Water Quality: Key EPA and State Decisions Limited by Inconsistent and
Incomplete Data (Chapter Report, 03/15/2000, GAO/RCED-00-54).

Pursuant to a congressional request, GAO provided information on the
Environmental Protection Agency's (EPA) National Water Quality
Inventory, focusing on whether: (1) the information in EPA's National
Water Quality Inventory is reliable and representative of water quality
conditions nationwide; and (2) available data are sufficient to allow
state officials to make key decisions about activities required by the
Clean Water Act, such as identifying waters that do not meet water
quality standards and developing strategies to address those waters.

GAO noted that: (1) the National Water Quality Inventory does not
accurately portray water quality conditions nationwide; (2) a major
reason is that it would be cost-prohibitive to physically monitor all of
the waters in the country, and, therefore, almost all states monitor a
subset of their waters; (3) most monitoring is not done in a way that
allows for statistically valid assessments of water quality conditions
in unmonitored waters; (4) the Inventory is not reliable because wide
variation exists in: (a) the way that individual states select their
monitoring sites; (b) the kinds of tests they perform on their waters
and how the results of these tests are interpreted; and (c) the methods
they use to determine the causes and sources of pollution in waters that
do not meet water quality standards; (5) as a result, the information in
the Inventory cannot be meaningfully compared across states; (6) EPA
used information from the Inventory as a basis for a number of important
decisions and activities, such as deciding how to allocate federal funds
for Clean Water Act programs to states and measuring and communicating
EPA's and states' progress in implementing the act; (7) the dearth of
the waters actually monitored, combined with the wide variation among
states' monitoring and assessment approaches, make the national
statistics unreliable and subject to misinterpretation and, therefore of
limited usefulness for these purposes; (8) there is little doubt among
experts that the remaining problems are considerable and that solutions
will entail significant expenditures; (9) what is uncertain is the
precise extent of water quality problems, where and what the most severe
problems are, and the location of high-quality waters that need to be
protected; (10) the limitations of the information in the Inventory
extend well beyond the problems associated with the national uses of the
report; (11) state-level activities, such as identifying water quality
problems and setting priorities among them to obtain the most
improvement in water quality for the dollar spent, are also affected by
data limitations; (12) GAO's survey shows that data gaps limit states'
abilities to carry out several key management and regulatory activities
on water quality; (13) only six states reported that they have a
majority of the data they need to assess whether their waters meet water
quality standards; and (14) a vast majority of states reported that they
have less than half the data they need to: (a) identify nonpoint sources
that are causing waters not to meet standards; and (b) develop total
maximum daily loads (TMDL) for those waters.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  RCED-00-54
     TITLE:  Water Quality: Key EPA and State Decisions Limited by
	     Inconsistent and Incomplete Data
      DATE:  03/15/2000
   SUBJECT:  Water pollution control
	     Environmental monitoring
	     Water quality
	     State-administered programs
	     Interagency relations
	     Data collection
	     Reporting requirements
	     Federal/state relations
IDENTIFIER:  EPA National Water Quality Inventory
	     USGS National Water Quality Assessment Program
	     EPA National Pollutant Discharge Elimination System
	     EPA Environmental Monitoring and Assessment Program
	     EPA Index of Watershed Indicators
	     USGS National Stream Quality Accounting Network
	     NOS National Status and Trends Program

******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO Testimony.                                               **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
******************************************************************
GAO/RCED-00-54
               
4

16

Represent Water Quality Conditions Nationwide

24

to Manage Water Quality

43

Appendix I: Copy of GAO Questionnaire

58

Appendix II: Comments From the Environmental Protection Agency

69

Appendix III: GAO Contacts and Staff Acknowledgments

73

Table 1: Monitoring Designs Described in EPA's Water Quality
Reporting Guidelines 28

Figure 1: States With a Majority of the Data Needed for Activities
to Manage Water Quality 7

Figure 2: Percentage of Waters Monitored, Evaluated, and Not
Assessed 9

Figure 3: Examples of Point and Nonpoint Sources of Pollution 17

Figure 4: Steps in Managing Water Quality 19

Figure 5: Percentage of Waters Monitored, Evaluated, and Not
Assessed in the 1996 305(b) Report 26

Figure 6: Monitoring Types and Pollutants or Conditions That They Measure 31

Figure 7: States With a Majority of the Data Needed for Activities
to Manage Water Quality 44

Figure 8: Number of States With a Rotating Basin Plan in Place or
Under Development 48

EPA Environmental Protection Agency

GAO General Accounting Office

IWI Index of Watershed Indicators

TMDL Total maximum daily load

Resources, Community, and
Economic Development Division

B-284241

March 15, 2000

The Honorable Sherwood L. Boehlert
Chairman
Subcommittee on Water Resources
and Environment
House of Representatives

Dear Mr. Chairman:

As requested, we are reporting on the Environmental Protection Agency's
(EPA) National Water Quality Inventory and the adequacy of water quality
data for key water quality decisions required by the Clean Water Act.

As arranged with your office, unless you publicly announce its contents
earlier, we plan no further distribution of this report until 7 days from
the date of this letter. At that time, we will send copies to the
appropriate congressional committees; the Honorable Carol M. Browner,
Administrator, EPA; and the Honorable Jacob J. Lew, Director, Office of
Management and Budget. We will make copies available to others upon request.

Please call me at (202) 512-6111 if you or your staff have any questions.
Major contributors to this report are listed in appendix III.

Sincerely yours,
Peter F. Guerrero
Director, Environmental
Protection Issues

Executive Summary

One of the states' key responsibilities under the Clean Water Act is to
collect data to monitor the condition of the nation's waters. The act
further requires states to submit reports on the condition of their waters
to the U.S. Environmental Protection Agency (EPA), which then compiles the
states' submissions into a report called the National Water Quality
Inventory. These data, and EPA's report, have become increasingly important
as national attention has focused on the most effective and efficient way to
target the nation's water pollution problems and, in particular, on the
soundness of the regulatory decisions that will be required to deal with the
nation's most heavily polluted waters.

Concerned about the adequacy of the data to support water quality policy and
key regulatory decisions, the Chairman, Subcommittee on Water Resources and
Environment, House Committee on Transportation and Infrastructure, asked GAO
to examine several issues relating to EPA's report on water quality and
water quality management. Specifically, GAO was asked to (1) determine
whether the information in EPA's National Water Quality Inventory is
reliable and representative of water quality conditions nationwide and (2)
determine if available data are sufficient to allow state officials to make
key decisions about activities required by the Clean Water Act, such as
identifying waters that do not meet water quality standards and developing
strategies to address those waters.

The Clean Water Act makes states responsible for developing programs to
manage water quality; the programs are intended to achieve the act's goals
of supporting aquatic communities, protecting human health, and sustaining
other uses and provides for funding to implement the act. The programs that
manage water quality generally include functions such as developing water
quality standards by which to assess whether waters are of acceptable
quality; monitoring water quality; and issuing permits to, and conducting
inspections of, facilities that discharge pollutants into bodies of water.
The Clean Water Act requires states to report to EPA biennially on the
quality of their waters and directs EPA to compile the states' reports and
analyze them. EPA's compilation, called the National Water Quality
Inventory, is the primary report for the public about the condition of the
nation's waters and is often used to characterize the nation's progress in
achieving the act's goals. The report's data are also used as a basis for
making key management decisions regarding water quality, such as how certain
Clean Water Act funds are to be allocated among the states.

In addition to providing a basis for satisfying the act's national reporting
requirements, the states need reliable and complete data to identify water
quality problems and develop cost-effective strategies to address them. Of
particular concern to EPA and the states are "nonpoint" sources of
pollution--diffuse sources that include a variety of land-based activities,
such as timber harvesting, agriculture, and urban development--which are
widely regarded as contributing to the largest share of remaining water
quality problems.

Sound data on water quality are becoming increasingly important as numerous
lawsuits are directing nationwide attention to the accelerated cleanup of
water quality problems through the development of total maximum daily loads
(TMDLs). TMDLs are used to restore water quality by identifying how much
pollution a body of water can receive and still meet standards and then
reducing the amount of pollution entering the water to that level. The Clean
Water Act requires states to develop TMDLs when other mechanisms to restore
water quality have not worked. The basis for many of the lawsuits is that
EPA and the states have not implemented this requirement of the act and have
not thoroughly identified all waters needing TMDLs. EPA proposed revisions
to its regulations on the management of water quality in August 1999 to
strengthen the TMDL program.

To address this review's objectives, GAO conducted a data reliability
evaluation of the National Water Quality Inventory by obtaining information
from EPA and state water quality officials on the processes used to generate
the report and by examining the controls in place to ensure that the
information was reliable. GAO also interviewed EPA officials in headquarters
and regions responsible for water quality issues to discuss the Clean Water
Act's requirements for managing water quality and the guidance provided to
states for implementing them. GAO conducted a mail survey of the 50 states
and the District of Columbia and met with water quality officials in 4
states to discuss their programs to manage water quality, the management and
regulatory decisions that rely on water quality data, and the issues that
states face in carrying out their programs.

The National Water Quality Inventory does not accurately portray water
quality conditions nationwide. A major reason is that states collectively
assess only a small percentage of waters in the United States. Specifically,
19 percent of the nation's rivers and streams were assessed for the 1996
Inventory (the latest report), as were 6 percent of ocean and other
shoreline waters--which have been the subject of an increasing number of
beach advisories and closures in recent years. It would be cost-prohibitive
to physically monitor all of the waters in the country, and, therefore,
almost all states monitor a subset of their waters. However, most monitoring
is not done in a way that allows for statistically valid assessments of
water quality conditions in unmonitored waters. In addition, the Inventory
is not reliable because wide variation exists in (1) the way that individual
states select their monitoring sites, (2) the kinds of tests they perform on
their waters and how the results of these tests are interpreted, and (3) the
methods they use to determine the causes and sources of pollution in waters
that do not meet water quality standards. As a result, the information in
the Inventory cannot be meaningfully compared across states.

EPA uses information from the Inventory as a basis for a number of important
decisions and activities, such as deciding how to allocate federal funds for
Clean Water Act programs to states and measuring and communicating EPA's and
states' progress in implementing the act. However, the dearth of the waters
actually monitored, combined with the wide variation among states'
monitoring and assessment approaches, make the national statistics
unreliable and subject to misinterpretation and, therefore, of limited
usefulness for these purposes. While these limitations may call into
question some of EPA's decisions that rely on these data, there is little
doubt among experts that the remaining problems are considerable and that
solutions will entail significant expenditures. What is uncertain, however,
is the precise extent of water quality problems, where and what the most
severe problems are, and the location of high-quality waters that need to be
protected.

The limitations of the information in the Inventory extend well beyond the
problems associated with the national uses of the report. State-level
activities, such as identifying water quality problems and setting
priorities among them to obtain the most improvement in water quality for
the dollar spent, are also affected by data limitations. GAO's survey shows
that data gaps limit states' abilities to carry out several key management
and regulatory activities on water quality. As figure 1 illustrates, these
data gaps are particularly serious in the case of nonpoint sources, which
are widely accepted as contributing to the majority of the nation's water
quality problems. Only six states reported that they have a majority of the
data they need to assess whether their waters meet water quality standards.
Moreover, a vast majority of states reported that they have less than half
the data they need to (1) identify nonpoint sources that are causing waters
not to meet standards and (2) develop TMDLs for those waters.

