This is the accessible text file for GAO report number GAO-14-80 
entitled 'Clean Water Act: Changes Needed If Key EPA Program Is to 
Help Fulfill the Nation's Water Quality Goals' which was released on 
January 13, 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: 

December 2013: 

Clean Water Act: 

Changes Needed If Key EPA Program Is to Help Fulfill the Nation's 
Water Quality Goals: 

GAO-14-80: 

GAO Highlights: 

Highlights of GAO-14-80, a report to congressional requesters. 

Why GAO Did This Study: 

The 1972 Clean Water Act aimed to “restore and maintain the chemical, 
physical, and biological integrity of the nation’s waters.” Under the 
act, states must establish water quality standards; for waters that do 
not meet these standards, states must develop TMDLs, which EPA 
approves. TMDLs set targeted limits for pollutants but are not self-
implementing; EPA and states help reduce pollutants by issuing permits 
for point sources, whereas they provide voluntary incentives to reduce 
nonpoint source pollution. 

GAO was asked to examine the TMDL program, specifically (1) EPA’s and 
states’ responsibilities in developing and implementing TMDLs, (2) 
what is known about the status of long-established TMDLs, (3) the 
extent to which such TMDLs contain features key to attaining water 
quality standards, and (4) the extent to which TMDLs exhibit factors 
that facilitate effective implementation. GAO asked water resource 
experts to review a random sample of 25 long-established TMDLs and 
surveyed state officials who are responsible for implementing a 
representative sample of 191 long-established TMDLs. 

What GAO Found: 

The Environmental Protection Agency (EPA) and the states each have 
responsibilities for developing and implementing pollution targets, 
known as total maximum daily loads (TMDL). EPA oversees states’ TMDL 
efforts by establishing in regulations minimum requirements TMDLs need 
for approval, providing funding, and furnishing technical assistance. 
States develop TMDLs and generally take the lead in implementing them 
by identifying pollutants that impair water quality and taking actions 
to reduce them. 

Of about 50,000 TMDLs developed and approved, nearly 35,000 were 
approved more than 5 years ago, long enough for GAO to consider them 
long established. State officials GAO surveyed in its representative 
sample of 191 TMDLs reported that pollutants had been reduced in many 
waters, but few impaired water bodies have fully attained water 
quality standards. 

The sample of 25 TMDLs reviewed by water resource experts GAO 
contacted seldom contained all features key to attaining water quality 
standards. According to the National Research Council and EPA, these 
features-—some that are beyond the scope of EPA’s existing 
regulations—-include identifying pollution-causing stressors and 
showing how addressing them would help attain such standards; 
specifying how and by whom TMDLs will be implemented; and ensuring 
periodic revisions as needed. The experts found, however, that 17 of 
25 long-established TMDLs they reviewed did not show that addressing 
identified stressors would help attain water quality standards; 12 
contained vague or no information on actions that need to be taken, or 
by whom, for implementation; and 15 did not contain features to help 
ensure that TMDLs are revised if need be. GAO’s review showed that 
EPA’s existing regulations do not explicitly require TMDLs to include 
these key features, and without such features in TMDLs—-or in addition 
to TMDLs—-impaired water bodies are unlikely to attain standards. 

In response to GAO’s survey, state officials reported that long-
established TMDLs generally do not exhibit factors most helpful for 
attaining water quality standards, particularly for nonpoint source 
pollution (e.g., farms and storm water runoff). The officials reported 
that landowner participation and adequate funding-—factors they viewed 
as among the most helpful in implementing TMDLs-—were not present in 
the implementation activities of at least two-thirds of long-
established TMDLs, particularly those of nonpoint source TMDLs. 
Because the Clean Water Act addresses nonpoint source pollution 
largely through voluntary means, EPA does not have direct authority to 
compel landowners to take prescribed actions to reduce such pollution. 
In GAO’s survey, state officials knowledgeable about TMDLs reported 
that 83 percent of TMDLs have achieved their targets for point source 
pollution (e.g., factories) through permits but that 20 percent 
achieved their targets for nonpoint source pollution. In 1987, when 
the act was amended to cover such pollution, some Members of Congress 
indicated that this provision was a starting point, to be changed if 
reliance on voluntary approaches did not significantly improve water 
quality. More than 40 years after Congress passed the Clean Water Act, 
however, EPA reported that many of the nation’s waters are still 
impaired, and the goals of the act are not being met. Without changes 
to the act’s approach to nonpoint source pollution, the act’s goals 
are likely to remain unfulfilled. 

What GAO Recommends: 

GAO recommends that EPA issue new regulations for TMDL development, 
adding key features. Further, Congress should consider revising the 
Clean Water Act’s approach to addressing nonpoint source pollution. 
EPA did not comment on the matter for Congress. The agency agreed with 
the need to add key features to TMDLs but did not agree to issue new 
regulations. GAO believes new regulations are needed. 

View [hyperlink, http://www.gao.gov/products/GAO-14-80]. For more 
information, contact J. Alfredo Gómez at (202) 512-3841 or 
gomezj@gao.gov. 

[End of section] 

Contents: 

Letter: 

Background: 

EPA and the States Have Certain Responsibilities for Developing and 
Implementing TMDLs: 

EPA Has Limited Information about the Status of Long-Established 
TMDLs, but States Report That Few Water Bodies with Such TMDLs Have 
Attained Water Quality Standards: 

Long-Established TMDLs Experts Reviewed Seldom Contained All Key 
Features, Some of Which Are Not Included in EPA's Guidance: 

Factors Reported as Facilitating the Implementation of Long-
Established TMDLs, Particularly Those for Nonpoint Source Pollution, 
Are Often Not in Place: 

Conclusions: 

Matter for Congressional Consideration: 

Recommendations for Executive Action: 

Agency Comments and Our Evaluation: 

Appendix I: Objectives, Scope, and Methodology: 

Appendix II: National Research Council Experts GAO Selected to Review 
TMDLs: 

Appendix III: Experts' Review of Long-Established TMDLs: 

Appendix IV: Survey of State TMDL Coordinators Responsible for 
Implementing Long-Established TMDLs: 

Appendix V: Comments from the Environmental Protection Agency: 

Appendix V: IGAO Contact and Staff Acknowledgments: 

Tables: 

Table 1: Elements for Approvable TMDLs: 

Table 2: Random Subsample of 25 TMDLs Reviewed by Water Resource 
Experts: 

Table 3: Number of Long-Established TMDLs in Which Experts Found Key 
Features Needed to Enable Water Bodies to Attain Water Quality 
Standards: 

Figures: 

Figure 1: Activities That Can Impair Water Bodies: 

Figure 2: An Eastern Washington Stream Channelized to Maximize Crop 
Yield: 

Figure 3: Extent of Impaired Freshwaters and TMDLs Developed since 
Passage of the Clean Water Act: 

Figure 4: Key Steps in Developing and Implementing TMDLs: 

Figure 5: Scales at Which USDA Aggregates Watershed Data, States 
Implement TMDLs, and EPA Needs GIS Data to Track TMDL Implementation 
Projects: 

Figure 6: Factors Facilitating TMDL Implementation and Extent to Which 
They Were Present in and Considered Helpful for Long-Established TMDLs: 

Abbreviations: 

EPA: Environmental Protection Agency: 

GIS: Geographic Information Systems: 

TMDL: Total Maximum Daily Load: 

USDA: U.S. Department of Agriculture: 

NRC: National Research Council: 

[End of section] 

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

December 5, 2013: 

The Honorable Barbara Boxer: 
Chairman: 
Committee on Environment and Public Works: 
United States Senate: 

The Honorable Benjamin L. Cardin: 
United States Senate: 

The Clean Water Act was enacted more than 40 years ago to "restore and 
maintain the chemical, physical, and biological integrity of the 
nation's waters."[Footnote 1] Toward this end, the act requires states 
to establish water quality standards protective of public health and 
the environment and considering aquatic wildlife and human consumption 
and recreation, among other uses; the act also requires the 
Environmental Protection Agency (EPA) to maintain and improve water 
quality by assisting and overseeing states' efforts, among other 
responsibilities. As of August 2013, EPA's assessment of national 
water quality, however, reported that more than half the nation's 
assessed waters do not meet these standards or their designated uses, 
such as fishing, swimming, or drinking.[Footnote 2] For example, the 
assessment identified more than 12 million acres of lakes (or 67 
percent of total assessed lake acres nationwide) and more than 500,000 
miles of rivers (or 53 percent of assessed river miles nationwide) as 
impaired; in all, the report shows, more water bodies that have been 
assessed are not meeting water quality standards than at any time in 
the past.[Footnote 3] Furthermore, the assessment cites pollutants 
from nonpoint sources, such as agricultural runoff and stormwater 
runoff from some urban and suburban areas, as leading causes of 
impairment of these water bodies. EPA has estimated that at historical 
funding levels and water body restoration rates, it would take longer 
than 1,000 years to restore all the water bodies that are now impaired 
by nonpoint source pollution. 

To improve the condition of water bodies that states identify as 
impaired, the Clean Water Act requires states to develop pollutant 
budgets, known as "total maximum daily loads" (TMDL), generally for 
each pollutant impairing a water body. A TMDL is essentially the 
numeric target for a specific pollutant, reflecting the maximum amount 
of the pollutant that a water body can contain and still be considered 
in compliance with water quality standards, and is described in a 
report that may also provide a general plan for how this target is to 
be achieved in the water body. According to EPA documents and 
officials, the agency's regulations refer to a TMDL generally as a 
calculation or formula used to address one pollutant in one particular 
part of a water body, but as the program has evolved, the concept of a 
TMDL has become more expansive. Overall, the goal of developing a TMDL 
is to end up with a plan, including the actions needed, to meet water 
quality standards and restore impaired water bodies. After states 
develop TMDLs, they take the lead in implementing these plans, and it 
may take many years to see actual improvements in water quality. 
[Footnote 4] 

EPA oversees states' TMDL development and implementation efforts under 
its TMDL program, which is a primary part of EPA's efforts to restore 
impaired waters. In addition, several key responsibilities related to 
TMDL development and implementation are found in other EPA water 
quality-based programs. For the purposes of this report, we use the 
term "TMDL program" to include the TMDL-related responsibilities in 
other EPA water quality-based programs. To guide the development of 
TMDLs, EPA issued regulations in 1985, and revised them in 1992, which 
states are to use to this day. These regulations establish minimum 
requirements for establishing TMDLs and, according to EPA officials, 
should lead to TMDLs that, if implemented, enable water bodies to 
attain water quality standards. Under the regulations, TMDLs are to 
identify the applicable water body, pollutant of concern, and 
pollutant sources--as well as the share of pollutant reduction to be 
achieved by point sources and nonpoint sources.[Footnote 5] 

According to EPA officials, states took little action under the Clean 
Water Act's requirement to develop TMDLs until the mid-1990s, when 
citizen groups sued EPA in several states for not doing so in the 
absence of state action. Since that time, states developed nearly 
50,000 TMDLs, in part as a result of EPA efforts to meet the schedules 
in consent decrees stemming from the lawsuits, according to EPA 
officials.[Footnote 6] EPA approved about 35,000 of these TMDLs more 
than 5 years ago, and, according to EPA officials, states should have 
implemented these TMDLs to some extent by the time of our review. For 
purposes of this report, we define TMDLs approved by EPA through 
December 31, 2007--that is, TMDLs developed more than 5 years ago--as 
long-established TMDLs. 

Since EPA last revised its TMDL regulations in 1992, the agency and 
others have been asked to evaluate states' TMDL development and 
implementation. In 1996, EPA convened a Federal Advisory Committee to 
examine the lack of progress in TMDL development and implementation 
and the pervasiveness of nonpoint source pollution.[Footnote 7] In 
2001, the National Research Council (NRC) of the U.S. National 
Academies issued a report assessing the scientific basis of the TMDL 
program, which described a number of features to strengthen TMDL 
development and implementation.[Footnote 8] For example, the report 
recommended that EPA's TMDL program focus on improving the condition 
of water bodies as measured by the attainment of designated uses, such 
as supporting fish populations or human recreation, rather than 
administrative outputs, such as numbers of permits issued or TMDLs 
developed. The report further recommended that TMDLs be implemented 
adaptively, that is, that periodic assessments should determine if 
water bodies are attaining water quality standards, and TMDLs should 
be revised if water quality is not improving. More recently, EPA 
published a series of papers in response to a 2007 EPA Office of 
Inspector General report, which found that the agency had limited data 
on TMDL implementation activities.[Footnote 9] In these papers, EPA 
identified particular factors associated with TMDLs that have been 
effectively implemented--that is, factors having helped the water 
bodies to which they apply partially or fully attain water quality 
standards--including the targeting of certain federal and state funds 
toward specific practices, such as keeping cattle out of streams, in 
locations where TMDLs have been developed.[Footnote 10] 

Nearly 20 years after developing the first TMDLs, states have largely 
developed those TMDLs required by consent decree, according to EPA 
officials, but face the task of implementing the projects and 
activities described in these TMDLs to control pollution and 
developing TMDLs for other waters that states have identified as 
impaired.[Footnote 11] EPA acknowledges that it must determine the 
extent to which the agency will encourage states to continue to direct 
their resources toward developing TMDLs for the tens of thousands of 
impaired water bodies that do not yet have them or focus on helping 
states implement the TMDLs that have already been developed. In this 
context, you asked us to study EPA's efforts to restore and maintain 
the nation's waters through its TMDL program. Our objectives were to 
examine (1) EPA's and states' responsibilities in the TMDL program, 
(2) what is known about the status of long-established TMDLs, (3) the 
extent to which long-established TMDLs contain key features that 
enable attainment of water quality standards, and (4) the extent to 
which such TMDLs exhibit factors that facilitate effective 
implementation. 

To address each of these objectives, we reviewed EPA reports and 
independent, peer-reviewed reports on the status of TMDLs and 
assessments of TMDL implementation projects and activities. In 
addition, we visited a nonprobability sample of four states in two of 
EPA's 10 regions to review EPA's and states' responsibilities in 
developing and implementing TMDLs, as well as any challenges the 
states may have faced in doing so.[Footnote 12] We selected these 
regions and states on the basis of the number of TMDLs each state had 
developed, types of water pollution, differences in state authorities 
to implement TMDLs, and extent of each state's TMDL implementation 
efforts. To examine what is known about the status of long-established 
TMDLs, we used an EPA database known as Assessment, TMDL Tracking and 
Implementation System. To assess the reliability of data from this 
database, we (1) electronically tested required data elements, (2) 
reviewed existing information about the data and the system that 
produced them, and (3) interviewed agency officials knowledgeable 
about the data. We found these data to be sufficiently reliable for 
purposes of this report. We also used the database to select a random, 
nationwide, representative sample of 191 long-established TMDLs, which 
we defined to be those approved by EPA through December 31, 2007. We 
used the sample of TMDLs in the following ways to assess TMDL 
development and implementation: 

* To examine the extent to which long-established TMDLs contain key 
features needed to enable a water body to attain water quality 
standards, we selected water resource experts to review a random 
subsample of 25 TMDLs drawn from the nationwide sample of 191 long-
established TMDLs.[Footnote 13] This subsample of 25 TMDLs is too 
small to be representative of all TMDLs, but it is an unbiased set 
that provides detailed examples of TMDL development in specific states 
and regions and helps illustrate TMDL development challenges and 
successes. We selected three experts to review these 25 TMDLs from 
among the eight authors of the 2001 NRC report on EPA's TMDL program, 
which is the most recent effort undertaken by the National Academies 
to evaluate the TMDL approach to water quality management. To increase 
the reliability of our findings, each TMDL was independently reviewed 
by two experts. 

* To examine the extent to which long-established TMDLs exhibit 
factors EPA identified that facilitate their implementation, we 
surveyed state TMDL coordinators responsible for implementing the 191 
long-established TMDLs in our representative sample; we received 177 
responses for a 93 percent response rate to our survey.[Footnote 14] 
To perform this analysis, we first identified factors EPA has 
determined to be necessary for facilitating effective TMDL 
implementation. We then developed a survey on whether and to what 
extent these factors were present in the representative sample of 191 
long-established TMDLs and whether and to what extent water quality 
had improved. We used the results of this survey to determine which 
factors were viewed by our survey respondents as helpful in TMDL 
implementation and, if so, to what extent TMDLs exhibited them. The 
estimates from our survey have a margin of error of no more than 10 
percent at the 95 percent confidence level, unless otherwise noted. 
[Footnote 15] 

We also interviewed officials in each of EPA's 10 regional offices 
responsible for reviewing and approving TMDLs. We discussed with these 
officials the extent to which regional EPA officials encourage state 
programs to ensure the presence of factors facilitating effective TMDL 
implementation. We obtained information on the regional offices' 
review and approval of state-developed TMDLs, as well as on their 
oversight of state programs, including how the offices encourage state 
programs to ensure the presence of factors needed for effective TMDL 
implementation. Appendix I describes our objectives, scope, and 
methodology in more detail. 

We conducted this performance audit from July 2012 to December 2013 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: 

The Clean Water Act establishes a nationwide approach to improve and 
maintain the quality of rivers, streams, lakes, and other water 
bodies, which involves different programs aimed at reducing pollution 
of the nation's waters. Under this approach, states are to set water 
quality standards, monitor water quality, and assess water quality 
against the applicable standards. Through the monitoring and 
assessment process, states are to identify water bodies that do not 
meet their standards, and these water bodies are considered to be 
impaired. Under the act, point source dischargers of pollutants must 
obtain permits, issued by authorized states or EPA regions. The 
permits limit the amount of pollutants according to specific 
guidelines set by EPA, as well as other considerations.[Footnote 16] 
Under the act and implementing regulations, states must develop a TMDL 
for each of the pollutants affecting each water body identified as 
impaired; in so doing, states typically identify pollutant sources and 
the actions needed to control pollution and restore water bodies, 
according to EPA documents. In identifying pollutants and their 
sources--both point and nonpoint, where applicable--a TMDL is to 
specify the amount by which one or more pollutants needs to be reduced 
to meet water quality standards. EPA's TMDL program has evolved over 
the last several decades as nationwide efforts have shifted from a 
focus on controlling point source pollutants to addressing nonpoint 
source pollution. Figure 1 shows some of the activities impairing 
water bodies. 

Figure 1: Activities That Can Impair Water Bodies: 

[Refer to PDF for image: illustration] 

Agriculture; 
Airborne pollution; 
Boating; 
Dam; 
Factory; 
Forestry; 
Irrigation; 
Irrigation diversion; 
Pump from Aquifer; 
Parking lots; 
Roofs and lawns; 
Tributaries; 
Wastewater treatment plant. 

Source: GAO. 

Note: The figure shows that impairment of water bodies may stem from 
pollution by point sources--which include industrial facilities, such 
as factories and wastewater treatment plants, and other sources that 
discharge wastewater from pipes or other discrete points--or nonpoint 
sources such as airborne pollution; agricultural fields; forestry; and 
runoff from roofs, lawns, parking lots, and roads. 

[End of figure] 

States Set Water Quality Standards, Assess Water Quality, and Identify 
Impaired Waters: 

Under the Clean Water Act, each state is to establish its own water 
quality standards and review and update these standards every 3 years. 
[Footnote 17] Water quality standards consist of designated uses and 
water quality criteria.[Footnote 18] Designated uses reflect the 
intended use and value of a water body, such as public drinking water 
supply; protection of fish, shellfish, and other wildlife; or 
recreational, agricultural, industrial, and navigational purposes. 
Water quality criteria generally describe the chemical, physical, and 
biological conditions necessary to achieve and protect designated 
uses, such as nitrogen concentrations, stream structure, and health of 
aquatic life, respectively.[Footnote 19] 

EPA approves states' water quality standards and establishes 
recommended national water quality criteria, which states can adopt, 
or states can develop their own criteria, subject to EPA approval. 
According to EPA documents, water quality standards should be 
comprehensive and strong enough to achieve the Clean Water Act's goal 
of "chemical, physical, and biological integrity" for the state's 
waters. According to the 2001 NRC report, it is not uncommon for water 
bodies to be largely free of chemical pollutants yet still vulnerable 
to other stressors, including stressors affecting habitat, stream 
morphology (structure), and streamflow, which are critical to 
supporting aquatic life. For example, a river or stream may be 
channelized, as when farmers straighten the waterway to maximize the 
amount of land that can be farmed and make it easier to move machinery 
across fields, thus altering the waterway's characteristics (see 
figure 2). Even if stream channeling does not contribute many chemical 
pollutants to a waterway, it typically removes vegetation along stream 
banks, alters streambed structure and water flows, and limits the 
stream's capacity to support aquatic life, according to EPA documents; 
such effects may extend far downstream. 

Figure 2: An Eastern Washington Stream Channelized to Maximize Crop 
Yield: 

[Refer to PDF for image: photograph] 

Source: GAO. 

[End of figure] 

After setting water quality standards, states are to monitor and 
assess water quality to determine the degree to which the standards 
are being met. According to EPA documents, states generally base their 
water quality assessments on three broad types of monitoring data: 
chemical, physical, and biological.[Footnote 20] States assess water 
bodies, often using each data type, and then integrate these data with 
other data types for an overall assessment. According to EPA 
documents, depending on the designated use for a given water body, one 
type of data may be more informative than others. 

* Chemical data include measurements of specific chemical constituents 
in water, sediments, and fish tissue. Examples include measuring the 
concentrations of metals, oils, pesticides, and nutrients such as 
nitrogen and phosphorus. Monitoring for specific chemicals helps 
states identify specific pollutants and helps trace pollutants to 
their sources. 

* Physical data include characteristics of water and water body 
configuration, such as temperature, flow, and physical structure. 
Physical attributes can be useful indicators of potential problems, 
often because they can modify the effects of chemicals. 

* Biological data are objective measurements of living aquatic 
communities (usually aquatic insects, fish, or algae) used to evaluate 
the condition of an aquatic ecosystem. Biological data are often the 
most appropriate measures for assessing if waters can support aquatic 
life. 

Using existing and readily available water quality-related data and 
other information, states identify impaired waters that do not meet 
applicable water quality standards and need TMDLs. EPA's regulations 
require that states report biennially to EPA on the quality of their 
waters, and states generally fulfill this requirement by submitting 
integrated reports to EPA every 2 years.[Footnote 21] In these 
reports, which list impaired water bodies, each water body is to be 
listed according to the following categories: 

* Category 1: all designated uses are supported, no use is threatened. 

* Category 2: available data and information indicate that some, but 
not all, designated uses are supported. 

* Category 3: there is insufficient available data and information to 
determine whether designated uses are supported. 

* Category 4a: a state developed TMDL has been approved by EPA or a 
TMDL has been established by EPA. 

* Category 4b: other required control measures are expected to result 
in the attainment of an applicable water quality standard in a 
reasonable period of time.[Footnote 22] 

* Category 4c: water quality is impaired by something other than a 
pollutant, such as a dam. 

* Category 5: available data and information indicate that at least 
one designated use is not being supported or is threatened, and a TMDL 
is needed. 

Evolution of the TMDL Program: 

The TMDL program has evolved over the last several decades as 
nationwide efforts have shifted from a focus on controlling point 
source pollutants to addressing nonpoint source pollution, according 
to EPA documents. Few TMDLs were developed during the first two 
decades after the TMDL program was created under the Clean Water Act. 
Figure 3 shows the extent of state-identified impaired water bodies 
(in acres of lakes and miles of rivers) and the cumulative number of 
TMDLs developed since passage of the act.[Footnote 23] 

Figure 3: Extent of Impaired Freshwaters and TMDLs Developed since 
Passage of the Clean Water Act: 

[Refer to PDF for image: combined vertical bar and line graph] 

Year: 1996; 
Cumulative TMDLs (number): 122. 

Year: 1997; 
Cumulative TMDLs (number): 459. 

Year: 1998; 
Cumulative TMDLs (number): 861. 

Year: 1999; 
Cumulative TMDLs (number): 1,191. 

Year: 2000; 
Cumulative TMDLs (number): 2,747. 

Year: 2001; 
Cumulative TMDLs (number): 5,326. 

Year: 2002; 
Cumulative TMDLs (number): 8,065; 
GAO estimate of impaired lakes (thousands of acres): 19,300; 
GAO estimate of impaired rivers (thousands of miles): 1,546; 
Previously impaired water bodies attaining all designated uses 
(number): 23. 

Year: 2003; 
Cumulative TMDLs (number): 11,068; 
Previously impaired water bodies attaining all designated uses 
(number): 4. 

Year: 2004; 
Cumulative TMDLs (number): 14,441; 
GAO estimate of impaired lakes (thousands of acres): 26,799; 
GAO estimate of impaired rivers (thousands of miles): 1,538; 
Previously impaired water bodies attaining all designated uses 
(number): 159. 

Year: 2005; 
Cumulative TMDLs (number): 18,754; 
Previously impaired water bodies attaining all designated uses 
(number): 107. 

Year: 2006; 
Cumulative TMDLs (number): 22,963; 
GAO estimate of impaired lakes (thousands of acres): 24,290; 
GAO estimate of impaired rivers (thousands of miles): 1,581; 
Previously impaired water bodies attaining all designated uses 
(number): 85. 

Year: 2007; 
Cumulative TMDLs (number): 27,284; 
Previously impaired water bodies attaining all designated uses 
(number): 233. 

Year: 2008; 
Cumulative TMDLs (number): 36,548; 
GAO estimate of impaired lakes (thousands of acres): 26,506; 
GAO estimate of impaired rivers (thousands of miles): 1,749; 
Previously impaired water bodies attaining all designated uses 
(number): 339. 

Year: 2009; 
Cumulative TMDLs (number): 40,950; 
Previously impaired water bodies attaining all designated uses 
(number): 318. 

Year: 2010; 
Cumulative TMDLs (number): 43,517; 
GAO estimate of impaired lakes (thousands of acres): 27,698; 
GAO estimate of impaired rivers (thousands of miles): 1,924; 
Previously impaired water bodies attaining all designated uses 
(number): 344. 

Year: 2011; 
Cumulative TMDLs (number): 46,346; 
Previously impaired water bodies attaining all designated uses 
(number): 252. 

Year: 2012; 
Cumulative TMDLs (number): 49,230; 
GAO estimate of impaired lakes (thousands of acres): 27,923; 
GAO estimate of impaired rivers (thousands of miles): 1,847; 
Previously impaired water bodies attaining all designated uses 
(number): 244. 

Source: GAO analysis of EPA data. 

Note: Before 2002, states identified impaired water bodies using 
varied methods and reported them to EPA in varied ways, making it 
difficult to reliably identify the number of impaired water bodies 
nationwide, according to EPA officials. 

[End of figure] 

A number of major events have occurred over the last 40 years, which 
have shaped the program as it evolved. After the Clean Water Act was 
passed in 1972, EPA and the states focused primarily on, and made 
progress in, reducing pollutants discharged from point sources, but 
they made less progress in controlling nonpoint source pollution. In 
1985, with minor amendments in 1992, EPA published regulations 
requiring the establishment of TMDLs and defining the minimum 
requirements TMDLs were to contain to qualify for approval, including 
the requirement that pollutant loads, or amounts, were to be set so as 
to attain water quality standards. Between these years, Congress in 
1987 amended the Clean Water Act, adding section 319 to explicitly 
address nonpoint source pollution through a cooperative, grant-based 
program with states.[Footnote 24] Among other provisions, the nonpoint 
source program was to fund voluntary projects aimed at curtailing 
nonpoint source pollution. In amending the act, Congress established 
the nonpoint source program to help states implement nonpoint source 
controls, largely through voluntary means and financial incentives, 
such as those set out in TMDLs. At the same time, some Members of 
Congress indicated that this provision was a starting point, to be 
changed if reliance on voluntary approaches did not significantly 
improve water quality. 

In 1996, EPA determined that it needed a comprehensive evaluation of 
the TMDL program, in large part because of limited progress states had 
made in developing TMDLs. Moreover, environmental and public-interest 
organizations had started filing lawsuits alleging that EPA should be 
held accountable under the Clean Water Act for failure to oversee TMDL 
development efforts and should establish TMDLs where states had not 
done so. EPA convened a Federal Advisory Committee, which developed 
more than 170 separate recommendations to strengthen the TMDL program, 
ultimately reporting in 1998 that (1) restoring impaired waters should 
be a high priority; (2) implementing TMDLs, particularly those for 
nonpoint source pollution, should be the key to program success; and 
(3) strengthening the federal government's capacity to carry out the 
TMDL program should be strengthened.[Footnote 25] 

Following on the heels of the Federal Advisory Committee's 
recommendations, EPA in 2000 promulgated significant revisions to the 
TMDL program in a rule published in the Federal Register. The preamble 
to the rule stated that although significant improvements had been 
made to water quality--particularly in stemming pollution from point 
sources--major challenges remained for addressing nonpoint source 
pollution. The rule specified elements that TMDLs needed to contain to 
be approved by EPA, including (1) implementation plans, which were to 
contain lists of specific actions to reduce pollution and schedules 
for reducing pollutant loads; (2) "reasonable assurances" that TMDLs 
for waters impaired by point and nonpoint source pollution show that 
nonpoint source pollution would actually be reduced; and (3) a plan 
for monitoring implementation projects and activities and TMDL 
revision procedures if such projects and activities did not have the 
anticipated effect on water quality. Finally, the rule also set a goal 
of 5 years to implement a given TMDL and 10 years for attaining water 
quality standards whenever practicable. In the preamble to the rule, 
EPA stated that the TMDL program needed new regulations because 
existing regulations had brought about insufficient improvement in 
water quality. 

