--Probably yes; 
Frequency: 19; 
Cumulative Frequency: 26. 

3--Probably no; 
Frequency: 8; 
Cumulative Frequency: 34. 

4--Definitely no; 
Frequency: 2; 
Cumulative Frequency: 36. 

5--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 39. 

6-No answer; 
Frequency: 1; 
Cumulative Frequency: 40. 

Section 4: Use of the UCMR Contaminant Occurrence Data: 

26. EPA uses UCMR data to make regulatory determinations for some of 
the contaminants listed on the CCL. Due to a time lag, the UCMR-3 
occurrence data will not be available to inform the regulatory 
determinations about the contaminants on CCL-3. As a result, EPA is 
using occurrence data from prior UCMRs to inform these regulatory 
determinations. CCL-3 includes 28 contaminants monitored under prior 
UCMRs. 

Is this likely to be an effective strategy for making regulatory 
determinations, given the time lag? 

Q26: 

1-Definitely yes; 
Frequency: 1; 
Cumulative Frequency: 1. 

2-Probably yes; 
Frequency: 9; 
Cumulative Frequency: 10. 

3-Probably no; 
Frequency: 18; 
Cumulative Frequency: 28. 

4-Definitely no; 
Frequency: 7; 
Cumulative Frequency: 35. 

5-Not enough information to judge; 
Frequency: 4; 
Cumulative Frequency: 39. 

6-No answer; 
Frequency: 1; 
Cumulative Frequency: 40. 

Please explain your answer to question 26: 

27. Sometimes EPA waits several years before it uses the UCMR 
occurrence data to inform its regulatory determinations. 

After the UCMR occurrence data has been collected, about how long is it 
likely to be a reliable indicator of the frequency and level of 
drinking water contamination? 

Q27: 

b.--Less than 5 years; 
Frequency: 11; 
Cumulative Frequency: 11. 

c.--5 to less than 11 years; 
Frequency: 8; 
Cumulative Frequency: 19. 

d.--11 to less than 16 years; 
Frequency: 2; 
21. 

e.--16 to less than 20 years; 
Frequency: 1; 
Cumulative Frequency: 22. 

g.--Not enough information to judge; 
Frequency: 18; 
Cumulative Frequency: 40. 

28. The primary purpose of the UCMR program is to collect contaminant 
occurrence data for informing regulatory determinations which are 
based, in part, on the frequency and level of drinking water 
contamination. 

Overall, how effective has the UCMR program been at collecting 
sufficient data to determine whether contaminants occur with a 
frequency and at levels of public health concern? 

Q28: 

1--Very effective; 
Frequency: 9; 
Cumulative Frequency: 9. 

2--Somewhat effective; 
Frequency: 26; 
Cumulative Frequency: 35. 

3--Not effective; 
Frequency: 1; 
Cumulative Frequency: 36. 

4--Not enough information to judge; 
Frequency: 3; 
Cumulative Frequency: 39. 

5--No answer; 
Frequency: 1; 
Cumulative Frequency: 40. 

29. Another purpose of the UCMR program is to inform EPA's development 
of future CCLs. EPA expects to complete its data collection and 
analysis for UCMR-3 in 2016; it expects to finalize its next CCL in 
2014. 

Given the tradeoff between data completeness and timeliness, would it 
be worthwhile for EPA to use partial UCMR data to inform its next CCL? 

Q29: 

1--Definitely yes; 
Frequency: 11; 
Cumulative Frequency: 11. 

2--Probably yes; 
Frequency: 24; 
Cumulative Frequency: 35. 

3--Probably no; 
Frequency: 5; 
Cumulative Frequency: 40. 

30. An additional purpose of the UCMR program is to inform EPA's health 
research planning. For example, EPA might prioritize the completion of 
health effects information for a specific contaminant over others if it 
believes the contaminant is likely to be found in drinking water. EPA 
expects to complete its data collection and analysis for UCMR-3 in 
2016; its health research planning is on-going. 

Given the tradeoff between data completeness and timeliness, would it 
be worthwhile for EPA to use partial UCMR data to inform its health 
research planning? 

Q30: 

1--Definitely yes; 
Frequency: 16; 
Cumulative Frequency: 16. 

2--Probably yes; 
Frequency: 20; 
Cumulative Frequency: 36. 

3--Probably no; 
Frequency: 2; 
Cumulative Frequency: 38. 

5--Not enough information to judge; 
Frequency: 2; 
Cumulative Frequency: 40. 