Figure 1: States With a Majority of the Data Needed for Activities to Manage
Water Quality
Among the recommendations GAO makes in this report are that EPA better
convey the limitations of the Inventory and that the agency reexamine the
implications of using the report for national decisions. GAO also recommends
a number of actions that EPA should take to provide the tools, information,
and assistance that states need to conduct critical water quality
activities.

Nationwide

As figure 2 illustrates, only a small percentage of U.S. waters are assessed
for the National Water Quality Inventory--19 percent in the case of rivers
and streams. Importantly, however, only about half of these assessments are
based on current, site-specific monitoring information. The states assess
conditions in the other waters by using monitoring data more than 5 years
old or by relying on other data, such as land use information or anecdotal
evidence of water quality, that serve as subjective indicators of whether a
body of water is meeting standards. (These assessments are referred to as
evaluated assessments.) Both EPA and state officials indicated that they are
less confident of some of these assessments' accuracy because they may not
reflect changes in water conditions and are sometimes based on unconfirmed
information. However, EPA did not distinguish between the two types of
assessments when presenting information on the percentage of waters meeting
standards in the 1996 Inventory.

Figure 2: Percentage of Waters Monitored, Evaluated, and Not Assessed
Source: 1996 National Water Quality Inventory.

EPA encourages states to use statistical monitoring in order to capture a
more comprehensive picture of water quality. GAO's survey of the 50 states
and the District of Columbia, however, revealed that few states employ such
a sampling design for a majority of their waters. Most states target their
limited monitoring resources on areas with known problems and areas that are
used by the public. The reliability of the data in the Inventory is limited
because there is considerable variation in the analytical methods that
states use to evaluate water quality. For example, states vary in the extent
to which they rely on monitoring the chemical and physical properties of
their waters (i.e., temperature, dissolved oxygen levels, or the presence of
chlorine, mercury, and other chemicals) versus biological properties (i.e.,
species diversity and habitat quality). In addition, as allowed by the Clean
Water Act, states may evaluate water quality against different standards. As
a result, the Inventory does not represent water quality conditions
nationwide, and the data are not reliable for comparisons across states.

EPA has attempted to address the inconsistencies among states primarily
through a workgroup whose members include EPA and state water quality staff,
and the agency has incorporated several of the workgroup's recommendations
into reporting guidelines for the Inventory. These are steps in the right
direction, but they will not guarantee improved consistency because states
are not legally required to comply with the reporting guidelines. In
addition, the Inventory does not integrate findings from other monitoring
programs to supplement state-reported information and help portray water
quality conditions nationwide.

Among the key activities that rely on data from the Inventory is the formula
that EPA uses to allocate funds to states to help them implement their water
quality programs. In the past, such funds were allocated on the basis of
several factors, such as the number of industrial dischargers and power
plants in a state, which were intended as indirect measures of potential
water quality problems. EPA recently revised this formula, in consultation
with the states, and the new formula includes the amount of waters not
meeting standards as reported in the Inventory by the states. Given the
differences in the methods that states use to assess their waters, however,
the resulting allocation may not reflect the relative magnitude of water
quality problems in states but, instead, reflect assessment and reporting
differences across them. While EPA acknowledged limitations with using the
Inventory and therefore decided to phase in the increased weighting of
waters not meeting standards beginning in fiscal year 1999, the agency does
not have a plan to resolve the limitations.

While the lack of reliable information may call into question some of EPA's
decisions that rely on data from the Inventory, there is widespread
agreement that many waters still do not meet standards and that addressing
these problems will be difficult and will require a considerable commitment
of resources. What is uncertain is the precise extent of the problems, where
and what the most severe problems are, and the location of high-quality
waters that need to be protected.

Quality Activities

The results of GAO's survey highlight the need for more comprehensive state
monitoring and call into question the extent to which unknown and
potentially serious problems are going undetected. Only six states reported
having a majority of the data needed to fully assess all their waters, and
less than half the states have a majority of the data needed to determine if
their waters that have been assessed should be placed on their lists of
waters that do not meet standards. While state officials believe they have
identified their most serious pollution problems, many acknowledge that more
comprehensive monitoring would reveal additional problems. A study of toxins
in Maine's lakes, for example, unexpectedly showed widespread elevated
levels of mercury. As a result of this more comprehensive monitoring effort,
Maine and other New England states issued advisories against the consumption
of fish for all their lakes.

In terms of addressing known pollution problems, there is a sharp division
between the adequacy of data for point sources and of that for nonpoint
sources. By and large, states have adequate data for identifying point
sources that cause waters not to meet standards and developing TMDLs for
these sources. Much of the Clean Water Act's implementation over the last 27
years has focused on addressing point sources. Additionally, it is
relatively easy to determine the point sources that are contributing to
water quality problems because they generally enter the water at a distinct
point. Accordingly, 40 states responding to GAO's survey reported having a
majority of the data they need to identify point sources that are causing
waters not to meet standards, and more than half reported having a majority
of the data needed to develop TMDLs for those waters.

Yet, while water quality officials responding to GAO's survey reported
having a majority of the data they need to deal with point sources of
pollution, few respondents reported having a majority of the data needed to
deal with nonpoint sources. Specifically, almost all states reported that
they did not have a majority of the data they need to identify nonpoint
sources of pollution and develop TMDLs for waters polluted primarily by
nonpoint sources. The data on nonpoint sources of water quality problems are
of particular importance because the nation's remaining water quality
problems are widely viewed as being caused, at least in part, by nonpoint
sources. In 1998, for example, Texas officials reported that nonpoint
sources contributed to almost all of the state's water quality problems.

States have always needed to obtain comprehensive and reliable water quality
data, but such data have become more important in recent years. The
increased emphasis on statutory and regulatory requirements for assessing
waters and developing TMDLs has heightened the need for states to have sound
information to make these decisions. Of particular concern are the thousands
of TMDLs that states must develop for their most heavily polluted waters, as
required by the Clean Water Act. GAO's survey indicated that states will
need additional tools, resources, and assistance in developing TMDLs for
their waters--a task that will significantly tax already limited resources
over a sustained period of time.

EPA has a number of initiatives under way that are designed to help states
deal with specific problems, such as guides for developing TMDLs for
specific pollutants that degrade waters and analytical methods for
determining airborne sources of water pollution. However, states identified
other areas in which they need assistance, such as additional tools for
evaluating nonpoint sources of pollution and technical assistance on using
analytical methods. EPA has initiated a strategy for supporting TMDL
development, but this effort has been put on hold because of other workload
demands. Until this strategy is finalized, it is not clear whether states'
needs will be addressed.

GAO makes a number of recommendations to the Administrator, EPA, to
appropriately characterize and improve the usefulness of the National Water
Quality Inventory. GAO also makes recommendations for EPA to work with the
states to finalize its strategy to support TMDL development.

GAO provided EPA with a draft of this report for review and comment. EPA's
comments and GAO's evaluation are discussed at the end of chapters 2 and 3.
EPA's comments are reproduced in appendix II.

EPA generally agreed with the report, noting that it underscores the need to
improve water quality monitoring and assessment across the country. EPA also
offered more specific comments dealing with the report's discussion of (1)
the sufficiency of state data for key water quality management activities,
(2) the National Water Quality Inventory, and (3) the Inventory's use in
grant allocation decisions.

First, EPA pointed out that states need not wait until "perfect data" are
collected before taking actions to improve water quality. GAO agrees that
perfect data are not needed before actions can be taken to improve water
quality but continues to believe that data of better quality are needed to
make more informed and defensible decisions. EPA also asked that GAO
acknowledge the agency's efforts that could help improve water quality data,
such as the modernization of its primary computer system that stores and
retrieves water quality data. GAO recognizes that EPA has many efforts aimed
at improving water quality data overall, but notes that some of these
efforts are limited in their ability to address the types of problems with
water quality data identified. EPA also recommended that GAO add language to
the report on the implications of inadequate data for identifying waters
that do not meet standards. GAO agrees with this suggestion and has added a
discussion of this issue to chapter 3.

Second, EPA made several points about the content and presentation of the
information in the National Water Quality Inventory. The agency states that
GAO's report should more strongly note that the Inventory is, by law, a
summary of states' assessments of whether their waters are meeting standards
and "is not intended to be a national water quality report on a selected,
consistent set of water quality parameters." GAO acknowledges this point and
has stated so in this report. However, the broader question that the
Subcommittee posed to GAO asked how well the report characterizes water
quality conditions nationwide. As such, this report discusses the utility of
the Inventory for characterizing water quality nationwide.

Moreover, because the Inventory is used to help make national policy and
regulatory decisions, GAO believes it is prudent for EPA to take steps to
ensure that the information presented is as reliable as possible. Where it
is not, EPA should clearly explain the reasons why and assess the possible
implications on national decision-making. EPA points out that past Inventory
reports have clearly explained that its information does not consistently
characterize nationwide water quality and that it has continued to highlight
this point in its draft 1998 report. GAO acknowledges the caution noted by
EPA in how the Inventory's data should be interpreted. However, by
aggregating data across states, EPA is implicitly suggesting that these data
can, in fact, be compared and in doing so is increasing the likelihood that
the data will be misused or misinterpreted. GAO believes that aggregating
data in different ways and adding language to characterize the potential
implications of similarities and differences among state programs would
present a clearer picture of how the state-reported information can be used.

In response to GAO's recommendation that the agency supplement the current
information in the Inventory by including findings from other programs
(e.g., the U.S. Geological Survey's National Water Quality Assessment
Program), EPA pointed out that it already includes descriptions of this
information in the Inventory. GAO believes, however, that EPA can improve
the presentation and usefulness of the report if it takes this descriptive
approach a step further by integrating the findings of these studies with
the report's own presentation of state data, where appropriate.

Third, EPA commented that it appears that GAO did not obtain a complete
description of the process used to develop and evaluate the agency's new
formula for allocating grant funds to states under section 106 of the Clean
Water Act. The agency stated that it conducted extensive data searches and
data source evaluations before selecting the Inventory as a source of
information on the extent of water pollution in each state. In addition, EPA
said that it evaluated numerous data scenarios to determine whether certain
data components unduly influenced distributions of grant funds to states.
GAO has incorporated a number of technical clarifications that EPA provided,
as appropriate. However, according to an EPA official, the data scenarios
referred to assessed the impact of using different data elements in the
formula--not how differences among state monitoring and assessment methods
may affect formula results, as GAO recommends.

Introduction

The Federal Water Pollution Control Act, commonly referred to as the Clean
Water Act, was enacted in 1972.1 One of its primary goals is to achieve and
maintain water quality for the protection and propagation of fish,
shellfish, and wildlife, and for recreation in and on the water. At the time
the act was passed, water quality in the United States had deteriorated to
the point where some rivers had severe pollution problems due to discharges
of raw sewage and others were severely contaminated by industrial chemicals.