In 2000, however, Congress delayed EPA's implementation of the rule by 
prohibiting the agency from using funds made available for fiscal 
years 2000 and 2001 to make a final determination on, or implement, 
the 2000 rule.[Footnote 26] Then, in association with the fiscal year 
2001 appropriations bill, the congressional appropriations committee 
directed EPA to contract with NRC to evaluate the adequacy of 
scientific methods and approaches available to support development and 
implementation of TMDLs. NRC issued its report in April 2001, 
identifying features to strengthen TMDL development and implementation. 

In 2002, EPA withdrew its rule, stating that significant changes would 
be needed before the rule could serve as a blueprint for an efficient 
and effective TMDL program. EPA also stated that it needed more time 
to evaluate whether and how to revise then-current regulations, 
reporting that the agency was not sure how long this effort would 
take. EPA cited significant controversy, pending litigation, and lack 
of stakeholder consensus on key aspects of the rule as the main 
challenges that a new rule would need to overcome. [Footnote 27] Since 
then, according to EPA officials, EPA has focused its efforts on 
implementing the TMDL program, including the review and approval of 
TMDLs submitted by states according to the 1985 regulations and 1992 
amendments. EPA has also issued additional guidance to states on a 
range of topics, including on specific elements that a TMDL should 
contain. 

In 2013, EPA drafted a long-term "vision" for its TMDL program, 
acknowledging that many TMDLs have been developed for impaired waters, 
although much remains to be done to restore water bodies to which 
TMDLs apply and to develop TMDLs for still more water bodies. EPA's 
vision seeks to help states set priorities for TMDLs to be developed 
and implemented, now that states are emerging from several decades of 
consent-decree-driven time frames for TMDL development and have more 
flexibility in addressing technical challenges, limited funding and 
other resources, state-specific legislative and regulatory frameworks, 
and public objectives. For example, by 2016, states are to review; 
systematically set priorities for; and, in their biennial integrated 
reports, identify waters for restoration and protection. Also by 2016, 
states are to identify and coordinate implementation of the water 
quality efforts of other federal agencies, such as the U.S. Department 
of Agriculture (USDA), to achieve each state's water quality goals. 

EPA and the States Have Certain Responsibilities for Developing and 
Implementing TMDLs: 

Under the Clean Water Act, EPA and the states each have certain 
responsibilities for developing and implementing TMDLs. EPA 
establishes elements for TMDLs for review and approval, funds some 
TMDL development and implementation, and provides technical assistance 
to states. States are responsible for developing TMDLs and 
implementing the pollution control actions prescribed by the plans for 
point sources, while they take the lead for implementing any actions 
for nonpoint sources.[Footnote 28] Figure 4 shows the key steps and 
parties involved in developing and implementing TMDLs. 

Figure 4: Key Steps in Developing and Implementing TMDLs: 

[Refer to PDF for image: illustration] 

1. State sets water quality standards, assesses water body, and 
determines if water quality is impaired. 

2. State develops total maximum daily load (TMDL) for impaired
water body. 

3. EPA reviews TMDL. 

4. EPA rejects TMDL; or EPA approves TMDL. 

5. If rejected, EPA itself develops TMDL within 30 days. 

6. State implements TMDL. 

Point sources: 

7. State or EPA incorporates limits from TMDLs into discharge permits. 

8. State monitors point sources for compliance with permits. 

9. Water quality changes[B]. 

Nonpoint sources: 

7. State uses federal and nonfederal funding to help landowners and
other stakeholders take actions prescribed by TMDL. 

8. Landowners or other stakeholders take action[A]. 

9. Water quality changes[B]. 

Source: GAO analysis of EPA information. 

[A] The Clean Water Act's approach to curtailing nonpoint source 
pollution depends in large part on voluntary participation by private 
landowners; other actors may include state or local regulatory or 
nonregulatory programs. 

[B] According to EPA officials, it may take years for changes to occur 
in water quality after implementation of best management practices or 
other projects and activities prescribed by TMDLs. 

[End of figure] 

EPA's Responsibilities: 

EPA has three main responsibilities for developing and implementing 
TMDLs. First, EPA establishes regulations and guidance for TMDL 
development and reviews TMDL documents that states submit. Under 
guidance implementing the agency's regulations promulgated in 1985 and 
amended in 1992, EPA reviews the adequacy of TMDLs on the basis of 11 
elements (see table 1). If EPA rejects a state's TMDL, EPA must itself 
develop one for the identified water body within 30 days. According to 
EPA program officials, its regional offices review TMDL documents and 
approve them when the information provided in those documents 
demonstrates that if the TMDL were implemented as written, it would 
enable an impaired water body to attain the relevant water quality 
standards. 

Table 1: Elements for Approvable TMDLs: 

TMDL element: 1. Identification of water body, pollutant of concern, 
pollutant sources, and priority ranking; 
Description: A TMDL should identify the water body and pollutant for 
which the TMDL is established; link between the pollutant of concern 
and applicable water quality standard; point and nonpoint sources of 
the pollutant of concern; and point source permits, if any. 

TMDL element: 2. Description of the applicable water quality standard 
and numeric water quality target; 
Description: A TMDL must describe the applicable water quality 
standard. The TMDL must also identify the numeric water quality target 
and how meeting the target will allow designated uses to be attained. 

TMDL element: 3. Loading capacity--linking water quality and pollutant 
sources; 
Description: A TMDL must identify the loading capacity of a water body 
for the applicable pollutant, that is, the greatest amount of a 
pollutant that the water body can receive and still attain the 
applicable water quality standard. 

TMDL element: 4. Load allocations; 
Description: A TMDL must include load allocations, which identify the 
portion of the loading capacity assigned to existing and future 
nonpoint sources--in other words, the share of the pollutant limit 
that nonpoint sources must meet. 

TMDL element: 5. Wasteload allocations; 
Description: A TMDL must include wasteload allocations, which identify 
the portion of the loading capacity assigned to each existing and 
future point source--in other words, the share of the pollutant limit 
that point sources must meet. 

TMDL element: 6. Margin of safety; 
Description: A TMDL must include a margin of safety--such as 
conservative assumptions--to account for any lack of knowledge 
concerning the relationship between load and wasteload allocations and 
water quality. 

TMDL element: 7. Seasonal variation; 
Description: A TMDL must consider how seasonal variations affect 
loading capacity, and it must describe the method that accounts for 
seasonal variation. 

TMDL element: 8. Reasonable assurances; 
Description: When a TMDL is developed for waters impaired by both 
point and nonpoint sources, a "TMDL should provide reasonable 
assurances that nonpoint source control measures will achieve the 
expected load reductions." This information is necessary for EPA to 
determine that the TMDL has been established at a level necessary to 
implement water quality standards. 

TMDL element: 9. Monitoring plan to track TMDL effectiveness; 
Description: EPA recommends that TMDLs include a monitoring plan to 
track effectiveness of implementation, particularly when a TMDL 
involves both point and nonpoint sources (emphasis added). The 
monitoring plan should describe data to be collected to determine if 
load reductions provided for in the TMDL are occurring and leading to 
attainment of water quality standards. 

TMDL element: 10. Implementation; 
Description: EPA encourages states to develop implementation plans 
that include reasonable assurances that nonpoint source pollution 
reductions will occur (emphasis added). 

TMDL element: 11. Public participation; 
Description: EPA recommends full and meaningful public participation 
in the TMDL development process (emphasis added). States must subject 
TMDL calculations to public review. Inadequate public participation 
may be a basis for disapproving a TMDL. 

Source: EPA. 

[End of table] 

EPA's second responsibility is to provide funding, which helps states 
develop and implement TMDLs. EPA has two grant programs under the 
Clean Water Act: grants for water pollution control programs (section 
106) and grants for state nonpoint source management programs (section 
319). Portions of grants from EPA for water pollution control give 
states funding to assist in administering programs for the prevention, 
reduction, and elimination of pollution, including, but not limited 
to, TMDLs. Similarly, portions of grants for state nonpoint source 
management programs fund state-selected projects to reduce nonpoint 
source pollution and restore impaired water bodies. These state grant 
programs generally fund implementation of TMDLs addressing nonpoint 
sources. In implementing section 319, EPA awards funds to states, 
provided they meet various minimum conditions, including targeting 
funds to geographic areas in need of restoration and demonstrating 
annual progress. EPA has discretion by statute to (1) add terms and 
conditions to grants; (2) require additional information on 
applications; and (3) request additional information, data, and 
reports that it considers necessary to determine continuing state 
eligibility for grants. For example, EPA's regional offices establish 
commitments from each state to develop a specific number of TMDLs each 
year, and each state is to report back to the relevant regional office 
how many it has developed. Under section 319, to receive funding, 
states must obtain a determination from EPA that they made 
satisfactory progress in meeting their nonpoint source program goals 
from the prior year.[Footnote 29] 

For fiscal years 2008 to 2012, EPA budget was $432 million per year, 
on average, for its grants supporting state efforts under its water 
pollution and nonpoint source management programs, including for 
developing and implementing TMDLs. During this time, EPA's water 
pollution control program was budgeted at $235.5 million per year, on 
average, and the nonpoint source management program was budgeted at 
$196.5 million per year, on average. Less than half of these amounts, 
however, are applied toward TMDL development and implementation 
because of the many other state water quality efforts they also 
support. For example, EPA officials told us, a majority of the 
agency's grants under the water pollution control program support 
states' permitting, monitoring, and assessment activities, not TMDL 
development per se. Moreover, officials estimated that about one-third 
of grants through the nonpoint source management program are used to 
implement TMDLs, whereas the remaining two-thirds support other 
aspects of states' nonpoint source management programs. 

EPA's third responsibility with regard to TMDLs is to provide 
technical assistance to help states restore water bodies through the 
TMDL process. For example, in addition to developing numerous models, 
monitoring methodologies, and databases over the past several decades, 
the agency has developed recent tools to assist states in (1) 
identifying the causes and effects of various impairments and (2) 
assessing different water bodies to identify those with the highest 
likelihood of successful restoration. Specifically, according to EPA 
documents, EPA has developed tools such as the Causal Analysis/ 
Diagnosis Decision Information System and the Recovery Potential 
Screening tools. The first is a web-based system that helps scientists 
and engineers in EPA regions and states assess water bodies and 
identify causes of impairment, with a focus on biological assessments. 
Biological assessments can evaluate the cumulative impacts of 
chemicals for which no water quality standards exist, as well as of 
nonchemical stressors such as flow alteration (e.g., stream 
channeling) and invasive species, according to EPA documents. The 
second tool is also web based and, according to EPA documents, assists 
state officials in considering where to invest their water quality 
restoration efforts for greater likelihood of success, given specific 
traits of their own geographic area and communities. The tool is to 
help states compare the relative restoration and recovery potential of 
large numbers of water bodies by measuring, for each water body or 
watershed, several indicators showing the likelihood that a 
restoration effort can succeed.[Footnote 30] 

States' Responsibilities: 

Under the Clean Water Act, states are responsible for developing TMDLs 
for waters they identify as impaired and, generally, for implementing 
the pollution control actions they prescribe for point sources; 
whereas for nonpoint sources, states take the lead in implementing 
projects and activities intended to achieve pollution 
reductions.[Footnote 31] To develop TMDLs, states identify individual 
pollutants or pollution impairing water quality; point and nonpoint 
sources of the pollution; and, generally, the actions needed to reduce 
pollution to an amount that meets water quality standards. TMDL 
development generally consists of the following steps: 

* identifying the pollutant responsible for listing a water body as 
impaired; 

* estimating the water body's loading capacity for that pollutant; 

* estimating pollutant loading in the water body from all sources; 

* determining pollutant reductions to meet the loading capacity; 

* allocating the allowable pollutant load among different pollutant 
sources--both point (i.e., wasteload allocation) and nonpoint (i.e., 
load allocation)--in a manner that achieves water quality; and: 

* describing a plan for the actions that should be implemented to 
reduce pollutants from entering the water body. 

As part of the TMDL development process, states use different 
approaches for point source and nonpoint source pollution. EPA and the 
states regulate point source discharges of pollutants by means of 
National Pollutant Discharge Elimination System permits, which are to 
incorporate overall pollutant loads--wasteload allocations--
established by TMDLs. Under the National Pollutant Discharge 
Elimination System, all facilities that discharge pollutants from any 
point source into U.S. waters must generally obtain a permit, 
typically from their state or EPA region. During TMDL development, a 
state identifies each point source discharging to the water body 
subject to a TMDL and determines specific limits for the pollutants 
addressed by the TMDL that would enable the water body to attain water 
quality standards. For example, in developing one California TMDL, the 
regional water quality control board determined that to meet water 
quality standards, point sources could discharge no more than 5.5 
pounds of nitrogen per day into the water body subject to the TMDL. 
States who issue these permits are thus responsible for implementing 
the point source control component of TMDLs. Unlike point source 
pollutants, nonpoint source pollution is not regulated through 
permitting and discharge limits. A state or EPA identifies pollution 
caused by adjacent or upstream lands contributing polluted runoff or 
other pollution to a water body and sets load allocations for 
particular pollutants of concern. The California TMDL, for example, 
states that nonpoint sources, which discharged about 84 pounds of 
nitrogen per day all together, were to collectively discharge no more 
than 8.5 pounds per day. 

Then, depending on the state or EPA office that develops it, a TMDL 
may identify, in more or less detail, the actions needed to control or 
reduce runoff into the water body. Some states' TMDL documents 
prescribe specific actions to reduce runoff from fields; for example, 
they include the location and type of projects and activities that may 
be needed to reduce pollution from different parcels of land. Such 
projects and activities include agricultural best management 
practices, such as installing fences to exclude cattle from stream 
banks, planting strips of grass along a stream bank to absorb nitrogen 
and phosphorus runoff from fertilized fields, or building in-stream 
structures to prevent sediment from flowing downstream. For example, 
the TMDL for the South Santiam River in Oregon describes and locates 
soil conditions, vegetation, and human uses affecting the river and 
its tributaries. It also describes in detail specific steps to address 
elevated temperature in the river, including restoring stream 
channels, native vegetation, and natural streamflow, and it aligns 
such steps with the specific conditions and areas within the TMDL's 
geographic boundary. Other state TMDLs simply direct that runoff from 
all nonpoint sources is to fall below the overall load allocated among 
them. For example, the TMDL for Mariposa Lake in Iowa does not 
identify specific locations from which nonpoint source pollution comes 
or the type and extent of actions needed to enable the lake to attain 
water quality standards. Runoff from 580 acres of nearby farmland 
flows into the lake, which is to support designated uses of swimming 
and aquatic wildlife, but the TMDL identifies a single maximum load of 
phosphorus to be collectively achieved from all these acres, without 
identifying the specific locations contributing the most polluted 
runoff or the extent to which actions are needed that could address 
polluted runoff. 

Once a TMDL is developed for a water body, states have certain 
responsibilities for implementing it--that is, for taking the actions 
prescribed by the TMDL to reduce point source pollution. To implement 
TMDLs for waters impaired by point sources, states are to ensure that 
pollutant discharge levels specified in permits are consistent with 
the assumptions and requirements of any available wasteload allocation 
for the facility specified in TMDL documents.[Footnote 32] 
Specifically, state staff determine whether the overall wasteload 
allocation for that pollutant can be met if each point source of that 
pollutant discharges no more than its permitted limit or whether a 
facility's permitted limit for the pollutant in question needs to be 
lower. If a TMDL has been developed for the receiving waters, the 
permitted limits for point sources must be consistent with the 
wasteload allocations in the TMDLs. 

In contrast, states take the lead in implementing TMDLs for waters 
impaired by nonpoint source pollution by relying on projects funded 
through their nonpoint source management programs and through a mix of 
other federal and state funds.[Footnote 33] In part because most land 
areas responsible for nonpoint source pollution are privately owned, 
EPA and the states use funding to try to secure voluntary landowner 
participation. In particular, states are encouraged by EPA (under 
guidance implementing section 319 of the Clean Water Act) to 
coordinate with other agencies, particularly USDA, that fund projects 
that may help reduce nonpoint source pollution. Under the Farm Bill 
and other legislation, USDA administers agricultural conservation 
programs that fund best management practices on agricultural land, 
providing funding assistance to landowners for projects aimed at 
reducing soil erosion and air pollution, as well as nonpoint source 
water pollution. Such programs, including USDA's Environmental Quality 
Incentives Program, have in the past received more funding than EPA's 
nonpoint source management program, but not all such funding is 
specifically directed toward projects in impaired or threatened 
watersheds. For example, of the $1.4 billion in fiscal year 2012 that 
the Environmental Quality Incentives Program received, $33 million was 
specifically set aside to help farmers and ranchers implement 
conservation practices to protect water quality in impaired or 
threatened watersheds. 

In addition, according to EPA documents, some states have taken steps 
through state law or guidance to further the implementation of TMDLs. 
Specifically, 10 states had taken steps to require formal TMDL 
implementation plans that specifically identified actions that could 
reduce nonpoint source pollution in a water body, according to a 2008 
report.[Footnote 34] Implementation plans may describe how nonpoint 
source load allocations are to be achieved for individual water 
bodies, for several water bodies within a watershed, or for all 
affected water bodies in the state, according to EPA. Moreover, a 
handful of states regulate nonpoint source pollution to some extent, 
according to the study.[Footnote 35] For example, farms in 
Pennsylvania must have a conservation plan for sediment;[Footnote 36] 
compliance with the plan can help farmers conserve soil. In 
Washington, dairies must have a nutrient management plan to prevent 
nutrients from entering nearby water bodies. Pennsylvania and 
Washington officials told us that the mechanisms are generally limited 
in scope, and state and local officials reported that success or 
failure to implement TMDLs addressing significant nonpoint source 
pollution depends largely on voluntary initiatives. 

EPA Has Limited Information about the Status of Long-Established 
TMDLs, but States Report That Few Water Bodies with Such TMDLs Have 
Attained Water Quality Standards: 

EPA tracks basic information on TMDL development, such as the number, 
location, and type of long-established TMDLs but, generally, does not 
have information on the extent to which the TMDLs have been 
implemented or have improved the quality of impaired water bodies. 
EPA's ability to track TMDL implementation is hindered by data system 
limitations and unavailable USDA data. In addition, results from our 
survey of state TMDL coordinators show that states have little 
information on TMDL implementation, and, where information exists, few 
water bodies to which long-established TMDLs apply have attained water 
quality standards. 

EPA Systematically Tracks Basic Information on TMDL Development but 
Has Limited Information on TMDL Implementation and Water Quality: 

Through its national databases, EPA systematically tracks basic 
information related to TMDL development, but these databases contain 
limited information on the extent of TMDL implementation and 
associated changes in water quality. This basic information includes 
the number of TMDLs developed, the name of water bodies to which TMDLs 
apply, pollutants contributing to impairment, and probable sources of 
the impairments, as well as information on the extent to which states 
use nonpoint source management grant funds to support TMDL development 
and implementation. Generally, EPA has information related to TMDL 
development in its Assessment, TMDL Tracking and Implementation 
System, much of which is entered by EPA regional offices with data 
provided by states. From this database, the agency knows the following 
facts, as of August 2013: 

* EPA has approved 50,184 TMDLs since fiscal year 1996, the year 
states generally began developing TMDLs; 34,794 of these, which we 
refer to as long-established TMDLs, were approved by December 31, 
2007. The information that EPA tracks shows that states have developed 
TMDLs for many water bodies the states have identified as impaired and 
that the probable causes of impairment include chemical, physical, and 
biological stressors. 

* The agency knows that 52 percent address only nonpoint source 
pollution, 5 percent address only point source pollution, and 41 
percent address both point and nonpoint source pollution.[Footnote 37] 

* The states with the largest number of approved TMDLs include 
Pennsylvania (7,084 TMDLs), New Hampshire (6,006 TMDLs), West Virginia 
(3,819 TMDLs), and Kansas (2,982 TMDLs).[Footnote 38] 

* Pollutant groups addressed by the largest number of TMDLs include 
pathogens, such as bacteria from feces (addressed in 11,595 TMDLs); 
metals other than mercury, such as arsenic, cadmium, and zinc (9,010 
TMDLs); mercury (7,046 TMDLs); nutrients such as nitrogen or 
phosphorus (5,466 TMDLs); and sediment (3,783 TMDLs). 

* The probable sources of stressors contributing to impairments may 
vary by type of water body and include agricultural runoff, municipal 
discharges (sewage), and atmospheric deposition[Footnote 39] into 
rivers, streams, lakes, ponds, bays, and estuaries. Other significant 
stressors include habitat alterations and physical modifications 
(e.g., altering water flow patterns by straightening streams or 
building dams). 

EPA has two additional databases that contain information about TMDL 
implementation, but these databases contain limited information on the 
extent of TMDL implementation and associated changes in water quality. 
As a result, EPA cannot use its different databases to assess the 
extent to which most TMDLs have been implemented, and it does not have 
comprehensive, nationwide information on whether and to what extent 
TMDLs have led to improved water quality. In its Permit Compliance 
System or Integrated Compliance Information System, EPA tracks data on 
point sources, including permitted pollutant levels associated with 
TMDLs.[Footnote 40] In its Grants Reporting and Tracking System, EPA 
tracks data on EPA-funded projects and activities that may help reduce 
nonpoint source pollution to meet levels called for in TMDLs. To track 
TMDL implementation and any associated effect on water quality, EPA 
and states need to know what actions have been taken to reduce 
pollution in a watershed with a TMDL and how much pollution has been 
reduced as a result, according to program officials. Through several 
studies, EPA has integrated information in its databases, where 
available, to assess the status of pollution control actions 
prescribed in TMDLs and the extent to which water quality improved 
from these efforts. Specifically, EPA conducted several studies, most 
in response to the 2007 EPA Office of Inspector General report, which 
found that the agency had limited data on TMDL implementation 
activities; these studies examined the following issues: 

* Whether sufficient information existed in the databases to assess 
TMDL implementation status. EPA determined that a limited number of 
TMDLs could be assessed for implementation status. For example, in a 
2010 report, the agency found that 17 percent of point source TMDLs 
and 8 percent of nonpoint source TMDLs had all of the data elements 
needed for TMDL implementation status to be addressed.[Footnote 41] 

* Whether point source permits had been revised to include TMDL 
limits. EPA found in a 2009 study that wasteload allocations for point 
sources were generally incorporated into discharge permits. [Footnote 
42] The study showed that of 399 wasteload allocations set by 100 
TMDLs, a majority (68 percent) was consistent with existing limits 
established in discharge permits, and no permit revisions were needed. 
Eleven percent required permit revisions and such revisions had been 
made, and 21 percent of wasteload allocations required permit 
revisions, but revisions had not been made. All the wasteload 
allocations for which permit revisions had not been made pertained to 
municipal stormwater or combined sewer overflows, for which permit 
revisions are not straightforward, according to EPA. 

* Whether TMDLs have implementation plans for carrying out actions to 
control pollution. In a 2007 study, EPA assessed how stakeholder 
participation and implementation planning, which may occur during TMDL 
development, affect implementation.[Footnote 43] As part of this 
study, EPA surveyed regional officials responsible for TMDLs and found 
that 37 percent of TMDLs often or always have implementation plans, 
and 46 percent reported that TMDLs never or seldom have implementation 
plans. 

* Whether TMDLs have helped improve water quality. A 2008 study funded 
by EPA on TMDL implementation status in Ohio and West Virginia, 
[Footnote 44] that officials responsible for implementing TMDLs were 
not aware of some TMDLs developed for watersheds under their 
jurisdiction and that incremental improvements in water quality had 
occurred in less than 20 percent of watersheds. For example, for the 
63 TMDLs sampled, the study found that agency officials did not know 
whether any TMDL implementation activities had occurred for 18 of 
them. The study found that implementation of pollution control actions 
had occurred in 29 of the 63 watersheds in the sample. Incremental 
improvements in water quality had occurred in 12 of the 63 watersheds, 
and full water body restoration had occurred in 2 of the 63 
watersheds.[Footnote 45] 

EPA's Ability to Track TMDL Implementation Is Hindered by Data System 
Limitations and Unavailable USDA Data: 

EPA is hindered in its ability to track TMDL implementation results 
because of data system limitations, as well as the fact that data on 
water quality projects paid for by USDA are generally unavailable. 
First, EPA's databases, which contain information uploaded from 
states' databases, function independently of one another. Because 
information relevant to TMDL implementation is not consistently tagged 
with geographic information system (GIS) data (i.e., latitude-
longitude coordinates on water body segments, permitted discharge 
locations, and the spatial extent of projects addressing nonpoint 
source pollution), as opposed to water body names and states, the 
information on TMDL implementation projects and impaired water bodies 
generally cannot be integrated across separate databases.[Footnote 46] 
As a result, EPA cannot link (1) information on TMDL implementation 
projects and activities from its Grants Reporting and Tracking System 
with (2) information on impaired water bodies or the causes of their 
impairment from the Assessment, TMDL Tracking and Implementation 
System and can therefore not assess the extent to which TMDLs have 
been implemented in watersheds with impaired water bodies. According 
to TMDL program officials, to link TMDL information in several 
databases, the agency must have GIS data--not just the water body 
name--for the extent of waters addressed by each data point for each 
TMDL. Nevertheless, the data that states enter into their own systems 
and upload to EPA's databases do not consistently include this 
information, in part because EPA does not require the information and 
in part because state officials developed the majority of TMDLs before 
the widespread use of GIS, according to program officials. Without 
consistently obtaining from states GIS data in addition to water body 
names, EPA cannot integrate information on TMDL implementation 
projects and impaired water bodies across separate databases and 
cannot assess whether and to what extent water quality has been 
affected by TMDL implementation. 

EPA's Office of Inspector General, in its 2007 report, recommended 
that EPA annually report on the progress of TMDL implementation 
activities and their effects on water quality. EPA officials explained 
to us that such an effort would be expensive (both to develop the 
system and to expand monitoring efforts to detect changes in water 
quality) and that limited funding should instead be invested in 
projects to control pollution. EPA program officials told us that in 
the present budget environment, they hesitate to ask states to invest 
additional resources or reallocate funding to track the water quality 
effects of TMDL implementation projects or update their databases. 
Nevertheless, EPA program officials told us, requiring states to 
report GIS data associated with TMDL implementation is more feasible 
than asking them to create new data systems or to merge existing ones 
with EPA's systems. EPA officials said that they are beginning to 
obtain GIS information associated with projects in the Grants 
Reporting and Tracking System for some nonpoint source TMDLs, but they 
acknowledged that this data-gathering effort will not provide 
information on the specific type, amount, or location of projects that 
have been implemented for all the TMDLs in the database--information 
that would generally be needed to ascertain the effect of TMDL 
implementation on water quality. 

A second reason that EPA has limited ability to track TMDL 
implementation results is that many USDA-funded projects undertaken on 
agricultural land adjacent to water bodies with TMDLs cannot be 
tracked by EPA or state water quality programs because of privacy 
provisions in the 2008 Farm Bill.[Footnote 47] In more than one 
thousand watersheds across the country, USDA funds conservation 
projects to implement best management practices that reduce soil 
erosion, water pollution, and air pollution, and many of these 
projects help curb pollutants like nutrient runoff from agricultural 
fields. EPA officials told us that data on the location, type, and 
number of these projects could help identify whether and to what 
extent TMDLs have been implemented and whether water quality has 
improved as a result. Statutory provisions enacted as part of the 2008 
Farm Bill, however, protect the identity and location of landowners 
and other agricultural producers who participate in USDA-funded 
conservation programs, including those implementing best management 
practices. Specifically, the provision prohibits USDA employees or 
contractors from sharing certain information concerning the land and 
operations of agricultural producers who participate in USDA 
conservation programs; the department also cannot share GIS 
information about agricultural land or operations.[Footnote 48] The 
provision, however, does not prohibit landowners and agricultural 
producers from consenting to disclosure or disclosing information on 
the specific best management practices that USDA has funded on their 
land. As a practical matter, according to EPA program officials, the 
biggest single information gap that limits tracking of implementation 
of nonpoint source TMDLs is the lack of information on the location of 
USDA-funded best management practices. 