[End of table] 

Important: Please continue to the next screen to submit your final 
responses to GAO. 

Submit Your Final Responses to GAO: 

31. Are you ready to submit your final completed survey to GAO? 

(This is equivalent to mailing a completed paper survey to us. It tells 
us that your answers are official and final): 

Yes, my survey is complete - To submit your final responses, please 
click on "Exit" below. 

No, my survey is not yet complete - To save your responses for later, 
please click on "Exit" below. 

[End of section] 

Appendix V: List of Experts Who Completed the Surveys: 

Scott Biernat, Association of Metropolitan Water Agencies:
Melissa Billman, Fairfax Water Utility:
Danielle Blacet, Association of California Water Agencies:
Mark Borchardt, Ph.D., U.S. Department of Agriculture:
Bryan Brooks, Ph.D., Baylor University:
Gary Burlingame, Philadelphia Water Department:
Peter L. deFur, Ph.D., Environmental Stewardship Concepts:
Gregory Delzer, U.S. Geological Survey:
Andy Eaton, Ph.D., Eurofins Eaton Analytical, Inc.
Stephen Estes-Smargiassi, Massachusetts Water Resources Authority:
Mike Focazio, Ph.D., U.S. Geological Survey:
David Fischer, American Chemistry Council:
Ann Marie Gebhart, Ph.D., ToxServices LLC:
Will Gonzalez, Coachella Valley Water District:
Annette Guiseppi-Elie, Ph.D., DuPont:
George Hallberg, Ph.D., The Cadmus Group, Inc.
Charles Hertz, Ph.D.,Aqua America, Inc.
Michael Hutcheson, Ph.D., Massachusetts Department of Environmental 
Protection:
Elston Johnson, Texas Commission on Environmental Quality:
Todd Johnson, Minnesota Department of Health:
Nancy Kim, Ph.D., State University of New York at Albany:
Rebecca D. Klaper, Ph.D., University of Wisconsin-Milwaukee:
Dana Kolpin, U.S. Geological Survey:
John H. Koon, Ph.D., Georgia Technical Institute:
Mark LeChevallier, Ph.D., American Water:
Gary R. Lynch, Park Water Company:
Charles Maddox, Austin Water Utility:
David Mazzera, Ph.D., California Department of Public Health:
Erik Olson, Pew Charitable Trusts:
Darrell Osterhoudt, Association of State Drinking Water Administrators:
Rebecca Parkin, Ph.D., George Washington University:
Gloria Post, Ph.D., New Jersey Department of Environmental Protection:
Alan Roberson, American Water Works Association:
Renee Sharp, Environmental Working Group:
Mary Skopec, Ph.D., Iowa Department of Natural Resources:
Shane Snyder, Ph.D., University of Arizona:
Mark D. Sobsey, Ph.D., University of North Carolina at Chapel Hill:
Lynn Thorp, Clean Water Action:
Patricia Toccalino, Ph.D., U.S. Geological Survey:
Lloyd Wilson, Ph.D., New York State Department of Health:
Mae Wu, Natural Resources Defense Council:
Nira Yamachika, Orange County Water District: 

[End of section] 

Appendix VI: Comments from the Environmental Protection Agency: 

United States Environmental Protection Agency: 
Office of Water: 
Washington, D.C. 20460: 

December 23, 2013: 

Mr. Alfredo Gomez: 
Acting Director: 
Natural Resources and Environment: 
U.S. Government Accountability Office: 
Washington, DC 20548: 

Dear Mr. Gomez: 

Thank you for the opportunity to review and comment on GAO's draft 
report, "EPA has Improved its Unregulated Contaminant Monitoring 
Program, but Additional Action is Needed" (GAO-14-103). The purpose of 
this letter is to provide the EPA's response to your recommendations. 

The EPA appreciates the report's acknowledgment that "The 
Environmental Protection Agency (EPA) has implemented all of the 
recommendations GAO made in its May 2011 report to improve the
Unregulated Contaminant Monitoring Rule (UCMR) program," and welcomes 
GAO's feedback on opportunities to further improve the program. We 
generally agree with the GAO's findings, conclusions, and 
recommendations and are committed to acting on those recommendations 
as described below. A number of detailed technical and editorial 
comments on various aspects of the draft report are also noted in the 
enclosure. 

GAO Recommendation #1: 

To help ensure that the UCMR data accurately reflect contaminant 
occurrence, vary the monitoring frequency for public water systems 
when the standard UCMR monitoring frequency is not expected to
accurately detect the presence of contaminants, such as sporadically 
occurring viruses or pesticides that fluctuate seasonally. 