The Clean Water Act has been credited with greatly improving the condition
of waters in the United States. Much of this improvement has been attributed
to the National Pollutant Discharge Elimination System, a program that
controls pollutant discharges from industrial facilities and wastewater
treatment plants, which are often referred to as "point sources" of
pollution. The program sets up a process for issuing permits to facilities
that discharge pollutants and imposes requirements for the pollution control
mechanisms they must use to reduce the amount and toxicity of the pollutants
they discharge. The Congress added provisions to the act in 1987 that called
attention to another source of pollution that was believed to be largely
responsible for continued water quality problems--nonpoint source pollution.

Nonpoint sources of pollution include many different types of activities
such as agriculture, timber harvesting, and urban development. Nonpoint
sources contribute pollution as rainwater, snowmelt, or irrigation water
moves over or through land surfaces. This causes pollutants, either
dissolved or solid, to be transported and eventually deposited into surface
waters. Airborne pollutants, sometimes carried through the air over long
distances and then deposited into bodies of water, are also considered
nonpoint pollution. The types of pollutants vary with the activity involved
and include sediment, nutrients, pesticides, pathogens (such as bacteria and
viruses), salts, oil, grease, toxic chemicals, and heavy metals. Figure 3
provides examples of point and nonpoint sources of pollutants.

Figure 3: Examples of Point and Nonpoint Sources of Pollution
Today, much attention is being focused on dealing with nonpoint sources of
pollution. States and the Environmental Protection Agency (EPA) are facing
increased pressure to address pollution problems as a result of numerous
lawsuits charging that they are violating the Clean Water Act by not doing
so. In 1998, the administration introduced the Clean Water Action Plan,
which is intended to strengthen federal and state actions to address
pollution problems; addressing nonpoint source pollution is a major
component. As part of this action plan, for example, funding for grants to
states to deal with nonpoint source pollution almost doubled from $105
million in fiscal year 1998 to $200 million in fiscal year 1999. This level
of funding was maintained for fiscal year 2000. In addition, EPA and the
U.S. Department of Agriculture jointly developed a national strategy for
minimizing environmental and public health impacts from nonpoint source
pollution coming from animal feeding operations. Furthermore, the plan
includes actions to ensure beaches--which have been subject to an increasing
number of advisories and closures in recent years--are safe for swimming.

Water Quality

While EPA is responsible for developing regulations and guidance for
implementing the Clean Water Act and ensuring that it is properly
implemented, states have primary responsibility for implementing programs to
manage water quality. States are provided with federal funds through EPA to
develop and implement these programs. In addition to the program that
regulates point sources of pollution, other activities to manage water
quality include developing water quality standards to protect public health
and aquatic life, monitoring water quality to assess whether standards are
attained, and implementing strategies to address pollution problems.

The Clean Water Act requires states to establish water quality standards as
a way to ensure that the goals of the act will be met. These goals include
the achievement and maintenance of water quality for the protection and
propagation of fish, shellfish, and wildlife and for recreation in and on
the water. To develop water quality standards, states classify waters
according to how they will be used and adopt water quality criteria to
protect those uses. Criteria--which can be numeric or narrative--indicate
the acceptable levels of chemicals allowed, or physical or biological
characteristics required, in a water in order for it to meet its designated
use. Waters that do not meet standards are often referred to as "impaired."
Without standards, a state has no basis to assert whether waters are of
acceptable quality. By law, states are allowed to set their own water
quality standards; as a result, standards differ across states.

Monitoring water quality is the linchpin for many management activities
regarding water quality. Figure 4 depicts this relationship, showing that
monitoring water quality is essential to identifying water quality problems
and determining whether actions taken to restore water quality have been
successful.

Figure 4: Steps in Managing Water Quality
States either monitor water quality, use data collected by other entities to
assess water quality, or rely on a combination of the two. Monitoring data
include information such as the presence of chemicals, physical
characteristics such as temperature, and biological characteristics such as
the health and abundance of fish and other aquatic species. As shown in
figure 4, states compare monitoring data with water quality criteria to
determine if waters meet standards. If a state's assessment of a body of
water indicates that it does not meet the criteria--for example, if it has
levels of chlorine that are too high to support aquatic life--then the water
is considered not to be supporting its intended use of aquatic life. This
indicates that the body of water needs pollution reduction strategies to
correct the problem.

The Clean Water Act sets forth a procedure for states to follow in
addressing waters that do not meet standards. This requires states to submit
to EPA a list of all of their waters that do not or are not expected to meet
applicable water quality standards.2 For the lists developed in 1998, states
identified about 21,000 waters that did not or were not expected to meet
standards. The Clean Water Act further requires that states take steps to
ensure that waters that do not meet standards are improved sufficiently to
support their designated uses. This is done through developing total maximum
daily loads (TMDLs), which refers to the maximum amount of a pollutant a
body of water can receive on a periodic basis and still support its
designated uses. Generally, TMDLs are developed by (1) analyzing the
pollutants and sources of those pollutants causing a water quality problem
and (2) determining how much the pollutants need to be reduced in order to
enable the body of water to meet standards. A key part of the TMDL is
assigning the pollutant reductions necessary among the various sources
contributing to the water quality problem.

Data used in the development of TMDLs are becoming increasingly important.
For example, legal actions have been taken in more than 30 states asserting
that the states and EPA have not adequately implemented water quality
standards. Most of these lawsuits have been filed in the last 5 years and
largely focus on the adequacy of states' lists identifying water quality
problems or on the lack of state and EPA action in developing TMDLs. In
August 1999, EPA proposed changes to water quality management regulations in
order to strengthen the TMDL program.

Another requirement of states under the Clean Water Act that relies on
monitoring and assessment information is found in section 305(b), which
requires states to report biennially to EPA on the quality of their waters
and EPA to produce a report analyzing the states' information. A further
requirement that the report be transmitted to Congress was recently
repealed.3 The information that states submit to EPA for this report, called
the National Water Quality Inventory but generally referred to as the 305(b)
report, includes (1) the amount of waters the state assessed; (2) whether
those waters meet water quality standards; (3) the pollutants, such as toxic
chemicals or bacteria, that cause waters to not meet standards; (4) the
sources, such as industrial facilities or wastewater treatment plants, that
contribute those pollutants; and (5) a description of the state's water
quality program.

The 305(b) report is the primary report for informing the public about the
conditions of the nation's waters. Therefore, it is often relied on to make
statements regarding water quality conditions nationwide and to support
national-level decisions. For example, the 1996 report's summary statistic
that forty percent of the waters assessed by states are unsafe for swimming
and fishing was a leading factor in the Administration's argument for the
need for the Clean Water Action Plan and associated requests for budget
increases. In addition, the report's summary information on causes and
sources of water quality impairments contributed to EPA's decision to
develop new regulations for animal feeding operations and criteria for
evaluating a group of pollutants referred to as nutrients.

As agreed with the Chairman, Subcommittee on Water Resources and
Environment, House Committee on Transportation and Infrastructure, the
objectives of this review were to (1) determine whether the information in
EPA's National Water Quality Inventory is reliable and representative of
water quality conditions nationwide and (2) determine if available data are
sufficient to allow state officials to make key decisions about activities
required by the Clean Water Act, such as identifying waters that do not meet
water quality standards and developing strategies to address those waters.

To determine the reliability and representativeness of EPA's National Water
Quality Inventory (the 305(b) report), we followed generally accepted data
reliability and information systems audit methodologies. We reviewed EPA's
guidelines for preparing the 305(b) report and other EPA guidance documents
regarding monitoring and assessment techniques. We reviewed the last two
305(b) reports, issued for 1994 and 1996. We interviewed EPA headquarters
staff and contractor personnel responsible for compiling state reports into
a national summary. We also interviewed EPA regional officials involved in
helping states to administer their monitoring programs and prepare their
reports on water quality.

We also conducted a mail survey of the agencies that administer water
quality management programs in the 50 states and the District of Columbia.
Our survey asked state officials to identify the primary monitoring design
and analytical methods used, the sources of data relied on, and their
opinions on the 305(b) assessment and reporting process and EPA's
information systems. We pretested our questionnaire with officials in
Illinois, Maryland, and North Carolina and also obtained comments from
officials in EPA's offices of Water and Inspector General. We received
survey responses from all 50 states and the District of Columbia.

In addition, we met with water quality management officials in Illinois,
Maine, North Carolina, and Texas to discuss the processes they use to assess
water quality and provide EPA with information for reporting purposes. We
judgmentally selected these states to represent diversity among (1) water
quality monitoring designs, (2) the percentage of waters assessed, (3)
geographical settings, (4) EPA regions, and (4) the type of information
systems used to store data. We spoke with EPA regional officials in each of
the regions responsible for the states we visited. We also obtained reports
issued by EPA's Office of Inspector General pertaining to water quality
management programs in five states and discussed their findings with staff
involved with those reviews.

We also used the survey and site visits to address our second objective. Our
survey included questions regarding the adequacy of data to conduct various
water quality management activities. We asked states to comment on issues
that help or hinder them in carrying out their water quality management
responsibilities. We used our site visits to the four states to obtain a
more detailed understanding of the primary water quality management and
regulatory decisions that rely on water quality data, the way in which
states obtain the data needed to carry out these activities, and the data
limitations that states face.

We met with several officials representing the monitoring and TMDL programs
in EPA's Office of Water to discuss guidance that states are provided with
in these areas and the activities under way that may address states' needs.
We also spoke with officials in the offices of Wastewater Management and
General Counsel to discuss EPA's authority under the Clean Water Act
regarding state programs for monitoring, assessing, and reporting on water
quality.

We conducted our work from April 1999 through February 2000 in accordance
with generally accepted government auditing standards.

National Water Quality Inventory Does Not Accurately Represent Water Quality
Conditions Nationwide

The 305(b) report is not a reliable representation of nationwide water
quality conditions.4 The assessments of water quality that states report
generally do not represent conditions in all the waters in their state but,
instead, represent only the subset of waters that the states assessed.5 An
underlying reason for the differences among state programs is that the Clean
Water Act allows states to have different standards by which to assess the
level of quality of their waters. In addition, there is considerable
variation among states in many aspects of monitoring and assessing water
quality. EPA uses information from the 305(b) report for many important
national decisions, such as deciding how to allocate federal funds for water
quality activities to states, measuring the progress that states and EPA are
making toward implementing the Clean Water Act, communicating water quality
issues to the public, and making major regulatory decisions. Given the lack
of the representativeness of the data in the report and their inconsistency,
they are of limited usefulness for these purposes.

EPA has taken steps to address problems with the 305(b) report, such as
modifying reporting guidelines. However, because states are not required to
follow the guidelines, their implementation is not guaranteed. In addition,
integrating information from other sources with state-reported information
in the 305(b) report may help provide a more accurate picture of water
quality conditions nationwide.

Across States

Assessment data in the 305(b) report are not complete because they do not
represent all of the states' waters, either through a census (i.e., the
monitoring of all waters in a state) or through statistical sampling that
would yield data that are projectable to all state waters. In addition,
there is substantial variation among states in virtually every aspect of
monitoring and assessment. Variations are found in (1) the standards states
use to assess water quality, (2) the way that states select their monitoring
sites, (3) the kinds of monitoring tests that states perform and how they
interpret the results, and (4) the methods that states use to determine the
causes and sources of pollution.