According to EPA officials, USDA has collected data on projects that 
it funded, but these data are summarized and presented on fairly large 
geographic scales--scales too large to match to watersheds with TMDLs. 
According to EPA program officials, such aggregated USDA data cannot 
be used to assess the effects of projects put in place to implement 
TMDLs because TMDLs typically apply to much smaller areas than 
represented by these data. To track the effectiveness of TMDL 
implementation, program officials told us, EPA generally needs data on 
specific best management practices implemented at specific sites on 
individual parts of a stream within a watershed. USDA's data, in 
contrast, are collected and aggregated for areas that cover, on 
average, 10,000 square miles, encompassing, for example, several 
rivers and their tributaries, according to EPA officials (see figure 
5).[Footnote 49] EPA TMDL program officials told us that their senior 
managers have not formally discussed with USDA the availability of the 
department's data to EPA, primarily because of the sensitivity of this 
issue with USDA and its need to abide by the privacy provisions of the 
Farm Bill. Nevertheless, without access to data on the location, type, 
and number of projects implemented by landowners who participate in 
conservation programs funded by the USDA in areas subject to a TMDL, 
EPA cannot track actions taken to implement TMDLs and subsequent 
changes in water quality associated with a core EPA program and a 
substantial federal investment. 

Figure 5: Scales at Which USDA Aggregates Watershed Data, States 
Implement TMDLs, and EPA Needs GIS Data to Track TMDL Implementation 
Projects: 

[Refer to PDF for image: map with 2 magnified sub maps] 

Depicted on map: 
Watershed of 10,000 square miles (average); 
Watershed of 1,500 square miles (average); 
Watershed of 40 square miles (average); 
Stream segment A; 
Stream segment B; 
Stream segment C; 
Streamside vegetation; 
Stream bank stabilization; 
Farm; 
Grass filter strips. 

Sources: GAO, EPA, and USDA. 

Note: USDA aggregates watershed data at the scales shown at top left; 
TMDLs are implemented at scales shown at top right (each of the 
colored stream segments A, B, and C is subject to a TMDL); and EPA 
needs data on the scale of TMDL implementation projects shown at 
bottom right. To track USDA-funded projects implemented near and in 
impaired water bodies, EPA would need project information tagged with 
GIS data (i.e., latitude-longitude coordinates), in addition to water 
body names and states. According to EPA officials, however, data that 
USDA has on TMDL projects it funded, such as projects to plant or 
protect vegetation along stream banks (bottom right), are summarized 
and presented on fairly large geographic scales (top left), which are 
too large to match to watersheds with TMDLs (top right). 

[End of figure] 

States Report Having Little Information on TMDL Implementation, as 
Well as Limited Improvement in Water Quality: 

Our survey of state TMDL coordinators responsible for implementing 
long-established TMDLs showed that state TMDL coordinators do not know 
the extent to which many long-established TMDLs have been implemented. 
For those TMDLs where information exists, state coordinators reported 
that pollutants had been reduced in many waters, but few TMDLs had 
helped water bodies attain water quality standards. State TMDL 
coordinators did not know the status of TMDL implementation for 
slightly less than half of the long-established TMDLs. Specifically 
state coordinators reported that they did not know (1) if wasteload 
allocations had been met for 36 percent of point source 
TMDLs,[Footnote 50] or (2) whether load allocations had been met for 
48 percent of nonpoint source TMDLs. Moreover, these coordinators did 
not know whether pollutant levels had changed in 35 percent of 
nonpoint source TMDLs. 

Among state TMDL coordinators who had information on the extent of 
TMDL implementation, coordinators reported that a higher proportion of 
long-established point source TMDLs helped water bodies attain water 
quality standards than did nonpoint source TMDLs. When state TMDL 
coordinators were knowledgeable about TMDLs addressing point sources, 
they reported that for 83 percent of long-established TMDLs, wasteload 
allocations the TMDLs called for had been met.[Footnote 51] When state 
TMDL coordinators were knowledgeable about TMDLs addressing nonpoint 
sources, however, they reported that 20 percent of load allocations 
had been met. According to state officials, this discrepancy exists 
primarily because actions called for in TMDLs for nonpoint source 
pollution either have not been implemented or have been implemented to 
a limited extent. Furthermore, state TMDL coordinators reported that 
they expected 13 percent of water bodies with TMDLs for nonpoint 
source pollution to attain water quality standards within the next 5 
years. 

Long-Established TMDLs Experts Reviewed Seldom Contained All Key 
Features, Some of Which Are Not Included in EPA's Guidance: 

Long-established TMDLs often do not contain key features that would 
help water bodies attain water quality standards, in part because 
EPA's regulations and guidance do not direct TMDLs to contain them. 
NRC and EPA have identified key features that TMDLs should contain to 
help impaired water bodies attain water quality standards. Water 
resource experts who reviewed a sample of 25 long-established TMDLs 
reported that these TMDLs do not contain many of these key features. 
In addition, in reviewing EPA's guidance on how regional offices are 
to apply EPA's 11 TMDL elements, we found that the guidance does not 
contain some of these key features and that some of these elements are 
vague, which has led EPA's regional offices to apply them in different 
ways and to approve TMDLs of variable quality. 

Certain Features Are Key If TMDLS Are to Help Water Bodies Attain 
Water Quality Standards: 

A 2001 NRC report and several EPA studies and guidance documents have 
identified key features for TMDL development, which would help water 
bodies attain water quality standards if TMDLs were 
implemented.[Footnote 52] We grouped these key features into three 
categories on the basis of the sequence of steps in the TMDL process: 
those that help ensure that (1) TMDLs accurately identify and address 
causes of impairment, (2) TMDLs can be implemented, and (3) TMDLs are 
revised if found to be ineffective in helping water bodies attain 
water quality standards. The steps in these three categories reflect a 
more expansive view of TMDLs than what EPA's existing regulations 
require TMDLs to include, according to EPA officials. 

Accurately Identifying and Addressing Causes of Impairment: 

The 2001 NRC report identified key features that help ensure that 
TMDLs accurately identify and address the causes of water body 
impairment. These key features include (1) evidence that impairment is 
driven by the stressor or stressors for which a TMDL is being 
developed and (2) evidence that addressing the stressor or stressors 
in question will be sufficient for the water body to attain its 
designated use or uses, as follows: 

* Evidence that impairment is caused by the stressors a TMDL is 
developed to address. The 2001 NRC report recommended that states 
conduct a thorough water body assessment before developing a TMDL to 
accurately and completely identify specific stressors or causes of 
impairment. NRC reported that thousands of water bodies had been 
placed on states' impaired waters lists on the basis of limited data, 
rather than a thorough assessment. Evidence from a thorough water body 
assessment before TMDL development establishes whether and to what 
extent the stressor or stressors prompting a state to list a water 
body as impaired are indeed causing impairment. 

* Evidence that addressing these stressors will be sufficient for a 
water body to attain designated uses. NRC recommended that TMDLs be 
developed to consider multiple stressors--be they pollutants entering 
a water body, such as sediment or nutrients, or other human-induced 
forms of pollution, such as stream straightening, altered flows, or 
removal of streamside vegetation--in large part because waters are 
often impaired by multiple stressors.[Footnote 53] 

Side bar: Miami Gold: 

A total maximum daily load (TMDL) plan for the lower Little Miami 
River watershed, Ohio approved by EPA in 2010, includes several key 
features making it likely that implementing the TMDL will help lead to 
attainment of designated uses, according to the water resource experts 
we selected. By including evidence from a comprehensive chemical, 
physical, and biological assessment, this TMDL clearly shows that the 
river and its tributaries are impaired by the stressors for which the 
TMDL was developed. The stressors include bacteria, nutrients, oxygen 
demand, sediment, and habitat quality. A model based on mathematical 
equations and watershed characteristics links habitat quality, one of 
the stressors, to impairment with respect to the water body’s 
designated use of supporting aquatic life. In addition, the TMDL 
identifies actors and actions needed to address both the point source 
and nonpoint source pollution impairing various sites in the 
watershed. For example, the TMDL identifies specific stream miles 
where buffers of native grasses and trees should be planted to help 
control nutrient and sediment runoff. 

[Photograph: Lower Little Miami River near a nature preserve] 

Source: Ohio Environmental Protection Agency. 

[End of side bar] 

Ensuring That Implementation Can Be Done: 

EPA studies and guidance documents have identified key features to 
help ensure that TMDLs can be implemented. These key features include 
(1) a plan for TMDL implementation that specifically identifies who 
must undertake what projects to control pollution (i.e., plan 
specifies actors) and on what land areas (i.e., plan specifies 
locations) and (2) a demonstration of reasonable assurances that 
projects to control nonpoint source pollution will actually be 
implemented, and to an extent that allows the water body to meet load 
allocations specified in the TMDL. These features may be described as 
follows: 

* Plan for TMDL implementation. An EPA study recommended that states 
prepare and submit implementation plans for TMDLs to help ensure that 
TMDLs, once developed, actually help lead to attainment of water 
quality standards.[Footnote 54] Other EPA studies have confirmed that 
effective TMDLs include or are accompanied by detailed plans for TMDL 
implementation that specify actors, as well as specific locations in 
need of remediation, to target efforts at those responsible for the 
problem, thus facilitating implementation.[Footnote 55] 

* Reasonable assurances that implementation activities will occur. EPA 
guidance states that when a TMDL is developed for waters impaired by 
both point and nonpoint sources, a TMDL should provide reasonable 
assurances that measures to control nonpoint source pollution will 
achieve the expected load reductions. According to an EPA study, such 
assurances may include identifying landowners who will actually 
implement necessary projects.[Footnote 56] Without reasonable 
assurances, EPA cannot be certain that limits in a TMDL for both point 
and nonpoint source pollution have been established at levels 
sufficient to attain water quality standards, according to EPA 
guidance. 

Revising TMDLs as Needed: 

Key features in the 2001 NRC report that help ensure that TMDLs are 
reviewed and revised if found to be ineffective in helping water 
bodies attain water quality standards include (1) a plan to monitor a 
TMDL's effect on water quality, including monitoring biological 
indicators, such as the diversity and abundance of aquatic organisms, 
and (2) a description of an adaptive approach to implementing the 
TMDL, whereby monitoring data will be used to periodically assess 
progress toward attaining water quality standards and to adjust the 
TMDL if needed, as follows: 

* Plan to monitor a TMDL's effect on water quality. The NRC report 
recommended that TMDLs include plans for post-implementation 
monitoring to ensure that actions taken as a result of a TMDL are 
effective at helping meet water quality standards. Further, according 
to the NRC report, such monitoring should include monitoring 
biological indicators, such as aquatic plants and animals, because 
such indicators provide a more accurate assessment of water body 
health with respect to designated uses, such as fishing or swimming, 
than chemical indicators alone. According to the report, aquatic 
plants and animals reflect the totality of their living conditions, 
including the chemical, physical, and biological stressors that impair 
specific water bodies. 

* Adaptive implementation approach. The NRC report recommended that 
TMDLs follow an adaptive approach, in which monitoring data are used 
to revise and improve a TMDL over time. An adaptive approach allows 
states to implement pollution control actions, monitor the effects of 
those actions, and then revise the TMDL as needed. According to the 
NRC report, comparing results predicted during TMDL development 
against empirical data from the water body serves to verify that water 
quality is improving and that progress is in fact being made toward 
attaining designated uses. 

Experts Reported That the 25 TMDLs They Reviewed Seldom Contained All 
Needed Key Features: 

Three water resource experts reviewed a sample of 25 long-established 
TMDLs and reported that many of them do not contain features 
identified as key if TMDLs are to help water bodies attain water 
quality standards.[Footnote 57] The experts reported that many of the 
TMDLs they reviewed do not contain key features that accurately 
identify and address the causes of water body impairment, about half 
do not contain key features that help ensure that implementation can 
be done, and few contain key features that help ensure that TMDLs are 
revised if they are found to be ineffective in helping a water body 
attain water quality standards. As a result, they expressed little 
confidence that most of the TMDLs they reviewed would--if implemented--
lead to attainment of water quality standards. 

Many of the TMDLs Experts Reviewed Do Not Accurately Identify and 
Address Causes of Impairment: 

The three water resource experts reported that many of the 25 TMDLs 
they reviewed do not contain key features to help ensure that TMDLs 
accurately identify and address the causes of water body impairment. 
Experts also reported that the TMDLs that do not diagnose and aim to 
treat the true causes of water body impairment are unlikely to lead to 
attainment of designated uses. 

* Evidence that impairment is caused by the stressors a TMDL is 
developed to address. Experts agreed that although 18 TMDLs they 
reviewed contain sufficient evidence that the water body in question 
was impaired by the stressor the TMDL was developed to address, 6 
other TMDLs do not contain sufficient evidence linking stressor to 
impairment.[Footnote 58] For example, the experts reported that a TMDL 
from an Alabama creek, approved by EPA in 2002, provides monitoring 
data showing that levels of the cause of impairment in question--
dissolved oxygen--did not indicate impairment. The TMDL states that 
samples drawn from the creek did not identify any violation of the 
water quality standard for dissolved oxygen. Nonetheless, because the 
creek was placed on the state's impaired waters list as a result of 
low dissolved oxygen, the TMDL was developed to address dissolved 
oxygen levels. The TMDL also refers to physical stressors that could 
have contributed to impairment, such as low water flow, but does not 
thoroughly assess these stressors or seek to treat them, according to 
the experts. The experts said that by targeting an unsubstantiated 
stressor and potentially overlooking other causes of impairment, the 
TMDL is unlikely to help the water body attain its designated use of 
supporting fish and wildlife. 

* Evidence that addressing stated stressors will be sufficient for a 
water body to attain designated uses. The experts agreed that 7 of the 
25 TMDLs they reviewed demonstrate that addressing the identified 
stressor would be sufficient to attain the applicable designated use 
or uses but that 17 other TMDLs do not demonstrate such a link. 
[Footnote 59] The experts said that in the water bodies for which most 
of these 17 TMDLs were developed, stressors other than the stressor or 
stressors addressed by the TMDL were at play and that the TMDLs do not 
address these stressors. For example, the experts agreed that in 8 of 
the 17 TMDLs, biological stressors were probably contributing to water 
body impairment, but the TMDLs do not address these biological 
stressors; in 9 other TMDLs, the experts reported that physical 
stressors were probably contributing but that they are not addressed 
in the TMDLs.[Footnote 60] For example, the experts reported that a 
TMDL approved by EPA in 2007 for a watershed in Maryland, which had 
active and abandoned mining operations, aims to treat low pH, 
indicating acidic water. According to the experts, however, mining is 
likely to result in stressors besides acidity, such as physical 
stressors including habitat degradation from stream channelization, 
and sedimentation caused by past and present land disturbances. The 
experts stated that the TMDL lays out a realistic approach for 
addressing the low pH caused by chemical pollutants and could thus 
likely help meet a numeric water quality target for pH. Nevertheless, 
they said, the TMDL would likely fall short of helping the water body 
attain its designated use--even if the numeric water quality target 
for pH were met--because the physical stressors they said are likely 
present, such as habitat degradation from stream channelization and 
other mining activities, would still be present. One expert said that 
the TMDL's narrow focus provides a false promise that addressing the 
chemical pollutants would solve the water quality problem. 

About Half of TMDLs Reviewed by Experts Do Not Have Features Ensuring 
That Implementation Can Be Done: 

The water resource experts reported that about half of the TMDLs they 
reviewed do not contain key features helping to ensure that 
implementation can be done, which leads to TMDLs that may do little to 
actually improve water quality, as follows: 

* Plan for TMDL implementation. The experts agreed that 13 of the 25 
TMDLs they reviewed contain detailed information about implementation 
but that the other 12 TMDLs contain vague or no information about 
implementation.[Footnote 61] Specifically, the experts agreed that 4 
of these 12 TMDLs do not discuss how the TMDL is to be implemented: 
the TMDLs contain no references to implementation or a stand-alone 
implementation plan. The experts also agreed that the other 8 of the 
12 TMDLs discuss implementation to some extent, but 6 of these 8 TMDLs 
do not identify specific actors who are to support implementation, and 
7 of these 8 do not identify actions that need to occur to attain 
water quality standards.[Footnote 62] For example, the experts 
reported that a South Dakota TMDL, approved by EPA in 2001, briefly 
discusses implementation but does not identify the actions or the 
actors needed to implement the TMDL. The experts told us that this 
brief discussion is unlikely to lead to attainment of water quality 
standards. A South Dakota water quality official who responded to our 
survey of states reported that as part of developing this TMDL, the 
state did not identify the landowners who needed to undertake projects 
to reduce nonpoint source pollution. The official added that nonpoint 
source load allocations have not been met and that nonpoint source 
loads are expected to stay the same over the next 5 years. 

* Reasonable assurances that implementation activities will occur. The 
experts agreed that for a different group of 13 of the 25 TMDLs they 
reviewed, which involved both point and nonpoint source pollution--
TMDLs that, according to the reasonable assurances element in EPA's 
guidance, are to provide reasonable assurances that projects to reduce 
nonpoint source pollution will achieve expected load reductions--5 
TMDLs do not include a reasonable assurances element but were 
nevertheless approved. Of the 8 TMDLs that the experts agreed discuss 
reasonable assurances, the experts reported that 6 do not provide 
sufficient evidence that reductions in nonpoint source pollution will 
actually occur, such as identification of landowners willing to 
implement necessary projects to reduce nonpoint source pollution. 
[Footnote 63] For example, the experts reported that an Indiana TMDL, 
approved by EPA in 2006, does not identify specific projects to reduce 
nonpoint E. coli bacteria stemming from sources such as manure 
applications on farmland and leaking septic tanks. The TMDL instead 
includes a generic list of projects and activities related to 
implementation, such as educating homeowners who have septic systems. 
According to the experts, such a list does not provide reasonable 
assurances that the projects and activities will be put in place or 
that load reductions laid out in the TMDL will be achieved. An Indiana 
water quality official responding to our survey of states reported 
that although nonpoint pollutant loads in this water body have 
decreased to some degree, load allocations have not been met and are 
not expected to be met within the next 5 years. 

Side bar: Specify, Specify: 

One key feature that can help ensure the implementation of total 
maximum daily loads (TMDL) is the presence of a detailed 
implementation plan. Three water resource experts we selected reviewed 
six TMDLs that include formal TMDL implementation plans and found that 
the plans generally specify actors, actions, and locations to be 
involved in implementation. For example, a TMDL from California 
approved in 2006 contains an implementation plan identifying actors, 
such as wastewater treatment plants and agricultural landowners, 
responsible for various types of pollution; specific remedial actions 
that need to occur, such as development and implementation of a 
program to collect banned pesticides; locations, such as agricultural 
land and construction sites, where such actions should be implemented; 
and completion dates for particular actions. Experts said that such a 
detailed implementation plan will likely help the water body meet 
numeric water quality targets laid out in the TMDL. 

[Photograph: A California river runs through agricultural land 
before entering the Pacific Ocean] 

Source: Bruce Perry, Department of Geological Sciences, California 
State University, Long Beach. 

[End of side bar] 

Few TMDLs Have Features That Ensure They Will Be Revised as Needed: 

The three experts agreed that almost all of the 25 TMDLs they reviewed 
are missing one or more of the key features that help ensure that 
TMDLs are revised if they are found to be ineffective in helping a 
water body attain water quality standards. Experts said that without 
adequate monitoring plans and an adaptive implementation approach to 
verify the accuracy of the TMDL and the effectiveness of remediation 
efforts, resources may be wasted developing and implementing TMDLs 
that do little to improve the condition of water bodies. 

* Plan to monitor a TMDL's effect on water quality. The experts agreed 
that 7 of the 25 TMDLs do not include a monitoring plan to track the 
TMDLs' effectiveness. Of the 18 TMDLs that include such a plan, the 
experts reported that 14 do not include indicators of biological 
outcomes, and 3 do.[Footnote 64] For example, the experts said that an 
Oregon TMDL for water temperature, approved by EPA in 2007, does not 
include plans to monitor biological indicators, such as the number and 
condition of fish and other aquatic organisms. The experts gave the 
TMDL credit for its plans to monitor temperature changes that might 
result from implementing the TMDL but said that without tracking 
biological conditions affected by temperature, it would be impossible 
to assess whether progress were being made toward the water body's 
designated use of supporting salmon, trout, and other cold-water fish 
species. 

* Adaptive implementation approach. The experts reported that of the 
18 TMDLs that include a monitoring plan, 8 do not describe an adaptive 
approach to TMDL implementation--that is, an approach that uses 
monitoring and data analysis to periodically assess progress toward 
attaining water quality standards and adjust the TMDL as needed. The 
experts agreed that 9 TMDLs describe such an approach.[Footnote 65] 
For example, experts reported that a Connecticut TMDL for phosphorus, 
approved by EPA in 2005, does not describe how monitoring data are to 
be used to assess progress toward water quality standards or to adjust 
the TMDL if implementation actions are found to be ineffective. 
Because the TMDL focuses only on phosphorus, a plant nutrient and 
chemical pollutant, one of the experts said that using monitoring data 
to assess progress toward the water body's designated uses is 
particularly important because such data could show that even while 
concentrations of phosphorus are declining, additional actions may be 
needed to attain the water body's designated uses of providing fish 
habitat and recreation. A Connecticut water quality official reported 
that for this TMDL, the official does not know whether point and 
nonpoint pollutant loads in the water body have changed or whether 
wasteload and load allocations have been met. 

Side bar: Sugar and Spice: 

According to the experts we selected, a total maximum daily load 
(TMDL) plan for 21 streams in the Sugar-Pecatonica River Basin, 
Wisconsin, contains several key features. It describes the 21 streams 
in detail, including evidence from physical and biological assessments 
that designated uses of supporting fish and other aquatic life are not 
being met. It explicitly links the effects of sediment, the targeted 
stressor, to the streams’ ability to support aquatic life. By showing 
that certain agricultural practices have successfully reduced sediment 
in streams with similar natural characteristics in the same watershed 
and enabled these streams to meet their designated uses, the TMDL 
demonstrates that addressing sediment in the impaired streams will 
likely achieve a comparable effect. The TMDL includes assessment of 
biological indicators, such as counts of the number and type of fish 
and insect populations, which, according to the National Research 
Council and the experts, inherently reflect the cumulative effects of 
many impairments. In addition, the TMDL identifies actions needed to 
reduce sediment in the streams and the particular locations where such 
actions should be taken. For example, it identifies a specific best 
management practice that should be adopted where livestock graze to 
minimize sediment runoff. 

[Photograph: A cattle crossing to reduce bank erosion; cattle
were later fenced out of the stream corridor] 

Source: Wisconsin Department of Natural Resources. 

[End of side bar] 

EPA's TMDL Guidance for Implementing Regulations Does Not Contain Some 
Key Features: 

In reviewing EPA's guidance on approvable TMDLs and how regional 
offices are to apply the 11 TMDL elements, we found that the guidance 
did not contain some key features and that some of these 11 elements 
are vague or optional, which has led EPA's regional offices to apply 
the elements in different ways and to approve TMDLs of variable 
quality. The 11 elements for approvable TMDLs described in EPA's 
guidance generally reflect the agency's interpretation of TMDL 
development requirements under existing regulations, according to 
EPA's guidance. In our review, we found that EPA's guidance does not 
include two key features--specifically, those that help ensure that 
TMDLs accurately identify and address causes of impairment with 
respect to a designated use or uses. First, the guidance does not 
specify that a TMDL should provide evidence that impairment is caused 
by a stressor or stressors for which the TMDL is being developed. 
Rather, according to the guidance, a TMDL need only identify a given 
pollutant. Second, although guidance states that a TMDL must provide 
evidence that pollutant reductions will meet numeric targets, it does 
not specify that a TMDL is to provide evidence that addressing the 
given stressor or stressors will be sufficient for a water body to 
attain its designated uses. Because these key features are not 
explicitly called for in EPA's guidance, chemical, physical, and 
biological stressors other than the pollutant identified in the TMDL 
may go unaddressed--unless additional TMDLs are developed for the 
water body--even if a numeric water quality target for the identified 
pollutant is met.[Footnote 66] 

EPA officials told us that the agency cannot include additional 
features in their guidance without issuing new regulations. Through 
its 2000 rule, EPA had planned to include additional features by 
conditioning TMDL approvals on more-rigorous requirements but withdrew 
the rule before its effective date. Instead, the agency developed 
guidance in 2002 for states and EPA regions to apply in implementing 
its existing regulations.[Footnote 67] According to EPA officials, the 
guidance contains weaker versions of many of the same elements 
included in the 2000 rule, and they have not reinitiated rulemaking, 
both because the agency wanted more time to assess the effectiveness 
of existing regulations and because rulemaking is costly and time-
consuming. But because the guidance does not contain certain features, 
states may not always have clear direction to develop TMDLs containing 
key features that can help water bodies attain water quality 
standards. Moreover, guidance does not have the force of regulation, 
and states can choose which and how much information to include under 
the 11 elements as currently stated. All three of the experts 
commented that TMDL development is a complex and costly endeavor and 
that developing TMDLs that do not contain key features--such as 
accurately identifying the causes of water body impairment--can waste 
effort and the limited funds available for the task. Absent the force 
of regulations to direct states to develop TMDLs containing key 
features, TMDLs are likely to continue to do little to attain water 
quality standards. Yet neither EPA's existing guidance nor the 
agency's 2013 long-term vision for the TMDL program has the detail on 
key features that can provide such direction. 

In addition, 3 of EPA's 11 elements in its TMDL guidance--those for 
(1) implementation, (2) reasonable assurances, and (3) a monitoring 
plan to track TMDL effectiveness--are either vague or stated as 
options (e.g., EPA encourages an implementation plan, EPA recommends 
inclusion of a monitoring plan) and are therefore subject to varying 
interpretation on the part of states developing a given TMDL and the 
regional offices applying the elements in approving TMDLs. Our 
analysis shows that EPA's regional offices do not apply the elements 
of the agency's existing guidance consistently. Specifically, several 
EPA regions have interpreted the guidance as calling for key features 
in state-developed TMDLs, while the remaining regions have not. The 
vagueness of certain elements in EPA's TMDL guidance--coupled with the 
absence of key features--results in regional offices' inconsistently 
interpreting what a TMDL is to contain and approving TMDLs of variable 
quality. EPA headquarters, however, has not worked with the regions to 
apply existing guidance consistently or to develop specific criteria 
on which to base TMDL approvals, as some regions do. EPA headquarters 
officials told us they have not done so because of differences among 
states' programs, including that some states have been under pressure 
from consent decree schedules to develop a certain number of TMDLs by 
specific dates. Nevertheless, without specific criteria on which to 
base TMDL approvals, some EPA regions are likely to continue to 
approve TMDLs in which the quality or quantity of information is 
insufficient to help water bodies attain water quality standards. 

Implementation: 

EPA's element for implementation encourages states to develop TMDL 
implementation plans but existing regulations do not require such 
plans or detail the information they should contain.[Footnote 68] For 
example, the guidance describing the element does not specify that 
implementation plans should include information such as who must 
undertake what projects to reduce pollution and at what locations 
these projects are to be put in place. Our analysis showed that EPA's 
regional offices vary in how they interpret this implementation 
element. Officials at most regional offices (6 of 10) told us they 
give little scrutiny to whether states include formal implementation 
plans in TMDLs. For example, Region 3 officials told us they do not 
review or approve implementation plans since EPA regulations do not 
require them. They said that most states do not submit such plans. On 
the other hand, officials from 4 of the 10 regional offices told us 
they encourage states to produce implementation plans or ask that 
states do so, particularly for TMDLs involving nonpoint source 
pollution. Region 2 officials, for example, told us they have asked 
states to complete implementation plans and that the office reviews 
and comments on these plans. The officials said that information in 
implementation plans helps them determine whether projects to control 
nonpoint source pollution will be implemented. 

Reasonable Assurances: 

When a TMDL is developed for waters impaired by both point and 
nonpoint sources, EPA guidance states that a TMDL should provide 
reasonable assurances that actions to control nonpoint source 
pollution will achieve expected load reductions. This information is 
necessary for EPA to determine that TMDLs have been set at levels 
sufficient to attain water quality standards. Nevertheless, EPA's 
element for reasonable assurances does not describe the type of 
information needed to demonstrate that projects to reduce nonpoint 
source pollution will achieve expected load reductions. For example, 
the guidance describing the element does not state that a TMDL is to 
detail any available regulatory authority over nonpoint source 
pollution or identify landowners who are willing to implement needed 
projects. Without such specificity, EPA has little assurance that 
projects to reduce nonpoint source pollution will in fact be 
implemented. 

We found that the regional offices vary in their review of the 
reasonable assurances element, with most regions (7 of 10) asking for 
little to no information on demonstrating that load reductions for 
nonpoint source pollution will occur. Officials from several regional 
offices told us they play a limited role in encouraging states to 
demonstrate reasonable assurances, although, according to EPA 
headquarters officials, EPA is beginning to put greater emphasis on 
demonstrating reasonable assurances in TMDLs. Officials from one 
regional office explained that many states lack data from which to 
develop reasonable assurances and that states without authority over 
nonpoint source pollution find it difficult to prove that sources of 
such pollution will take action to implement a TMDL. In the absence of 
specific guidance from EPA headquarters, officials from one regional 
office said that they encourage states to do the best they can to 
demonstrate reasonable assurances and that they do not reject TMDLs if 
reasonable assurances have not been demonstrated, in part because they 
do not have the resources to develop TMDL themselves. 