EPA Response #1: 

Beginning with the development of the rule for the next UCMR cycle 
(UCMR 4), the EPA will give greater consideration to the potential for 
particular contaminants to occur sporadically and/or seasonally. To 
the extent practical (taking into account, for example, cost and 
implementation/logistical considerations) the Agency will seek to 
adjust the monitoring frequency for contaminants that are anticipated 
to have significant sporadic/seasonal occurrence. 

GAO Recommendation #2: 

To help address potential limitations with the UCMR data. when 
considering if adequate data had been collected under a prior UCMR 
during the contaminant selection process, take into account whether: 

* the sampling methodology the EPA used in a prior UCMR for a 
particular contaminant provided sufficient occurrence data to make a 
regulatory determination. 

* the contaminant's health information is influx or has been updated 
and now indicates that adverse health effects occur at a lower level 
than that at which UCMR data were originally collected, or 

* the data collected in an earlier cycle have become outdated and no 
longer present an accurate picture of contaminant occurrence, for 
example, because industrial or agricultural chemical use has changed. 

EPA Response #2: 

Beginning with the development of the rule for UCMR 4, the EPA will 
give greater consideration to whether additional occurrence data 
gathering is warranted for contaminants monitored during prior UCMR 
cycles. In doing so, the Agency will specifically consider (1) the 
sampling methodology used in the prior cycle; (2) changes in health 
information (e.g., establishment of a lower health reference level)
that may warrant additional monitoring at a lower level; and (3) the 
degree to which the available data are expected to be representative 
of current contaminant occurrence. These factors will be considered in
the Agency's establishment of relative priorities for UCMR contaminant 
selection. As to the use of the UCMR data from prior cycles to support 
the Agency's regulatory determinations, the EPA will continue to judge 
the appropriateness of the data on a case-by-case basis. As 
acknowledged in the report, the Agency uses the "best available" data 
to support its decisions. 

GAO Recommendation #3: 

To optimize the ability of the UCMR program to support regulatory 
determinations, convene an internal workgroup of officials responsible 
for the UCMR, regulatory determinations, and CCL to fully examine
opportunities to improve the timeliness of the UCMR program. Among 
other things, the workgroup should consider: 

* whether a shorter process for selecting contaminants would yield a 
high quality list. or whether the existing process can be started 
sooner, and; 

* whether monitoring for contaminants over a shorter period, instead 
of the 3-year period, is feasible, given data quality, logistical 
needs, and the burden on public water systems. 

EPA Response #3: 

The EPA will convene the recommended workgroup and will consider, 
among the potential improvements, the practicality of a shorter 
process for contaminant selection and a shorter period for contaminant 
monitoring. Workgroup recommendations will be incorporated into our 
approach to the extent practical. 

Again, we appreciate the opportunity to review the draft report. We 
also appreciate the constructive recommendations and the 
professionalism with which the review was conducted. Please direct
questions regarding this response to Greg Carroll, 
carroll.gregory@epa.gov or (513)569-7948. 

Sincerely, 

Signed by: 

Nancy K. Stoner: 
Acting Assistant Administrator: 

Enclosure: 

[End of section] 

Appendix VII: GAO Contact and Staff Acknowledgments: 

GAO Contact: 

J. Alfredo Gómez, (202) 512-3841 or gomezj@gao.gov: 

Staff Acknowledgments: 

In addition to the individual named above, Diane B. Raynes (Assistant 
Director), Elizabeth Beardsley, Mark Braza, Justin Fisher, Annamarie 
Lopata, Rebecca Makar, Christopher Murray, Alison O'Neill, Dan Royer, 
Kelly Rubin, and Esther Toledo made significant contributions to this 
report. 

[End of section] 

Related GAO Products: 

Chemical Assessments: An Agencywide Strategy May Help EPA Address 
Unmet Needs for Integrated Risk Information System Assessments. 
[hyperlink, http://www.gao.gov/products/GAO-13-369]. Washington, D.C.: 
May 10, 2013. 

Environmental Health: Action Needed to Sustain Agencies' Collaboration 
on Pharmaceuticals in Drinking Water. [hyperlink, 
http://www.gao.gov/products/GAO-11-346]. Washington, D.C.: August 8, 
2011. 

Safe Drinking Water Act: Improvements in Implementation Are Needed to 
Better Assure the Public of Safe Drinking Water. [hyperlink, 
http://www.gao.gov/products/GAO-11-803T]. Washington, D.C.: July 12, 
2011. 