Quality Conditions Nationwide

A major reason why the 305(b) report does not accurately portray nationwide
water quality conditions is that only a small percentage of waters are
actually assessed. This small percentage would not necessarily be a problem
for nationwide reporting if the waters were selected for monitoring in a
statistically projectable manner. However, the waters that are monitored are
not generally selected in a way that would allow for projecting water
quality conditions in other waters. In addition, the conditions in many of
the waters reported on in the 305(b) report are not based on current water
quality monitoring but, instead, are assessed on the basis of dated
information or subjective determinations.

A Small Percentage of U.S. Waters Are Assessed

Given that there are over 3.5 million miles of rivers and streams and 42
million acres of lakes, ponds, and reservoirs in the United States, it is
cost prohibitive for states to assess all of their waters. As a practical
matter, therefore, states almost always limit their assessments to a subset
of total waters; the result is that a majority of U.S. waters remain
unassessed. For example, the percentage of waters assessed in the 1996
report was 6 percent for ocean shoreline waters; 19 percent for rivers and
streams; 40 percent for lakes, ponds, and reservoirs; and 72 percent for
estuaries.

For many years, EPA has encouraged states, through 305(b) reporting
guidelines, to improve the coverage of state waters in their monitoring
efforts. The guidelines for the 1998 305(b) report suggest that states
should develop a strategy for achieving the goal of comprehensively
assessing all state waters. The monitoring techniques suggested by the
guidance include a combination of monitoring known problem waters, as well
as monitoring randomly selected sites, which would allow inferences to be
made about entire categories of waters. The guidelines also suggest that
states include information from federal agencies and other relevant
organizations in their 305(b) reports to increase the breadth or extent of
assessments. State responses to our survey indicate that many states
currently consider information from various sources in their assessments of
water quality. EPA requests that states describe their progress toward the
goal of comprehensive coverage in their 305(b) reports.

Of the waters that are assessed, only a portion are done so in a fashion
that meets EPA's definition of a "monitored" assessment--that which is based
on water quality monitoring data that are less than 5 years old. As figure 5
illustrates, of the 19 percent of river and stream miles that were assessed
for the 1996 report, half were assessed using these types of data.

Figure 5: Percentage of Waters Monitored, Evaluated, and Not Assessed in the
1996 305(b) Report
Source: 1996 National Water Quality Inventory.

The remaining assessments were based on data that are either older than data
used for monitored assessments, data calculated using water quality models,
or qualitative information, including sources such as land use data, the
location of pollution sources, or evaluations made by fish and game
biologists of water quality or wildlife conditions. This type of information
is used as an indicator of potential water quality in waters that are not
actually monitored. Assessments based on such information are called
evaluated assessments and can vary in quality and reliability. EPA officials
said that while some types of evaluated assessments can be used with a
relatively high degree of confidence, others are more open to question
because they may not reflect changes in water conditions or may be based on
unconfirmed information. Thirty-nine states used evaluated data for their
assessments of rivers and streams in the 1996 report, and 19 of them used
such data for half or more of their assessments.6 In the 1996 report, EPA
identified the percentage of evaluated and monitored assessments. However,
it did not distinguish between the two types when presenting information on
the percentage of waters not meeting water quality standards.

In addition to using older and potentially subjective information as a basis
for water quality assessments, some states also assess waters by
"extrapolating" assessments from one monitoring site to other unmonitored
waters. EPA's 305(b) report guidelines state that this practice is
acceptable if the monitored site is representative of other parts of the
body of water and specify conditions present in the water that might change
water quality or habitat and therefore preclude extrapolation. Officials in
several states we visited said they use extrapolated assessments in
preparing their 305(b) report. Additionally, a few states make presumed
assessments, which are assumptions that water quality is of a certain level
in the absence of any quantitative data to make a determination of quality.
EPA's guidelines for the 305(b) report advise against using such
assessments.

States' Selection of Monitoring Sites Does Not Yield Representative Data

The methods that states use to select monitoring sites affect how the
resulting data can be used. Monitoring sites that are chosen on the basis of
a random or statistical sample will result in data that are representative
of the condition of all waters in a population, such as all streams of a
certain size or type. EPA encourages states to include this type of design
in their monitoring programs in order to capture a more comprehensive
picture of water quality. Conversely, monitoring sites that are targeted
toward waters with specific characteristics, such as known pollution
problems or high levels of public use, result in data that can provide
information only on those particular waters that were monitored. Additional
information on monitoring designs is provided in table 1.

Table 1: Monitoring Designs Described in EPA's Water Quality Reporting
Guidelines

 Design             EPA's description            GAO's comments
                    This method allows questions
                    to be addressed that are
                    focused on site-specific
                    problems. The aggregation of
                                                 This design is intended to
 Targeted           these site-specific results  represent only the site
 (problem-based)    to make comprehensive        itself--not other bodies
                    assessments is open to
                    question regarding the       of water.
                    representativeness of those
                    sites to the resource as a
                    whole.
                    This method is intended to
                    produce snapshots of the
                    condition of an entire
                    resource when that resource
                    cannot be subject to a       Randomization is the only
                    census. Sample surveys rely  method that allows
 Sample survey or   on the selection of          statistically valid
 probabilistic      monitoring sites that are    inferences to be drawn to
                    representative of the        a population as a whole.
                    resource. Randomization in
                    the site selection process
                    is one way to ensure that
                    the sites represent the
                    resource of interest.
                    This design is based on some
                    criterion other than
                    randomness. Judgmental       We put EPA's assertion in
                    selection of sites is based  italics because, in
                    on the judgment that the     general, using nonrandom
                    sites are representative of  judgmental samples to
                    the target resource. Such    project results beyond the
                    judgmentally based sample    actual sites surveyed is
                    surveys require strong       questionable. If strong,
                    defense regarding the        compelling evidence
                    representativeness of the    demonstrates that the
                                                 sites selected to sample
 Judgmental         sites selected, and it may   are like the non-selected
                    not be possible to estimate
                    the uncertainty with which   sites in the area,
                    inferences are made as it is perhaps--and with clearly
                    when using probability-based stated, strong
                    sample surveys. This method  caveats--limited
                    is a nonrandom selection of  inferences could be made
                    sampling sites with the      beyond just the sample
                    intent of using assessment   sites. No clear,
                    results for drawing          statistically defensible
                    inferences on a population   statement is possible
                    as a whole. [Emphasis        under this method.
                    added.]

Source for EPA's description Guidelines for Preparation of the Comprehensive
State Water Quality Assessments (305(b) Reports) and Electronic Updates
(Sept. 1997).

By and large, states use a targeted approach to select monitoring sites and
focus monitoring efforts in problem areas or publicly used waters; very few
states use a statistical approach. According to our survey, only nine states
reported using a statistical approach for a majority of their water quality
monitoring of biological conditions, and five states reported doing so for
the monitoring of chemical conditions. As a result of the limited use of
statistical sampling by states, most of the data they gather are not
suitable for representing statewide conditions. State officials told us that
information on the condition of specific waters is more essential to state
programs for managing water quality than the kind of general information
provided by statistical designs. Because states have limited resources for
monitoring and assessment activities, they tend to use statistical
monitoring in limited cases or relegate it to a lower priority.

States

There are significant differences among states in virtually all aspects of
the 305(b) assessment and reporting process. An underlying difference is
that states can have different standards by which to assess the level of
quality of their waters. In addition, states vary in how they select their
monitoring sites, the kinds of monitoring tests they perform and how the
results of these tests are interpreted, and the methods they use to
determine the causes and sources of pollution.

States Vary in the Standards They Use to Assess Water Quality

The Clean Water Act allows states to set their own water quality standards.
Specifically, states can designate waters to support multiple uses, such as
for public water supplies, the propagation of fish and wildlife,
recreational purposes, and agricultural, industrial, and navigational
purposes. States can also set their own criteria (i.e., acceptable levels of
pollutants or conditions required in a water in order for it to support its
intended uses) as long as they are at least as restrictive as EPA's. As a
result, standards vary across the states. 7

According to EPA, some differences in water quality standards are based on
sound science. For example, criteria to protect aquatic life should reflect
the different natural ranges of biological communities. Similarly,
acceptable physical conditions can vary depending on geographic and other
factors. For example, natural fluctuations in the temperature and clarity of
lakes are different in states in the northeast from those in states such as
Florida and therefore, state standards may vary to accommodate these
differences. Even criteria set to prevent concentrations of metals from
causing toxic effects can be different from state to state because the way
that metals affect a body of water varies, depending on the water's natural
mineral content and acidity. Because of differences such as these, EPA
believes it is appropriate for water quality standards to vary from state to
state.

What is not appropriate, according to EPA, is the uneven development and
implementation of standards among states. Some states assess water quality
using a range of measures, such as physical, chemical, and biological
measures, while other states may base their assessments on only one or two
chemical measures. The rigor of state programs for water quality standards,
both in terms of the types of different measures that are included and the
monitoring programs used, should be consistent from state to state so that
the findings of water quality assessments are comparable.

Currently, however, most states' water quality standards and monitoring
programs are not comparable, yet in the 305(b) report, EPA does not provide
details on these differences, nor does it discuss their potential
implications on resulting national information. While one state may appear
to have many more water quality problems than another state, it could
actually be the result of more restrictive or more thorough standards.
Likewise, if one state monitors a high percentage of its waters, it may
identify more problems than a state that monitors fewer waters.

States Vary in the Types of Monitoring Tests That They Perform

The Clean Water Act outlines goals for attaining water quality in terms of
the biological, chemical, and physical conditions of waters. EPA's 305(b)
reporting guidelines discuss the different types of monitoring tests in each
of these areas--each of which yields data about particular aspects of bodies
of water. Physical monitoring tests the physical characteristics of bodies
of water such as temperature and the amount of suspended solids in the
water. Chemical monitoring tests for chemicals that may be present, such as
chlorine or ammonia, and metals, such as mercury. Biological monitoring
measures the health of aquatic communities and includes a variety of
techniques, such as assessing species' health and abundance. These
monitoring types and the additional pollutants or conditions that they
measure are described in figure 6.

Figure 6: Monitoring Types and Pollutants or Conditions That They Measure
States compare water quality monitoring data--chemical, physical,
biological, or a combination--with water quality criteria in order to
determine whether waters are meeting the standards. Officials in the states
we visited said that they tend to rely more heavily on one type of
monitoring test to make these assessments. For example, Illinois and Maine
rely primarily on biological monitoring to determine if rivers and streams
meet water quality standards, while Texas relies on chemical and physical
monitoring. This may cause differences in reporting on water quality because
as a study conducted by Ohio showed, different monitoring tests may yield
different results as to whether waters are of acceptable quality. The
study's assessment of 645 waters revealed that 50 percent of the waters met
chemical but not biological criteria.

Assessment results can also vary within a single type of monitoring. Many
different pollutants or conditions can be monitored within a type of
monitoring, but states generally do not monitor for all of them. Different
assessments of water quality in a single body of water can result if
different pollutants or conditions are monitored. An example given in EPA's
305(b) guidelines shows that in one body of water, toxicity levels were
within the standards, but the levels of certain metals exceeded the
standards. This is also the case for biological monitoring. Officials in
North Carolina told us that because the types of biological tests performed
vary from state to state, the resulting assessments should be compared with
caution. Maine and Illinois both use biological tests to assess rivers and
streams, for example, but their methods measure different aspects of the
biological health.