Conversely, officials from Regions 1, 5, and 9 told us they ask for 
substantial information demonstrating that load reductions for 
nonpoint sources will actually occur. For example, in describing a 
decision to withdraw approval of a TMDL for Lake Champlain in Vermont, 
Region 1 officials explained that they expect reasonable assurances to 
show quantitative reductions anticipated from specific projects to 
reduce nonpoint source pollution and that they expect regulatory or 
other mechanisms to be in place to ensure compliance with the actions 
and reductions called for in the TMDL. In addition, Region 5 officials 
told us that, as part of providing reasonable assurances that nonpoint 
source pollution will be reduced, they ask states to document specific 
state and local regulatory authority over such pollution. By calling 
for this type of information during TMDL development, the officials 
said that the region gains leverage it can call upon later in 
encouraging states to apply those authorities when a TMDL is 
implemented. EPA has acknowledged that the regional offices vary in 
how exacting they are when reviewing reasonable assurances. 

Monitoring Plan: 

EPA's monitoring plan element recommends that TMDLs include plans for 
monitoring but does not specify that such plans should include 
indicators reflecting biological conditions or adaptive approaches for 
TMDL revision in light of monitoring results. We found that EPA's 
regional offices again vary in how they interpret and apply the 
monitoring plan element. Officials at most regional offices (7 of 10) 
told us they do not ask for TMDLs to include much, if any, information 
about monitoring. For example, Region 6 officials told us monitoring 
information is not a focus in their review of TMDLs and that only 
about 1 percent of TMDLs they review contain monitoring plans. On the 
other hand, officials in the other three regions ask for monitoring 
plans to provide assurance that TMDLs will lead to improved water 
quality as intended. For example, Region 8 officials told us that they 
ask for monitoring plans for a handful of TMDLs, particularly for 
those in which a high degree of uncertainty exists about proposed 
pollution limits or other elements. They said that asking for a 
monitoring plan gives some assurance that a state will examine whether 
projects to control pollution are working as predicted and will adjust 
a TMDL if needed. 

Factors Reported as Facilitating the Implementation of Long-
Established TMDLs, Particularly Those for Nonpoint Source Pollution, 
Are Often Not in Place: 

State TMDL coordinators reported that many factors EPA identified as 
helpful for effective TMDL implementation were often not in place in 
the long-established TMDLs they reviewed. EPA has identified a number 
of factors associated with effective TMDL implementation--that is, 
factors that help state TMDL coordinators and other stakeholders take 
actions, prescribed in TMDLs, that ultimately result in attainment of 
water quality standards. State TMDL coordinators reported that many 
factors, including those viewed as most helpful, were often not in 
place. Some factors are absent because EPA has not always asked states 
to make such factors a priority when funding TMDL implementation, and 
others--including those viewed as most helpful for implementation--are 
beyond EPA's and states' existing authorities to put in place, 
according to state TMDL coordinators. 

A Number of Factors Facilitate Effective TMDL Implementation: 

Through its studies and state conferences on TMDLs, EPA and the states 
have identified numerous factors associated with effective TMDL 
implementation, particularly TMDLs for nonpoint source pollution 
(since point sources are subject to regulation through permits and 
discharge limits). These facilitating factors can be organized into 
the following three categories: (1) those securing landowners' and 
others' participation in implementing TMDLs, (2) those providing 
external assistance to implement TMDLs, and (3) those improving 
states' TMDL program functions. According to EPA documents, the agency 
expects some implementation actions to take place within 5 years of 
TMDL approval and all implementation actions to have taken place 
within 10 years of approval, and facilitating factors generally help 
states meet these goals. These facilitating factors--15 in all--can be 
organized as follows: 

* Securing participation includes: 

- landowners willing to implement projects[Footnote 69] 

- landowners willing to implement projects in most-effective places: 

- landowners willing to implement most-effective projects: 

- citizen participation: 

- supportive local government leaders: 

* Providing external assistance includes: 

- funding enough or more than enough: 

- assistance from state nonpoint source program: 

- assistance from USDA: 

- assistance from state agricultural program: 

- assistance from EPA: 

- assistance from state or local planning agency: 

- assistance from state stormwater program: 

* Improving states' TMDL program functions includes: 

- presence of TMDL implementation plan[Footnote 70] 

- higher funding priority for water bodies with TMDLs: 

- consolidation of TMDL responsibilities: 

Securing Landowners' and Others' Participation in Implementing TMDLs: 

The first category of factors aims to secure participation by 
landowners and other stakeholders in implementing TMDLs. Without a 
federal regulatory framework similar to the one governing point source 
pollution, which requires that dischargers meet pollution limits or 
face penalties, most states generally must persuade landowners to 
implement best management practices to reduce nonpoint source 
pollution. For TMDL implementation to address nonpoint source 
pollution, an EPA study found, engaging owners of private land is 
crucial.[Footnote 71] According to the study, without landowners' 
cooperation--including willingness to implement projects that are the 
most effective at reducing pollution in the places where they are most 
needed--projects to reduce such pollution may not be implemented. 

In a 2007 study, EPA found that citizen-based groups, such as local 
watershed associations and supportive local governmental leaders, can 
have a positive effect on TMDL implementation.[Footnote 72] Citizen-
based groups often have strong local support, are well informed on 
watershed issues, and have knowledge and experience that enables them 
to aid implementation, according to the study. When present, these 
citizen-based groups may lead in implementing a TMDL, such as by 
pursuing funding for projects and monitoring changes in water quality. 
Without a citizen-based group involved in TMDL implementation efforts, 
no entity may be in charge of implementing a TMDL. 

Providing External Assistance to Implement TMDLs: 

The second category of factors aims to provide states with external 
assistance to implement TMDLs. EPA found in the 2007 study that 
implementing TMDLs requires adequate funding to undertake projects 
that reduce pollution. For example, states may target funds received 
as grants under EPA's nonpoint source management program to implement 
best management practices on agricultural lands near water bodies 
subject to TMDLs, or states themselves may provide additional funding 
through other state water resource programs. According to a 2007 
study, assistance from external agencies and programs (e.g., technical 
assistance) can also be a critical factor in TMDL implementation. 
[Footnote 73] Particularly important is the involvement of agencies 
and programs--such as state nonpoint source programs, USDA programs, 
state agricultural agencies, EPA, state and local planning agencies, 
and state stormwater programs--whose participation may be necessary to 
improve water quality. 

Improving States' TMDL Program Functions: 

The third category of factors aims to improve states' TMDL program 
functions. In several recent conferences sponsored by EPA, state TMDL 
program managers identified ways to foster better TMDL implementation, 
including requiring or providing guidance for development of detailed 
implementation plans, administering funding to target water body 
restoration projects in areas where TMDLs have been developed, and 
consolidating TMDL development and implementation activities within 
the same state agency or program. 

Five states have laws or regulations requiring development of TMDL 
implementation plans, and five others provide guidance for such plans, 
according to a 2008 study. In a 2009 conference sponsored by EPA, 
state water quality program officials said that formal implementation 
plans enhance TMDL implementation for nonpoint source pollution by 
providing explicit direction for implementation, such as specific 
information on existing water quality conditions, necessary load 
reductions for pollutants, land areas likely to be contributing to 
pollution, programs that may help provide technical or financial 
assistance, best management practices and where they are to be 
implemented, and an accurate timeline over which implementation will 
be accomplished and water quality would improve. 

In a 2009 conference, state TMDL program managers supported putting a 
high priority on using federal and state funding for reducing nonpoint 
source pollution to support TMDL implementation. Some states have 
revised funding criteria under their nonpoint source management 
programs to give higher priority to projects designed to implement 
best management practices where a water body was already identified by 
the state as impaired and had an associated TMDL. 

States have also identified the consolidation of their 
responsibilities for TMDL implementation and development activities as 
a factor that facilitates TMDL implementation. Some states have 
recently integrated their TMDL development and implementation 
responsibilities with their nonpoint source management programs, which 
fosters collaboration among staff working toward the same goal who 
have complementary knowledge and skills. Washington State, for 
example, merged its TMDL development and implementation programs into 
its Department of Ecology, and departmental staff reported that both 
programs are more effective. In Virginia, the state's general assembly 
recently passed legislation that consolidated the agencies responsible 
for developing and implementing TMDLs. 

State TMDL Coordinators Reported That Many Factors, Including Those 
Viewed as Most Helpful, Were Often Not in Place for Long-Established 
TMDLs: 

In responding to our survey, state TMDL coordinators reported that the 
factors aiming to secure participation by landowners and other 
stakeholders are among the most helpful in implementing long-
established TMDLs. They also reported that external assistance, such 
as financial or technical assistance from a state nonpoint source 
program, is also very helpful, whereas factors improving state TMDL 
program functions were sometimes helpful. The results of our survey, 
however, showed that many factors, including those viewed as most 
helpful, were often not in place for long-established TMDLs (see 
figure 6).[Footnote 74] 

Figure 6: Factors Facilitating TMDL Implementation and Extent to Which 
They Were Present in and Considered Helpful for Long-Established TMDLs: 

[Refer to PDF for image: horizontal bar graph] 

Securing participation: 

Landowners willing to implement projects: 
Percentage of long-established TMDLs: 
Considered helpful: 98%; 
Present: 36%. 

Landowners willing to implement projects in most-effective places: 
Percentage of long-established TMDLs: 
Considered helpful: 100%; 
Present: 33%. 

Landowners willing to implement most-effective projects: 
Percentage of long-established TMDLs: 
Considered helpful: 100%; 
Present: 35%. 

Citizen participation: 
Percentage of long-established TMDLs: 
Considered helpful: 97%; 
Present: 54%. 

Supportive local governmental leaders: 
Percentage of long-established TMDLs: 
Considered helpful: 93%; 
Present: 36%. 

Providing external assistance: 

Funding enough or more than enough: 
Percentage of long-established TMDLs: 
Considered helpful: 94%; 
Present: 14%. 

Assistance from state nonpoint source program: 
Percentage of long-established TMDLs: 
Considered helpful: 95%; 
Present: 70%. 

Assistance from USDA: 
Percentage of long-established TMDLs: 
Considered helpful: 89%; 
Present: 43%. 

Assistance from state agricultural program: 
Percentage of long-established TMDLs: 
Considered helpful: 91%; 
Present: 43%. 

Assistance from EPA: 
Percentage of long-established TMDLs: 
Considered helpful: 81%; 
Present: 37%. 

Assistance from state or local planning agency: 
Percentage of long-established TMDLs: 
Considered helpful: 83%; 
Present: 37%. 

Assistance from state stormwater program: 
Percentage of long-established TMDLs: 
Considered helpful: 67%; 
Present: 28%. 

Improving states’ total maximum daily loads (TMDL) program functions: 

Presence of TMDL implementation plan: 
Percentage of long-established TMDLs: 
Considered helpful: 89%; 
Present: 44%. 

Higher funding priority for water bodies with TMDLs: 
Percentage of long-established TMDLs: 
Considered helpful: 93%; 
Present: 54%. 

Consolidation of TMDL responsibilities: 
Percentage of long-established TMDLs: 
Considered helpful: 54%; 
Present: 92%. 

Source: GAO survey of state TMDL coordinators responsible for 
implementing a nationally representative sample of long-established 
TMDLs. 

Note: "Present" (light blue bars) denotes the percentage of long-
established TMDLs in which the given factor on the y-axis was present 
during implementation of the TMDL in question. "Considered helpful" 
(teal bars) denotes the percentage of these long-established TMDLs in 
which the state TMDL coordinators responding to our survey indicated 
that the factor either was helpful (if present) or (if factor was not 
present or TMDL was not yet implemented) would have been or would be 
helpful in implementing the TMDL in question. In general, estimates 
have a margin of error of no more than 12 percent. For reporting 
purposes, we describe a factor as helpful (or less helpful) on the 
basis of the percentages of long-established TMDLs for which state 
coordinators responding to our survey identified that factor as 
helpful or less helpful in TMDL implementation. 

[End of figure] 

Factors Securing Landowners' and Others' Participation Are Always 
Helpful but Are Often Not in Place: 

Virtually all state TMDL coordinators reported that factors aiming to 
secure landowners' and others' participation in implementing long-
established TMDLs are among the most helpful but, they reported, these 
factors were generally absent from about two-thirds of long-
established TMDLs, and where they were present, they were not always 
present to an extent needed for effective TMDL implementation. 
[Footnote 75] For example, state coordinators reported that landowner 
willingness to implement projects was considered helpful in TMDL 
implementation for 98 percent of long-established TMDLs. Similarly, 
according to state TMDL coordinators, assistance from other 
stakeholders was considered helpful: the coordinators reported that 
the assistance of citizen-based groups was considered helpful for 97 
percent of long-established TMDLs and that the assistance of local 
government leaders was considered helpful for 93 percent of long-
established TMDLs. Yet factors such as these were often not present. 

Specifically, state TMDL coordinators told us that landowners were 
willing to implement projects, including practices most effective at 
reducing pollution and in the locations where they were most needed, 
for about one-third of long-established TMDLs.[Footnote 76] For 
example, officials with Idaho's Department of Environmental Quality 
allocated nonpoint source funds to a restoration project to reduce 
stream bank erosion and sediment in a particular creek because the 
landowner was willing to install a gate to redirect water flow and 
alter the streambed to help reestablish the banks. The creek where 
these actions were taken, however, contributed relatively little 
pollution to the larger impaired watershed. According to a local 
conservation district official, other projects the district wanted to 
pursue, which would have had a much greater benefit in the watershed, 
stalled because landowners at those locations that contributed 
pollution had not agreed to implement the proposed projects because 
they lacked agricultural value. Moreover, state TMDL coordinators 
reported that relatively small proportions of long-established TMDLs 
had enough landowners who were willing to implement projects to 
achieve load allocations called for in these TMDLs. For example, 
landowners had implemented all needed best management practices for 1 
percent of long-established TMDLs, and landowners had implemented more 
than half of needed best management practices for 19 percent of long-
established TMDLs. For 79 percent of long-established TMDLs, 
landowners had implemented half or less than half of needed best 
management practices.[Footnote 77] 

In addition, state TMDL coordinators reported that citizen-based 
groups assisted in implementing 54 percent of long-established TMDLs 
and that local government officials were supportive in implementing 36 
percent of long-established TMDLs. For example, citizen-based groups 
are key to implementing many TMDLs in Pennsylvania, according to 
officials from the state's Department of Environmental Protection. 
Volunteers we met with from one stream restoration association--
consisting of residents in a community directly affected by drainage 
from an abandoned mine--have helped establish and maintain treatment 
systems that remove metals and neutralize acidic water draining from 
the mine.[Footnote 78] State officials told us, however, that if this 
organization had not been involved, they would not have put a high 
priority on funding implementation of this TMDL because they would 
have had little confidence that needed treatment systems would have 
been built or maintained. 

Factors Providing External Assistance Are Almost Always Helpful but 
Are Often Not in Place: 

State TMDL coordinators responding to our survey reported that factors 
providing external assistance to implement projects to reduce nonpoint 
source pollution for long-established TMDLs are also helpful. 
Specifically, state TMDL coordinators reported that the provision of 
funding was considered helpful in implementing 97 percent of long-
established TMDLs. Involvement of agencies other than the one 
responsible for implementing a particular TMDL was also considered 
helpful. For example, technical or financial assistance from state 
nonpoint source pollution programs was considered helpful in 95 
percent of long-established TMDLs, according to our survey 
respondents, and the assistance of other external agencies ranged from 
67 percent (for state stormwater programs)[Footnote 79] to 91 percent 
(for state agriculture programs). With the exception of assistance 
from state nonpoint source pollution programs, however, such factors 
were present for less than half of long-established TMDLs. 

Specifically, state TMDL coordinators told us that enough funding was 
available to implement best management practices for 14 percent of 
long-established TMDLs. The need for enough funding is shown by one 
project in Pennsylvania, where funding was available from two sources-
-grants under EPA's nonpoint source management program and the state's 
cleanup program for nonpoint source pollution--for projects aimed at 
reducing the amount of phosphorus entering a lake used primarily for 
recreation. During the first year after funding was received, best 
management practices that were implemented brought the lake within 25 
percent of the load allocation for phosphorus established in the TMDL. 
With another year's funding received, Pennsylvania water quality 
officials expect other projects to be implemented that will enable the 
lake to meet its load allocation for phosphorus. 

State TMDL coordinators also reported that involvement from external 
agencies or programs varied by agency or program. For example, 
according to these coordinators, state nonpoint source management 
programs helped with implementation for 70 percent of long-established 
TMDLs. State TMDL coordinators reported that other agencies or 
programs (e.g., EPA, USDA, state agricultural programs, state and 
local planning agencies, and state stormwater management programs) 
helped with the implementation of less than half of long-established 
TMDLs. 

Factors Improving State Program Functions Were Sometimes Helpful and 
More Often in Place than Other Factors: 

State TMDL coordinators reported that one factor aiming to improve 
program function was considered helpful in implementing projects to 
reduce pollution for nonpoint sources for most long-established TMDLs, 
and two others were considered helpful for about half of them. 
Specifically, according to state TMDL coordinators, a formal TMDL 
implementation plan was considered helpful for 89 percent of long-
established TMDLs. Coordinators reported, however, that less than half 
(44 percent) of long-established TMDLs have implementation plans to 
identify who must undertake what actions to reduce pollution and on 
what land areas such actions are to be taken to restore impaired water 
bodies. For example, a TMDL for a largely rural watershed in 
Pennsylvania that we visited has an implementation plan that has, 
since 2007, resulted in implementation of best management practices at 
130 different locations within the watershed, including one 300-acre 
farm that has terraced its land and implemented other best management 
practices to minimize soil loss into neighboring streams. 

On the other hand, two factors--funding priority and consolidation of 
TMDL responsibilities in a single agency--were generally present, but 
state TMDL coordinators considered them helpful for about half of the 
long-established TMDLs. For example, in determining how to allocate 
grant funding from EPA's nonpoint source management programs to reduce 
nonpoint source pollution, state TMDL coordinators reported that, for 
93 percent of long-established TMDLs, they assigned water bodies 
associated with a TMDL a higher priority for funding than water bodies 
without a TMDL. State TMDL coordinators reported that allocating grant 
funding in this way was helpful for about half (54 percent) of long-
established TMDLs.[Footnote 80] Similarly, state TMDL coordinators 
reported that for 92 percent of long-established TMDLs, TMDL 
development and implementation responsibilities were consolidated 
within a single state agency. State TMDL coordinators reported that 
the consolidation of TMDL programs within the same agency was 
considered helpful for about half (54 percent) of long-established 
TMDLs in improving water quality in the water body to which the TMDL 
applied.[Footnote 81] 

Some Factors Associated with Effective TMDL Implementation Are Often 
Not In Place: 

Some of the factors facilitating TMDL implementation are not often in 
place, either because (1) EPA has not always asked states to make such 
factors a priority when funding TMDL implementation projects or, 
according to state TMDL coordinators, (2) implementing certain TMDLs 
is a low priority. Two other factors, which are those viewed as most 
helpful, are generally beyond EPA's and states' existing authorities 
to put in place. Specifically, states (1) may find it difficult to 
convince landowners that actions need to be taken on their properties 
to reduce nonpoint source pollution and (2) may have limited ability 
to provide funding to implement more than a small fraction of TMDLs. 

When making grants to states through its nonpoint source management or 
water pollution control programs, EPA has not always asked states to 
make various factors facilitating implementation a priority in 
selecting TMDL implementation projects for funding, particularly for 
the relatively older TMDLs in the scope of our review. Specifically, 
EPA headquarters officials said that in the past the agency rarely 
asked states to focus on more specific conditions, such as 
demonstrating local stakeholder involvement, gaining external 
assistance, or developing formal TMDL implementation plans. Recently, 
however, EPA has taken steps to target nonpoint source management 
grants toward projects demonstrating some of these factors--for 
example, encouraging states to fund projects that target nonpoint 
source management grant funds to water bodies with TMDLs. In 2013, EPA 
issued new guidance on the use of these grants, encouraging states to 
fund projects with local stakeholder involvement, external agency 
assistance, and formal TMDL implementation plans containing specific 
information on locations likely to be contributing to pollution and 
the projects needed to reduce such pollution.[Footnote 82] For 
nonpoint source management grants, states must use 50 percent of their 
annual grant funding for implementing plans in watersheds with 
impaired waters; these plans may identify the presence of some of 
these factors. Although the guidance targets funding to where some of 
these factors may be present, the guidance does not require states to 
follow recommendations for selecting projects to fund or include these 
factors among programmatic conditions on annual nonpoint source 
grants. Without incorporating these factors as conditions into annual 
grants, regarding funding specific projects, rather than only plans, 
EPA may not be targeting its limited funds toward TMDL implementation 
projects that exhibit factors facilitating effective TMDL 
implementation or ensuring that the funds are spent on a mix of 
projects likely to help attain water quality standards. Without such 
targeting, it is likely that the limited improvement in attaining 
water quality standards reported by state coordinators, especially for 
water bodies with associated TMDLs for nonpoint source pollution, will 
continue. 

In addition, certain long-established TMDLs may not be a high priority 
for implementation, and circumstances may have discouraged EPA and 
states from focusing on factors facilitating successful TMDL 
implementation. EPA and state officials said that because many TMDLs 
were developed in response to schedules laid out in court-ordered 
consent decrees--not necessarily because they applied to water bodies 
at high priority for restoration--implementing many of these TMDLs is 
not a high priority and that states may therefore be reluctant to 
invest limited resources in drawing up implementation plans for TMDLs 
they do not plan to implement. For example, officials from Virginia's 
Department of Environmental Quality told us that about one-third of 
TMDLs developed in response to consent decrees have implementation 
plans, whereas a greater share of TMDLs developed without the pressure 
of a consent decree have implementation plans. On the other hand, 
where consent decrees did not drive TMDL development, TMDLs have 
sometimes reflected certain factors that facilitate implementation. 
For example, according to regional officials, states in EPA's Region 5 
developed their TMDL programs without constraint from court-ordered 
schedules, and the region has placed strong emphasis on TMDLs that can 
be implemented. As a result, the region's TMDLs typically contain 
details that facilitate implementation, including identification of 
individual sources of nonpoint pollution and the specific actions that 
each needs to undertake to achieve required load allocations. 
According to agency officials, EPA's long-term vision, which the 
agency plans to implement in fiscal year 2014, seeks to allow states 
to rank TMDLs for development and implementation now that states are 
emerging from several decades of consent-decree-driven time frames and 
have more flexibility in allocating their limited resources. 

Certain other factors, however, including those state TMDL 
coordinators considered most helpful, are beyond EPA's and states' 
existing authorities to put in place, particularly for nonpoint source 
pollution. Specifically, the Clean Water Act addresses nonpoint 
pollution through largely voluntary means and EPA does not have direct 
authority to require landowners to implement activities to reduce 
nonpoint source pollution. As such authority is absent and where 
additional effective state authority is also limited, the inability to 
find enough landowners willing to implement projects to reduce 
nonpoint source pollution has resulted in limited improvements in 
water quality for waters impaired by such pollution, according to our 
survey results. The act does not provide states with the authority to 
require landowner implementation of projects to control nonpoint 
source pollution, and state TMDL coordinators cited a lack of 
authority as the main reason why nonpoint source TMDLs had not been 
implemented. For example, of the 67 state TMDL coordinators who 
submitted narrative responses to our survey, 46 (69 percent) said that 
long-established TMDLs faced implementation challenges because of the 
inability to secure landowners' and others' participation, such as by 
not having authority over nonpoint sources. One survey respondent 
stated that the state had no effective way to compel landowners to 
implement the best management practices called for in a TMDL, and as a 
result, no practices prescribed by that TMDL had been implemented in 
the 10 years since EPA approved it. EPA officials concurred with the 
state coordinator results, saying that the agency works within the 
authority it has to encourage landowners to carry out projects. The 
officials said that the Clean Water Act does not give the agency or 
states the authority to regulate nonpoint source pollution. As noted 
by some congressional members in 1987, when the act was amended to 
create the section 319 program, the program would be revisited if the 
voluntary approaches did not significantly address nonpoint source 
pollution. 

For those states with some specific authority over nonpoint source 
pollution, according to state coordinators, the authority may be 
limited in scope or not routinely used. For example, nearly 1,000 
Pennsylvania TMDLs identify sediment as impairing water quality, and 
runoff from farms is the source of sediment pollution for many of 
these TMDLs. Pennsylvania law requires that all farms have a plan to 
control or reduce sediment entering waterways from fields and animal 
use areas over a certain threshold. Nevertheless, officials said, the 
state's Department of Environmental Protection has never enforced this 
law and required farms to have, and implement, these plans, even 
though the law has been in effect for more than 40 years. The 
department's manager of conservation programs told us that the 
department has not been strict with the agricultural community over 
the years. Of the nearly 1,000 portions of water bodies identified as 
impaired by sediment, 2 have been restored, according to data from 
EPA's Assessment, TMDL Tracking and Implementation System. 

EPA and states also are limited by their respective budgets, and, as a 
result, they cannot provide funding to ensure implementation of a 
sufficient number of best management practices, according to state 
TMDL coordinators. Since section 319 was added to the Clean Water Act 
in 1987, EPA has awarded about $150 million per year, on average, in 
grants to the 50 states and others (the District of Columbia, U.S. 
territories, and Native American tribes) to address the primary cause 
of pollution in more than 33,000 water bodies. Most (86 percent) of 
long-established TMDLs addressing nonpoint source pollution did not 
have adequate funding, according to state TMDL coordinators, to 
implement the best management practices the TMDLs called for. State 
TMDL coordinators reported that although TMDL implementation is a 
priority for funding under the state's nonpoint source management 
grants, funding is typically not in proportion with the amount needed 
to implement projects in all areas or at all specific sites where they 
are needed to reduce nonpoint source pollution and enable water bodies 
to one day attain water quality standards. Moreover, 42 of 67 (63 
percent) state TMDL coordinators reported in narrative responses that 
limited availability of funding impeded TMDL implementation. EPA 
officials said that they work with available funding to leverage 
funding from other agencies, such as USDA grants, for conservation 
projects in the areas subject to TMDLs and have begun to work with 
states under the 2013 long-term vision to prioritize TMDLs for 
development and implementation. 

Conclusions: 

EPA has worked with the states to develop more than 50,000 TMDLs over 
the past two decades, and these TMDLs have proven useful in 
identifying and reducing pollutant loads in many waters impaired by 
pollutants from point sources. Yet progress toward the Clean Water 
Act's goals of restoring and maintaining "the chemical, physical, and 
biological integrity of the nation's waters"--including designated 
uses of fishing, swimming, and drinking--has stalled, largely because 
nonpoint source pollution has not been controlled. As our survey of 
state TMDL program coordinators indicates, few TMDLs have been 
implemented for nonpoint source pollution, and for those that have 
been implemented, progress has generally been incremental, in large 
part because of long-recognized limits to nonpoint source authority 
and funding. In 1987, when Congress amended the Clean Water Act to 
create a voluntary, incentive-based program to address nonpoint source 
pollution, Members of Congress also indicated that this provision was 
a starting point, subject to change if the reliance on voluntary 
participation did not significantly improve water quality. As reported 
by state TMDL coordinators, the absence of two key factors--
specifically, legal authority and sufficient funding--has generally 
stymied the implementation of TMDLs meant to curtail nonpoint source 
pollution. The approach for abating nonpoint source pollution 
established by the 1987 Clean Water Act amendments has not shown much 
progress toward achieving the goals of the act and likely will not do 
so in the foreseeable future. Consequently, more than 40 years after 
enactment of the Clean Water Act, and more than 25 years after 
Congress amended it to institute a program to control nonpoint source 
pollution, a majority of our nation's waters continue to be impaired. 
The preamble to EPA's 2000 rule recognized this fact, stating that 
although significant progress had been made--particularly in stemming 
pollution from point sources--major challenges remained for addressing 
nonpoint source pollution. EPA officials have said that they are 
working within existing authorities to make TMDL implementation a 
priority and to target and leverage funding; even so, many years will 
be needed to attain water quality standards. Without changes to the 
program's voluntary approach to implement projects in waters impaired 
by nonpoint source pollution, the act's goals are likely to remain 
unfulfilled. 

Whether or not changes are made to the Clean Water Act, our review has 
uncovered several shortcomings in EPA's current implementation of the 
TMDL program, as follows: 

* First, the experts' review shows that, where developed properly, 
TMDLs can be useful tools to identify impaired waters and plan actions 
to reduce water pollution. Yet the 11 elements in EPA's guidance, 
which generally reflect the agency's interpretation of TMDL 
development requirements in existing regulations, do not include 
certain features identified by the 2001 NRC report and several EPA 
studies as key for TMDLs to help water bodies attain water quality 
standards, such as comprehensive identification of impairment and 
plans to monitor water bodies to verify that water quality is 
improving. EPA officials said that the agency cannot require TMDLs to 
include additional features without issuing new regulations. Unless 
key features are incorporated into the agency's regulations--upon 
which future guidance can be based--states may include them in TMDLs 
only sporadically, if at all. Furthermore, without the force of 
regulations to direct states to develop TMDLs containing key features, 
TMDLs are likely to do little to attain water quality standards, 
particularly the designated uses of fishing, swimming, and drinking. 
EPA withdrew the 2000 rule before its effective date, and EPA 
officials have not taken action to reinitiate it, they said, both 
because the agency had wanted more time to assess the effectiveness of 
existing regulations and because rulemaking is costly and time-
consuming. Nevertheless, TMDLs without certain key features may be 
unlikely to help water bodies attain water quality standards and may 
potentially waste states' limited resources. 

