-------
NHEERL-COR-2369A
NHFFRT m* ^fiQA TECHNICAL REPORT DATA
Htuiiiiu, IAJK -ijoyA (Please read instructions on the reverse before complef—' PR9000.in?q44
1. REPORT NO. 2.
EPA. 600/A-99/096
4. TITLE AND SUBTITLE Going against the current: expanding the inland aquatic
monitoring culture of federal and state agencies.
7. AUTHOR(S) Anthony R. Olsen
9. PERFORMING ORGANIZATION NAME AND ADDRESS
US EPA NHEERL
200 SVV 35lh Street
Corvallis, OR 97333
12. SPONSORING AGENCY NAME AND ADDRESS
US EPA ENVIRONMENTAL RESEARCH LABORATORY
200 SW 35th Street
Corvallis, OR 97333
3- niiinuiiiiiiiiiiiinn
5. REPORT DATE
6. PERFORMING ORGANIZATION
CODE
8. PERFORMING ORGANIZATION REPORT
NO.
10. PROGRAM ELEMENT NO.
1 1 . CONTRACT/GRANT NO.
1 3. TYPE OF REPORT AND PERIOD
COVERED
14. SPONSORING AGENCY CODE
EPA/600/02
' 15. SUPPLEMENTARY NOTES:
16. Abstract: My interest concerns the design of lake and stream monitoring programs implemented by federal and state
agencies in response to the Clean Water Act. Past, and most current, monitoring designs select sites using judgement criteria.
Such designs provide valuable data for assessing impacts of point source discharges. The Clean Water Act also requires a
national assessment of all waters within the United States. The assessment is accomplished by compiling information obtained
from state monitoring programs. Such data do not provide a defensible assessment. An impetus for change began in the late
1980's. The U.S. Geological Survey and the U.S. EPA initiated two independent programs: the National Water Quality
Assessment Program (NAWQA) and the Environmental Monitoring and Assessment Program (EMAP). These programs use two
different scientific approaches.for monitoring. In addition, states have competing monitoring requirements. When the academic
community's scientific research perspective is added, the result is a natural adversity among the groups. The paper discusses
the underlying cultural conflicts associated with monitoring and present mechanisms that have been used to expand the culture,
especially with states.
1 7. KEY WORDS AND DOCUMENT ANALYSIS
a. DESCRIPTORS
Environmental monitoring, Survey design,
Social culture of organizations, Clean Water
Act, Probability sampling
18. DISTRIBUTION STATEMENT
b. IDENTIFIERS/OPEN
TERMS
ENDED
19. SECURITY CLASS
20. SECURITY CLASS
(This Report}
(This page)
c. COSATI Field/Group
21. NO. OF PAGES: 8
22. PRICE
EPA Form 2220-1 (Rev. 4-77) PREVIOUS EDITION IS OBSOLETE
-------
GOING AGAINST THE CURRENT: EXPANDING THE INLAND AQUATIC
MONITORING CULTURE OF FEDERAL AND STATE AGENCIES
Anthony R. Olsen, US Environmental Protection Agency
Western Ecology Division, 200 S. W. 35th Street, Corvallis, OR 97333
Key Words: Environmental monitoring, survey
design, social culture of organizations, Clean Water
Act, probability sampling,
ABSTRACT
My interest concerns the design of lake and stream
monitoring programs implemented by federal and state
agencies in response to the Clean Water Act. Past, and
most current, monitoring designs select sites using
judgement criteria. Such designs provide valuable data
for assessing impacts of point source discharges. The
Clean Water Act also requires a national assessment of
all waters within the United States. The assessment is
accomplished by compiling information obtained from
state monitoring programs. Such data do not provide a
defensible assessment. An impetus for change began in
the late 1980s. The U.S. Geological Survey and the U.S.
EPA initiated two independent programs: the National
Water Quality Assessment program (NAWQA) and the
Environmental Monitoring and Assessment Program
(EMAP). These programs use two different scientific
approaches for monitoring. In addition, States have
competing monitoring requirements. When the academic
community's scientific research perspective is added, the
result is a natural adversity among the groups. The paper
discusses the underlying cultural conflicts associated with
monitoring and present mechanisms that have been used
to expand the culture, especially with states.