Safe Drinking Water Act: EPA Should Improve Implementation of 
Requirements on Whether to Regulate Additional Contaminants. 
[hyperlink, http://www.gao.gov/products/GAO-11-254]. Washington, D.C.: 
May, 27, 2011. 

Perchlorate: Occurrence is Widespread but at Varying Levels; Federal 
Agencies Have Taken Some Actions to Respond to and Lessen Releases. 
[hyperlink, http://www.gao.gov/products/GAO-10-769]. Washington, D.C.: 
August 12, 2010. 

Safe Drinking Water Act: Progress and Future Challenges in 
Implementing the 1996 Amendments. [hyperlink, 
http://www.gao.gov/products/GAO/RCED-99-31]. Washington, D.C.: January 
14, 1999. 

[End of section] 

Footnotes: 

[1] Pub. L. No. 93-523 (1974), codified as amended at 42 U.S.C. §§300f-
300j-26 (2013). 

[2] Under SDWA, a drinking water contaminant is defined as any 
physical, chemical, biological, or radiological substance or matter in 
water. 

[3] In the 1996 amendments, Congress removed a previous requirement 
that EPA regulate 25 new contaminants every 3 years. 

[4] EPA decided to regulate perchlorate in 2011 and is currently 
developing a regulatory proposal for it. Perchlorate is both a 
naturally occurring and man-made chemical that is used to produce 
rocket fuel, fireworks, flares and explosives. It can also be present 
in bleach and in some fertilizers. According to EPA, perchlorate may 
have adverse health effects because scientific research indicates that 
this contaminant can disrupt the thyroid's ability to produce hormones 
needed for normal growth and development. 

[5] The current Contaminant Candidate List is referred to as CCL3. 

[6] For purposes of the UCMR, EPA defines a small water system as one 
serving 10,000 or fewer people. EPA defines a large water system as 
one that serves more than 10,000 people. 

[7] GAO, Safe Drinking Water Act: EPA Should Improve Implementation of 
Requirements on Whether to Regulate Additional Contaminants, 
[hyperlink, http://www.gao.gov/products/GAO-11-254] (Washington, D.C.: 
May 27, 2011). 

[8] The Delphi method consists of surveying experts in two or more 
rounds to obtain group agreement or consensus on key topics. 

[9] To protect drinking water from disease-causing organisms, or 
pathogens, water suppliers may add a disinfectant, such as chlorine, 
to drinking water. However, disinfectants can react with naturally 
occurring materials in the water to form by-products, some of which 
may pose health risks. 

[10] States may regulate some of these contaminants. Under state laws, 
some state environmental agencies have the authority to regulate 
additional contaminants or establish more stringent standards than 
federal regulations, while others do not have these authorities. 

[11] [hyperlink, http://www.gao.gov/products/GAO-11-254]. 

[12] EPA defines a community water system as a public water system 
that has at least 15 service connections used by year-round residents 
or regularly serves an average of at least 25 year-round residents. 
EPA defines a nontransient, noncommunity water system as a public 
water system that is not a community water system and regularly serves 
at least 25 of the same people over 6 months per year. There are more 
than 52,000 community water systems and more than 19,000 nontransient, 
noncommunity water systems. 

[13] EPA defines a transient noncommunity water system as a public 
water system that does not regularly serve at least 25 of the same 
people over 6 months per year. There are more than 87,000 transient, 
noncommunity water systems. 

[14] 42 U.S.C. § 300g-1(b)(1)(C) (2013). 

[15] Under the time frames set out in the statute, EPA would have 
issued CCL3 in 2008, the list of UCMR3 contaminants in 2009, and the 
regulatory determinations for CCL3 in 2011. 

[16] EPA has issued or updated health advisories for the following 9 
contaminants: (1) 1,1,2,2-tetrachloroethane; (2) 2,4-dinitrotoluene; 
(3) 2,6-dinitrotoluene; (4) boron; (5) tetrachloroterephthalic acid; 
(6) monomethyl tetrachloroterephthalic acid; (7) manganese; (8) 
sodium; and (9) sulfate. EPA also developed consumer information for 
Acanthamoeba. 

[17] Laboratories conducting the pre-screen testing analyses for 
viruses and pathogen indicators are under direct contracts with EPA 
and are subject to an analogous laboratory approval process. 