* Second, because some elements in the agency's existing 11-element 
guidance for TMDL development are either vague or stated as options, 
EPA's regional offices do not apply them consistently. A few regional 
offices have demonstrated, however, that asking states to provide more-
detailed information about pollution causes and abatement actions 
provides greater assurance that TMDLs, if implemented, can achieve 
tangible water quality results. In the absence of clear guidance from 
headquarters directing all regional offices to review TMDLs on the 
basis of specific criteria, EPA's regional offices are unlikely to 
systematically ask states for more than the limited information 
described in the existing elements, including those that are vague or 
optional. As a result, some EPA regions may continue to approve TMDLs 
in which the quality or quantity of information is insufficient to 
help water bodies attain water quality standards, even if the TMDLs 
are implemented. 

* Third, EPA has studied TMDL implementation and identified specific 
factors facilitating such implementation, but it has generally not 
placed on state grants conditions reflecting these factors for TMDL 
implementation plans and projects. In 2013, the agency issued guidance 
targeting nonpoint source management grant funds to states and 
projects that demonstrate some of these factors for effective TMDL 
implementation (e.g., targeting grant funds to projects where 
implementation plans have been developed and where external agency 
assistance is available). The guidance, however, neither requires 
states to follow these recommendations for selecting projects to fund, 
nor requires EPA regions to include these factors among programmatic 
conditions on annual nonpoint source grants. Without incorporating 
these factors as conditions into annual grants, the agency may neither 
target its limited funds toward effective TMDL implementation 
projects, nor ensure that the funds are spent on the mix of projects 
most likely to attain water quality standards or that they meet the 
purposes for which they are awarded. 

* Fourth, the agency has little nationwide information to determine 
whether and to what extent TMDLs have been implemented or to what 
extent implemented TMDLs have helped impaired waters attain water 
quality standards. EPA recognizes that obtaining specific GIS data for 
TMDL implementation projects could help track the status of 
implementation efforts, and the agency has taken steps to gather some 
of these data. According to EPA officials, however, the agency has not 
required states to provide specific GIS data for TMDL projects funded 
in part by EPA that would enable it to link key pieces of such data in 
its own separate databases. Without consistently obtaining from states 
GIS data in addition to water body names, EPA cannot integrate 
information on TMDL implementation projects and impaired water bodies 
across separate databases and cannot assess whether and to what extent 
water quality has been affected by TMDL implementation. Regardless, 
important data related to TMDL implementation are collected by USDA on 
conservation projects it funds and, under the 2008 Farm Bill, remain 
outside of EPA's authority to obtain from USDA without landowners' 
consent. Without access to data from landowners who participate in 
conservation programs funded by USDA in areas subject to a TMDL (e.g., 
on agricultural land adjacent to water bodies with TMDLs)--data 
including the location, type, and number of projects implemented under 
these programs--EPA cannot track actions taken to implement TMDLs or 
any subsequent changes in water quality associated with a core EPA 
program and a substantial federal investment. 

Matter for Congressional Consideration: 

To help ensure effective TMDL implementation in water bodies impaired 
by nonpoint source pollution and to hasten progress toward the Clean 
Water Act's goals of restoring and maintaining the chemical, physical, 
and biological integrity of the nation's waters, Congress should 
consider revising the act's largely voluntary approach to restoring 
waters impaired by such pollution. Specifically, Congress could 
consider ways to address factors, such as limited authority, which 
currently impede attainment of water quality standards, particularly 
the designated uses of fishing, swimming, and drinking. 

Recommendations for Executive Action: 

To enhance the likelihood that TMDLs support the nation's waters' 
attainment of water quality standards and to strengthen water quality 
management, we recommend that the Administrator of EPA take the 
following four actions: 

* Develop and issue new regulations requiring that TMDLs include 
additional elements--and consider requiring the elements that are now 
optional--specifically, elements reflecting key features identified by 
NRC as necessary for attaining water quality standards, such as 
comprehensive identification of impairment and plans to monitor water 
bodies to verify that water quality is improving. 

* To ensure more consistent application of existing TMDL elements and 
to provide greater assurance that TMDLs, if implemented, can achieve 
tangible water quality results, identify regional offices with 
criteria for interpreting and applying such elements in reviewing and 
approving state-developed TMDLs and issue guidance with more 
specificity, directing all regional offices to follow the same 
criteria, including requesting that states provide more-detailed 
information about pollution causes and abatement actions. 

* Place conditions on states' annual use of nonpoint source management 
and water pollution control grants to ensure that the funds meet the 
purposes for which they are awarded and achieve greater reductions in 
nonpoint source pollution associated with TMDL implementation, such as 
by targeting funds to states and projects that incorporate factors 
needed for effective TMDL implementation (e.g., targeting grant funds 
to projects where implementation plans have been developed and where 
external agency assistance is available). 

* Obtain missing data that currently impede EPA's efforts to determine 
whether and to what extent TMDLs have been implemented or to what 
extent implemented TMDLs have helped impaired waters attain water 
quality standards by: 

- directing states to use and report specific GIS data when 
implementing projects to which TMDLs apply and: 

- requesting that USDA ask landowners who participate in conservation 
programs funded by the department in areas subject to a TMDL to 
disclose information on the location, type, and number of projects 
implemented under these programs. 

Agency Comments and Our Evaluation: 

We provided a draft of this report to the Administrator of the 
Environmental Protection Agency for review and comment. EPA provided 
written comments, reproduced in appendix V, in which the agency 
expressed general agreement with the report's findings and partial 
agreement with its conclusions and recommendations. Overall, the 
letter cited agreement with the report's characterization of the 
challenges facing EPA's water quality-based programs and agreement in 
principle with the kinds of changes that would be able to increase 
success under the Clean Water Act. EPA did not comment on the matter 
for congressional consideration, agreed with one recommendation, 
agreed with the findings of another recommendation but did not agree 
to take the recommended action, and partially agreed with two other 
recommendations. EPA also provided separate technical comments, which 
we incorporated in the report as appropriate. 

In response to our recommendation that EPA develop and issue new 
regulations requiring that TMDLs include additional elements, EPA 
stated that it agrees that broader implementation and monitoring 
requirements would be helpful and pointed out that the agency 
attempted to revise the TMDL regulations to include such provisions in 
2000. EPA stated that since Congress prevented those regulations from 
going into effect, this recommendation might be better directed to 
Congress than to EPA. To help address the recommendation, EPA stated 
that it plans to continue its long-standing promotion of comprehensive 
watershed management well beyond the TMDL program. However, these 
plans do not include initiating a rulemaking. We note that while 
Congress temporarily prevented the 2000 rule from going into effect, 
EPA ultimately withdrew the rule in 2002, stating that it needed more 
time to evaluate the effectiveness of existing regulations. In 
withdrawing the rule, EPA stated that the rule would need significant 
changes before it could serve as a blueprint for an efficient and 
effective TMDL program. As we describe in the report, key events have 
transpired since development of the rule that provide EPA and state 
and other water quality managers with a wealth of information on the 
characteristics of effective TMDLs. In addition, and as we reported, 
ample time has passed for EPA to evaluate the effectiveness of 
existing regulations and substantial data shows that EPA's existing 
TMDL regulations--written nearly 30 years ago--do not go far enough to 
ensure they effectively support attainment of water quality standards. 
Specifically, the agency has gained new information on the 
characteristics of effective TMDLs. For these reasons, we continue to 
believe that EPA--with its expertise and decades of experience in 
water quality-based programs--should prepare updated regulations for 
the TMDL program and create a blueprint for an effective TMDL program 
for the future. 

In response to our recommendation that EPA identify regional offices 
with criteria for interpreting and applying TMDL elements in reviewing 
and approving state-developed TMDLs and issue guidance with more 
specificity, directing all regional offices to follow the same 
criteria, EPA states that it will initiate a dialogue with its 
regional offices on overall TMDL review and approval guidance. EPA 
further stated that it would undertake this dialogue in 2014 and that 
this dialogue will consider the recommended changes in specificity and 
content that we raised. 

In response to our recommendation that EPA place conditions on states' 
annual use of nonpoint source management and water pollution control 
grants to ensure that the funds meet the purposes for which they are 
awarded, EPA agreed with the recommendation as it relates to one of 
two grant programs, Section 106, saying that it will review the use of 
funds from one of two grant programs that we examined, and several 
others where our findings may be helpful, to address the 
recommendation. EPA stated that the agency is already implementing the 
recommendation for the second program, Section 319; however, we do not 
agree that this is the case. In particular, EPA stated that it has 
recently issued new guidance for the nonpoint source program that 
substantially addresses much of GAO's recommendation and that there is 
no value added in revising the guidance for the program. We note, 
however, that the guidance targets funds to watersheds where 
comprehensive planning has taken place, whereas we recommend that EPA 
further target funds to specific on-the-ground TMDL projects for which 
certain factors--those that facilitate effective implementation--have 
been confirmed. We do not agree that these are the same actions, as 
the existence of a watershed based plan does not mean that the 
projects ultimately chosen to implement the plan and the TMDLs within 
the watershed will, in fact, reflect those factors that we and EPA 
have identified as necessary for helping to attain water quality 
standards. For this reason, we continue to believe that EPA should 
place conditions on grants to states to ensure the presence of such 
factors. 

In response to our recommendation that EPA obtain missing data that 
currently impede its efforts to determine whether and to what extent 
TMDLs have been implemented or have helped impaired waters attain 
water quality standards, EPA stated that it generally agrees that it 
could direct states to use and report specific GIS data when 
implementing projects to which TMDLs apply. The agency disagreed with 
the second part of the recommendation, however, that it should request 
that USDA ask certain landowners--those participating in conservation 
programs funded by the department in areas subject to a TMDL--for 
permission to disclose information on the location, type, and number 
of projects implemented under these programs. EPA agreed that data on 
the location and type of USDA pollution control projects would be 
useful data for tracking implementation and evaluating water quality 
effects but urged GAO to direct this part of the recommendation to 
USDA. EPA stated that as currently written, the recommendation implies 
that EPA would use USDA data to assess individual actions to implement 
individual TMDLs. Recognizing that the state programs that develop 
TMDLs are partially funded by EPA and that many TMDL projects in 
waters impaired by agricultural runoff are also funded in part by EPA, 
we believe it is incumbent upon the agency to try to obtain data that 
it needs to evaluate the effectiveness of a key water quality program. 
The agency stated that it prefers to rely on states' oversight and 
tracking programs to assess states' collective progress in TMDL 
implementation, which may reflect data shared with the state by the 
department. We understand from our discussions with EPA and state 
officials, however, that this approach provides data aggregated at 
levels that are not precise enough to track actions taken to implement 
TMDLs. As we state in the report, without access to data on the 
location, type, and number of projects implemented in areas subject to 
a TMDL, EPA cannot track possible changes in water quality associated 
with a core EPA program and a substantial federal investment. For 
these reasons, we continue to believe that--at the least--EPA should 
ask for these data from those who may be willing to provide it. 

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 to the 
Administrator of EPA, the appropriate congressional committees, and 
other interested parties. 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) 512-3841 or 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 major 
contributions to this report are listed in appendix VI. 

Signed by: 

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

[End of section] 

Appendix I: Objectives, Scope, and Methodology: 

The objectives of our work were to examine (1) the Environmental 
Protection Agency's (EPA) and states' responsibilities in the total 
maximum daily load (TMDL) program, (2) what is known about the status 
of long-established TMDLs, (3) the extent to which long-established 
TMDLs contain key features that enable attainment of water quality 
standards, and (4) the extent to which such TMDLs exhibit factors that 
facilitate effective implementation. 

To conduct this work, we reviewed EPA reports and independent, peer-
reviewed reports on the status of TMDLs and assessments of TMDL 
implementation projects and activities. We interviewed EPA officials 
in the agency's impaired waters and nonpoint source management 
programs. We also interviewed officials from the Association of Clean 
Water Administrators, National Association of Clean Water Agencies, 
and the Environmental Law Institute. In addition, we visited four 
states in two EPA regions to review the states' responsibilities in 
developing and implementing TMDLs, as well as any challenges the 
states may have faced in doing so. We selected these regions and 
states on the basis of the number of TMDLs each state had developed, 
types of water pollution, differences in state authorities to 
implement TMDLs, and the extent of each state's TMDL implementation 
efforts. 

To examine EPA's and states' responsibilities in developing and 
implementing TMDLs, we reviewed documents and published reports to 
identify and describe the responsibilities of EPA and the states. Such 
documents and reports included relevant provisions of the Clean Water 
Act, including sections 303(d), 305(b), 319, and 402, which describe 
the responsibilities of EPA and the states with regard to identifying 
and restoring impaired waters. We reviewed EPA's 1985 regulations and 
1992 amendments, which describe EPA's implementation of its Clean 
Water Act requirements, as well as the agency's 2002 "Guidelines for 
Reviewing TMDLs under Existing Regulations Issued in 1992." We also 
reviewed EPA's long-term vision for its impaired waters program under 
section 303(d) of the Clean Water Act. We reviewed other relevant 
reports addressing responsibilities within the TMDL program, including 
a 2001 report by the National Research Council (NRC) of the National 
Academies titled Assessing the TMDL Approach to Water Quality 
Management, which is the most recent effort undertaken by the National 
Academies to evaluate the TMDL approach to water quality management. 
We also reviewed a 2007 report by EPA's Office of Inspector General 
titled Total Maximum Daily Load Program Needs Better Data and Measures 
to Demonstrate Environmental Results and a 2008 report by the 
Congressional Research Service titled Clean Water Act and Total 
Maximum Daily Loads (TMDLs) of Pollutants. 

To examine what is known about the status of long-established TMDLs, 
we used information from the Assessment, TMDL Tracking and 
Implementation System on the number of impaired water bodies subject 
to TMDLs and the types of TMDLs by source (point, nonpoint, mixed), 
pollutant, and cause of impairment. To assess the reliability of the 
data from this system, we (1) electronically tested required data 
elements, (2) reviewed existing information about the data and the 
system that produced them, and (3) interviewed agency officials 
knowledgeable about the data. We found these data to be sufficiently 
reliable for purposes of this report. In addition, we reviewed EPA 
reports that examined the status of TMDLs, including the following: 

* FY2010 National Report on Implementing Total Maximum Daily Loads 
(TMDLs), EPA841-R-11-002 (March 2011): 

* Analysis of TMDL Implementation Rates in EPA Region 5, EPA841-R-09-
005 (December 2009): 

* John Hoornbeek et al., Implementing Total Maximum Daily Loads: 
Understanding and Fostering Successful Results (Kent, Ohio: Kent State 
University, December 2008). 

* Developing Effective TMDLs: An Evaluation of the TMDL Process 
(Washington, D.C.: 2007): 

* Developing Effective Nonpoint Source TMDLs: An Evaluation of the 
TMDL Development Process (Washington, D.C.: January 2007). 

To examine the extent to which long-established TMDLs contain key 
features needed to enable a water body to attain water quality 
standards, as well as to examine the extent to which long-established 
TMDLs exhibit factors facilitating effective implementation, we used 
the Assessment, TMDL Tracking and Implementation System database to 
select a generalizable, nationwide sample of 191 long-established 
TMDLs, which we defined to be those approved by EPA through December 
31, 2007.[Footnote 83] From this sample, we drew a random subsample of 
25 TMDLs, which we asked three authors of the 2001 NRC report to 
review (see appendix II). In addition, we used the whole sample of 191 
long-established TMDLs in a survey of state TMDL coordinators on the 
water quality status associated with long-established TMDLs, such as 
whether and to what extent water quality had improved in the water 
body subject to the TMDL and whether water quality standards had been 
attained. 

Experts' Review of Long-Established TMDLs: 

To examine the extent to which long-established TMDLs contain key 
features needed to enable a water body to attain water quality 
standards, we hired three water resource experts who had contributed 
to NRC's 2001 report. They reviewed our random subsample of 25 TMDLs 
drawn from our generalizable nationwide sample of 191 TMDLs. Unlike 
the sample as a whole, the subsample is not generalizable to all 
TMDLs, although it is an unbiased set that provides detailed examples 
of TMDL development in specific states and regions, which help 
illustrate challenges and successes in TMDL development. The 25 TMDLs 
in the subsample include TMDLs from all EPA regions except Region 2 
and were approved by EPA from 1997 to 2007; see appendix III for 
additional information about these TMDLs. 

We developed a questionnaire for the experts to use in examining the 
25 TMDLs, and we charged them with identifying those features that are 
most important, or key, if a TMDL is to enable a water body to attain 
water quality standards. We began by reviewing the 2001 NRC report and 
EPA studies and guidance documents, such as the agency's 2002 
guidelines for reviewing TMDLs, for features important to TMDL 
development and implementation (see appendix III for results on 
features). We used the resulting list of features to then develop a 
total of 82 survey questions, which we pretested with the three 
experts to ensure that (1) the questions were clear and unambiguous, 
(2) the experts interpreted terms and questions similarly, (3) 
terminology was used correctly, (4) the information could feasibly be 
obtained from TMDL documents, and (5) the questionnaire was 
comprehensive and unbiased. We changed the questionnaire's content and 
format after the pretest according to the feedback we received. 
Independent GAO survey experts also reviewed a draft of the 
questionnaire, and we further revised the questionnaire to reflect 
that review. 

To administer the questionnaire, we asked each expert to independently 
review 16 or 17 of the 25 TMDLs, fill out the questionnaire, and then 
compare and discuss responses with one another to reach consensus on 
as many questions as possible; we collected the questionnaire data 
from March through May 2013. We first randomly assigned each expert a 
set of TMDLs to review; we then assigned half of each expert's set to 
another expert to review, such that each half of one expert's 
assignment overlapped with half of another expert's assignments, and, 
thus, each TMDL was ultimately reviewed by two experts. The experts 
reviewed their assigned TMDLs using a self-administered electronic 
version of the questionnaire posted on the World Wide Web. We e-mailed 
the experts unique usernames and passwords to ensure that only they 
could participate in the questionnaire. To minimize potential error, 
we facilitated a Microsoft Word-based process for the experts to 
compare and discuss questionnaire responses and to reach agreement on 
as many responses as possible. For each TMDL, we compiled the 
responses from the two experts who reviewed it and sent a copy of the 
compiled responses to the experts. We identified questions to which 
the experts provided substantially divergent responses, which we 
defined as responses where the experts disagreed about the presence or 
absence of a feature. We then asked the experts to exchange comments 
electronically and, if possible, to come to agreement on the best 
response. Once experts reached agreement, we coded their questionnaire 
responses to reflect changes, if any, and to enable us to tally the 
final responses to each of our questions. In the body of our report, 
we present only those questionnaire responses that experts reached 
agreement on. Two independent GAO analysts checked and verified that 
responses were compiled correctly and that any changes were coded 
correctly. 

As a result of our steps to minimize survey error, experts reached 
agreement on 75 percent or more of the survey questions for 24 of the 
25 TMDLs they reviewed.[Footnote 84] The remaining TMDL had a lower 
agreement rate, in large part because the experts disagreed about 
whether this TMDL adequately demonstrates that sediment was the cause 
of biological impairment; their differing views on this issue affected 
their responses to many survey questions about this TMDL. Another 
source of divergent answers was the size and scope of some TMDLs. A 
few TMDLs in our sample are part of large TMDL documents encompassing 
numerous water bodies. The experts said that some of these large TMDL 
documents include little to no information about the specific water 
bodies to which the TMDLs in our sample apply, which made it difficult 
to ensure that each expert was basing his evaluations on the same 
portion of the documents. 

For 73 out of 82 questions, experts reached agreement on 75 percent or 
more of the TMDLs they reviewed. The 9 questions with lower rates of 
agreement were mostly of the same type: they asked whether the experts 
thought particular elements of a TMDL contain key features that, if 
the TMDL were fully implemented, would lead to attainment of a given 
numeric water quality target. One expert consistently expressed less 
confidence than the other experts that the models and other analytical 
methods used to develop most TMDLs would lead to attainment of numeric 
water quality targets. The experts generally agreed in their responses 
to a set of related questions: they responded that, for most TMDLs 
they reviewed, elements in the TMDLs do not contain key features that, 
if the TMDL were fully implemented, would lead to attainment of a 
designated use. 

Survey of State TMDL Coordinators: 

To examine the extent to which long-established TMDLs exhibit factors 
that facilitate effective implementation, we surveyed state TMDL 
coordinators responsible for implementing the 191 long-established 
TMDLs in our generalizable sample. The purpose of this survey was to 
examine topics such as the extent of TMDL implementation, the extent 
of water quality improvement; factors that were helpful in 
implementing each TMDL; and factors that would have been helpful in 
implementing a given TMDL if they had been present or, if a given TMDL 
had not yet been implemented, would be helpful if it were present. To 
identify factors pertaining to TMDL implementation and to develop the 
survey questions, we reviewed EPA reports identifying particular 
factors associated with TMDLs that have been effectively implemented--
that is, have helped the water bodies to which they apply partially or 
fully attain water quality standards (see appendix IV for survey 
results). 

We drew an initial simple random sample of TMDLs for state 
coordinators to examine from EPA's Assessment, TMDL Tracking and 
Implementation System database, out of a universe of 35,705 TMDLs 
approved by EPA from January 1, 1997, to December 31, 2007. To 
minimize the burden on prospective survey respondents, we drew a 
stratified random sample of 191 TMDLs--the smallest number we could 
have drawn for a nationally representative sample--to reflect their 
proportion in the universe of 35,075 TMDLs. Stratifying the sample in 
this way allowed us to reduce the number of TMDLs that any one state 
TMDL coordinator would receive for review. From this initial sample of 
191 TMDLs, we removed TMDLs that were developed for entire states or 
multistate regions, such as those developed for atmospheric pollutants 
such as mercury or acid rain. We then drew a supplemental random 
sample to replace the TMDLs we had removed, so that we still had a 
representative sample of 191 TMDLs nationwide that applied to water 
bodies. The results of our survey are generalizable to the population 
of TMDLs that were developed by December 31, 2007, which we defined as 
long-established TMDLs, and do not apply at statewide or regional 
scales. 

We conducted the survey by means of self-administered electronic 
questionnaires posted on the World Wide Web, and we collected survey 
data from February through May 2013. We pretested the survey questions 
with officials from four states--Illinois, Kansas, Minnesota, and 
North Carolina--to check that (1) the questions were clear and 
unambiguous, (2) terminology was used correctly, (3) the questionnaire 
did not place an undue burden on state agency officials, (4) the 
information could feasibly be obtained, and (5) the survey was 
comprehensive and unbiased. We made changes to the content or format 
of the questionnaire after each pretest according to the feedback we 
received. Independent GAO survey experts also reviewed a draft of the 
survey, and we further revised the survey to reflect that review. We 
contacted survey respondents by e-mail, and we gave each potential 
respondent a unique password and username to ensure that only members 
of the target population could participate in the survey. We sent 
follow-up e-mail messages to those who had not responded by the 
deadline, and, beginning in May 2013, we telephoned all remaining 
nonrespondents for whom contact information was available. We received 
a total of 177 survey responses, which, after certain adjustments 
needed to analyze the data, gave an overall response rate of 93 
percent. 

To analyze the results of our survey, we first weighted the TMDLs from 
each state coordinator in proportion to the number of TMDLs developed 
by each state; in other words, a TMDL from Pennsylvania, for example--
which had developed roughly five times as many TMDLs as, say, 
Washington State--had more weight in our analyses than a TMDL from 
Washington. 

Estimates produced from the sample of projects are subject to sampling 
error. We express our confidence in the precision of our results as a 
95 percent confidence interval. This interval would contain the actual 
population value for 95 percent of the samples we could have drawn. As 
a result, we are 95 percent confident that each of the confidence 
intervals in this report includes the true values in the study 
population. The estimates from our survey generally have a margin of 
error of no more than 12 percent at the 95 percent confidence level, 
unless otherwise noted. Additionally, to encourage honest and open 
responses, we pledged in the introduction to the survey that we would 
report information in the aggregate and not report data that would 
identify a particular respondent. This report does not contain all the 
results from the survey; the survey and a more complete tabulation of 
the results are provided in appendix IV. To eliminate data-processing 
errors, we independently verified the computer program that generated 
the survey results. 

In addition to tabulating and analyzing the frequencies of survey 
responses, we conducted a content analysis of all of the open-ended 
narrative responses received to survey questions 7, 14, 24, 26B, 27, 
36, 39G, 42, 48, and 49. We analyzed the content of the 111 responses 
to question 14, the 115 responses to question 24, the 33 responses to 
question 26B, the 72 responses to question 27, the 93 responses to 
question 36, the 51 responses to question 39G, the 52 responses to 
question 42, the 39 responses to question 45G, the 82 responses to 
question 48, and the 90 responses to question 49. Question 7 was coded 
using the following categories: additional data needed, EPA-developed 
TMDL, natural condition TMDL, new TMDLs are better, and TMDL no longer 
applies. Question 14 was coded using the following categories: 
additional data needed, legal authority lacking, limited verification 
data, low priority, and TMDL was not needed. Question 24 was coded 
using the following categories: additional priorities weighted, data 
available, lack of project sponsorship, not applicable, priority on 
listing, and too many impaired water bodies. Question 26B was coded 
using the following categories: authority rests with another agency, 
cumbersome or difficulty authority to use, improvements too costly, 
prefer voluntary approach, vague authority, and addressed point 
sources and municipal separate storm sewer systems first. Question 27 
was coded using the following categories: forestry regulation, manure 
management, septic system, and wetland and riparian. Question 36 was 
coded using the following categories: additional data needed, consent 
decree TMDL, funding availability, low priority, more outreach needed, 
and poor targeting. Question 39G was coded using the following 
categories: discharge monitoring reports and geographic information 
system data. Question 48 was coded using the following categories: 
additional data needed, funding availability, more time needed, no 
change expected, no longer applies, low priority, and now a municipal 
separate storm sewer system. Question 49 was coded using the following 
categories: funding availability, legal authority lacking, 
stakeholders lacking, new TMDLs are better, resources misspent, and 
TMDL was not the appropriate tool. Coding was performed independently 
by two coders; team members then met to discuss the coding categories 
and reached consensus on the final coding category assignment for each 
response. Measures of reliability between GAO coders were calculated 
before codes were reconciled and found to be sufficiently high for 
purposes of this analysis. The numbers of responses in each content 
category were then summarized and tallied. 

Regional Office Review: 

We also interviewed officials in each of EPA's 10 regional offices 
responsible for reviewing and approving TMDLs. We discussed with these 
officials the extent to which they require states' TMDLs to 
demonstrate key features as conditions of approving TMDLs. Further, we 
discussed with them the extent to which they encourage state programs 
to ensure the presence of factors necessary for effective TMDL 
implementation. We obtained information on the offices' review and 
approval of state-developed TMDLs, as well as on their oversight of 
state programs, including whether and to what extent regional offices 
include in annual grants to states conditions that encourage presence 
of factors associated with effective TMDL implementation. 

We conducted this performance audit from July 2012 to December 2013 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: National Research Council Experts GAO Selected to Review 
TMDLs: 

James R. Karr; 
Professor Emeritus, University of Washington, Seattle: 

Vladimir Novotny; 
Professor Emeritus, Marquette University, Milwaukee, Wisconsin, and 
Northeastern University, Boston, Massachusetts: 

Chris O. Yoder; 
Research Director, Midwest Biodiversity Institute, Columbus, Ohio: 

[End of section] 

Appendix III: Experts' Review of Long-Established TMDLs: 

To examine the extent to which long-established TMDLs contain key 
features needed to enable a water body to attain water quality 
standards, we appointed water resource experts to review a random, non-
generalizable subsample of 25 TMDLs drawn from a random, generalizable 
nationwide sample (see appendix I). Table 2 gives more detail about 
these 25 TMDLs. 

Table 2: Random Subsample of 25 TMDLs Reviewed by Water Resource 
Experts: 

State: Alabama; 
Year EPA approved: 1997; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Ammonia (nitrogen). 

State: Alabama; 
Year EPA approved: 2002; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Biochemical oxygen demand. 

State: Arkansas; 
Year EPA approved: 2006; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Copper. 

State: California; 
Year EPA approved: 2006; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Siltation. 

State: Colorado; 
Year EPA approved: 2002; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Lead. 

State: Connecticut; 
Year EPA approved: 2005; 
Water body classification: Lake, reservoir, or pond; 
Pollution source: Point and nonpoint; 
Stressor: Phosphorus. 

State: Georgia; 
Year EPA approved: 2002; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Sediment. 