Introduction
The United States interest in water quality began in
the late nineteenth century with a now obscure law called
the Rivers and Harbors Act of 1899. This law remained
the principal basis for protection of water quality until the
mid-twentieth century. The Water Pollution Control Act
of 1948 (P.L. 80-845) provided the first federal funds for
state water pollution control programs and started
subsidies for the construction of sewage treatment plants.
A key aspect was that all the details were left to the
states. This process continued under the Federal Water
Pollution Control Act of 1956 (P.L. 84-660) with
increasing commitments in 1961 (P.L. 87-88), 1965 (P.L.
89-234), and 1966 (P.L. 89-753). For the most part no
enforceable standards were imposed. In 1965 Congress
created the Federal Water Pollution Control
Administration and required the states to develop water
quality standards for interstate waters. Adlcr ct. al.
(1993) state that "even then, enforcers had to prove that
a particular polluter caused violations of these instream
standards - no small task given the primitive state of
water quality monitoring and science and the crowd of
dischargers to most polluted waters."
An event on June 22, 1969, dramatized the state of
water pollution at the time - the Cuyahoga River in
Cleveland, Ohio, burst into flames, fueled by oil and
other industrial wastes. In 1972, Congress passed the
Clean Water Act, overriding President Nixon's veto.
Significantly, the act begins with a statement of
underlying visions and goals. Congress declared, "The
objective of this Act is to restore and maintain the
chemical, physical, and biological integrity of the
Nation's waters." For the first time, Congress framed the
issue in terms of the entire aquatic ecosystem, not just
chemical pollution. The Act also insisted that we restore
and maintain aquatic ecosystems, and that we actively
protect waters that were currently clean. The Clean
Water Act was significantly modified in 1977 and 1987.
Section 305(b) of the Clean Water Act mandated that the
states monitor their waters and report their findings to
EPA. EPA was required to submit the findings in a
National Water Quality Inventory to Congress every two
years. Hence a basis exists in the Clean Water Act for
national statistics on the status of aquatic ecosystems.
Aquatic Monitoring 1960-1988
Monitoring of lakes and rivers during the period
1960-1988 can be characterized as follows. First, site
selection was dominated by judgment selection of fixed
sites. Since states were directly responsible for
monitoring, their monitoring programs reflected their
major priorities. Consequently, compliance monitoring
of point source discharges constituted the primary
objective of monitoring. Compliance monitoring initially
focused on monitoring water quality in streams or lakes
near point discharges. Subsequently, dischargers were
required to monitor effluent directly, prior to its
discharge. Regulatory monitoring related to drinking
water was also a focus. Public and private drinking water
facilities were required to monitor the chemical quality of
their finished water.
-------
Beginning in 1972, the states also were required to
provide reports to US EPA for its biennial report to
Congress on the status of aquatic waters in the United
States, informally termed the 305(b) report. The 305(b)
report provides estimates of the number of stream miles,
lake area, and estuarinc area that meet specific designated
beneficial uses. Designated uses include aquatic life
support, fish consumption, shellfish harvesting, drinking
water supply, primary contact recreation swimming,
secondary contact recreation, agriculture, ground water
recharge, and wildlife habitat. Section 305(b) of the
Clean Water Act requires that the states biennially survey
their water quality for attainment of designated beneficial
uses. EPA's Office of Water provides guidelines to
states on the information to be reported and also funds
monitoring by the states (USEPA 1995). The process
over the years can best be described as quasi-laissez-faire
with EPA providing general guidelines to States.
Consequently, the national report is a collection of
information that can not be easily compared across states
or from one report to the next. Adler et. al. (1993) report
that "the percentage of waters assessed varies wildly
among states", rarely is sampling done.
In 1987, the Clean Water Act increased its focus on
the National Pollution Discharge Elimination System
(NPDES) permitting process. These permits are
supposed to be based on a detailed total maximum daily
load (TMDL) and wasteload allocation, which take into
account information about all sources of a pollutant into
a given receiving waterbody. Section 303(1) institutes a
listing of all state waters that fail to meet water quality
standards or otherwise do not support designated uses.