States Do Not Use Consistent Data Analysis Methods

States vary in how they use information on water quality to make use
assessments for individual waters, particularly the extent to which they use
monitored or evaluated data. For the 1996 305(b) report, 39 states used
evaluated data for assessing their rivers and streams, and 19 of these used
such data for half or more of their assessments. Seven states reported that
they did not use any evaluated assessments. Texas officials, for example,
told us they assess only those waters for which they have current,
site-specific data. Even though a majority of states do use evaluated
assessments when preparing their 305(b) reports, some of the same states do
not use these assessments as a basis for other decisions, such as whether to
place waters on their list of waters that do not meet standards, because
they consider the soundness of evaluated assessments to be questionable. For
example, officials in Illinois said that when evaluated assessments show
that a body of water does not meet standards, they first carry out
additional monitoring at that site to determine with greater certainty
whether the body of water is, in fact, below standards.

Further differences among states may arise as a result of whether and how
they use assessments from one section of a body of water to represent
conditions in other unassessed waters--a practice called extrapolation.
EPA's 305(b) reporting guidelines state that this practice is acceptable as
long as certain, similar conditions exist in the affected waters. Among the
states we visited, we found variations in how states extrapolate water
quality assessments. Officials in Texas do not use this practice at all,
while officials in Illinois, Maine, and North Carolina do. However, North
Carolina extrapolates only when assessments indicate that waters meet
standards, while Illinois extrapolates all assessments, whether they
indicate that waters meet standards or not.

States Use Inconsistent Methods to Determine Causes and Sources of Pollution

EPA's 305(b) reporting guidelines ask states to identify the
causes--pollutants or stressors, such as sediment, nutrients, and chemical
contaminants--that prevent a body of water from meeting standards. They also
are to identify the sources--specific activities or entities such as
agriculture, wastewater treatment plants, and industrial facilities--that
contribute pollutants to waters. EPA's guidelines provide general
definitions and list some of the types of information useful in determining
the sources of water quality impairment, including categories of pollution
sources, such as agriculture, storm sewers, and industrial point sources.
However, the guidelines lack specific instructions on the methods for
determining causes and sources.

The identification of causes and sources is an area in which officials in
several of the states we visited said there is considerable subjectivity. An
official in one state, for example, said that determining causes and sources
for lakes is generally based on the best professional judgment of the
monitoring staff members, and that while it may be a "reasonable guess," it
is nonetheless a subjective process. Officials in another state told us they
identify sources only if there are sufficient data to provide reasonable
assurance that the source identification is accurate. One state official
said that the reason for subjectivity in the identification of causes and
sources is because EPA has provided little guidance in this area.

In the 1996 305(b) report, EPA acknowledged that states do not assign
sources to all water quality problems identified. Despite this, EPA also
states in the report that it rates the significance of sources of pollution
by the amount of waters that states report is affected by each source. As a
result, when comparing one state with another, or when aggregating
state-reported causes and sources, some sources may appear to be less
significant than others, when these differences might actually be a result
of inconsistent or incomplete reporting.

Weaknesses

The flaws in the 305(b) report--incomplete data and differences in how
states monitor and assess their waters--greatly limit its utility for
national decisions and activities. Nevertheless, EPA uses the 305(b) report
for several national-level activities because it is the only source of
national information on whether waters are meeting water quality standards.
There is widespread agreement that many waters are still not meeting
standards and that addressing these problems will be challenging. What is
not clear is the full extent of these problems, where and what the most
severe problems are, and the location of high-quality waters that need to be
protected.

The Clean Water Act authorizes funds for grants to states to assist them in
administering programs for the prevention, reduction, and elimination of
pollution. These grants are the chief source of federal funding to states
for water quality management activities and are used to support the
monitoring and assessment of water quality and the regulation of pollution
dischargers, among other things.

The Clean Water Act indicates that these funds should be allocated to the
states on the basis of the extent of water pollution in each state. Prior to
1999, EPA used a formula that included point source factors, such as the
number of industrial dischargers and power plants in a state, as well as
other factors, such as population and land area. EPA revised this formula,
effective fiscal year 1999, to be more in line with the act's intent to
allocate grant funds on the basis of the extent of pollution and to improve
the quality and consistency of the data used for the allocation of funds.

One of the components in the new formula is the number of waters identified
in the 305(b) report as not meeting standards.8 To account for the fact that
not all of the formula components contribute equally to the extent of the
pollution problem within the states and to address concerns regarding the
current consistency of the 305(b) information, each component is weighted
individually. Component weights are being phased in over the course of 5
years, and final component weights are to be phased in by fiscal year 2004.
The 305(b) information was given a weight of 13 percent in fiscal year 1999
and will reach its maximum weight of 35 percent in fiscal year 2004. Also,
as a part of the new formula, EPA instituted mechanisms to routinely update
the data used.

Yet given the differences in the methods that states use to assess their
waters, the information from the 305(b) report may not reflect the relative
magnitude of water quality problems in states but, rather, assessment and
reporting differences across them. Moreover, EPA has no plans to ensure that
there will be greater consistency in the 305(b) reporting process by the
time that the formula component reaches its maximum weighting in 2004,
beyond encouraging states to adopt practices that would lead to more
comparable data. In response to our survey, half of the states reported that
they oppose using 305(b) data as a factor in allocating Clean Water Act
funds. A few states wrote specific comments stating that they oppose EPA's
use of 305(b) data in the formula because states' monitoring and assessment
methods are not comparable.

The Government Performance and Results Act of 1993 requires EPA and other
agencies to clearly define their missions, establish long-term strategic
goals (and annual goals linked to them) against which to measure their
performance, and report this information to the Congress. Importantly, the
statute emphasizes the need for agencies to focus on and achieve measurable
program results.

Some of EPA's performance measures are tied to the percentage of waters
meeting standards as identified in the 305(b) report. In EPA's fiscal year
2000 annual performance plan, a key strategic objective is for 75 percent of
the nation's waters to support healthy aquatic communities by 2005.9
However, the problems with the completeness and consistency of the data in
the 305(b) report undermine its usefulness in tracking progress toward this
goal. In fact, the agency acknowledged the problem in its fiscal year 2000
budget proposal, stating that "differing processes and methods among states
can result in varying depictions of the nation's water quality" and
concluding that "due to the manner by which data are currently collected,
305(b) data cannot be used to establish trends."

Information from the 305(b) report is often used to characterize conditions
in the nation's waters in statements asserting, for example, that a specific
percentage of the nation's waters do not meet standards. Given the problems
with the underlying data, however, the report does not provide a complete or
reliable estimate of the extent of water quality problems, the types of
these problems, and the identification of the primary causes and sources
nationwide. An EPA official acknowledged that EPA and other users of the
report do not always communicate that the 305(b) report is based on
assessments of a subset of the nation's waters.

Many states question EPA's use of the 305(b) report to represent water
quality conditions nationwide. Fourteen states responded that they oppose
the use of the report in this manner. While 30 states responded that they
support the use of 305(b), nearly half of these qualified their support,
commenting that consistency in state reporting needs to be improved and that
data should be held to higher standards.

EPA also uses information from the 305(b) report, among other information,
in its Index of Watershed Indicators (IWI), as presented through EPA's "Surf
Your Watershed" application, a map-based tool available on the Internet. EPA
designed the IWI to be used by individuals interested in learning about the
overall quality of the watersheds in which they live. The IWI is also
intended to provide a national baseline of the condition and vulnerability
of aquatic resources that, over time, could be used to measure progress
toward the goal that all watersheds be healthy and productive places.

The 305(b) water quality assessment information is 1 of 15 different
"layers" of information in the IWI, but it is weighted more heavily than
other layers in calculating an overall indicator of water quality for each
watershed. An indicator of water quality (such as good water quality and low
potential for problems) is provided for watersheds, which gives the
impression that conditions can be compared across watersheds. However, given
the considerable differences in the methods used to generate the underlying
data, such comparisons are not well founded. The IWI could be particularly
misleading, considering that most users are likely to be unaware that these
inconsistencies exist.

Nineteen states reported that they do not support EPA's use of 305(b) data
in the IWI. Several of them noted in particular that the IWI extrapolates
water quality assessments far beyond a level for which they were ever
intended. Officials in one state told us that when they reviewed the IWI
data for their watersheds, they found many errors, such as dams that were
located in the wrong watershed. In addition, several of the other 14 layers
in the IWI are also data sources for states' assessments of whether waters
meet standards. For example, physical-chemical data and fish consumption
advisories are often used by states to make their 305(b) assessments of
water quality. These data sources are also separate layers within the IWI,
so they are essentially double counted, as some states observed.

to Address Them

EPA recognizes that inconsistencies exist in how states monitor and assess
their waters and that these inconsistencies impair the 305(b) report's
utility for making comparisons across states. The agency has been working
toward a long-term goal of improving the accuracy and consistency of the
report. A key activity in this area is EPA's 305(b) consistency workgroup,
whose members include representatives from states, tribes, federal agencies,
and EPA regions and headquarters. EPA has incorporated several of the
workgroup's recommendations for improving the accuracy and consistency of
reporting into its guidelines for the 1998 report.10 The guidelines include
the goals of documenting and improving the quality of states' assessments
and the states achieving comprehensive assessment coverage of their waters.
However, because states are not legally required to adhere to EPA's
guidelines, there is no assurance that the states will carry out the tasks
necessary to meet these long-term goals.

To encourage consistency in states' 305(b) reporting, EPA developed an
information system called the Assessment Database for states to store and
transmit their water quality assessment data. This database is a relatively
new system that replaced EPA's prior data storage system. Although EPA
encourages states to use this database, it is not mandatory, and states are
free to use another system. One of EPA's goals for using an automated system
to store 305(b) data was for the agency to improve the quality and
consistency of 305(b) reporting. While there appear to be adequate controls
over the entry of data into the system, the considerable differences in the
methods that states use to generate the data going into the database
severely constrain the system's utility in improving the quality and
consistency of the data presented in the 305(b) report.

Water Quality Information in the National 305(b) Report

Providing information that is representative of water quality conditions
nationwide would require the use of statistical monitoring designs and
consistent assessment and reporting methods. Such an approach would require
changes to state and federal monitoring programs, and additional resources
would be needed to preserve states' water quality management programs while
also conducting more comprehensive monitoring.

Although the availability of additional funding is uncertain, other sources
of water quality information could be useful to EPA in compiling the 305(b)
report. According to EPA, 18 federal agencies--including the U.S. Geological
Survey, the National Oceanic and Atmospheric Administration, and EPA
itself--conduct 141 monitoring programs on regional or national scales. None
of these programs provide information that is representative of water
quality conditions across the nation. Instead, they focus on a particular
geographic area, a certain type of body of water, or specific pollutants;
studies may also have other discrete purposes. They differ from states'
assessments of water quality in that the states focus on water quality in
relation to state-specific water quality standards, whereas these programs
focus on broader program objectives. The individual programs generally have
monitoring and assessment standards and procedures to ensure that consistent
and comparable data are collected within a particular program.