State: Hawaii; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Nutrients (nitrogen and phosphorus). 

State: Indiana; 
Year EPA approved: 2006; Water body classification: 
River or stream; 
Pollution source: Point and nonpoint; 
Stressor: E. coli bacteria. 

State: Iowa; 
Year EPA approved: 2004; 
Water body classification: Lake, reservoir, or pond; 
Pollution source: Nonpoint; 
Stressor: Algal growth. 

State: Kansas; 
Year EPA approved: 2004; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Sulfate. 

State: Maryland; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Sulfate. 

State: Massachusetts; 
Year EPA approved: 2002; 
Water body classification: Lake, reservoir, or pond; 
Pollution source: Nonpoint; 
Stressor: Phosphorus. 

State: Mississippi; 
Year EPA approved: 2003; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Biochemical oxygen demand. 

State: Mississippi; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Sediment. 

State: New Mexico; 
Year EPA approved: 2006; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Turbidity. 

State: Oregon; 
Year EPA approved: 2006; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Temperature. 

State: Oregon; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Temperature. 

State: Pennsylvania; 
Year EPA approved: 2003; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Phosphorus. 

State: South Dakota; 
Year EPA approved: 2001; 
Water body classification: Lake, reservoir, or pond; 
Pollution source: Nonpoint; 
Stressor: Eutrophicationa. 

State: South Dakota; 
Year EPA approved: 2003; 
Water body classification: River or stream; 
Pollution source: Point; 
Stressor: Ammonia (nitrogen). 

State: Utah; 
Year EPA approved: 2002; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Phosphorus. 

State: Washington; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Point and nonpoint; 
Stressor: Temperature. 

State: Wisconsin; 
Year EPA approved: 2005; 
Water body classification: River or stream; 
Pollution source: Nonpoint; 
Stressor: Sediment. 

State: Wyoming; 
Year EPA approved: 2007; 
Water body classification: River or stream; 
Pollution source: Point; 
Stressor: Chlorine. 

Source: GAO analysis of EPA data. 

[A] Eutrophication is the natural process by which lakes and ponds age 
as minerals and organic matter accumulate in the water over thousands 
of years. These nutrients promote blooms of algae, which drop to the 
bottom when they die, in turn increasing oxygen-demanding 
decomposition rates and depleting oxygen needed by other forms of life 
in the lake. As a result, overall diversity of life in the lake 
declines. Excess nutrients contributed by human activities, such as 
the application of nitrogen-rich fertilizers, hastens this process. 

[End of table] 

Table 3 shows the results of the experts' review of the TMDLs listed 
in table 2. These results are expressed as the number of TMDLs 
containing the feature listed in each table row. 

Table 3: Number of Long-Established TMDLs in Which Experts Found Key 
Features Needed to Enable Water Bodies to Attain Water Quality 
Standards: 

EPA element: 1; 

TMDL key feature: Provides sufficient evidence that water quality is 
impaired by the selected pollutant of concern; 
Experts agreed: 
Contain feature: 18; 
Do not contain feature: 6; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 1. 

TMDL key feature: Addresses chemical stressors that may be affecting 
attainment of water quality standards; 
Experts agreed: 
Contain feature: 14; 
Do not contain feature: 8; 
Not enough information: 0; 
Feature not applicable: 3; 
Experts did not agree: 0. 

TMDL key feature: Addresses biological stressors that may be 
affecting attainment of water quality standards; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 11; 
Not enough information: 0; 
Feature not applicable: 2; 
Experts did not agree: 4. 

TMDL key feature: Addresses physical stressors that may be affecting 
attainment of water quality standards; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 13; 
Not enough information: 0; 
Feature not applicable: 3; 
Experts did not agree: 1. 

EPA element: 2; 

TMDL key feature: Demonstrates that achieving numeric 
water quality target for given pollutant will lead to attainment of 
designated use; 
Experts agreed: 
Contain feature: 7; 
Do not contain feature: 17; 
Not enough information: 1; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Describes scientific linkage between numeric water 
quality target and pollutant of concern; 
Experts agreed: 
Contain feature: 15; 
Do not contain feature: 4; 
Not enough information: 0; 
Feature not applicable: 3; 
Experts did not agree: 3. 

TMDL key feature: Describes how numeric water quality target was 
calculated; 
Experts agreed: 
Contain feature: 22; 
Do not contain feature: 3; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Represents numeric water quality target in terms of 
magnitude, or how much pollutant of concern can be present without loss 
of water's designated use; 
Experts agreed: 
Contain feature: 22; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 1; 
Experts did not agree: 1. 

TMDL key feature: Represents numeric water quality target in terms of 
duration, or time period over which pollutant levels are to be measured 
to determine how much pollutant can be present without loss of 
designated use; 
Experts agreed: 
Contain feature: 18; 
Do not contain feature: 2; 
Not enough information: 0; 
Feature not applicable: 2; 
Experts did not agree: 3. 

TMDL key feature: Represents numeric water quality target in terms of 
frequency, or how often numeric water quality target may be exceeded 
without loss of designated use; 
Experts agreed: 
Contain feature: 14; 
Do not contain feature: 5; 
Not enough information: 0; 
Feature not applicable: 3; 
Experts did not agree: 3. 

TMDL key feature: Numeric water quality target is measurable with 
reasonably obtainable monitoring data; 
Experts agreed: 
Contain feature: 25; 
Do not contain feature: 0; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Numeric water quality target accounts for pollutant 
loads flowing into or out of other water bodies or segments not part 
of geographic area of water body to which TMDL applies; 
Experts agreed: 
Contain feature: 15; 
Do not contain feature: 2; 
Not enough information: 5; 
Feature not applicable: 0; 
Experts did not agree: 3. 

EPA element: 3; 

TMDL key feature: Describes "critical conditions," or worst-case 
scenario of environmental conditions in water body under which 
relevant pollutant loading will still allow water body to attain 
water quality standards; 
Experts agreed: 
Contain feature: 17; 
Do not contain feature: 4; 
Not enough information: 1; 
Feature not applicable: 0; 
Experts did not agree: 3. 

TMDL key feature: Model or other method used to calculate loading 
capacity includes natural processes taking place in the watershed; 
Experts agreed: 
Contain feature: 19; 
Do not contain feature: 2; 
Not enough information: 1; 
Feature not applicable: 0; 
Experts did not agree: 3. 

TMDL key feature: Model or other method used to calculate loading 
capacity is appropriately complex; 
Experts agreed: 
Contain feature: 14; 
Do not contain feature: 4; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 7. 

TMDL key feature: Explains any assumptions in the TMDL analysis; 
Experts agreed: 
Contain feature: 22; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 2. 

TMDL key feature: Describes strengths and weaknesses in analytical 
process used to calculate loading capacity; 
Experts agreed: 
Contain feature: 11; 
Do not contain feature: 13; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 1. 

TMDL key feature: Describes results from any water quality modeling 
used in the TMDL analysis; 
Experts agreed: 
Contain feature: 17; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 2; 
Experts did not agree: 5. 

EPA element: 4; 

TMDL key feature: Demonstrates that meeting load allocations will have 
desired effect on water quality conditions needed to attain designated 
use; 
Experts agreed: 
Contain feature: 12; 
Do not contain feature: 8; 
Not enough information: 0; 
Feature not applicable: 2; 
Experts did not agree: 3. 

TMDL key feature:Identifies the nonpoint sources contributing to 
impairment precisely enough for load allocations to be implemented; 
Experts agreed: 
Contain feature: 12; 
Do not contain feature: 5; 
Not enough information: 2; 
Feature not applicable: 2; 
Experts did not agree: 4. 

EPA element: 5; 

TMDL key feature: Specifies pollutant limits in 
individual permits for point source facilities[A]; 
Experts agreed: 
Contain feature: 15; 
Do not contain feature: 1; 
Not enough information: 1; 
Feature not applicable: 6; 
Experts did not agree: 2. 

TMDL key feature: Demonstrates that wasteload allocations for 
individual point source permit holders will not cause localized 
impairments; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 5; 
Not enough information: 0; 
Feature not applicable: 6; 
Experts did not agree: 6. 

TMDL key feature: Demonstrates that pollutant loads have not been 
reallocated from nonpoint sources to point sources; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 2; 
Not enough information: 4; 
Feature not applicable: 6; 
Experts did not agree: 5. 

EPA element: 6; 

TMDL key feature: Refers to a margin of safety; 
Experts agreed: 
Contain feature: 23; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 1. 

TMDL key feature: Includes an explicit uncertainty analysis as basis 
for determining margin of safety; 
Experts agreed: 
Contain feature: 2; 
Do not contain feature: 19; 
Not enough information: 0; 
Feature not applicable: 2; 
Experts did not agree: 2. 

TMDL key feature: Margin of safety adequately accommodates 
variability that might impair attainment of numeric water quality 
target; 
Experts agreed: 
Contain feature: 9; 
Do not contain feature: 7; 
Not enough information: 5; 
Feature not applicable: 2; 
Experts did not agree: 2. 

EPA element: 7; 

TMDL key feature: Refers to seasonal variation; 
Experts agreed: 
Contain feature: 23; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 1. 

TMDL key feature: Includes information about how seasonal variations 
in stressors influence water quality;
Experts agreed: 
Contain feature: 17; 
Do not contain feature: 3; 
Not enough information: 0; 
Feature not applicable: 4; 
Experts did not agree: 1. 

TMDL key feature: Describes method chosen for incorporating seasonal 
variation; 
Experts agreed: 
Contain feature: 18; 
Do not contain feature: 1; 
Not enough information: 1; 
Feature not applicable: 4; 
Experts did not agree: 1. 

EPA element: 8; 

TMDL key feature: Refers to reasonable assurances; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 5; 
Not enough information: 0; 
Feature not applicable: 12; 
Experts did not agree: 0. 

TMDL key feature: Provides sufficient evidence that reductions in 
nonpoint source pollution will occur; 
Experts agreed: 
Contain feature: 0; 
Do not contain feature: 6; 
Not enough information: 0; 
Feature not applicable: 17; 
Experts did not agree: 2. 

TMDL key feature: Identifies specific actions that need to occur to 
attain a water quality standard; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 0; 
Not enough information: 0; 
Feature not applicable: 17; 
Experts did not agree: 0. 

TMDL key feature: Identifies land areas where actions need to occur 
to attain a water quality standard; 
Experts agreed: 
Contain feature: 6; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 17; 
Experts did not agree: 1. 

EPA element: 9; 

TMDL key feature: Refers to a monitoring plan to track TMDL 
effectiveness; 
Experts agreed: 
Contain feature: 18; 
Do not contain feature: 7; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Monitoring plan describes data to be collected to 
determine if TMDL's targeted load reductions are occurring; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 5; 
Not enough information: 0; 
Feature not applicable: 7; 
Experts did not agree: : 5. 

TMDL key feature: Monitoring plan includes indicators reflecting 
biological responses, such as measures of biological diversity and 
relative abundance of particular groups of aquatic organisms; 
Experts agreed: 
Contain feature: 3; 
Do not contain feature: 14; 
Not enough information: 0; 
Feature not applicable: 7; 
Experts did not agree: 1. 

TMDL key feature: Includes information about combining monitoring 
data and modeling so as to assess modeling error; 
Experts agreed: 
Contain feature: 0; 
Do not contain feature: 13; 
Not enough information: 0; 
Feature not applicable: 8; 
Experts did not agree: 4. 

TMDL key feature: Describes an adaptive approach to implementing 
the TMDL, whereby monitoring and data analysis will be used to 
periodically assess progress toward water quality standards and TMDL 
will be adjusted as needed; 
Experts agreed: 
Contain feature: 9; 
Do not contain feature: 8; 
Not enough information: 0; 
Feature not applicable: 7; 
Experts did not agree: 1. 

EPA element: 10; 

TMDL key feature: Refers to implementation; 
Experts agreed: 
Contain feature: 21; 
Do not contain feature: 4; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Identifies actors to help support TMDL 
implementation; 
Experts agreed: 
Contain feature: 14; 
Do not contain feature: 7; 
Not enough information: 0; 
Feature not applicable: 4; 
Experts did not agree: 0. 

TMDL key feature: Identifies actions that can help attain water 
quality standards; 
Experts agreed: 
Contain feature: 12; 
Do not contain feature: 7; 
Not enough information: 0; 
Feature not applicable: 4; 
Experts did not agree: 2. 

TMDL key feature: Includes a formal TMDL implementation plan; 
Experts agreed: 
Contain feature: 6; 
Do not contain feature: 19; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Implementation plan identifies actors to help 
support TMDL implementation; 
Experts agreed: 
Contain feature: 5; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 19; 
Experts did not agree: 0. 

TMDL key feature: Implementation plan identifies specific actions 
that need to occur to attain water quality standards; 
Experts agreed: 
Contain feature: 5; 
Do not contain feature: 1; 
Not enough information: 0; 
Feature not applicable: 19; 
Experts did not agree: : 0. 

TMDL key feature: Implementation plan identifies land areas where 
actions need to occur to attain water quality standards; 
Experts agreed: 
Contain feature: 3; 
Do not contain feature: 2; 
Not enough information: 0; 
Feature not applicable: 19; 
Experts did not agree: : 1. 

TMDL key feature: Implementation plan provides sufficient evidence 
that reductions in nonpoint source pollution will occur; 
Experts agreed: 
Contain feature: 0; 
Do not contain feature: 5; 
Not enough information: 0; 
Feature not applicable: 19; 
Experts did not agree: 1. 

EPA element: 11; 

TMDL key feature: Refers to public participation; 
Experts agreed: 
Contain feature: 21; 
Do not contain feature: 4; 
Not enough information: 0; 
Feature not applicable: 0; 
Experts did not agree: 0. 

TMDL key feature: Describes extent of public participation, such as 
number or types of public meetings; 
Experts agreed: 
Contain feature: 8; 
Do not contain feature: 10; 
Not enough information: 0; 
Feature not applicable: 4; 
Experts did not agree: 3. 

TMDL key feature: TMDL key feature: Includes a summary of significant 
public comments and state's response; 
Experts agreed: 
Contain feature: 3; 
Do not contain feature: 17; 
Not enough information: 0; 
Feature not applicable: 4; 
Experts did not agree: 1. 

Source: GAO. 

Notes: Numbers in this table represent numbers of TMDLs in a random 
subsample of 25 TMDLs. Each TMDL was independently reviewed by two 
experts to increase the reliability of our findings. For each column 
under "Experts agreed," the two experts agreed that the given feature 
in a row was present in that number of TMDLs. For the TMDLs tallied in 
the last column at right, the two experts who reviewed them did not 
reach agreement on the presence or absence of the given feature. 

[A] Permits are those issued to point source facilities under the 
National Pollutant Discharge Elimination system permit program. 

[End of table] 

[End of section] 

Appendix IV: Survey of State TMDL Coordinators Responsible for 
Implementing Long-Established TMDLs: 

The questions we asked of state TMDL coordinators in a nationwide 
survey of long-established TMDLs are shown below. In our survey, we 
asked both closed and open-ended questions about specific TMDLs. For 
closed questions, we have included the frequency for each response 
option that state TMDL coordinators selected and a representative 
estimate for each response when it was generalizable. Generally, we 
considered a response to be generalizable when the bounds of the 95 
percent confidence interval did not exceed 12 percent. For the open-
ended questions in our survey, we have provided the three most 
prevalent response themes when we could identify them. For a more 
detailed discussion of our survey and sampling methodology see 
appendix I. 

TMDL Development: 

Question: Q1. Was the TMDL for [pollutant] in [water body] developed in 
response to a consent decree or other settlement order? 

Response: Yes; 
Count: 79; 
Estimated percentage: 62.7; 
95 percent confidence interval: 54-71.4. 

Response: No; 
Count: 89; 
Estimated percentage: 36.0; 
95 percent confidence interval: 27.4-44.5. 

Response: Don't know; 
Count: 6; 
Estimated percentage: 1.3; 
95 percent confidence interval: 0-2.7. 

Question: Q2. Generally, were land use data available when developing 
the TMDL for [pollutant] in [water body]? 

Response: Yes; 
Count: 154; 
Estimated percentage: 92.4; 
95 percent confidence interval: 88.4-96.4. 

Response: No; 
Count: 10; 
Estimated percentage: 4.9; 
95 percent confidence interval: 1.5-8.3. 

Response: Don't know; 
Count: 9; 
Estimated percentage: 2.7; 
95 percent confidence interval: 0.5-4.9. 

Question: Q2A. If yes, how helpful was the availability of land use 
data when developing the TMDL? 

Response: Very helpful; 
Count: 77; 
Estimated percentage: 52.3; 
95 percent confidence interval: 42-62.6. 

Response: Somewhat helpful; 
Count: 51; 
Estimated percentage: 28.9; 
95 percent confidence interval: 20-37.7. 

Response: Not helpful; 
Count: 23; 
Estimated percentage: 18.8; 
95 percent confidence interval: 9.4-28.3. 

Question: Q2B. If no, how helpful would the availability of land use 
data have been when developing the TMDL? 

Response: Very helpful; 
Count: 2; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q3. Generally, were land cover data available when developing 
the TMDL for [pollutant] in [water body]? 

Response: Yes; 
Count: 150; 
Estimated percentage: 89.7; 
95 percent confidence interval: 84.9-94.5. 

Response: No; 
Count: 12; 
Estimated percentage: 6.4; 
95 percent confidence interval: 2.5-10.4. 

Response: Don't know; 
Count: 11; 
Estimated percentage: 3.9; 
95 percent confidence interval: 1.2-6.6. 

Question: Q3A. If yes, how helpful was the availability of land cover 
data when developing the TMDL? 

Response: Very helpful; 
Count: 71; 
Estimated percentage: 49.8; 
95 percent confidence interval: 39.3-60.3. 

Response: Somewhat helpful; 
Count: 51; 
Estimated percentage: 30.5; 
95 percent confidence interval: 21.3-39.7. 

Response: Not helpful; 
Count: 25; 
Estimated percentage: 19.7; 
95 percent confidence interval: 10-29.3. 

Question: Q3B. If no, how helpful would the availability of land cover 
data have been when developing the TMDL? 
Response: Very helpful; 

Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 2; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q4. Generally, were ambient water quality data available when 
developing the TMDL for [pollutant] in [water body]? 

Response: Yes; 
Count: 150; 
Estimated percentage: 89.8; 
95 percent confidence interval: 84.5-95.2. 

Response: No; 
Count: 11; 
Estimated percentage: 5.8; 
95 percent confidence interval: 2.2-9.3. 

Response: Don't know; 
Count: 11; 
Estimated percentage: 4.4; 
95 percent confidence interval: 0.2-8.6. 

Question: Q4A. If yes, how helpful was the availability of ambient 
water quality data when developing the TMDL? 

Response: Very helpful; 
Count: 122; 
Estimated percentage: 84.8; 
95 percent confidence interval: 78.1-91.5. 

Response: Somewhat helpful; 
Count: 22; 
Estimated percentage: 14.2; 
95 percent confidence interval: 7.6-20.7. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: 1.1; 
95 percent confidence interval: 0-2.5. 

Question: Q4B. If no, how helpful would the availability of ambient 
water quality data have been when developing the TMDL? 

Response: Very helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q5. Generally, were source loading data available when 
developing the TMDL for [pollutant] in [water body]? 

Response: Yes; 
Count: 94; 
Estimated percentage: 63.3; 
95 percent confidence interval: 54.5-72.1. 

Response: No; 
Count: 63; 
Estimated percentage: 33.5; 
95 percent confidence interval: 24.9-42.2. 

Response: Don't know; 
Count: 15; 
Estimated percentage: 3.2; 
95 percent confidence interval: 1.2-5.2. 

Question: Q5A. If yes, how helpful was the availability of source 
loading data when developing the TMDL? 

Response: Very helpful; 
Count: 72; 
Estimated percentage: 79.5; 
95 percent confidence interval: 70.2-88.9. 

Response: Somewhat helpful; 
Count: 20; 
Estimated percentage: 20.5; 
95 percent confidence interval: 11.1-29.8. 

Question: Q5B. If no, how helpful would the availability of source 
loading data have been when developing the TMDL? 

Response: Very helpful; 
Count: 27; 
Estimated percentage: [Empty]; 
95 percent confidence 
interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 4; 
Estimated percentage: 3.5; 
95 percent confidence interval: 0-7.8. 

Question: Q6. Were runoff quality data available when developing the 
TMDL for [pollutant] in [water body]? 

Response: Yes; 
Count: 36; 
Estimated percentage: 21.6; 
95 percent confidence interval: 13.7-29.4. 

Response: No; 
Count: 110; 
Estimated percentage: 68.7; 
95 percent confidence interval: 60-77.4. 

Response: Don't know; 
Count: 25; 
Estimated percentage: 9.7; 
95 percent confidence interval: 4.8-14.6. 

Question: Q6A. If yes, how helpful was the availability of runoff 
quality data in developing the TMDL? 

Response: Very helpful; 
Count: 29; 
Estimated percentage: 95.6; 
95 percent confidence interval: 91.6-99.6. 

Response: Somewhat helpful; 
Count: 3; 
Estimated percentage: 2.2; 
95 percent confidence interval: 0-4.9. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: 2.2; 
95 percent confidence interval: 0-4.9. 

Question: Q6B. If no, how helpful would the availability of runoff 
quality data have been when developing the TMDL? 

Response: Very helpful; 
Count: 31; 
Estimated percentage: 23.8; 
95 percent confidence interval: 14.4-33.1. 

Response: Somewhat helpful; 
Count: 47; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 17; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q7. Is there anything further we should know about the development of 
the TMDL for [pollutant] in [water body]? 

Open-ended response theme: Older TMDLs are not as comprehensive or as 
accurate as newer ones; 
Number of responses: 14. 

Open-ended response theme: The state TMDL coordinator had limited 
information on the TMDL because EPA developed the TMDL; 
Number of 
responses: 12. 

Open-ended response theme: Additional data or specificity would have 
made the TMDL more implementable; 
Number of responses: 10. 

Open-ended response theme: Total coded responses; 
Number of responses: 
54. 

[End of table] 

Status of TMDL Implementation: 

Question: Q8. Was it necessary to revise National Pollutant Discharge 
Elimination System (NPDES) permits to meet the wasteload allocation in 
the TMDL for [pollutant] in [WATERBODY]? 

Response: Yes; 
Count: 36; 
Estimated percentage: 20.7; 
95 percent confidence interval: 12.7-28.7. 

Response: No; 
Count: 138; 
Estimated percentage: 79.3; 
95 percent confidence interval: 71.3-87.3. 

Question: Q9. Since EPA approved the TMDL in [year], how many of the 
NPDES permits for [pollutant] discharge that needed revision has your 
state been able to revise? 

Response: All; 
Count: 24; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: More than half; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: About half; 
Count: 1; 
Estimated percentage: 2.9; 
95 percent confidence interval: 0-8.7. 

Response: Less than half; 
Count: 3; 
Estimated percentage: 4.3; 
95 percent confidence interval: 0-10.6. 

Response: None; 
Count: 4; 
Estimated percentage: 2.9; 
95 percent confidence interval: 0-6.1. 

Question: Q10. How confident are you in the accuracy of your answer to 
question 9? 

Response: Extremely confident; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Very confident; 
Count: 11; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Moderately confident; 
Count: 2; 
Estimated percentage: 5.1; 
95 percent confidence interval: 0-14.1. 

Response: Somewhat confident; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q11. Was it necessary to install best management practices to 
meet the load allocation for [pollutant] in [WATERBODY]? 

Response: Yes; 
Count: 109; 
Estimated percentage: 66.7; 
95 percent confidence interval: 57.4-76.1. 

Response: No; 
Count: 52; 
Estimated percentage: 33.3; 
95 percent confidence interval: 23.9-42.6. 

Question: Q12. Since EPA approved the TMDL in [year], about how many of 
the landowners who needed to install best management practices for 
[pollutant] had them installed? 

Response: Not checked; 
Count: 1; 
Estimated percentage: 1.1; 
95 percent confidence interval: 0-3.3. 

Response: All; 
Count: 1; 
Estimated percentage: 1.1; 
95 percent confidence interval: 0-3.3. 

Response: More than half; 
Count: 15; 
Estimated percentage: 19.2; 
95 percent confidence interval: 8.2-30.2. 

Response: About half; 
Count: 12; 
Estimated percentage: 14.0; 
95 percent confidence interval: 4.7-23.4. 

Response: Less than half; 
Count: 50; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: None; 
Count: 17; 
Estimated percentage: 10.2; 
95 percent confidence interval: 4.1-16.4. 

Question: Q13. How confident are you in the accuracy of your answer to 
question 12? 

Response: Extremely confident; 
Count: 11; 
Estimated percentage: 9.8; 
95 percent confidence interval: 3.2-16.5. 

Response: Very confident; 
Count: 19; 
Estimated percentage: 18.3; 
95 percent confidence interval: 8.5-28.1. 

Response: Moderately confident; 
Count: 27; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat confident; 
Count: 22; 
Estimated percentage: 22.3; 
95 percent confidence interval: 11.1-33.4. 

Response: Not confident; 
Count: 18; 
Estimated percentage: 16.2; 
95 percent confidence interval: 7.8-24.6. 

Q14. Is there anything else we should know about the implementation 
status of the TMDL for [pollutant] in [water body]? 

Open-ended response theme: Limited data exist to verify 
implementation; 
Number of responses: 17. 

Open-ended response theme: No legal authority to compel polluters or 
landowners to abate nonpoint source pollution; 
Number of responses: 
14. 

Open-ended response theme: Additional data or specificity would have 
made the TMDL more implementable; 
Number of responses: 9. 

Open-ended response theme: Total coded responses; 
Number of responses: 50. 

Factors That Helped or Could Have Helped Implementation: 

Question: Q15. Since EPA approved the TMDL in [year], was your agency 
able to find enough landowners in the area covered by the TMDL who were 
willing to install best management practices on their properties? 

Response: Yes; 
Count: 31; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 30; 
Estimated percentage: 19.0; 
95 percent confidence interval: 9.9-28.1. 

Response: Don't know; 
Count: 40; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q15A. If yes, how helpful was the willingness of landowners 
to install best management practices in implementing the TMDL? 

Response: Very helpful; 
Count: 17; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable [A]. 

Response: Somewhat helpful; 
Count: 14; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q15B. If no, how helpful would the willingness of landowners 
to install best management practices have been in implementing the 
TMDL? 

Response: Very helpful; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 9; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: 5.2; 
95 percent confidence interval: 0-15.9. 

Question: Q16. Since EPA approved the TMDL in [year], were landowners 
in the area covered by the TMDL generally willing to install the types 
of best management practices that are most-effective at reducing 
[pollutant] loads? 

Response: Yes; 
Count: 35; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 14; 
Estimated percentage: 11.2; 
95 percent confidence interval: 3.2-19.3. 

Response: Don't know; 
Count: 53; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q16A. If yes, how helpful was the willingness of landowners 
to install the most-effective types of best management practices in 
implementing the TMDL? 

Response: Very helpful; 
Count: 24; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q16B. If no, how helpful would the willingness of landowners 
to install the most-effective types of best management practices have 
been in implementing the TMDL? 

Response: Very helpful; 
Count: 11; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 2; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q17. Since EPA approved the TMDL in [year], were landowners 
generally willing to install best management practices in the area 
covered by the TMDL where they could most effectively reduce 
[pollutant] loads? 

Response: Yes; 
Count: 31; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 15; 
Estimated percentage: 10.7; 
95 percent confidence interval: 2.9-18.6. 

Response: Don't know; 
Count: 56; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q17A. If yes, how helpful was the willingness of landowners 
to install best management practices in the places where they could 
most effectively reduce [pollutant] loads in implementing the TMDL? 

Response: Very helpful; 
Count: 24; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 7; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q17B. If no, how helpful would the willingness of landowners 
to install best management practices in places where they could most 
effectively reduce [pollutant] loads have been in implementing the 
TMDL? 

Response: Very helpful; 
Count: 11; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q18. Since EPA approved the TMDL in [year], was enough 
funding available to implement best management practices to reduce 
[pollutant] in [WATERBODY]? 

Response: Enough funding was available to implement best management 
practices; 
Count: 15; 
Estimated percentage: 14.1; 
95 percent confidence interval: 5.5-22.7. 

Response: Some funding was available to implement best management 
practices; 
Count: 40; 
Estimated percentage: [Empty]; 
95 percent confidence 
interval: Data not generalizable. 

Response: Not enough funding was available to implement best management 
practices; 
Count: 45; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q18A. If more than enough funding was available or enough 
funding was available to implement best management practices, how 
helpful was available funding in implementing the TMDL? 

Response: Very helpful; 
Count: 9; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 1; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q18B. If some funding was available or not enough funding was 
available, how helpful would the availability of additional funding 
have been in implementing the TMDL? 

Response: Very helpful; 
Count: 44; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 25; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 4; 
Estimated percentage: 4.1; 
95 percent confidence interval: 0-9. 