The TMDL process requires monitoring data to support
its implementation and has been the cause of significant
monitoring activity by the states.
During this period, state agencies and US EPA
focused on monitoring the impacts of point sources to
aquatic systems, i.e., compliance monitoring. This period
can be characterized as a period of inadequate and
inconsistent monitoring at all levels. Many individual
point source discharges were adequately monitored for
contaminants.
Impetus for Change
In the late 1980s several changes in our approach to
monitoring began to take place in the United States. One
major change was an increased interest in the biological
or ecological condition of lakes and streams. Another
was the realization that it was not feasible to monitor all
lakes and rivers in the United States. This realization was
partly the result of states being faced with declining funds
with which to conduct monitoring.
Two major federal initiatives had their beginning in
the late 1980s: the National Water Quality Assessment
program (NAWQA) and the Environmental Monitoring
and Assessment Program (EMAP).
In 1991, the U.S. Geological Survey began the
NAWQA program. NAWQA provides consistent and
comparable information on water resources in 60 of the
largest and most important hydrologic systems in the
United States. It covers about one-half of the
conterminous United States and approximately 60-70% of
national water use. The program focuses on 20 units at
one time with each unit proceeding through three phases:
initial planning and retrospective analysis of existing data
(2 years), intensive data collection and analysis (3 years),
and report preparation and low-level assessment activity
(6 years). One third of the 60 units are in the intensive
phase at any given time. The objectives of NAWQA are
(Helsel 1995): "To describe current water-quality
conditions for a large part of the Nation's freshwater
streams, rivers and aquifers. To describe for those areas
how water quality is changing over time. To improve
understanding of the primary natural and human factors
that affect water-quality conditions." The stream
sampling design focuses on selecting sites to represent the
major land uses and physiographic regions. A
combination of fixed sites and synoptic surveys are used
to collect data appropriate to the building of relational
models of water-quality and environmental factors. The
models form the basis for generalizing the results to entire
geographic regions. No attempt is made to estimate the
stream length that meets specific designated uses or other
criteria.
EPA's Environmental Monitoring and Assessment
Program (EMAP) began in 1990 as an interagency,
interdisciplinary program that would contribute to
decisions on environmental protection and management
by integrating research, monitoring, and assessment.
EMAP's goal was to monitor and assess the condition of
the ecological resources of the United States. To
accomplish this goal EMAP proposed (1) to estimate the
current status, trends, and change in the extent of selected
indicators of the Nation's ecological resources on a
regional basis with known confidence and (2) to seek
associations among selected indicators of natural and
anthropogenic stresses and indicators of condition of
ecological resources. The intent was to provide annual
reports on regional status. (Messer et. al. 1991). The
program was designed to address the information needs
for EPA's relative risk paradigm (Thornton et. al. 1993).
-------
EMAP has evolved into a research program to develop
the tools necessary to monitor and assess the status and
trends of national ecological resources (USEPA 1997).
EMAP's research goal is to develop the scientific
understanding for translating environmental monitoring
data from multiple spatial and temporal scales into
assessments of ecological condition and forecasts of the
future risks to the sustainability of our natural resources.
A major part of the program is the conduct of large-scale
regional geographic studies. These studies, lasting
approximately five years enable researchers to investigate
new monitoring methods at these scales and to provide
monitoring data that are used in regional assessments of
the status of the ecological systems. Major emphasis now
is given almost exclusively to inland aquatic and
estuarine ecosystems and to landscape analyses.
In 1992 the Intergovernmental Task Force on
Monitoring Water Quality (ITFM) was formed to review
and evaluate water-quality monitoring activities
nationwide and to recommend improvements. The ITFM
issued a strategy for improving water-quality monitoring
in 1995 (ITFM 1995). It concluded that "it has become
clear that monitoring activities need to be improved and
integrated better to meet the full range of needs more
effectively and economically." The committee
determined that a new approach to monitoring is required
to target water-pollution-control resources to prioritize
concerns and to evaluate the effectiveness of actions
taken to prevent or remediate problems. In particular, a
better balance of ambient and compliance monitoring was
identified as a need. Their recommendations are
organized around the topics of working together, sharing
data, using comparable methods, designing monitoring
programs, and reporting findings. They identified
institutional and technical changes as necessary parts of
a "strategy for nationwide, integrated, voluntary water-
quality monitoring."