EPA's Environmental Monitoring and Assessment Program, for example, is a
research program that develops the tools necessary to monitor and assess the
status and trends of national ecological resources. This program has
undertaken statistically based monitoring projects to report on ecological
conditions in several regions around the country. The Geological Survey's
National Stream Quality Accounting Network currently focuses on water
quality in four of the nation's largest river basins. The National Water
Quality Assessment Program, also administered by the Survey, is a more
comprehensive program, investigating water quality in major river basins
that cover about one-half of the land area of the conterminous United
States. This program seeks not only to determine water quality status and
trends but also to identify and explain the major causes of observed
conditions and changes. The National Oceanic and Atmospheric
Administration's National Status and Trends Program is an effort to
determine the extent and impact of contaminants on coastal and estuarine
areas throughout the nation through monitoring conducted jointly by federal
and state partners.

EPA included information on some of these programs in the 1996 305(b)
report, such as responsible agency, program objectives, scope of coverage,
and, in some cases, program findings. However, it did not aggregate the
findings into a national assessment of water quality because of
inconsistencies in the methods used across the programs, nor did it
integrate the findings into the 305(b) report to support or explain
state-reported information. If this information were assembled and
comprehensively analyzed, however, it could provide useful insights into
water quality conditions across the country and complement individual state
reports.

EPA considers the 305(b) report to be an essential tool for reporting on the
condition of the nation's waters and on the states' progress in meeting the
goals of the Clean Water Act. The agency also uses information from the
report to support certain national-level decisions. However, for a variety
of reasons, the report does not portray nationwide water quality conditions.
Of particular concern is that the report's results are based on more limited
monitoring data than is implied by its presentation, and the report
aggregates data from states that have different water quality standards and
use widely different monitoring and assessment methods. We believe these
weaknesses seriously limit the report's utility for the activities and
decisions for which it is presently used.

EPA has acknowledged some of the report's flaws and has taken steps to
improve it, including the establishment of a working group to promote
greater consistency of states' reporting and the issuing of revised
reporting guidelines. However, the prospects for significantly improving the
report are limited in the near term by a number of practical barriers,
including the limited resources among states to monitor a greater percentage
of their waters. In the long run, it will be up to the Congress and the
administration to determine whether the public's interest is better served
by devoting higher priority to monitoring activities. In the meantime,
however, we believe EPA should undertake a more holistic approach in
addressing the report's current deficiencies.

We recommend that the Administrator of EPA take actions to more
appropriately characterize state-reported information and more effectively
use other available water quality data in the National Water Quality
Inventory. Specifically, the Administrator should do the following:

ï¿½ Identify other ways of aggregating state-reported information that take
into consideration the inconsistencies between states' programs in order to
minimize the potential for misuse and misinterpretation.

ï¿½ Better convey the report's limitations in characterizing the nation's
water quality by more clearly identifying (1) the percentage of waters that
are actually monitored and the resulting assessments of quality and (2) the
similarities and differences among states' water quality standards and
monitoring and assessment methods, and the implications of these differences
on nationally aggregated information.

ï¿½ Supplement the current information in the Inventory by integrating, where
appropriate, the findings from other programs to support or explain
state-reported information. These programs include EPA's Environmental
Monitoring and Assessment Program, the U.S. Geological Survey's National
Water Quality Assessment Program, and others.

ï¿½ Reexamine the implications of relying on the report's data for use in
national decisions and communications to the public. Specifically, the
Administrator should do the following:

ï¿½ Evaluate the potential impacts of inconsistent data on the extent of
pollution in states on the allocation of the Clean Water Act's section 106
grant funds to states, determine what data are needed to reliably support
allocations, and develop a plan for ensuring that such data will be
developed and used.

ï¿½ Identify the information that EPA will use for establishing a baseline of
water quality conditions and measuring progress toward the goal of having 75
percent of U.S. waters supporting healthy aquatic communities and develop a
plan to ensure that reliable information will be obtained and used.

ï¿½ Evaluate the implications of using these data to generate an indicator of
the quality of watersheds in the Index of Watershed Indicators and take any
corrective actions needed to reliably represent water quality conditions.

EPA shares our concern that a majority of the nation's waters have not been
monitored. The agency notes, however, that states can be taking actions
stemming from the data they have and that it would be inappropriate to wait
until "perfect data" are collected on all waters before initiating actions
on waters with known problems. We do not intend for the report to convey the
message that actions cannot be taken to address known pollution problems in
the absence of complete data. In fact, we address this issue by discussing
states' opinions that they can take steps to improve water quality without
complete data, for example, by implementing pollution control strategies for
nonpoint sources. While states needn't be inactive with the data they have,
we continue to believe that it is important to pursue the collection of more
complete data in order to make more informed and defensible decisions.

EPA recognizes the need for more complete and reliable data to ensure that
decision-making and resource allocations are better supported and states
that it is working hard to achieve that goal. EPA suggested that we
acknowledge additional activities that the agency feels will help improve
water quality data--the modernization of its primary system for storing
water data and work done through the National Water Quality Monitoring
Council. In addition, EPA states that although we presented concerns
regarding the Index of Watershed Indicators, the agency believes it is a
very useful tool because, for the first time, multiple data layers are
integrated to characterize water quality. We recognize that EPA has many
efforts under way to improve water quality monitoring and associated
activities. Some of these efforts, however, do not specifically address the
problems we identified regarding the completeness and consistency of state
data but instead are focused on broader issues. For example, while
modernization of EPA's primary system for storing water data should help
states store, access, and maintain water data, only one state commented that
data storage and retrieval is a limiting factor in its ability to conduct
more comprehensive assessments or carry out other related activities. In
addition, while we agree that the Index of Watershed Indicators provides a
powerful presentation of water quality information, given the many concerns
with the underlying data, we believe it is important for EPA to investigate
the potential implications of data limitations on the information presented.
Simply providing information in an easy-to-use format is just half the game;
the information should be reliable as well.

EPA states that our report should more strongly note that the national
305(b) report is, by law, a summary of states' assessments of whether their
waters are meeting standards and "is not intended to be a national water
quality report on a selected, consistent set of water quality parameters."
We recognize that section 305(b) of the Clean Water Act specifically
provides that the 305(b) report include states' assessments of water quality
in terms of state standards and EPA's analysis of them, as well as other
information, and have stated this in our report. However, the broader
question that the Subcommittee posed to us asked how well the 305(b) report
characterizes water quality conditions nationwide. As such, our report
discusses the utility of the Inventory for characterizing water quality
nationwide.

Moreover, because the 305(b) report is used to help make national policy and
regulatory decisions, we believe it is prudent for EPA to take steps to
ensure that the information presented is as reliable as possible. Where it
is not, EPA should clearly explain the reasons why and assess the possible
implications on national decision-making. For example, as EPA points out in
its letter, a major use of the 305(b) report is the allocation of grant
funds to states, which is to be based on the extent of pollution in states.
However, given the lack of complete information on water quality conditions
and differences among state programs, EPA cannot be certain that the extent
of pollution across states is accurately reflected in the report.

EPA points out that past 305(b) reports have clearly explained that its
information does not consistently characterize nationwide water quality, and
that it has continued to highlight this point in its draft 1998 305(b)
report. We recognize that EPA has included statements in past reports to
advise readers that the data should be used with caution because of
differences in state water quality standards and monitoring and assessment
methods. However, by aggregating the data across states, EPA is implicitly
suggesting that these data can, in fact, be compared and in doing so is
increasing the likelihood that the data will be misused or misinterpreted.
For example, the statistic in the 1996 305(b) report that agriculture
contributes pollution to 25 percent of the assessed river miles that do not
meet standards represents an aggregation of data from states using different
assessment methods. While EPA states that this information should be used
with caution because of differences across states' programs, users have no
context from which to interpret the information because the common baseline
is river miles−not states, where the differences occur.

Regarding our recommendation that EPA make greater use of water quality
information from other sources, EPA states that it has included information
from other studies in past 305(b) reports and plans to do so for the 1998
report. We acknowledge that EPA has included information from other programs
in past reports. However, for the most part, this has been limited to a
discussion about a specific program and its findings. We intended our
recommendation to take this a step a further and correlate findings, where
appropriate, with state data. Including other sources of information in the
305(b) report to support and/or explain state-reported information may be
helpful because the report is often criticized for being of questionable
quality. We have clarified our recommendation to indicate that EPA should
integrate findings from other studies, where appropriate, to supplement
state-reported information. EPA also pointed out that it has encouraged
states to use data from other sources for their assessments for the 305(b)
report. We recognize this and have added information from our survey on
states' use of data from other sources.

Finally, EPA commented that apparently, we did not obtain a complete
description of the process used to develop and evaluate the agency's new
formula for allocating grant funds to states under section 106 of the Clean
Water Act. The agency stated that it conducted extensive data searches and
data source evaluations before selecting the Inventory as a source of
information on the extent of water pollution in each state. In addition, EPA
said that it evaluated numerous data scenarios to determine whether certain
data components unduly influenced the distributions of grant funds to
states. We incorporated a number of technical clarifications that EPA
provided, as appropriate. However, according to an EPA program official,
these scenarios only assessed the impact of using different data elements in
the formula and did not assess how differences among state monitoring and
assessment methods may affect formula results, as we recommend.

States Have Insufficient Data for Several Key Activities to Manage Water
Quality

Data limitations extend well beyond the problems associated with the
National Water Quality Inventory and that report's ability to characterize
water quality on a nationwide basis. States need water quality data to make
critical regulatory decisions, as well as decisions on where to focus
limited resources to obtain the most water quality improvement for the
dollar. States' responses to our survey show that data gaps limit states'
abilities to carry out several key activities to manage water quality.

As figure 7 illustrates, few states have sufficient data to assess all their
waters for pollution problems. Even for the waters that have been assessed,
most states reported that they did not have the data they need to determine
if the waters meet standards. Where water quality problems have been
identified, many states reported that they have a majority of the data they
need to deal with those caused by point sources because much of the last 27
years of the Clean Water Act's implementation has focused on such sources.
The picture is bleaker, however, for states' abilities to deal with nonpoint
sources. Very few states have sufficient data to identify nonpoint sources
of pollution and to develop total maximum daily loads for waters that do not
meet standards due to such sources.

Figure 7: States With a Majority of the Data Needed for Activities to Manage
Water Quality
Standards

States report that they lack the data needed for two activities essential to
the process of managing water quality: comprehensively assessing all state
waters and compiling a list of waters that do not meet standards. EPA
recognizes that these data limitations may prevent states from fully
addressing their polluted waters and is encouraging states to take steps
toward more comprehensive monitoring.

Only six states responded to our survey that they have a majority of the
data they need to fully assess all their waters. This response is consistent
with the relatively low percentage of waters that states reported assessing
for the 1996 305(b) report. For that report, for example, states assessed 19
percent of the nation's rivers and streams and 40 percent of the lakes and
reservoirs. The thorough assessment of state waters is essential because it
enables states to identify polluted waters and prioritize them for
remediation through the development of TMDLs or other means.