Question: Q19. Since EPA approved the TMDL in [year], was an 
implementation plan developed to reduce [pollutant] in [water body] 
that contained reasonable assurances with specific implementation 
activities and milestones? 

Response: Yes; 
Count: 43; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 48; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 13; 
Estimated percentage: 14.9; 
95 percent confidence interval: 5.1-24.7. 

Question: Q19A. If yes, how helpful was an implementation plan in 
implementing the TMDL? 

Response: Very helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 15; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 5; 
Estimated percentage: 7.4; 
95 percent confidence interval: 0-15.8. 

Question: Q19B. If no, how helpful would an implementation plan have 
been in implementing the TMDL? 

Response: Very helpful; 
Count: 16; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 7; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q20. Since EPA approved the TMDL in [year], did land in the 
area covered by the TMDL for [pollutant] in [water body] become more 
developed? 

Response: Yes; 
Count: 32; 
Estimated percentage: 32.8; 
95 percent confidence interval: 21-44.6. 

Response: No; 
Count: 48; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 25; 
Estimated percentage: 20.3; 
95 percent confidence interval: 10.3-30.2. 

Question: Q20A. If yes, how helpful was land development in 
implementing the TMDL? 

Response: Very helpful; 
Count: 4; 
Estimated percentage: 11.8; 
95 percent confidence interval: 0-23.7. 

Response: Somewhat helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 22; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q20B. If no, how helpful would land development have been in 
implementing the TMDL? 

Response: Somewhat helpful; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 35; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q21. Since EPA approved the TMDL in [year], has land in the 
area covered by the TMDL for [pollutant] in [water body] been retired? 

Response: Yes; 
Count: 16; 
Estimated percentage: 15.2; 
95 percent confidence interval: 6.6-23.8. 

Response: No; 
Count: 45; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 44; 
Estimated percentage: 38.8; 
95 percent confidence interval: 26.9-50.6. 

Question: Q21A. If yes, how helpful was land retirement in implementing 
the TMDL? 

Response: Very helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 8; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q21B. If no, how helpful would land retirement have been in 
implementing the TMDL? 

Response: Very helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 14; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q22. Does your agency administer EPA's section 319 nonpoint 
source management Program in your state? 

Response: Yes; 
Count: 93; 
Estimated percentage: 92.0; 
95 percent confidence interval: 86.8-97.2. 

Response: No; 
Count: 12; 
Estimated percentage: 8.0; 
95 percent confidence interval: 2.8-13.2. 

Question: Q22A. If yes, how helpful was your agency's administration of 
the section 319 program in implementing the TMDL for [pollutant] in 
[water body]? 

Response: Very helpful; 
Count: 34; 
Estimated percentage: 32.1; 
95 percent confidence interval: 20.3-43.9. 

Response: Somewhat helpful; 
Count: 21; 
Estimated percentage: 18.8; 
95 percent confidence interval: 9.6-28.1. 

Response: Not helpful; 
Count: 33; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q22B. If no, how helpful would administering the section 319 
program have been in implementing the TMDL for [pollutant] in [water 
body]? 

Response: Very helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q23. When awarding section 319 funding, does your state a 
give a higher priority to water bodies with TMDLs, as compared with 
water bodies without TMDLs? 

Response: Yes; 
Count: 95; 
Estimated percentage: 93.2; 
95 percent confidence interval: 88.4-98. 

Response: No; 
Count: 9; 
Estimated percentage: 6.5; 
95 percent confidence interval: 1.8-11.3. 

Response: Don't know; 
Count: 1; 
Estimated percentage: 0.2; 
95 percent confidence interval: 0-0.7. 

Question: Q23A. If yes, how helpful has giving a higher priority to 
water bodies with TMDLs, as compared with water bodies without TMDLs 
been in implementing the TMDL for [pollutant] in [water body]? 

Response: Very helpful; 
Count: 34; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 21; 
Estimated percentage: 17.7; 
95 percent confidence interval: 8.2-27.1. 

Response: Not helpful; 
Count: 39; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q23B. If no, how helpful would giving a higher priority to 
water bodies with TMDLs, as compared with water bodies without TMDLs 
have been in implementing the TMDL for [pollutant] in [water body]? 

Response: Very helpful; 
Count: 3; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

[A] In the report, we combined the responses to those questions in 
which the state TMDL coordinators responding to our survey indicated 
that the factor either was helpful (if present) or (if factor was not 
present or TMDL was not yet implemented) would have been or would be 
helpful in implementing the TMDL in question. Unless otherwise noted, 
those combined estimates had margins of error of no more than 12 
percent. When presented separately in this appendix, the questions had 
fewer responses and more often had margins of error that exceeded 12 
percent, and we do report them as not generalizable. 

[End of table] 

Q24. If not addressed by question 23, how, if at all, does your state 
prioritize funding for water bodies with TMDLs? 

Open-ended response theme: The state considers additional or other 
priorities when prioritizing funding; 
Number of responses: 22. 

Open-ended response theme: Lack of a sponsor, such as a citizen-based 
group, prevents the awarding of 319 funds; 
Number of responses: 15. 

Open-ended response theme: Funding is prioritized according to whether 
a water body is on the state's 303(d) list, rather than if there's a 
TMDL to address the listing; 
Number of responses: 10. 

Open-ended response theme: Total coded responses; 
Number of responses: 48. 

Q25. Does your state have legal authority to compel landowners to 
reduce nonpoint source [pollutant] discharge from their properties? 

Response: Yes; 
Count: 43; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 61; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q26. If you answered yes to question 25, since EPA approved the TMDL 
in [year], has your state exercised its legal authority to compel 
landowners to reduce nonpoint source pollution from their properties? 

Response: Yes; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 16; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 14; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

[End of table] 

Q26A. If yes, how many times has your state exercised its legal 
authority to compel landowners to reduce nonpoint source pollution from 
their properties in the TMDL for [pollutant] in [water body]? 

Respondents reported that they used their state legal authority 7 times 
in the long-established TMDLs in our sample. 

26b. If no, why did your state choose not to use its legal authority to 
compel landowners to reduce nonpoint source pollution from their 
properties? 

Open-ended response theme: The state prefers to use a voluntary 
approach; 
Number of responses: 15. 

Open-ended response theme: Using the state's legal authority is 
cumbersome or difficult because it requires identification of specific 
violators; 
Number of responses: 6. 

Open-ended response theme: The state's legal authority to address 
nonpoint source pollution is vague; 
Number of responses: 5. 

Open-ended response theme: Total coded responses; 
Number of responses: 29. 

[End of table] 

Q27. If local government agencies in your state have legal authority to 
compel landowners to reduce nonpoint source pollution from their 
properties, please describe this authority and how it applied to the 
TMDL for [pollutant] in [water body], if at all. 

Open-ended response theme: Livestock management act, county manure 
management laws, or local waste storage laws; 
Number of responses: 6. 

Open-ended response theme: Septic system enforcement, typically through 
county health departments; 
Number of responses: 5. 

Open-ended response theme: Wetland protection and riparian buffers; 
Number of responses: 4. 

Open-ended response theme: Total coded responses; 
Number of responses: 14. 

Q28. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has the U.S. Department of Agriculture assisted your agency in 
implementing the TMDL? 

Response: Yes; 
Count: 45; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 27; 
Estimated percentage: 21.5; 
95 percent confidence interval: 12-31. 

Response: No assistance was necessary; 
Count: 21; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 11; 
Estimated percentage: 7.9; 
95 percent confidence interval: 2.1-13.7. 

Q28A. If yes, how helpful was the assistance 
of the U.S. Department of Agriculture in implementing the TMDL? 

Response: Very helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 2; 
Estimated percentage: 6.9; 
95 percent confidence interval: 0-16.6. 

Q28B. If no, how helpful would the assistance of the U.S. Department 
of Agriculture have been in implementing the TMDL? 

Response: Very helpful; 
Count: 8; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 12; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 6; 
Estimated percentage: Empty]; 
95 percent confidence interval: Data not generalizable. 

Q29. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has the U.S. Environmental Protection Agency assisted your 
agency to implementing the TMDL? 

Response: Not checked; 
Count: 1; 
Estimated percentage: 1.0; 
95 percent confidence interval: 0-2.9. 

Response: Yes; 
Count: 43; 
Estimated percentage: 37.1; 
95 percent confidence interval: 25.6-48.6. 

Response: No; 
Count: 43; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No assistance was necessary; 
Count: 13; 
Estimated percentage: 9.3; 
95 percent confidence interval: 3.3-15.2. 

Response: Don't know; 
Count: 5; 
Estimated percentage: 5.3; 
95 percent confidence interval: 0-11.9. 

Q29A. If yes, how helpful was the assistance of the U.S. Environmental 
Protection Agency in implementing the TMDL? 

Response: Very helpful; 
Count: 28; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 13; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 2; 
Estimated percentage: 6.6; 
95 percent confidence interval: 0-15.9. 

Q29B. If no, how helpful would the assistance of the U.S. 
Environmental Protection Agency have been in implementing the TMDL? 

Response: Very helpful; 
Count: 4; 
Estimated percentage: 9.7; 
95 percent confidence interval: 0-20.6. 

Response: Somewhat helpful; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not Estimated percentage: 
generalizable. 

Response: Not helpful; 
Count: 13; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q30. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has a state government agricultural program assisted your 
agency to implementing the TMDL? 

Response: Yes; 
Count: 43; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 28; 
Estimated percentage: 17.6; 
95 percent confidence interval: 9.6-25.5. 

Response: No assistance was necessary; 
Count: 24; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 10; 
Estimated percentage: 8.5; 
95 percent confidence interval: 1.1-15.9. 

Q30A. If yes, how helpful was the assistance of a state government 
agricultural program in implementing the TMDL? 

Response: Very helpful; 
Count: 24; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 18; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: 0.6; 
95 percent confidence interval: 0-1.8. 

Q30B. If no, how helpful would the assistance of a state government 
agricultural program have been in implementing the TMDL? 

Response: Very helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 13; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q31. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has a state government stormwater program assisted your agency 
in implementing the TMDL? 

Response: Yes; 
Count: 32; 
Estimated percentage: 28.2; 
95 percent confidence interval: 18.1-38.3. 

Response: No; 
Count: 32; 
Estimated percentage: 23.7; 
95 percent confidence interval: 13.3-34.1. 

Response: No assistance was necessary; 
Count: 37; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 4; 
Estimated percentage: 4.6; 
95 percent confidence interval: 0-10.8. 

Q31A. If yes, how helpful was the assistance of a state government 
stormwater program in implementing the TMDL? 

Response: Very helpful; 
Count: 15; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 14; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not Estimated percentage: 
generalizable. 

Response: Not helpful; 
Count: 3; 
Estimated percentage: 7.7; 
95 percent confidence interval: 0-17.6. 

Q31B. If no, how helpful would the assistance of a state government 
stormwater program have been in implementing the TMDL? 

Response: Very helpful; 
Count: 1; 
Estimated percentage: 4.2; 
Number of Estimated percentage: 
responses: 0-13. 

Response: Somewhat helpful; 
Count: 8; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not Estimated percentage: 
generalizable. 

Response: Not helpful; 
Count: 22; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not Estimated percentage: 
generalizable. 

Q32. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has a state government nonpoint source program assisted your 
agency in implementing the TMDL? 

Response: Yes; 
Count: 65; 
Estimated percentage: 69.6; 
95 percent confidence interval: 58.2-81. 

Response: No; 
Count: 28; 
Estimated percentage: 22.8; 
95 percent confidence interval: 12.1-33.5. 

Response: No assistance was necessary; 
Count: 9; 
Estimated percentage: 7.0; 
95 percent confidence interval: 1.6-12.5. 

Response: Don't know; 
Count: 2; 
Estimated percentage: 0.5; 
95 percent confidence interval: 0-1.2. 

Q32A. If yes, how helpful was the assistance of a state government 
nonpoint source program in implementing the TMDL? 

Response: Very helpful; 
Count: 43; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: 2.3; 
95 percent confidence interval: 0-6.8. 

Q32B. If no, how helpful would the assistance of a state government 
nonpoint source program have been in implementing the TMDL? 

Response: Very helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 14; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q33. Since EPA approved the TMDL in [year] for [pollutant] in [water 
body], has a state or local government planning agency assisted your 
agency in implementing the TMDL? 

Response: Yes; 
Count: 34; 
Estimated percentage: 37.0; 
95 percent confidence interval: 25.1-49. 

Response: No; 
Count: 53; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 17; 
Estimated percentage: 19.5; 
95 percent confidence interval: 8.9-30.1. 

Q33A. If yes, how helpful was the assistance of a state or local 
government planning agency in implementing the TMDL? 

Response: Very helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q33B. If no, how helpful would the assistance of a state or local 
government planning agency have been in implementing the TMDL? 

Response: Very helpful; 
Count: 7; 
Estimated percentage: 14.3; 
95 percent confidence interval: 2.6-26. 

Response: Somewhat helpful; 
Count: 26; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 17; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q34. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], has a citizen-based group such as a watershed organization or 
other nongovernmental organization assisted your agency in 
implementing the TMDL? 

Response: Yes; 
Count: 51; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 34; 
Estimated percentage: 28.6; 
95 percent confidence interval: 17.8-39.4. 

Response: No assistance was necessary; 
Count: 4; 
Estimated percentage: 3.0; 
95 percent confidence interval: 0-6.7. 

Response: Don't know; 
Count: 15; 
Estimated percentage: 14.6; 
95 percent confidence interval: 6.6-22.6. 

Q34A. If yes, how helpful was the assistance of a citizen-based group 
in implementing the TMDL? 

Response: Very helpful; 
Count: 34; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 15; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not Estimated percentage: 
generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: 0.5; 
95 percent confidence interval: 0-1.5. 

Q34B. If no, how helpful would the assistance of a citizen-based group 
have been in implementing the TMDL? 

Response: Very helpful; 
Count: 10; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 16; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q35. Since EPA approved the TMDL for [pollutant] in [water body] in 
[year], have local governmental leaders generally been supportive of 
implementing the TMDL? 

Response: Yes; 
Count: 37; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 20; 
Estimated percentage: 18.7; 
95 percent confidence interval: 8.6-28.8. 

Response: Don't know; 
Count: 47; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q35A. If yes, how helpful was the support of local governmental 
leaders in implementing the TMDL? 

Response: Very helpful; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 11; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 1; 
Estimated percentage: 0.8; 
95 percent confidence interval: 0-2.3. 

Q35B. If no, how helpful would the support of local governmental 
leaders have been in implementing the TMDL? 

Response: Very helpful; 
Count: 7; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Somewhat helpful; 
Count: 6; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Not helpful; 
Count: 5; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

[End of table] 

Q36. Is there anything else we should know about the factors that 
helped or could have helped in implementing the TMDL for [pollutant] in 
[water body]? 

Open-ended response theme: The availability of funding was an obstacle 
to implementing the TMDL; 
Number of responses: 18. 

Open-ended response theme: The TMDL was not a priority for 
implementation; 
Number of responses: 10. 

Open-ended response theme: More public outreach is needed to implement 
the TMDL successfully; 
Number of responses: 9. 

Open-ended response theme: Total coded responses; 
Number of responses: 44. 

[End of table] 

Pollutant Loads from Point Sources: 

Question: Q37. Since EPA approved the TMDL for [pollutant] in [water 
body] in [year], has the pollutant load from point sources decreased, 
stayed the same, or increased? 

Response: Decreased; 
Count: 18; 
Estimated percentage: 22.3; 
95 percent confidence interval: 11.6-32.9. 

Response: Stayed the same; 
Count: 19; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Don't know; 
Count: 23; 
Estimated percentage: 24.7; 
95 percent confidence interval: 14.1-35.2. 

Response: There were no point sources; 
Count: 19; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q38. Since EPA approved the TMDL for [pollutant] in [water 
body] in [year], has the wasteload allocation for point sources been 
met? 

Response: Yes; 
Count: 31; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 7; 
Estimated percentage: 10.9; 
95 percent confidence interval: 2.5-19.2. 

Response: Don't know; 
Count: 21; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: There was no wasteload allocation for point sources in the 
TMDL; 
Count: 1; 
Estimated percentage: 0.4; 
95 percent confidence interval: 0-1.3. 

[End of table] 

Q39. What is the basis for your response in question 38? 

Question: Q39A. Water quality monitoring; 

Response: Yes; 
Count: 18; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 16; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39B. Visual inspection; 
Response: Yes; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 27; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39C. Revised NPDES permits; 

Response: Yes; 
Count: 17; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 19; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39D. Model estimates; 

Response: Yes; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 25; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39E. Land use changes; 

Response: Yes; 
Count: 4; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 26; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39F. Discharge changes; 

Response: Yes; 
Count: 15; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 19; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q39G. Other. Please specify below; 

Response: Yes; 
Count: 28; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 7; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

[End of table] 

If you checked "Other" above, what was the other source? 

Open-ended response theme: Discharge monitoring reports; 
Number of responses: 11. 

Open-ended response theme: Geographic information system data; 
Number of responses: 3. 

Open-ended response theme: Total coded responses; 
Number of responses: 15. 

[End of table] 

Question: Q40. Within the next 5 years, do you expect the pollutant 
load to decrease, stay the same, or increase as a result of actions 
that have already been taken for point sources in the TMDL for 
[pollutant] in [water body]? 

Response: Decrease; 
Count: 15; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Stay the same; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: Increase; 
Count: 1; 
Estimated percentage: 1.7; 
95 percent confidence interval: 0-5.2. 

Response: Don't know; 
Count: 20; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q41. Within the next 5 years, do you expect the wasteload 
allocation for [pollutant] in [water body] to be met? Response: Yes; 

Count: 31; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 5; 
Estimated percentage: 7.4; 
95 percent confidence interval: 0.5-14.4. 

Response: Don't know; 
Count: 23; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Q42. Is there anything else that you think we should know about point 
source pollutant loads in the TMDL for [pollutant] in [water body]? 

State TMDL coordinators offered no substantive responses to this 
question. 

[End of table] 

Pollutant Loads from Nonpoint Sources: 

Question: Q43. Since EPA approved the TMDL for [pollutant] in [water 
body] in [year], has the pollutant load from nonpoint sources 
decreased, stayed the same, or increased? 

Response: Decreased; 
Count: 42; 
Estimated percentage: 32.4; 
95 percent confidence interval: 22.7-42.2. 

Response: Stayed the same; 
Count: 38; 
Estimated percentage: 25.4; 
95 percent confidence interval: 16.5-34.2. 

Response: Increased; 
Count: 1; 
Estimated percentage: 0.6; 
95 percent confidence interval: 0-1.9. 

Response: Don't know; 
Count: 67; 
Estimated percentage: 34.7; 
95 percent confidence interval: 25.6-43.8. 

Response: There were no nonpoint sources; 
Count: 11; 
Estimated percentage: 6.8; 
95 percent confidence interval: 2.4-11.3. 

Question: Q44. Since EPA approved the TMDL for [pollutant] in [water 
body] in [year], has the load allocation for nonpoint sources been 
met? 

Response: Yes; 
Count: 15; 
Estimated percentage: 9.1; 
95 percent confidence interval: 3.9-14.3. 

Response: No; 
Count: 51; 
Estimated percentage: 36.6; 
95 percent confidence interval: 26.3-47. 

Response: Don't know; 
Count: 71; 
Estimated percentage: 48.2; 
95 percent confidence interval: 37.7-58.7. 

Response: There was no load allocation for nonpoint sources in the 
TMDL; 
Count: 9; 
Estimated percentage: 6.0; 
95 percent confidence interval: 1.6-10.5. 

[End of table] 

Q45. What is the basis for your response in 44? 

Question: Q45A. Water quality monitoring; 

Response: Yes; 
Count: 65; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 35; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q45B. Visual inspection; 
Response: Yes; 
Count: 33; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 49; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q45C. Implemented best management practices; 

Response: Yes; 
Count: 52; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 39; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q45D. Model estimates; 

Response: Yes; 
Count: 16; 
Estimated percentage: 15.9; 
95 percent confidence interval: 5.3-26.4. 

Response: No; 
Count: 63; 
Estimated percentage: 84.1; 
95 percent confidence interval: 73.6-94.7. 

Question: Q45E. Land use changes; 

Response: Yes; 
Count: 28; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 56; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Question: Q45F. Discharge changes; 

Response: Yes; 
Count: 12; 
Estimated percentage: 14.5; 
95 percent confidence interval: 5.8-23.2. 

Response: No; 
Count: 68; 
Estimated percentage: 85.5; 
95 percent confidence interval: 76.8-94.2. 

Question: Q45G. Other. Please specify below; 

Response: Yes; 
Count: 28; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

Response: No; 
Count: 51; 
Estimated percentage: [Empty]; 
95 percent confidence interval: Data not generalizable. 

[End of table] 

If you checked "Other" above, what was the other source? 

State TMDL coordinators offered no substantive responses to this 
question. 

Question: Q46. Within the next 5 years, do you expect that the 
pollutant load for nonpoint sources will decrease, stay the same, or 
increase as a result of actions that have already been taken for 
nonpoint sources in the TMDL for [pollutant] in [water body]? 

Response: Decrease; 
Count: 41; 
Estimated percentage: 39.9; 
95 percent confidence interval: 28.8-50.9. 

Response: Stay the same; 
Count: 45; 
Estimated percentage: 30.1; 
95 percent confidence interval: 20.4-39.8. 

Response: Increase; 
Count: 2; 
Estimated percentage: 0.9; 
95 percent confidence interval: 0-2.4. 

Response: Don't know; 
Count: 50; 
Estimated percentage: 29.1; 
95 percent confidence interval: 19.7-38.5. 

Question: Q47. Within the next 5 years, do you expect the load 
allocation for [pollutant] in [water body] to be met? 

Response: Yes; 
Count: 18; 
Estimated percentage: 13.0; 
95 percent confidence interval: 5.7-20.4. 

Response: No; 
Count: 43; 
Estimated percentage: 30.1; 
95 percent confidence interval: 20-40.2. 

Response: Don't know; 
Count: 77; 
Estimated percentage: 56.9; 
95 percent confidence interval: 46-67.7. 

[End of table] 

Q48. Is there anything else that you think we should know about 
nonpoint source pollutant loads in the TMDL for [pollutant] in [water 
body]? 

Open-ended response theme: No changes were expected in the water body; 
Number of responses: 16. 

Open-ended response theme: It will take more time to see changes in 
water quality; 
Number of responses: 11. 

Open-ended response theme: The TMDL was not a priority for 
implementation; 
Number of responses: 10. 

Open-ended response theme: Total coded responses; 
Number of responses: 49. 

[End of table] 

Additional Comments: 

Q49. Thank you for answering these questions. Please use the following 
space to share with us any additional thoughts you have about the 
challenges and opportunities that face your state's TMDL program. 

Open-ended response theme: The availability of funding was an obstacle 
to implementing the TMDL; 
Number of responses: 42. 

Open-ended response theme: No legal authority to compel polluters or 
landowners to abate nonpoint source pollution; 
Number of responses: 25. 

Open-ended response theme: A lack of stakeholders prevents the TMDL 
being implemented; 
Number of responses: 21. 

Open-ended response theme: Total coded responses; 
Number of responses: 67. 

[End of table] 

[End of section] 

Appendix V: Comments from the Environmental Protection Agency: 

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 the Government 
Accountability Office's draft report GAO·14-80, "Changes Needed if Key 
EPA Program is to Help Fulfill the Nation's Water Quality Goals." The 
purpose of this letter is to provide the Environmental Protection 
Agency' s response to your findings, conclusions and recommendations. 

The EPA generally agrees with the GAO's findings and partially agrees 
with your conclusions and recommendations. To a great extent, we agree 
with the report's characterization of the challenges facing the EPA's 
water quality-based programs (i.e., the much broader context within 
which the TMDL program operates) under the Clean Water Act. The GAO 
thoroughly analyzed and understands the abundance and complexity of 
impaired waters across the Nation, the limited funding and authorities 
for the CWA programs of the EPA and our state partners, and the 
limited achievement of CWA goals to date. Further, we agree in 
principle on the kinds of changes that would be able to increase CWA
success. We wish to use this response letter to help clarify, to the 
extent possible, how the scope of the EPA's TMDL program figures into 
GAO's otherwise well-compiled findings, conclusions and 
recommendations. We provide our primary responses within this letter 
and offer technical details via the enclosed draft document file copy 
with embedded comments. In addition, we respond to the report's
individual recommendations below. 

GAO Review Findings Are Consistent with the EPA Understanding of Water 
Quality Issues and Needs: 

The draft report adequately recognizes the magnitude of the impairment 
problems addressed by the EPA and the states, and the general status 
of nationwide water quality. The description of the broader context
of CW A programs, interacting programmatic components, and the EPA and 
states' respective roles is generally accurate. The EPA commends the 
GAO's substantial investment of effort in designing and carrying out a 
statistical survey of state water quality managers that has 
contributed new empirical data and insights into state TMDL 
development and implementation, including some findings on impaired
waters restoration status and outcomes. 

On the whole, the GAO's analysis identified key issues and driving 
factors as well as logical options for solutions. Although these 
issues include long-term challenges that remain unresolved, they have 
been and continue to be addressed as high priorities by our programs 
within the limits of our resources and statutory authorities. 
Specifically, the GAO found that whereas in a high proportion (83%) of 
TMDLs the required point source load reduction is accomplished, a 
minority (20%) of TMDLs needing nonpoint load reduction achieve this 
goal. These findings are consistent with our experience that the 
regulatory permitting approach addressing point sources is more 
effective at meeting water quality targets than the voluntary nonpoint 
source approach - even though over 70% of the EPA's Section 319 nonpoint
pollution control projects achieve all objectives. Further, the draft 
opines that achievement of the CWA goals might be improved by more 
comprehensive assessment; better causal association of stressor
reduction and designated use support; greater implementation, 
particularly involving landowner participation; more post-TMDL 
monitoring and adaptive revision where needed; higher funding 
commensurate with restoration needs; and consideration of regulatory, 
guidance, or statutory changes as they relate to accomplishing these 
needed improvements. The EPA agrees in general with these potential
options for action, but many require actions by Congress or other 
entities other than the EPA. 

GAO Recommendations and EPA Responses: 

To enhance the likelihood that the TMDL program will enable the 
nation's waters to attain water quality standards and to strengthen 
water quality management; the GAO recommended that the Administrator 
of the EPA takes the following four actions: 

Recommendation 1: Develop and issue new regulations requiring that 
TMDLs include additional elements--and considering the elements that 
are now optional--specifically, elements reflecting key features 
identified by the National Research Council as necessary for attaining 
water quality standards, such as comprehensive identification of 
impairment and plans to monitor water bodies to verify that water 
quality is improving. 

EPA Response: The EPA agrees that broader implementation and 
monitoring requirements would be helpful and, in fact, attempted to 
revise the TMDL regulations to include such strengthening provisions
in 2000. Since Congress prevented those regulations from going into 
effect, this recommendation might be better directed to Congress than 
to the EPA. 

To help address this recommendation, the EPA plans to continue its 
long-standing promotion of comprehensive watershed management well 
beyond just the TMDL program. Within the bounds of 40 CFR 130.7 as 
currently written, implementation is not a required component for TMDL 
approval. However, the State Water Quality Management Plan (WQMP), 
required at 40 CFR 130.6, is the existing mechanism providing the 
authority for the State to implement TMDLs. 40 CFR 130.6(b) and
130.6(c)(1), read together, clearly establish that TMDLs are 
"elements" of the Water Quality Management (WQM) Plan and that "WQM 
plans are used to direct implementation." TMDLs should be the piece of 
these watershed management plans that identifies the load reductions 
necessary to achieve water quality standards. The watershed management 
plan should manage point and nonpoint control implementation, identify 
controls such as BMPs, develop a schedule for implementation of 
controls, water quality assessment monitoring, and post-TMDL 
monitoring of the effectiveness of point and nonpoint source controls, 
and adjust the implementation plan and/or TMDL accordingly (adaptive
implementation). 

Recommendation 2: To ensure more consistent application of existing 
TMDL elements and to provide greater assurance that TMDLs, if 
implemented, can achieve tangible water quality results, identify 
regional offices with criteria for interpreting and applying such 
elements in reviewing and approving state~developed TMDLs and issue 
guidance with more specificity, directing all regional offices to 
fOllow the same criteria, including requesting states provide 
more~detailed information about pollution causes and abatement actions. 

EPA Response: The EPA generally agrees with this recommendation. The 
EPA appreciates the GAO's effort to interact with all the EPA Regions 
and compile their findings on a national, comprehensive basis that 
reveals consistencies as well as differences. We see national 
coordination and guidance on procedures for the EPA Regional TMDL 
review and approval as an ongoing responsibility. Informally, we 
consistently maintain dialogue across the Regional TMDL programs via 
monthly conference calls that frequently include specific case studies 
or TMDL approval issues in practice. In recent years, our approach to 
Regional coordination on TMDL guidance has focused on clarifying 
specific elements as needs arise, rather than an overall relook. For 
example, recently the EPA Headquarters has invested substantially in 
analyzing issues pertaining to reasonable assurance and coordination 
with the EPA Regions on their ramifications for state partnerships and 
TMDL approvals. We will plan to initiate a dialogue with our Regional 
offices on overall TMDL review and approval guidance during 2014, and
this dialogue will be able to consider the recommended changes in 
specificity and content raised by the GAO. 