[18] These data are included in EPA's National Contaminant Occurrence 
Database. EPA developed the database to satisfy the statutory 
requirements set by Congress in the 1996 amendments to SDWA to 
maintain a national drinking water contaminant occurrence database 
using sampling data for both regulated and unregulated contaminants in 
public water systems. 

[19] Holding time is the time between the collection of a water sample 
by a public water system and the laboratory analysis of that water 
sample. 

[20] Three of the 17 recommendations we made in [hyperlink, 
http://www.gao.gov/products/GAO-11-254] applied to the UCMR program. 
Other recommendations concern regulatory determinations and could 
affect the UCMR program due to the relationship between the programs. 

[21] MRLs represent an estimate of the lowest concentration of a 
compound that can be quantitatively measured by members of a group of 
experienced drinking water laboratories. Under the UCMR, laboratories 
are to report all analyses of listed contaminants. Laboratories report 
values when they are equal to or greater than the established MRL, or 
report that the analysis was below the MRL. 

[22] [hyperlink, http://www.gao.gov/products/GAO-11-254]. 

[23] One of the 30 contaminants to be monitored under UCMR3 is total 
chromium, a regulated contaminant. EPA included total chromium in 
UCMR3 to collect data on it concurrent with data for chromium-6, an 
associated unregulated contaminant. EPA used a different statutory 
authority, section 1445(a)(1)(A) of SDWA, to require monitoring for 
total chromium. 

[24] The 9 contaminants include 4 monitored under, or in association 
with UCMR1, and 5 monitored under the program that preceded the UCMR. 
Under the earlier program, public water systems were required to 
monitor for up to 48 unregulated contaminants and to report monitoring 
results to the states or to EPA. 

[25] [hyperlink, http://www.gao.gov/products/GAO-11-254]. 

[26] EPA set most of the MRLs for the UCMR3 contaminants at the level 
of micrograms per liter (µg/L), generally ranging from 0.0001 µg/L to 
1 µg/L. Some contaminants are able to be detected at much lower 
concentrations than others, due to their chemical properties, among 
other factors. The levels set for the UCMR3 contaminants are generally 
more sensitive than the levels set for UCMR1 and UCMR2 contaminants, 
which generally ranged from 0.5 µg/L to 10 µg/L and from 0.002 µg/L to 
2 µg/L, respectively. 

[27] Groundwater under the direct influence of surface water means any 
water beneath the surface of the ground with significant occurrence of 
insects or other macroorganisms, algae, or large-diameter pathogens 
such as Giardia lamblia or Cryptosporidium, or significant and 
relatively rapid shifts in water characteristics such as turbidity, 
temperature, conductivity, or pH that closely correlate to 
climatological or surface water conditions. For example, a groundwater 
source located close enough to nearby surface water, such as a river 
or lake, to receive direct surface water recharge is considered at 
risk of contamination from pathogens that are not normally found in 
true groundwaters. 

[28] Environmental Protection Agency, "Revisions to the Unregulated 
Contaminant Monitoring Regulation for Public Water Systems," 64 Fed. 
Reg. 50,556, 50,570, 50,588 (Sept. 17, 1999). 

[29] Id. at 50,571 ("For much of the United States, east of the Rocky 
Mountains, many studies have shown that the season of greatest 
vulnerability for contaminant occurrence is the late-spring, early-
summer runoff-recharge period, particularly for contaminants such as 
pesticides and nitrate (e.g., Larson et al., 1997; Barbash and Resek, 
1996; Hallberg, 1989a,b). For deeper, more confined ground water 
systems, defining vulnerable periods is much more difficult."). 

[30] Systems that are unable to monitor within the assigned time frame 
must contact EPA to reschedule their monitoring. 

[31] Environmental Protection Agency, UCMR2 Categorized Public 
Comments (Washington, D.C.: June 2006); and Environmental Protection 
Agency, Response to Comments Document for the Unregulated Contaminant 
Monitoring Regulations (UCMR3) (Washington, D.C.: January 2012). 

[32] The 800 systems use groundwater wells that do not treat the water 
and are located in areas of vulnerable geology, such as karst aquifers. 

[33] According to an EPA fact sheet, the primary means of virus-
related disease transmission is exposure to fecally contaminated water 
or consumption of food exposed to the contaminated water. Exposure to 
enterovirus-contaminated drinking water can cause fever, respiratory 
illness, rash, conjunctivitis, hand-foot-mouth disease, and 
meningitis, among other things. Exposure to norovirus-contaminated 
drinking water can cause mild fever, vomiting, diarrhea, and cramps. 