Despite not having assessed some of their waters, the state officials we
interviewed said they feel confident that they have identified most of their
serious water quality problems. States tend to focus their monitoring on
waters with suspected pollution problems and areas used by the public in
order to direct scarce resources to areas that could pose the greatest risk.
However, studies that have more thoroughly monitored water quality
conditions--either through monitoring previously untested waters or
conducting different types of monitoring tests--have identified additional
pollution problems. For example, a 1993 EPA-funded study of toxins showed
widespread elevated levels of mercury in Maine's lakes. This finding ran
counter to the expectations of Maine's water quality officials, who had
assumed that these waters were likely meeting standards because they are in
areas with little or no human activity. As a result of these findings, Maine
issued advisories against the consumption of fish for all the state's lakes.
Several New England states with presumably clean lakes considered their
waters' vulnerability to mercury and also posted fish consumption warnings.
In addition, a study conducted by Ohio's environmental protection agency
found that using additional types of monitoring tests identified a
significant number of pollution problems in waters that had been shown by
other monitoring efforts to be meeting standards.

Even though the state officials we interviewed are confident that they have
identified their most serious pollution problems, they nonetheless
acknowledge that more thorough monitoring would likely reveal additional
waters that do not meet standards. Likewise, 41 states reported in our
survey that increasing the percentage of state waters that are monitored
would be very helpful in carrying out their responsibilities for managing
water quality.

Data limitations also affect states' abilities to accurately compile for
submission to EPA a list of the waters that do not or are not expected to
meet standards. States must develop TMDLs for waters on this list. Less than
half the states reported that they have a majority of the data they need to
determine whether they should list waters they have assessed. State
officials said that their inability to make a listing decision stems, in
part, from the fact that they are less confident of the accuracy of some of
their assessments because they are based on evaluated data. Evaluated data
include site-specific monitoring data more than 5 years old and information
that serves as an indicator of potential water quality conditions, such as
land use or reports on wildlife or habitat conditions. EPA and state
officials acknowledge that some of these data sources are less reliable than
current, site-specific monitoring data. Several state officials told us that
while they may use this information to make an assessment of water quality
conditions for the 305(b) report, they prefer not to use it for making
decisions about whether to list these waters because of the requirement to
develop a TMDL once a water is listed. State officials said that they prefer
to conduct additional monitoring of these waters to determine whether they
are meeting standards.

While state officials allow that they may not have identified some waters
that need TMDLs, they also told us that some waters that have been listed do
not need TMDLs. The reasons for this varied widely among states. For
example, officials in one state said that they mistakenly assessed some
waters against higher standards than necessary; consequently, these waters
were placed on their list of polluted waters. In another state, officials
told us that about half of the waters on their list were placed there on the
basis of evaluated data. Upon additional monitoring of these waters, the
state found that many meet standards and, therefore, do not need TMDLs.

EPA officials are also concerned by the states' lack of comprehensive
monitoring. Officials told us that data limitations may cause some water
quality problems to go undetected and existing high-quality waters to be
degraded because actions were not taken to protect them (the Clean Water Act
requires states to protect existing good water quality). One regional
official believes that states would identify many more waters that do not
meet standards if more comprehensive monitoring were done.

EPA has been encouraging states in its guidelines for developing the 305(b)
report to move toward more comprehensive assessments of their waters. EPA
discusses the use of targeted monitoring as well as statistically based
monitoring, which would allow inferences to be made about entire categories
of waters, to achieve this goal. The guidelines also suggest that states use
a rotating basin approach to help achieve more comprehensive assessments. A
rotating basin approach essentially divides a state into major river basins
and provides for data collection in selected basins each year. The
monitoring rotates among the basins so that waters in all basins will be
monitored on a periodic basis, such as every 5 years.

While many states are incorporating a rotating basin approach into their
water quality management programs (see fig. 8), simply using this approach
does not ensure that comprehensive data for statewide water quality
conditions will result. The comprehensiveness of the data depends on the
type of monitoring conducted within the rotating plan. An EPA regional
official said, for example, that many states focus on the same monitoring
locations within their rotating basin plans because of resource constraints
and, therefore, do not significantly increase their monitoring coverage.
Several states are, however, incorporating some statistical monitoring in
order to capture a picture of water quality for waters that have not
typically been monitored in the past.

Figure 8: Number of States With a Rotating Basin Plan in Place or Under
Development
Source-Related Activities

In the waters that states have identified as not meeting standards, many
states reported that they have a majority of the data they need to carry out
management activities for problems that are caused primarily by point
sources. States generally have better data for point source problems because
much of the last 27 years of Clean Water Act implementation has focused on
addressing point sources of pollution through pollution control programs. In
addition, much of EPA's guidance on developing TMDLs, which dates back to
the 1980s, has focused on point sources.

Forty states reported having a majority of the data they need to identify
point sources of pollution. Point sources are easy to identify because they
generally discharge pollutants into bodies of water from some discrete
conveyance, such as a pipe. Accordingly, it is relatively easy to determine
the point sources that are contributing to water quality impairments.

In order to develop a TMDL for a pollution problem caused by point sources,
states must identify how much of a particular pollutant the sources
contribute to a body of water. This information is either already collected
as a condition of a discharge permit or is directly measurable at the
facilities' discharge point(s). Therefore, it is also relatively easy to
identify how much pollution the individual sources are contributing.

Twenty-nine states reported having a majority of the data they need to
develop TMDLs for water quality problems caused primarily by point sources.
In fact, officials in the states we visited told us that they have been
following the TMDL process for point sources for many years as a way of
achieving water quality standards and developing appropriate discharge
limits. These officials said, however, that EPA's review and approval of
TMDLs was not a priority until the recent onset of TMDL lawsuits.

Nonpoint Source- Related Activities

While water quality officials responding to our survey reported having a
majority of the data they need to deal with point sources of pollution, few
respondents reported having a majority of the data needed to deal with
nonpoint sources. Specifically, almost all states reported that they did not
have a majority of the data they needed to identify nonpoint sources of
pollution and develop TMDLs for waters polluted primarily by nonpoint
sources.

Forty-seven states reported that they have half or less of the data they
need to identify nonpoint sources of water quality problems, and 29 of those
states reported having much less than half or almost none of the data
needed. Information on nonpoint sources is particularly important because
there is general agreement that most remaining water quality problems are
caused, at least in part, by nonpoint sources. For example, 35 percent of
the waters that Texas officials identified as not meeting standards in 1998
were polluted solely by nonpoint sources, as compared with 5 percent that
were polluted solely by point sources. Slightly more than half of Texas's
reported water quality problems were caused by a combination of point and
nonpoint sources. Similarly, in Maine, nonpoint sources contributed
pollution to about half of the state's rivers and streams that were
identified as not meeting standards.

Because of the diffuse nature of nonpoint sources, it is difficult to
identify specific sources contributing to a particular water quality
problem. Unlike point sources, where pollutant contributions can be directly
measured as they come out of a pipe, nonpoint source pollution may come from
many disparate sources. For example, rainwater may carry fertilizer, manure,
pesticides, and soil with it as it runs off of farm fields into bodies of
water. Urban sources may contribute oil and grease, animal waste,
fertilizer, and pesticides as rainwater runs off of literally thousands of
individual residences, businesses, and roads into nearby streams, or storm
drains that discharge into streams or rivers.

Beyond the problems with simply identifying the nonpoint sources causing
waters not to meet standards, it is difficult to directly measure pollutant
contributions from individual nonpoint sources and, therefore, assign
specific loadings to sources in order to develop TMDLs. Accordingly, only
three states reported having a majority of the data they need to develop
TMDLs for water quality problems that are caused primarily by nonpoint
sources. Dealing with water quality problems that are caused by a
combination of point and nonpoint sources is difficult because progress can
be impeded by the weakest link--the inability to identify and quantify
nonpoint source contributions.

Developing TMDLs generally involves data collection and analysis beyond what
is done by routine water quality monitoring, especially for nonpoint source
problems. An EPA study of 14 TMDL development efforts--all but one of which
included nonpoint sources of pollution--found that each entailed additional
data collection that averaged about 40 percent of the total cost of
developing the TMDL.11 Officials in one state said that it takes about 3 to
4 years to do the data collection, analysis, and other activities needed to
prepare a TMDL for EPA's review. Officials in another state told us that
because they lack the data for certain TMDL projects and they are uncertain
of what level of data EPA will accept for a TMDL, they are focusing on TMDLs
that are relatively easy to develop rather than those that are the highest
priority. These officials said this is due to the pressure they feel to show
that they are making progress on TMDL development.

Several state officials told us that because most of the TMDLs they must
develop are for pollution caused by nonpoint sources, they prefer to use
methods that require less initial data collection prior to the
implementation of pollution control strategies. Two-thirds of the state
officials responded in our survey that using a phased TMDL approach--a
process recognized in EPA's current guidance--is very helpful for addressing
pollution problems. The state officials whom we spoke with said that a
phased approach enables them to apply best management practices to nonpoint
source activities identified as contributing to a problem, while at the same
time, gathering additional monitoring data to better understand the relative
contributions of sources.12 Several state officials said they see this as a
way to more quickly address water quality problems, rather than studying the
problem extensively before taking any remediation actions.

As states develop more TMDLs, the need for additional water quality
monitoring will increase accordingly. Current EPA guidance and proposed TMDL
regulations require that states conduct monitoring after pollutant controls
or other activities are implemented to determine if the TMDL is working and
the body of water is attaining water quality standards. This means that
significant new monitoring efforts will be needed, particularly for TMDLs
addressing nonpoint sources of pollution. Nonpoint sources are largely
addressed through the use of best management practices, which are generally
changes in the behaviors of how individuals use land. The effectiveness of
best management practices on improving water quality varies on the basis of
site-specific conditions, such as soil type and climate. Therefore,
additional monitoring is needed to see if the practices are having the
desired effect of improving water quality.

Carry Out Key Activities

Several factors were repeatedly identified by states as limitations to their
ability to conduct monitoring and analyses to fully address their water
quality management needs. Almost all states identified a need for additional
resources, such as funding and staff, to carry out their duties. There was
also strong evidence of the need for additional analytical methods and
technical assistance to help states analyze complex pollution problems and
develop TMDLs.

The limitation most often cited by states was that the amount of
resources--in terms of funding and staff--devoted to monitoring, assessment,
and TMDL development is far short of the task at hand. Forty-five states
reported that the lack of resources was a key limitation to making more
progress on water quality issues. In addition, several states pointed out
that they are operating under a state-imposed staffing level ceiling, and
other states said they are limited in how many samples they can analyze
because of shortages in lab funding.

EPA water officials told us that overall, less resources are being devoted
to monitoring and assessment at the state level than ever before. In
addition, EPA and state officials told us that more requirements have been
placed on state programs for managing water quality, such as issuing permits
to implement new storm-water and sludge regulations. Of particular note,
officials in two states said that the increased focus on addressing waters
that do not meet standards is constraining their abilities to carry out
monitoring and assessment activities.

EPA is conducting a study of funding shortfalls in states' water quality
management programs and plans to identify alternative approaches for
addressing the anticipated gap. The agency plans to finalize its methodology
for estimating these shortfalls in the spring of 2000. On the basis of a
preliminary analysis of 10 state programs, EPA found that states have
shortfalls in most areas of water quality management, including water
quality monitoring and TMDL development programs.