Recommendation 3: Place conditions on states' annual use of nonpoint 
source management and water pollution control grants to ensure that 
the funds meet the purposes for which they are awarded and achieve 
greater reductions in nonpoint source pollution associated with TMDL 
implementation, such as by targeting funds to states and projects that 
incorporate factors needed for effective TMDL implementation (e.g., 
targeting grant funds to projects where implementation plans have been
developed and where external agency assistance is available). 

EPA Response: We agree with this recommendation and are doing this 
already in the Section 319 program by requiring nine~element watershed 
plans before 319 funds are spent on watershed restoration projects, 
including those that implement TMDLs. We will share this 
recommendation with other sources of funding, including geographic 
programs such as the Great Lakes Restoration Initiative and the
Chesapeake Bay Program where much of the program funding resides. 

The TMDL program agrees to review the possible uses of Section 106 
funding to address this recommendation. Section 106 funds frequently 
support the development of TMDLs, and thus may have some capability to 
increase the detail of within-TMDL implementation sections and/or 
plans. We note, however, that although this Section might be able to 
increase implementation planning, it is not appropriately used to 
support on-the-ground implementation activities. 

Both implementation planning and on-the-ground support are within the 
scope of Section 319 funding. In April 2013 Section 319 guidance was 
reissued by the EPA's Non-Point Source Control Program, with several 
modifications that strengthen TMDL-related implementation and on-the-
ground control actions. Among the provisions in the new guidance were 
several elements related to the GAO's recommendation, including in 
particular "an increased emphasis on watershed project implementation in
watersheds with impaired waters, and greater accountability measures." 
More emphasis is placed on watershed implementation by requiring a 
full 50 percent of 319 funds to go for that purpose. The guidance also 
requires supplemental information for TMDLs developed with 319 funds, 
including detailed identification of causes/sources, existing loads 
and load reductions needed by source, and management measures to 
achieve load reduction. These improvements are already in the form of a
Section 319 grant condition, not a new TMDL element, and are 
structured to make TMDLs more useful to watershed based plans and vice 
versa. In summary, we believe that the recently revised Section 319
guidance substantially addresses much of what this GAO recommendation 
seeks to accomplish, and further revision may not be value-added under 
the current funding picture. 

Recommendation 4: Obtain missing data that currently impede EPA's 
efforts to determine whether and to what extent TMDLs have been 
implemented or to what extent implemented TMDLs have helped impaired 
waters attain water quality standards by: 

* directing states to use and report specific GIS data when 
implementing projects to which TMDLs apply and, 

* requesting that USDA ask landowners who participate in conservation 
programs funded by the department in areas subject to a TMDL to 
disclose information on the location, type, and number of projects 
implemented under these programs. 

EPA Response: The EPA generally agrees with the first part of this 
recommendation and suggests that the GAO direct its second part to the 
USDA. Understanding TMDL implementation rates and patterns and their 
related driving factors is at once essential to effective program 
management while also cost-prohibitively complex to accomplish fully 
and nationally. The practical question is, what specifics of
implementation tracking and data are feasible and essential to program 
success? The TMDL program has an established track record of nearly 
ten years of analyzing implementation, within our limited
resources and against a backdrop of continually changing 
implementation activity among a wide slate of restoration 
participants, and with no EPA authority to centrally track all these 
actions. The TMDL program earlier estimated an annual cost of over $10 
million to attempt to track all implementation actions (likely with 
limited success) for all TMDLs when about 30,000 TMDLs existed - not 
to mention the higher costs of relating these actions definitively to 
all water quality effects. We emphasize that a complete inventory of 
implementation actions and their influence on impaired waters 
recoveries will remain infeasible as a course of action, but share the 
GAO's keen interest in obtaining key data and developing more 
insightful analyses. 

We agree that, due to its gradual but continued emergence as a 
routinely available element in environmental programs, geospatial data 
provides a highly valuable basis for data integration from multiple 
information sources. Although far from enabling complete inventory, 
geospatial data on NPDES permits and Section 319 projects enabled the 
EPA to summarize some basic implementation activities for thousands of 
TMDLs nationwide in 2011. Permits and Section 319 tracking continues to
accumulate georeferenced information for newer permits and projects. 
In 2013, the EPA is actively engaging with states on a geospatial data 
requirements and applications study that seeks to exploit greater 
efficiencies in impaired waters and TMDL reporting and data 
management. Although we are uncertain as to the breadth of project 
types for which state implementation reporting could be required,
we are actively engaging with states to develop a productive consensus 
on geospatial data requirements and uses. We expect this project and a 
broader effort on an improved national EPA geospatial framework, 
including a significant investment of effort in catchment-level data 
management, to remain a priority activity during 2014 and beyond. 

Regarding the second element of the GAO's recommendation, we agree 
that location and type of USDA pollution control projects would be 
useful data for tracking implementation and evaluating water quality
effects, and urge GAO to direct this recommendation to USDA. As 
currently written, the recommendation implies that the EPA would use 
the USDA data to assess individual actions to implement individual 
TMDLs. Instead, we envision using the EPA's state oversight and 
tracking to assess states' collective progress in TMDL implementation, 
which may in tum reflect data shared with the state by USDA through 
conservation cooperator agreements. 

Again, thank you for the opportunity to comment. The EPA expresses its 
appreciation for the thoughtful work of the GAO staff members 
throughout this review and their engagement with the EPA headquarters, 
the EPA Regional, and State water program staff. 

Sincerely, 

Signed by: 

Nancy K. Stoner: 
Acting Assistant Administrator: 

Enclosure: 

cc: EPA GAO Liaison Team. 

[End of section] 

Appendix VI: 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, Susan Iott (Assistant 
Director), Nathan Anderson, Elizabeth Beardsley, Mark Braza, Ellen W. 
Chu, Catherine Hurley, Justin Fisher, Paul Kazemersky, Katherine 
Killebrew, Dan C. Royer, Kelly Rubin, and Kiki Theodoropoulos made key 
contributions to this report. 

[End of section] 

Footnotes: 

[1] The Federal Water Pollution Control Act Amendments of 1972, Pub. 
L. No. 92-500, § 2, 86 Stat. 816, codified as amended at 33 U.S.C. §§ 
1251-1387 (2013) (commonly referred to as the Clean Water Act). For 
consistency throughout this report, we refer to the statute and its 
amendments as the Clean Water Act. 

[2] EPA, "National Summary of State Information" in Water Quality 
Assessment and TMDL Information, accessed August 19, 2013, [hyperlink, 
http://ofmpub.epa.gov/waters10/attains_nation_cy.control]. 

[3] Under the Clean Water Act, states are to report every 2 years to 
EPA on the quality of the water bodies within their boundaries. In 
2012, states assessed about 18 million lake acres, or 43 percent of 
lake acres nationwide, and almost 1 million river miles, or 28 percent 
of river miles nationwide. 

[4] The Clean Water Act does not expressly require implementation of a 
TMDL, but states are required to engage in a continuous planning 
process that includes TMDLs, among other things, and is to achieve 
water quality standards. 33 U.S.C. § 1313(e) (2013). 

[5] A point source discharges pollutants from a discrete point, such 
as a pipe carrying effluent from a wastewater treatment plant or an 
industrial facility. Under Clean Water Act regulations, limits in 
TMDLs for pollutants from point sources are to be implemented through 
National Pollutant Discharge Elimination System permits. Under the 
Clean Water Act, the discharge of pollutants from a point source into 
U.S. waters is prohibited without a permit; facilities obtain permits 
from authorized states or else from the applicable EPA region. If 
TMDLs have been developed for these waters in question, the permitted 
limits for point sources must be consistent with the pollutant levels 
called for in the TMDLs. 

[6] Consent decrees are settlement agreements signed by the parties 
and entered, or approved, by a court; they are therefore enforceable 
by the courts. 

[7] EPA, Report of the Federal Advisory Committee on the Total Maximum 
Daily Load (TMDL) Program, EPA-100-R-98-006 (Washington, D.C.: July 
1998). 

[8] National Research Council, Assessing the TMDL Approach to Water 
Quality Management (Washington, D.C.: National Academies Press, 2001). 
The National Academies comprises four organizations: the National 
Academy of Sciences, National Academy of Engineering, Institute of 
Medicine, and National Research Council. 

[9] EPA, Office of Inspector General, Total Maximum Daily Load Program 
Needs Better Data and Measures to Demonstrate Environmental Results, 
Report No. 2007-P-00036 (Washington, D.C.: Sept. 19, 2007). 

[10] EPA, Developing Effective Nonpoint Source TMDLs: An Evaluation of 
the TMDL Development Process (Washington, D.C.: January 2007); EPA, 
Developing Effective TMDLs: An Evaluation of the TMDL Process 
(Washington, D.C.: 2007); and Center for TMDL and Watershed Studies, 
TMDL Implementation: Characteristics of Successful Projects, Final 
Report, VT-BSE doc. no. 2006-0003 (Blacksburg, VA: 2006). 

[11] Projects described in TMDLs may include implementing best 
management practices on agricultural land to reduce nutrient-laden 
runoff, such as planting riparian buffers, which are vegetative areas 
along stream banks that can absorb excess nutrients before they enter 
water bodies. Activities described in TMDLs may include efforts such 
as nutrient management planning, which describes a process to develop 
a coordinated combination of projects that help farmers manage the 
amount, form, placement, and timing of fertilizer to support crop 
production while also minimizing polluted runoff. 

[12] Because the states we visited made up a nonprobability sample, 
the information we collected from those states is not generalizable to 
all states but provides examples of how the selected states developed 
and implemented TMDLs. 

[13] These key features were identified in the 2001 NRC report on 
EPA's TMDL program and in various EPA studies and guidance documents. 

[14] The population of long-established TMDLs is 34,794. 

[15] Survey results based on probability samples are subject to 
sampling error. The sample we drew for our survey is only one of a 
large number of samples we might have drawn. Because different samples 
could have provided different estimates, we express our confidence in 
the precision of our particular sample results as a 95 percent 
confidence interval. This interval would contain the actual population 
values for 95 percent of the samples we could have drawn. As a result, 
we are 95 percent confident that each of the confidence intervals in 
this report will include the true values in the target population, 
unless otherwise noted. 

[16] Section 502(6) of the Clean Water Act defines the term pollutant 
as "dredged spoil, solid waste, incinerator residue, sewage, garbage, 
sewage sludge, munitions, chemical wastes, biological materials, 
radioactive materials, heat, wrecked or discarded equipment, rock, 
sand, cellar dirt and industrial, municipal, and agricultural waste 
discharged into water." Section 502(19) of the act defines pollution 
as "the man-made or man-induced alteration of the chemical, physical, 
biological, and radiological integrity of water." According to the 
2001 NRC report on EPA's TMDL program, pollution goes beyond 
pollutants; it includes pollutants and other stressors, such as 
habitat degradation, flow alteration, channelization, and loss of 
riparian areas. 

[17] According to EPA documents and officials, the triennial review is 
to consider all the scientific information EPA has issued for specific 
pollutants since the state last updated its standards, adding or 
revising standards as necessary. EPA is to review new or revised water 
quality standards consistent with Clean Water Act requirements in 
connection with its approval role. EPA can make a determination that 
new or revised water quality standards are necessary to meet the act's 
requirements. If EPA makes such a determination and the state fails to 
adopt such standards, EPA must promulgate the standards for the state. 

[18] EPA's regulations provide that the minimum requirements for state 
water quality standards also include an antidegradation policy to 
maintain and protect the existing uses of water bodies. 

[19] Water quality criteria are constituent concentrations, levels, or 
narrative statements representing a quality of water that supports a 
particular use. EPA-recommended criteria are to reflect the latest 
scientific information on the effect of the constituent on health and 
welfare, including aquatic life. States must use EPA-recommended 
values or other scientifically defensible methods for numeric 
criteria. Where numeric criteria cannot be established, states may 
establish narrative criteria. 

[20] According to EPA documents, to supplement these three data types, 
states may also use habitat assessments and toxicity analyses. Habitat 
assessments include descriptions of sites and surrounding land uses; 
condition of streamside vegetation; and measurement of features such 
as stream width, depth, flow, and substrate (bottom). States can use 
habitat assessments to supplement and interpret other kinds of data. 
States can generate toxicity analyses by exposing selected organisms, 
such as fathead minnows or daphnia (water fleas), to known dilutions 
of water taken from a sampling location. These tests can help 
determine whether poor water quality stems from toxins or another 
cause, such as degraded habitat. 

[21] Under the act, states are required to monitor and assess waters 
under section 305(b) of the Clean Water Act and to identify impaired 
waters under section 303(d). In 2002, EPA recommended that states 
submit "integrated reports," which include information gathered under 
each of these provisions. According to EPA, integrated reports result 
in a more comprehensive and consistent description of states' waters, 
including impaired waters. Not all states submit integrated reports, 
however, so the data EPA receives from all states are not uniform, 
according to EPA officials. 

[22] According to EPA guidance, alternative measures are reviewed by 
EPA on a case-by-case basis and may include (1) permits that are 
stringent enough to meet water quality standards by the end of permit 
terms in water bodies impaired solely by point sources and (2) state 
regulations requiring nonpoint sources to implement certain projects, 
along with a demonstration that such projects will result in meeting 
water quality standards in a reasonable time. 

[23] The measures used by states and EPA for impaired water bodies and 
for water bodies attaining designated uses differ by water body type: 
specifically, acres of impaired lakes and miles of impaired rivers but 
numbers of water bodies attaining designated uses. 

[24] Water Quality Act of 1987, Pub. L. No. 100-4 § 316 (1987), adding 
§ 319 to the Clean Water Act, codified at 33 U.S.C. § 1329 (2013). 

[25] The TMDL Federal Advisory Committee comprised 20 members with 
diverse geographic, policy, and professional perspectives, including 
state and local governments, tribal organizations, environmental 
groups, industry, agriculture, forestry, and academia. In addition, 
three federal agency officials--from EPA and the Department of 
Agriculture's Natural Resources Conservation Service and Forest 
Service--served as nonvoting members. 

[26] EPA issued the final rule but delayed its effective date until 
after the spending prohibition would end. 

[27] According to the Congressional Research Service, EPA's proposal 
had few strong supporters, for varied reasons. States, which would 
have been directly affected by the rule, criticized the burdens that 
new requirements would place on them. They were concerned that they 
did not have resources to meet tight deadlines for developing and 
implementing TMDLs. Further, states said that TMDLs should not 
necessarily take higher priority over other elements of existing water 
quality management programs. Industry groups expressed concern about 
impacts of new pollution control requirements. For their part, 
municipal and industrial point source groups urged states and EPA to 
ensure that TMDL requirements did not fall disproportionately on their 
discharges, while possibly failing to address nonpoint source 
contributions to impaired waters. Farm groups and others with nonpoint 
discharges questioned EPA's authority to include nonpoint source 
pollution in the TMDL program. The forestry industry vigorously 
criticized potential impacts of the rule. Environmentalists, who 
supported the need for a stronger and more comprehensive TMDL program, 
objected to the long periods in the rule before water quality 
improvements were likely to occur. They criticized the lack of 
aggressive implementation of a program that has existed in law since 
1972. Congressional Research Service, Clean Water Act and Total 
Maximum Daily Loads (TMDLs) of Pollutants, Order Code 97-831 
(Washington, D.C.: August 2008). 

[28] According to EPA, "it is the responsibility of the states, 
through the exercise of state law, to implement TMDLs that establish 
load allocations for nonpoint sources." See Brief for Defendant-
Appellants at 9, Sierra Club v. Meiburg, 296 F.3d 1021 (11th Cir. 
2002) (No. 01-14587-G). 

[29] Nonpoint source grant funding under section 319 is allocated to 
states using a formula established by EPA, which incorporates states' 
population, cropland, pasture, forest harvest acreage, and other 
factors, set when the grant program began. 

[30] For example, according to EPA documents, ecological indicators 
measure natural watershed features, such as the relative abundance of 
aquatic animals or plants, and the tool assigns an "ecological index 
score" that reflects the watershed's restoration potential. Further, a 
"social context score" includes factors such as community involvement, 
local regulations, and whether any plans to improve water quality have 
already been developed, which often strongly influences how difficult 
restoration may be. 

[31] See, e.g., Sierra Club v. Meiburg, 296 F.3d 1021, 1031 (11th Cir. 
2002) (explaining that "… the national policy and objectives relating 
to clean water are most reliably embodied in the act itself which puts 
the responsibility for implementation of TMDLs on the states"). As we 
previously noted, the act does not expressly require implementation of 
a TMDL, but states are required to engage in a continuous planning 
process which includes TMDLs, among other things, and which will 
achieve water quality standards. With respect to state implementation 
of nonpoint source controls, the act "provides no direct mechanism to 
control nonpoint source pollution but rather uses the 'threat and 
promise' of federal grants to the states to accomplish this task." 
Pronsolino v. Nastri, 291 F.3d 1123, 1126-27 (9th Cir. 2002). 

[32] Forty-six states are approved by EPA to issue permits setting 
pollutant limits for point sources. For the four states that do not 
have this approval--Idaho, Massachusetts, New Hampshire, and New 
Mexico--EPA is responsible for incorporating TMDL limits for point 
sources into their permits. 

[33] States develop their own project selection processes and the 
criteria their nonpoint source management programs are to consider 
when determining what projects to fund. Annually, each state submits 
its list of selected projects to the applicable EPA regional office 
for incorporation into the state's work plan, which describes what 
projects will be funded through an EPA program established under 
section 319 of the Clean Water Act. Organizations that apply for 
funding through this program--often including conservation districts, 
local governments, and nonprofit organizations--submit project 
proposals to states' nonpoint source management programs and, if 
selected, are responsible for implementing their proposed nonpoint 
source pollution projects under an agreement with the state. 

[34] "State Implementation Information," compiled by Virginia 
Polytechnic Institute and State University under EPA grant no. 
83156301 (September 2008). The states include Arizona, California, 
Delaware, Hawaii, Idaho, Kansas, Minnesota, Oregon, Texas, and 
Virginia. 

[35] These states include California, Florida, Hawaii, Oregon, 
Pennsylvania, Washington, and Wisconsin. 

[36] Specifically, agricultural plowing or tilling activities and 
animal heavy use areas disturbing over 5,000 square feet are required 
to have a written erosion and sedimentation control plan, and to 
implement best management practices. 

[37] Some TMDLs in the database did not have a TMDL type listed, so 
the percentages do not add to 100. 

[38] These data do not necessarily mean that states with the most 
TMDLs have the most impaired waters. Pennsylvania's water body 
assessment units are relatively small, meaning that one stream may 
have dozens of assessment units and, if impaired, require dozens of 
TMDLs. In addition, New Hampshire was credited with thousands of TMDLs 
for completing one TMDL document--the New England Mercury TMDL--that 
applied to all water bodies in the state. 

[39] Atmospheric deposition--a process that transfers pollutants from 
the air to the earth's surface--can significantly impair water quality 
in the nation's rivers, lakes, bays, and estuaries and harm human 
health and aquatic ecosystems. In particular, airborne pollutants can 
fall to the ground in precipitation or as a gas or particle and be 
deposited either directly onto the surface of a water body or onto 
land and then transported into a water body through runoff. 

[40] EPA and the states are making a transition from the national 
database known as the Permit Compliance System to a database called 
Integrated Compliance Information System: NPDES. The states are 
divided in their use of the two databases. 

[41] EPA, FY2010 National Report on Implementing Total Maximum Daily 
Loads (TMDLs), Document No. EPA841-R-11-002 (Washington, D.C.: March 
2011). 

[42] EPA, Analysis of TMDL Implementation Rates in EPA Region 5, 
EPA841-R-09-005 (Washington, D.C.: Dec. 1, 2009). 

[43] EPA, Developing Effective Nonpoint Source TMDLs: An Evaluation of 
the TMDL Development Process. (Washington, D.C.: January 2007). 

[44] John Hoornbeek et al., Implementing Total Maximum Daily Loads: 
Understanding and Fostering Successful Results (Kent, Ohio: Kent State 
University, December 2008). 

[45] In the watersheds where water quality improved, it was unclear 
whether the TMDLs caused these incremental improvements or whether the 
improvements would have occurred even in the absence of the TMDLs. 

[46] GISs manipulate, analyze, and graphically present an array of 
information associated with geographic locations. 

[47] Pub. L. No. 110-246 § 1619(b)(2), codified at 7 U.S.C. § 
8791(b)(2) (2013). 

[48] The provision has a limited exception for aggregated data and 
certain statistical information. 

[49] In some cases, USDA has aggregated and shared with states data 
for areas that cover, on average, 1,500 square miles, according to EPA 
officials. 

[50] For certain survey questions, we received fewer than 100 
responses. Where the margin of error for such responses is greater 
than 12 percent, we list the 95 confidence interval as (X, Y). For 
this estimate, the 95 percent confidence interval is (21, 50). 

[51] The 95 percent confidence interval for this estimate is (70, 96). 

[52] NRC and EPA identified key features in addition to the ones we 
discuss (see appendix III). 

[53] For example, a water body may no longer support native fish 
populations because of phosphorus runoff from nearby farms, the 
introduction of invasive species, and degraded streamside habitat. 
These stressors interact with one another, making it difficult to 
isolate and treat one stressor at a time. TMDLs that do not address 
the multiple stressors causing impairment of the water body in 
question--for example, a TMDL developed on the assumption that 
phosphorus alone is responsible for the absence of native fish--does 
not provide evidence that addressing the stressor or stressors in 
question will be sufficient for the water body to attain its 
designated use or uses. 

[54] EPA, Report of the Federal Advisory Committee. (1998). 

[55] Brian Benham et al., "TMDL Implementation: Lessons Learned," 
Proceedings of the Water Environment Federation, TMDL 2007, vol. 2007, 
pp. 428-442 (2007) and Center for TMDL and Watershed Studies, TMDL 
Implementation: Characteristics of Successful Projects, Final Report, 
VT-BSE doc. no. 2006-0003 (Blacksburg, Va: 2006). 

[56] EPA, Report of the Federal Advisory Committee (1998). 

[57] Each TMDL was independently reviewed by two experts to increase 
the reliability of our findings. The water resource experts reviewed a 
random subsample of 25 TMDLs that we drew from the nationwide sample 
of 191 of long-established TMDLs. 

[58] Experts did not agree about whether the remaining TMDL contains 
sufficient evidence. In this section of our report, we generally 
present only the total number of TMDLs that experts agreed do and do 
not contain particular key features, with the remainder equaling the 
number of TMDLs for which experts (1) did not agree about whether a 
key feature is present or absent, (2) agreed that not enough 
information was contained in the TMDL to determine whether a key 
feature is present or absent, or (3) agreed that a key feature does 
not apply to the TMDL. 

[59] Experts agreed that the remaining 1 of the 25 TMDLs does not 
contain enough information to determine whether this key feature is 
present or absent. 

[60] Overall, 12 of the 17 TMDLs apply to water bodies that are 
subject to additional TMDLs the experts did not review, but most of 
these other TMDLs also address only chemical pollutants and therefore 
may not mitigate physical and biological stressors. The experts 
reported that the contribution of multiple, nonchemical stressors 
makes it highly unlikely that designated uses will be attained by 
addressing one or a few chemical pollutants. 

[61] Experts reviewed implementation information contained within the 
25 TMDL documents as well as in accompanying implementation plans, 
where present, that were issued as separate documents. 

[62] Of the 8 TMDLs that discuss implementation to some extent, the 
experts agreed that 2 of them identify actors who are to support 
implementation but are missing other key features related to 
implementation. They agreed that 1 of these 8 TMDLs identifies actions 
that need to occur to attain water quality standards but is missing 
another key feature related to implementation. 

[63] Experts did not agree on whether the remaining 2 of the 8 TMDLs 
that discuss reasonable assurances provide sufficient evidence that 
reductions from nonpoint sources will occur. 

[64] The experts disagreed about whether the remaining 1 of the 18 
TMDLs includes indicators of biological outcomes. 

[65] Of the 18 TMDLs with a monitoring plan, experts disagreed about 
whether 1 of them describes an adaptive approach. 

[66] EPA has a process to accommodate a nonchemical approach to water 
body assessment and listing, but it appears to be underused by states, 
according to our analysis. States may list water bodies impaired by 
nonchemical stressors as category 4c (water quality is impaired by 
something other than a pollutant, such as a dam), as opposed to 
category 5 (water quality is impaired, and the water body is waiting 
for a TMDL to be developed). Of water bodies nationwide, about 1,000 
were listed as category 4c in recent integrated reports out of about 
128,000 total assessed water bodies. 

[67] After EPA withdrew the rule in 2002, it conducted dozens of 
studies evaluating and describing characteristics of TMDLs that enable 
water bodies to attain water quality standards. According to EPA 
officials, these studies were meant to provide information and 
guidance to states, and certain EPA regions applied study findings in 
their decisions to approve or disapprove TMDLs. EPA has also developed 
draft guidance, which was not final at the time of our review, 
outlining situations that may warrant TMDL revision, such as 
monitoring results showing that TMDL implementation is not achieving 
anticipated pollution reductions. 

[68] See 40 C.F.R. §§ 130.7(c), 130.2(i) (2013) (EPA's TMDL 
regulations). A court has held that neither the Clean Water Act nor 
EPA regulations define TMDLs as including an implementation plan. 
Sierra Club v. Meiburg, 296 F.3d 1021, 1030-31 (11th Cir. 2002). See 
also Amigos Bravos v. Green, 306 F. Supp. 2d 48, 57 (D.D.C. 2004). 

[69] Projects prescribed in TMDLs may include landowners' 
implementation of best management practices on agricultural land to 
reduce nutrient-laden runoff. 

[70] According to EPA studies, detailed implementation plans are 
associated with TMDLs that have been implemented effectively. Thus, 
detailed implementation plans are both a key feature in TMDL 
development and a factor facilitating effective TMDL implementation. 

[71] Center for TMDL and Watershed Studies, TMDL Implementation: 
Characteristics of Successful Projects, Final Report, VT-BSE doc. no. 
2006-0003 (Blacksburg, Va: 2006). 

[72] Center for TMDL and Watershed Studies and Biological Systems 
Engineering, TMDL Implementation: Lessons Learned (Blacksburg, Va: 
2007). 

[73] The 2007 EPA study found that external agency officials 
responsible for projects or activities related to TMDL implementation 
were generally unaware of the projects or activities identified in 
TMDLs that are needed to achieve water quality standards. The study 
noted that TMDLs are implemented by many stakeholders, but less than 
50 percent of external agency stakeholders always or often had 
knowledge of TMDLs and the projects or activities needed to restore 
water bodies. These data underscore the importance of involving 
representatives from external agencies in the TMDL implementation 
process, to obtain their buy-in and encourage participants in their 
programs to implement best management practices or projects that can 
help improve water quality. EPA, Developing Effective TMDLs (2007). 

[74] Appendix IV presents the full results of our survey, including 
the helpfulness of two implementation factors--land development and 
land retirement--that state TMDL coordinators told us were less 
helpful than the implementation factors we discuss in the body of this 
report. 

[75] Our survey asked if state TMDL coordinators found specific 
factors very helpful, somewhat helpful, not helpful, or not applicable 
for each of 191 TMDLs. For a given TMDL factor, "considered helpful" 
denotes the percentage of these long-established TMDLs in which a 
survey respondent indicated that the factor either was helpful (if 
present) or (if factor was not present or TMDL was not yet 
implemented) would have been or would be helpful in implementing the 
TMDL in question (see appendix IV). 

[76] State TMDL coordinators responding to our survey reported that 
landowners were willing to (1) implement best management practices for 
36 percent of long-established TMDLs (the 95 percent confidence 
interval for this estimate is [24, 49]), (2) implement the most 
effective best management practices for 35 percent of long-established 
TMDLs, and (3) implement best management practices in the most-
effective places for 33 percent of long-established TMDLs. 

[77] One state TMDL coordinator did not indicate in our survey the 
extent to which best management practices were installed. Other 
factors, such as limited funding, could also help explain the limited 
extent to which landowners had implemented all necessary best 
management practices. 

[78] Treatment systems for drainage from abandoned mines, which 
usually consists of acidic water, channel the drainage into ponds, 
where the acidic water is neutralized through contact with limestone, 
and metal pollutants are removed. 

[79] The 95 percent confidence interval for this estimate is (52, 82). 

[80] The 95 percent confidence interval for this estimate is (41, 66). 

[81] The 95 percent confidence interval for this estimate is (41, 67). 

[82] EPA, Nonpoint Source Program and Grants Guidelines for States and 
Territories (Washington, D.C.: April 2013). 

[83] The population of long-established TMDLs was 34,794. 

[84] Overall, experts agreed on 88 percent of the 1,917 questions they 
answered. 

[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]