[34] Wisconsin Water and Health Trial for Enteric Risks (WAHTER 
Study)--Part 1: Risk of Illness from Municipal Groundwater 
Consumption. EPA Office of Water Symposium on Groundwater-Borne 
Infectious Disease, Etiologic Agents and Indicators (Washington, D.C.: 
December 2-4, 2008). 

[35] Environmental Protection Agency, Public Comment and Response 
Summary for the Unregulated Contaminant Monitoring Regulation 
(Washington, D.C.: August 1999). 

[36] According to EPA's website, Aeromonas is a genus of bacteria with 
the potential to contaminate drinking water by evading treatment 
(e.g., chlorine disinfection) and to cause human disease such as 
gastroenteritis and bacterial septicemia, which are typically acquired 
through ingestion or though exposure of open wounds. 

[37] Specifically, the three herbicides included in UCMR2 are 
acetochlor, alachlor, and metolachlor. Such pesticides have the 
potential to contaminate drinking water by making their way into 
surface water or groundwater systems that feed drinking water 
supplies, for example, through runoff or drainage resulting from 
rainfall or irrigation. EPA classified all three contaminants as 
"possible" or "probable" carcinogens. The six degradates included in 
UCMR2 are acetochlor ethane sulfonic acid; acetochlor oxanilic acid; 
alachlor ethane sulfonic acid; alachlor oxanilic acid; metolachlor 
ethane sulfonic acid; and metolachlor oxanilic acid. Degradates are 
the result of pesticides breaking down in the environment as they 
interact with water, soil, air, sunlight, or microorganisms. 

[38] The USGS National Water-Quality Assessment Program (NAWQA) 
conducted a national-scale assessment of pesticide occurrence and 
concentrations in streams and groundwater during 1992-2001. The 
assessment's findings were published in a March 2006 report entitled, 
The Quality of Our Nation's Waters--Pesticides in the Nation's Streams 
and Ground Water, 1992-2001. The report notes that "although NAWQA did 
not measure pesticide concentrations at drinking-water intakes, NAWQA 
results for the wide range of streams sampled indicate that seasonal 
pulses of pesticide concentrations probably occur in some streams that 
are used as sources of drinking water." 

[39] A contaminant of emerging concern is a contaminant characterized 
by a perceived, potential or real threat to human health or the 
environment or by lack of published health standards. 

[40] Environmental Protection Agency, "Revisions to the Unregulated 
Contaminant Monitoring Regulation (UCMR 3) for Public Water Systems," 
Proposed Rule, 76 Fed. Reg. 11,713 (Mar. 3, 2011). 

[41] EPA's Office of Science and Technology, part of the Office of 
Water, provides scientific support for drinking water regulatory 
determinations, information for drinking water health advisories, and 
other human health risk assessments, among other duties. 

[42] Total chromium, while already regulated in drinking water, was 
included in conjunction with UCMR3 monitoring to determine its 
relationship with chromium-6, a contaminant of emerging concern. 

[43] EPA classified the UCMR program as a Tier 3 action under its 
Action Development Process. Tier 3 actions allow the lead office to 
design their own review processes and do not require formal EPA 
management reviews or OMB reviews. 

[44] Given continuing budget pressures combined with the focus on 
performance envisioned in the GPRA Modernization Act of 2010, we have 
reported that federal agencies need to identify ways to operate more 
efficiently. We have defined efficiency as maintaining federal 
government services or outcomes using fewer resources (such as time 
and money) or improving or increasing the quality or quantity of 
services or outcomes while maintaining (or reducing) resources. See 
GAO, Streamlining Government: Key Practices from Select Efficiency 
Initiatives Should be Shared Governmentwide, [hyperlink, 
http://www.gao.gov/products/GAO-11-908] (Washington, D.C.: Sept. 30, 
2011). 

[45] PFCs are man-made chemicals that are used in a wide variety of 
industrial and commercial products. EPA has reported that PFCs have 
significant adverse effects on laboratory animals and wildlife and 
anticipates that continued exposure to PFCs could result in adverse 
effects on humans. 

[46] Environmental Protection Agency, Long-Chain Perfluorinated 
Chemical's (PFCs) Action Plan (Washington, D.C.: Dec. 30, 2009). 

[47] EPA has funded the small public water system's UCMR monitoring 
from EPA's State Revolving Fund appropriation. 