In July, 1995, the National Science and Technology
Council through its Committee on Environment and
Natural Resources created the Environmental Monitoring
Team with the charge "to develop a national framework
for integration and coordination of environmental
monitoring and related research through collaboration and
building upon existing networks and programs." The
team proposed a conceptual framework described by
CENR (1997). The framework proposed three levels of
monitoring: intensive monitoring and research sites,
national and regional resource surveys, and complete
coverage programs (mainly remote sensing based). At
each level existing programs were to be coordinated to
meet the requirements. Neither ITFM nor CENR
proposed the development of a new monitoring strategy
to replace existing, diverse, overlapping, and incomplete
programs.
A Preferred State for Aquatic Monitoring?
Aquatic monitoring encompasses many different
measurement activities on multiple types of aquatic
ecosystems. Ecosystems include lakes, reservoirs,
streams, rivers, estuaries, coastal waters, wetlands as well
as their associated riparian areas and upland watersheds.
Monitoring may be intensive (e.g., daily at a single site),
watershed (multiple sites with frequent measurements), or
national (e.g., large number of sites measured annually).
Aquatic monitoring involve only a single contaminant or
may encompass a wide suite of ecological measurements.
The ITFM report gives a general set of
recommendations for aquatic monitoring within the
United States. It does not propose a specific vision nor
specific organizational structure for monitoring. In that
sense, it falls far short of outlining a federal/state
environmental monitoring statistical framework.
Although a number of changes have been made in the
Clean Water Act, the basic requirements for monitoring
have not changed much during the last 25 years. Ambient
and compliance monitoring have always been part of the
Act. However, the historical focus has been on
compliance monitoring. Section 305(b) of the Clean
Water Act requires EPA to complete "a description of the
water quality of all navigable waters in such State during
the preceding year" and " an analysis of the extent to
which all navigable waters of such State provide for the
protection and propagation of a balanced population of
shellfish, fish, and wildlife and allows recreational
activities in and on the water." Ambient monitoring that
would enable quantitative, scientifically-defensible
estimates as required by Section 305(b) has never been
achieved. This difficult task requires either complete
coverage of all navigable waters or a probability survey.
Equally important is the requirement that monitoring must
include more than water column chemistry, contaminants
in fish tissue, and contaminants in sediment. Monitoring
must include water column physical properties,
biological measures, in-stream habitat, riparian habitat,
and upland (or watershed) habitat to meet the
requirements.
Prioritizing the monitoring effort is a daunting task.
As in any statistical reporting effort, priorities must be set
to meet the most important information needs of society
within the budgets available. No organized effort to do
this appears to be taking place. The National Water-
Quality Monitoring Council (NWQMC) was established
-------
as the successor to the 1TFM in 1997. It is charged with
enhancing the collaboration and coordination of water
resource quality monitoring at the national, state, tribal,
and local levels, as well as similar activities involving
business and industry, academia, agriculture, and
environmental groups. Its role is described on their web
page: http://h2o.usgs.gov/public/wicp/nwqmc.hlTnl. No
indication is given that a goal is to develop a federal/state
environmental statistics framework, except as it may
develop through voluntary cooperation. The terms of
reference for the NWQMC do not include any reference
to statistical survey design, or even the importance of site
selection in monitoring.
What would be a preferred state for inland aquatic
monitoring in the United. States? Should we ex-pect to
have a federal environmental statistical agency with data
collection, analysis and reporting responsibility such as
we have for national agricultural economic statistics,
economic statistics, transportation statistics? Would it be
cost-effective to centralize all inland aquatic monitoring
through an integration of monitoring conducting by a
single federal agency and a single agency within each
State? Should an entirely new monitoring effort be
planned, with the transition of existing programs into the
new program? How closely linked should monitoring be
to providing information for supplemental accounts tc the
National System of Accounts? At this time, I do not
believe a preferred state for monitoring exists.