Assistance to Help Develop TMDLs

To assist with the development of TMDLs, EPA has issued multiple guidance
documents since 1984. These, however, have largely focused on point source
TMDLs. EPA has also been developing a watershed model and analysis tool
called BASINS, which is intended for use in the development of TMDLs,
particularly for the analysis of nonpoint sources.13 The model was released
about 3 years ago, although its capabilities have been evolving over that
time. EPA is working on simplifying the model in order to make it easier to
use. EPA has also provided a compendium of models--including those developed
by other agencies and organizations--that are available for states to use in
analyzing pollution problems.

In response to our survey, however, a majority of states indicated that they
need additional assistance for TMDL development. Specifically, 31 states
reported that additional technical tools, such as models and analytical
methods, would be very or extremely helpful for TMDL development; 28 states
reported that additional technical assistance in these areas would also be
helpful. According to our survey, states are most concerned about their
abilities to develop TMDLs for nonpoint sources. Analyzing these sources is
much more complex than the analysis of point source impairments and often
relies on watershed models. However, several states pointed out that they
are on the front end of the learning curve when it comes to using watershed
models such as BASINS and developing TMDLs for nonpoint sources of
pollution. While state officials told us they have in-house expertise with
models used for analyzing point sources, they need assistance in selecting
and using appropriate watershed models for nonpoint sources.

Some state officials told us that they are unsure of where to go for
technical assistance or that experts they called on in the past were no
longer available. For example, officials in two states told us that they
previously obtained model development, troubleshooting, and refinement
assistance from experts in EPA's modeling lab in Athens, Georgia. This
assistance, however, is no longer available because of reductions in
funding, according to an official in EPA's TMDL office. Moreover, this
official told us that there is no systematic process for providing
assistance to states on the development of TMDLs. He said that assistance is
provided largely in an ad-hoc fashion by EPA staff in headquarters, regions,
and labs. EPA regional officials identified several areas where they have
provided states with assistance on TMDL issues. However, EPA does not keep a
national inventory of state requests for assistance and what is done to
address those requests.

Some states suggested that EPA should develop sample or standardized
approaches, such as templates that states could use to guide them through
certain types of TMDLs. In addition, several states pointed out the need for
efficiency in developing TMDLs. For example, one state noted that states
should be benefiting from others' experiences in developing TMDLs, rather
than "reinventing the wheel."

Activities currently under way at EPA could help states in some of the areas
identified above. Perhaps most directly relevant to states' needs are EPA's
efforts to develop protocols for some of the more common pollutants causing
waters not to meet standards--sediment, nutrients, and pathogens. (According
to EPA, these pollutants were selected because the states report that they
are the leading causes of water quality impairments.) The protocols are
intended to provide states with an organizational framework for completing
the technical and programmatic steps in the development of TMDLs for
specific pollutants. EPA issued the sediment and nutrient protocols in
October and November 1999, respectively. These documents appear to provide
some of the information and specific guidance that states identified as
needed, such as the suggestions for the kinds of data and analyses necessary
to develop specific TMDLs. How helpful the protocols are will become clearer
after they have been used in several TMDL development efforts.

In addition, EPA is conducting two pilot studies to investigate the
relationship between air emissions of mercury and water quality impacts. The
goal of the pilots is to examine the methods for taking air sources into
account when determining TMDLs. In each of the two study areas--one is in a
small lake in Wisconsin and the other covers a 700-square-mile area of the
Florida Everglades--techniques will be evaluated for determining (1) the
amount of mercury reductions needed to meet water quality standards, (2) the
relative contributions of mercury from various sources, and (3) the
geographic extent of sources contributing mercury. A legal analysis of
federal and state programs to address airborne sources of mercury deposited
in bodies of water is also being conducted. EPA plans to issue a "lessons
learned" report on the findings of the pilots in the spring of 2000 to
assist states in the development of TMDLs involving mercury from air
sources.

EPA officials told us that they have also started a library of approved
TMDLs to help facilitate information sharing among states. Currently, the
library consists of a manual file of approved TMDLs and is maintained in
headquarters. EPA regions receive a list of the TMDLs in the library and are
to use it for referring states to specific TMDLs when they are working on
similar water quality problems. In addition, EPA regions are in the process
of assessing states' TMDL programs to help develop regional strategies for
deciding how best to allocate available resources to support them. These
assessments will include examining states' programs for water quality
monitoring, prioritizing water quality problems, and developing and
implementing TMDLs to identify areas needing additional resources.

EPA recognizes, however, that overall, its research and tools are not as
fully developed as they need to be to support TMDL development. Therefore,
the agency initiated the development of a strategy in the summer of 1999 to
identify the critical needs for developing scientifically credible TMDLs and
the actions necessary to address these needs. According to EPA, states and
other stakeholder groups have provided input into the strategy. EPA has
tentatively identified three broad areas in which it will address needs--the
capacity to develop TMDLs, tools to support TMDL decisions, and the
management of TMDL-related information. However, EPA staff working on the
strategy have been diverted to work on other issues. The official in charge
of developing the strategy told us that he hopes work can resume on the
strategy in order to release a draft for public comment in the summer of
2000.

While it is encouraging that EPA has initiated the development of an overall
strategy to support TMDL development, until complete, it is uncertain
whether states' needs for additional assistance will be addressed. The
regional assessments of states' programs, as well as states' requests for
assistance from EPA headquarters, regions, and labs, could both provide
valuable insight into the states' needs as EPA finalizes its strategy.

While states have always needed to obtain comprehensive and reliable water
quality data, such data have become more important in recent years. An
increased emphasis on regulatory requirements for assessing waters and
developing TMDLs has heightened the need for states to have sound
information to make these decisions. Of particular concern, states will be
required to develop thousands of TMDLs for their most heavily polluted
waters.

Our survey results show that states are missing much of the information they
need to comprehensively assess their waters and develop TMDLs for some of
the waters they have already identified as not meeting standards. They have
said they need additional resources, tools and assistance in developing
TMDLs for thousands of their waters--a task that will significantly tax
already limited resources over a sustained period of time. EPA's analysis of
funding gaps in states' programs should shed light on the resources needed
to conduct these and other activities. It will then be up to the
administration and the Congress to determine whether and how these needs
will be met.

EPA has a number of targeted initiatives under way that are designed to help
states deal with specific problems, such as providing protocols for
pollutants that frequently cause water quality problems and developing
information on airborne pollutants. In addition, EPA has initiated work on a
strategy for supporting TMDL development, although it is not certain when
this strategy will be finalized and what kind of support will be provided.
As states face long-term challenges in carrying out important regulatory
decisions, it is critical that EPA finalize its strategy and take actions to
provide the support necessary to develop scientifically defensible TMDLs as
efficiently as possible.

We recommend that the Administrator, EPA, finalize the agency's strategy for
supporting TMDL development. As a part of this process, EPA should continue
working with states and other stakeholder groups, as well as gather
information on states' requests for technical assistance from headquarters,
regions, and labs to obtain the most complete picture of states' needs as
possible. We also recommend that as a part of its overall strategy, EPA
establish a process for systematically tracking states' requests for
technical tools and assistance and how these requests are addressed, on a
routine basis, to ensure that it is addressing needs efficiently and to stay
abreast of changing needs.

To facilitate more efficient information sharing regarding TMDL development,
we recommend that the Administrator, EPA, expand the dissemination of its
TMDL library. A first step should be for EPA to routinely send the list of
TMDLs in the library to states, as well as to regions and to consider
automating the library to facilitate more efficient access to the
information.

EPA commented that our draft report did not discuss states' lists of waters
that do not meet standards, which serves as the link between the 305(b)
report and the TMDL program, and the adequacy of state data for the list
development process. We agree that this is an important issue and have added
a discussion of this to our report. EPA also provided several clarifications
on the TMDL program that we have incorporated, as appropriate.

Copy of GAO Questionnaire

Comments From the Environmental Protection Agency

GAO Contacts and Staff Acknowledgments

Steve Elstein, Assistant Director, (202) 512-6111

Gillian Friedrichs, Heather Halliwell, Barbara Johnson, Karen Keegan, Trish
McClure, Luann Moy, Dena Owens, and Judy Pagano made significant
contributions to this report.

(160480)

Table 1: Monitoring Designs Described in EPA's Water Quality
Reporting Guidelines 28

Figure 1: States With a Majority of the Data Needed for Activities
to Manage Water Quality 7

Figure 2: Percentage of Waters Monitored, Evaluated, and Not
Assessed 9

Figure 3: Examples of Point and Nonpoint Sources of Pollution 17

Figure 4: Steps in Managing Water Quality 19

Figure 5: Percentage of Waters Monitored, Evaluated, and Not
Assessed in the 1996 305(b) Report 26

Figure 6: Monitoring Types and Pollutants or Conditions That They Measure 31

Figure 7: States With a Majority of the Data Needed for Activities
to Manage Water Quality 44

Figure 8: Number of States With a Rotating Basin Plan in Place or
Under Development 48
  

1. 33 U.S.C. sections 1251-1387.

2. This list is commonly referred to as the 303(d) list.

3. The Federal Reports Elimination and Sunset Act of 1995, Pub. L. No.
104-66, 109 Stat. 707 (Dec. 21, 1995), provides that most statutory
requirements for annual reports by federal agencies to the Congress are
eliminated as of December 21, 1999. However, according to EPA, the states
are still required to submit reports on their waters every 2 years.
Accordingly, EPA officials told us they intend to continue to produce, in
some form, an analysis of the state reports and to use it to measure states'
progress and communicate to the public.

4. As defined in GAO's Assessing the Reliability of Computer-Processed Data,
data reliability exists when data are sufficiently complete and error free
to be convincing for their purpose and context.

5. For the purpose of this report, "states" refers to the 50 states and the
District of Columbia.

6. Two states did not specify whether their assessments were evaluated or
monitored, and three states did not report any assessments at all.

7. In 1998 and 1999, EPA's Office of Inspector General issued reports on
five states' water quality programs that, among other things, identified
weaknesses in state water quality standards and the regional review of state
standards. For example, some states had less restrictive criteria than EPA's
and some did not adopt all applicable uses. In addition, EPA regions were
not always reviewing state standards in the required time frames. In 1998,
EPA issued a plan to strengthen water quality standards programs for both
EPA and the states.

8. The other components are surface water area, ground water use, point
sources, nonpoint sources, and the population of urbanized areas.

9. In this context, "waters" refers to rivers and streams, lakes, wetlands,
estuaries, coastal areas, oceans, and ground waters.

10. According to EPA officials, the 1998 305(b) report is scheduled to be
issued in early 2000.

11. See TMDL Development Cost Estimates: Case Studies of 14 TMDLs, USEPA
(1996). One of the TMDL projects did not provide separate cost data on
additional data collection.

12. Examples of best management practices are (1) leaving strips of farmland
next to bodies of water uncultivated to minimize erosion and (2) using
manmade ponds or basins to detain storm-water runoff from roads to minimize
the velocity of water reaching bodies of water during storms and to allow
sediment and other pollutants to settle.

13. Watershed models are often used to analyze nonpoint source pollution
impairments because they can take into account many of the factors that
influence such pollution, such as land use, climate, and geographic
features.
*** End of document. ***