[48] EPA has only once remonitored for a contaminant: namely, the 
pesticide acetochlor, which was monitored under UCMR1 and UCMR2. MTBE 
was monitored under UCMR1 and was considered for, but not included, in 
UCMR3. EPA's practice when selecting contaminants for UCMR3 was to 
exclude those monitored under a prior UCMR, as stated in the Federal 
Register. 

[49] Environmental Protection Agency, "Drinking Water: Regulatory 
Determinations Regarding Contaminants on the Second Drinking Water 
Contaminant Candidate List," Final Rule, 73 Fed. Reg. 44,251 (July 30, 
2008). 

[50] Environmental Protection Agency, "Revisions to the Unregulated 
Contaminant Monitoring Regulation for Public Water Systems," Final 
Rule, 64 Fed. Reg. 50,556, 50,568 (Sept. 17, 1999). 

[51] The 300 public water systems included 120 systems serving more 
than 10,000 people and 180 systems serving 10,000 or fewer people. 

[52] The 800 public water systems included 320 systems serving 10,001 
to 100,000 people; and a representative sample of 480 systems serving 
10,000 or fewer people. Approximately 400 systems serve more than 
100,000 people. 

[53] In 2009, EPA canceled the registration for disulfoton. In 2004, 
EPA canceled certain uses of molinate. Molinate is still registered 
but, as of 2012, there were no active uses, according to EPA documents. 

[54] EPA's data do not indicate whether the decrease in nitrobenzene 
use occurred before, during, or after the UCMR1 data were collected. 

[55] In May 2011, we reported that EPA had not defined what data are 
adequate to support a regulatory determination. In that report, we 
also recommended that EPA develop guidance for making regulatory 
determinations that included factors for determining when data provide 
adequate support. In its response to the recommendation, EPA stated 
that it did not believe establishing such guidance was practicable, 
given the many combinations of health effects factors and potential 
ranges of frequencies and levels of contaminants measured in drinking 
water. 

[56] SDWA requires EPA, in carrying out the provisions concerning 
listing, selecting, and regulating contaminants, to the degree that an 
action is based on science, to use the best available peer-reviewed 
science and data collected by accepted or best available methods. 

[57] SDWA requires EPA to make at least 5 regulatory determinations 
every 5 years, and hence we refer to the 5-year mark as a "deadline." 
See 42 U.S.C. § 300g-1(b)(1)(B)(ii)(I) (2013). 

[58] EPA has made 20 negative regulatory determinations and 1 positive 
regulatory determination (perchlorate) since the SDWA Amendments of 
1996 established the UCMR program. 

[59] EPA also needs information on a contaminant's health effects to 
make a regulatory determination. In 2011, GAO reported that EPA did 
not have health effects information for many of the contaminants on 
the CCL, which limited EPA's ability to make regulatory 
determinations. See [hyperlink, 
http://www.gao.gov/products/GAO-11-254]. 

[60] In its comments on a draft of this report, EPA noted that the 
agency makes judgments on a case by case basis to determine whether 
data are too old. 

[61] SDWA limits EPA to listing no more than 30 contaminants every 5 
years and, as a result, EPA cannot accelerate the UCMR schedule by 
issuing the next list earlier-that is, before 5 years from the last 
one-under current law. Each time EPA publishes a UCMR list or set of 
regulatory determinations, the agency adds 5 years to determine its 
next deadline. 

[62] EPA had considered shortening the time to report, review, and 
approve the data by about half (90 days total) but decided not to, 
based on public comments on the proposed UCMR3 rule and OMB's review 
of the final rule. 

[63] Final regulatory determinations for the CCL3 contaminants were 
originally expected to be published in 2013, according to EPA 
documents. However, EPA officials told us they have been delayed until 
2015 for reasons unrelated to the UCMR program. 

[64] Survey on EPA's Unregulated Contaminant Monitoring Program, 
Question 19B: For the 30 contaminants being monitored under UCMR-3, 
would a 2-year monitoring timeframe instead of the 3-year timeframe be 
likely to provide EPA sufficient information on seasonal fluctuations 
in contaminant occurrence? (See appendix IV for full survey.) 

[65] As noted earlier, EPA may have an opportunity to use partial 
UCMR3 data to inform its CCL3 regulatory determinations, but not 
enough UCMR3 data will be available to serve as a primary source. 

[66] Eighteen of the contaminants on CCL3 were carried over from CCL1 
and CCL2. 

[67] EPA can make regulatory determinations outside of the 5-year 
cycle but has done so only once. In 2011, EPA made an out of cycle 
determination to regulate perchlorate. 