Changing the Culture
Since 1990, several efforts mentioned above have
begun to change the culture of aquatic monitoring in the
United States. How have these efforts fared? What
resistance have they faced? Are they coordinated? Are
they making a difference? What is the prospect that the
United States will have a scientifically defensible national
water quality report in the near-future? I consider these
questions below under the topics of statistical
methodology, statistical education, and sociology of
organizations. My discussion focuses on statistical
aspects, while ignoring equally important areas of
ecological science, information management, and others.
Statistical Methodology
My discussion of statistical methodology will be
limited to site selection issues. Other areas are important,
such as response design (Stevens and Urquhart 1999) and
determining cause-effect. However, much of the change
in culture centers on site selection.
stream monitoring in the United States has been
conducted through fixed station networks. Note that our
use of the term lake will refer to both natural lakes and
reservoirs and that our use of the term stream will refer to
all streams and rivers fromsmall headwater streams to the
great rivers such as the Mississippi. Periodically, USEPA
has conducted national or regional probability-based
surveys of lakes and streams, but in most cases the
surveys targeted a subset of lakes or streams. To the best
of my knowledge, no state had integrated a probability-
based survey as part of their basic monitoring program
prior to 1990. The book used in a prominent monitoring
design course, taught at least annually, includes no
reference to probability-based survey design for site
selection (Sanders et. al. 1983). The authors state that
locating a site consists of selecting a river reach to
sample, locating a specific site within the reach, and
selecting representative locations on the cross-section at
the site. They focus on the hydrology of the stream
network and known locations of point sources as
important factors in selecting the reach and site within the
reach. Their network design approach does not appear to
have an objective to provide quantitative estimates of the
status of aquatic resources summarized for states or other
geographic delineations.
EMAP proposed a radically different approach to
sampling lakes and streams (Messer et. al. 1991). As
stated previously, a primary objective was to produce
quantitative starur and trend information over large
geographic regions. It was based on a two-stage
probability-based survey design. The first stage consisted
of a systematic area sample of 40 km2 hexagons. The
second stage constructed a list frame of all lakes or
streams within each hexagon. A complex unequal
probability weighting and spatial clustering algorithm was
used to spatially disperse the sample and select an
expected one site per hexagon (Overton et. al. 1990). The
entire design used a newly created discrete global grid
(White etal 1992). EMAP attempted to communicate the
monitoring design to the monitoring and ecological
community but with limited success. It was too much of
a cultural change for the current way of doing business.
Rather than recognizing the overall benefits of EMAP
survey design, the details of the design became the focus.
The statistical community also expressed concern about
the design and how well it would meet the needs of
EMAP. Some of their concerns were justified. The
design required additional research on statistical methods
before it could be supported theoretically. Since then, the
necessary theoretical foundations for probability-based
With rare exceptions, the entire history of lake and
-------
survey designs of lakes and streams has been completed
(Stevens 1997, Olsen and Stevens 1998, Stevens and
Olsen 1999, Stevens and Urquhart 1999).
New theoretical developments were necessary to
address the problems presented by large-scale
probability-based survey designs for lakes and streams, as
well as other ecological resources. The wealth of survey
methodology and experience that exists for sampling
hum;tn populo''}ns and institutions does not exist for
ecological resources. Finding or extracting suitable
sampling frames is only beginning to occur. Although
EMAP relies heavily on design-based estimation
methods, over time model-based methods will almost
certainly become more important. Since spatial location
is such an integral part of ecological resource monitoring,
all statistical methods must incorporate the spatial
dimension.
Statistical Education
The inland aquatic monitoring community has a
history of fixed-station, judgement-based site selection.
Data from these sites are summarized routinely as if they
were a simple random sample, i.e., sample means,
proportions, variances are calculated. If scientifically
defensible national estimates of the status of lakes and
streams is to be attained (Paulsen et. al. 1998), this
culture must be changed. What steps can be taken to
change the culture?
Most members of the inland aquatic scientific
community have an introduction to statistics as part of the
undergraduate or graduate education. Typically, they
have taken one or two courses for non-statistics majors
and a number of years has elapsed since taking the
courses. They understand the basic concepts of statistics,
or at least know they should be concerned about
particular issues. Most have been exposed to
experimental designs and the need for randomization. It
seems that all have heard that stratification is critical - so
much so that stratification is used excessively even when
it's not necessary. Seldom do they remember being
introduced to probability-based survey design. If they
have, they know simple random sampling, stratified
random sampling, and systematic sampling. My
conclusion is that the academic community (statistics,
environmental science, ecology) must emphasize survey
sampling more, so that the next generation of scientists
has a broader base of statistical methods.