[68] Environmental Protection Agency, "Regulatory Determinations for 
the Third Drinking Water Contaminant Candidate List, Presentation for 
Stakeholder Meeting" (Washington, D.C.: June 16, 2011). 

[69] EPA first listed these contaminants on CCL1 and carried them over 
to CCL2 and CCL3. They are (1) acetochlor, (2) alachlor ESA, (3) 
disulfoton, (4) diuron, (5) MTBE, (6) metolachlor, (7) molinate, (8) 
nitrobenzene, (9) RDX, (10) terbufos, and (11) vanadium. 

[70] The other four contaminants are microbes and have not been 
monitored yet because EPA does not have analytical methods available 
for them. Methyl bromide is listed by the name bromomethane in UCMR3. 

[71] While EPA is authorized to make regulatory determinations outside 
of the 5-year cycle, in practice, it has done so only once. 

[72] These contaminants are (1) disulfoton, (2) diuron, (3) MTBE, (4) 
molinate, (5) nitrobenzene, and (6) terbufos. 

[73] We have previously reported that gaps in health effects data 
limit EPA's ability to make regulatory determinations and recommended 
that EPA develop a coordinated process for obtaining the necessary 
health effects and occurrence data. EPA's response to our 
recommendation did not acknowledge that any further steps are needed. 
See [hyperlink, http://www.gao.gov/products/GAO-11-254]. 

[74] Environmental Protection Agency, "Drinking Water: Regulatory 
Determinations Regarding Contaminants on the Second Drinking Water 
Contaminant Candidate List," 73 Fed. Reg. 44,251 (July 30, 2008). 

[75] Drinking water types in the United States include groundwater 
sources, such as underground aquifers, and surface water sources, such 
as streams, rivers, and lakes. 

[76] Volatile organic compounds, or VOCs, may occur naturally in the 
environment or occur only as a result of manmade activities; some VOCs 
have both origins. VOC contamination of drinking water supplies is a 
human-health concern because many are toxic and are known or suspected 
human carcinogens. EPA has issued drinking water regulations for eight 
VOCs. 

[End of section] 

GAO's Mission: 

The Government Accountability Office, the audit, evaluation, and 
investigative arm of Congress, exists to support Congress in meeting 
its constitutional responsibilities and to help improve the 
performance and accountability of the federal government for the 
American people. GAO examines the use of public funds; evaluates 
federal programs and policies; and provides analyses, recommendations, 
and other assistance to help Congress make informed oversight, policy, 
and funding decisions. GAO's commitment to good government is 
reflected in its core values of accountability, integrity, and 
reliability. 

Obtaining Copies of GAO Reports and Testimony: 

The fastest and easiest way to obtain copies of GAO documents at no 
cost is through GAO's website [hyperlink, http://www.gao.gov]. Each 
weekday afternoon, GAO posts on its website newly released reports, 
testimony, and correspondence. To have GAO e-mail you a list of newly 
posted products, go to [hyperlink, http://www.gao.gov] and select 
“E-mail Updates.” 

Order by Phone: 

The price of each GAO publication reflects GAO's actual cost of 
production and distribution and depends on the number of pages in the 
publication and whether the publication is printed in color or black 
and white. Pricing and ordering information is posted on GAO's 
website, [hyperlink, http://www.gao.gov/ordering.htm]. 

Place orders by calling (202) 512-6000, toll free (866) 801-7077, or 
TDD (202) 512-2537. 

Orders may be paid for using American Express, Discover Card, 
MasterCard, Visa, check, or money order. Call for additional 
information. 

Connect with GAO: 

Connect with GAO on facebook, flickr, twitter, and YouTube.
Subscribe to our RSS Feeds or E mail Updates. Listen to our Podcasts.
Visit GAO on the web at [hyperlink, http://www.gao.gov]. 

To Report Fraud, Waste, and Abuse in Federal Programs: 

Contact: 
Website: [hyperlink, http://www.gao.gov/fraudnet/fraudnet.htm]; 
E-mail: fraudnet@gao.gov; 
Automated answering system: (800) 424-5454 or (202) 512-7470. 

Congressional Relations: 

Katherine Siggerud, Managing Director, siggerudk@gao.gov: 
(202) 512-4400: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7125: 
Washington, DC 20548. 

Public Affairs: 

Chuck Young, Managing Director, youngc1@gao.gov: 
(202) 512-4800: 
U.S. Government Accountability Office: 
441 G Street NW, Room 7149: 
Washington, DC 20548. 

[End of document]