EMAP seriously misjudged the gap that existed
between the design they proposed and the knowledge of
the monitoring community. Based on our experience
over the past ten years, we have found that the gap can be
closed by acknowledging that statistical education is part
of the process of designing a monitoring program. No
textbook exists that focuses on survey designs for
ecological resources, particularly aquatic ecosystems.
Useful references are Gilbert (1987), Thompson (1992),
and Schreuder et. al. (1993).
In 1994 EMAP began an effort, called Regional
EMAP, that contributes to the statistical education of the
monitoring community. Regional EMAP, or REMAP, is
a program where every two years each of the ten EPA
Regions is asked to participate in a monitoring study of
an aquatic resource. EMAP provides (partial) funding for
the study, but the Region selects the study area, resource,
and objectives for the study. EMAP also contributes its
research staff time to help the Region formulate the study,
determine measurement protocols, and, most important,
select a survey design. EMAP not only works with the
study team to select the probability-based survey design,
but actually selects the sample. An integral part of the
process are discussions concerning the clarity of the study
objectives, explicit definitions of the target populations,
importance of the sampling frame, concepts of
probability-based surveys, and importance of tying the
statistical analysis to the survey design. Over 100 studies
have begun under this program, involving hundreds of
individual researchers and government agencies.
Sociology of Organizations
Changing culture always involves the sociology of
organizations. What organizations are directly involved
in the process? For inland aquatic resources, the
principal organizational entities are EPA Office of Water,
EPA Office of Research and Development, EPA
Regional Offices, State monitoring agencies, U.S.
Geological Survey, ecological academic community,
statistical academic community. Other federal agencies
and professional organizations are also part of the
community, but the above illustrates the breadth of
organizational entities.
Three organizations within EPA actively participate
in pursuing the status of aquatic resources in the United
States: Regional Offices, Office of Water, and Office of
Research and Development. Regional Offices work
directly with the states to implement the Clean Water Act
monitoring requirements. Office of Water has
responsibility for compiling the National Water Quality
Inventory report to Congress and develops the federal
regulations and guidelines for state monitoring. Office of
Research and Development (ORD) conducts research in
support of the Agency's mission to restore and maintain
-------
the chemical, physical, and biological integrity of the
Nation's waters. EMAP is an ORD initiative. A change
in the culture of aquatic monitoring must have the
agreement of the Office of Water. As in any
organization, priorities differ among the three EPA
offices. ORD's EMAP program was a radical initiative
that could have been viewed as a direct threat to the
current monitoring environment established by the Office
of Water. Any radical change implies a measure of chaos
during implementation. Keeping a firm handle on
communications among the EPA offices would be critical
to its success. Within ORD the budget for EMAP would
likely have a significant negative budget impact on other
research programs, given the likely constant budget of
ORD. These and other considerations make changes in
culture extremely difficult within an agency unless the
change is mandated by Congress. This was not the case
with EMAP.
EMAP and NAWQA, a program of USGS, began
about the same time. Both stated that their objectives
were to measure the status and trends of the inland
aquatic resources. Confusion arose about the programs
and a perception that they duplicated each other
developed. Although EMAP was intended to be an
interagency effort, the agencies found it difficult to
collaborate.
EPA regional offices and state monitoring agencies
have enormous investments in compliance and TMDL
monitoring. Although the states are required to
contribute to EP A's national report, most do not view that
reporting as critical to meeting their primary monitoring
objectives. EMAP focused on obtaining data to meet the
requirements for the national report. Even more critical,
the monitoring was designed to provide national and
regional estimates, not state-level estimates. Many states
did not see what benefit EMAP would bring to them.
EMAP envisioned a future where the sample size and
survey design would address state-level concerns about
status and trends. However, the cost to do so prohibited
proposing a state-level program. The consequence was
that states found it difficult, if not impossible, to support
the change in culture.
Some in the academic ecological community are
involved directly or indirectly; some are not. A typical
view is that monitoring is not research and consequently
does little to further our understanding of the aquatic
ecosystems. Most ecological research occurs in small
watersheds and in limited geographic extent. The studies
appropriately use concepts associated with experimental
design. The measurements focus more on ecological
processes and can be labor and time intensive; variability
is a major factor. Consequently, when a large-scale
monitoring program, such as EMAP, states that it will
make biological measurements over large geographic
areas and generally restrict the measurements to be
obtainable within a single day, the community is naturally
apprehensive. Their concerns arc legitimate and must be
addressed as part of the research necessary to implement
a monitoring program. EPA, like most federal agencies,
actively solicits external peer review. EMAP was no
exception and initiated a review by the National Research
Council (NRC) during the development and planning
phase of the program. Their final report (NRC 1995)
concluded "that EMAP's goals arc laudable. However,
because achieving the goals of this ambitious program
will require that EMAP successfully meet many difficult
scientific, practical, and management challenges, the
committee continues to question whether and how well all
these goals can be achieved." They were concerned that
the design was at too coarse a scale in time and space to
detect meaningful changes, that further development of
biological indicators of ecological health was required,
and that the retrospective monitoring approach may not
match needs of assessment. From a statistical
perspective, they recommended stratified random
sampling by ecoregion, linking with site-specific
intensive research study sites, and undertaking more
power analyses of ability to detect change. These
concerns can not be dismissed and were not dismissed by
management within EPA. The social culture of EPA
responded to the reviews, in some cases as providing
guidance on how to improve the program and in others as
justification for reducing or eliminating it. Peer review
panels must understand the social culture of an agency in
terms of how it interprets external reviews.
The academic statistics community raised many of
the same scientific concerns as their ecological
counterparts. Although the ecological community
questioned the basic use of probability-based survey
designs, the statistical community accepted that basic
concept. Their concerns focused on the particular survey
design choice. What was the basis for EMAP's decision
to use a systematic grid area sample for the first stage?
Would the design provide the required precision? Why
rely on design-based estimation, rather than incorporate
model-based estimation? A committee from the American
Statistical Association reviewed the statistical aspects of
EMAP during the same time period as the NRC review.
Although die reviews were supportive of EMAP's use of
survey sampling and the statistical underpinnings of the
program, they also expressed concerns, as mentioned
above. These reviews were valuable inputs to the
statistical team designing EMAP. They provided
-------
ammunition to organizations and individuals who were
not in favor of EMAP.
Understanding the social culture of all organizations
that are involved in or can influence decisions on the
development and implementation of a monitoring
program is critical. At times, it may take more effort to
manage the social culture and its expectations than to
address the myriad of scientific issues. This is
particularly true when the monitoring program is a radical
change from the status quo.
Conclusions
Some of the issues involved in changing the culture
of monitoring aquatic ecosystems have been given. After
ten years working within EPA as part of the
Environmental Monitoring and Assessment Program
statistical design team, I believe that the monitoring
culture is changing - but more slowly than I expected.
Although having a solid scientific foundation is
important, understanding the social culture that affects
the monitoring program may be more important.
Significant advances in aquatic monitoring have occurred
as the result of EMAP. In 1990, no state used survey
designs as an integral part of their monitoring program.
In 1996, two states used survey designs; by 1999 the
number increased to 11 states . Other states continue to
express interest and most states have been involved in
one or more survey designs developed by EMAP. By
working with individual states on problems of direct
interest to them, EMAP has the opportunity to introduce
probability-based surveys. The states have the
opportunity to see how such surveys can. help them
answer some of their questions. Gradually, support is
building for a cultural change in the use of survey designs
nationally.
Many questions remain on what a federal/state
national environmental statistical program might look
like. However, it appears certain that progress toward
such a goal is being made.
Acknowledgments
The information in this article has been funded in
part by the United States Environmental Protection
Agency. It has been subjected to Agency review and
approved for publication. The conclusions and opinions
are solely those of the author and are not necessarily the
views of the Agency.
References
Adler, R. W., J. C. Landman, et al. (1993). The Clean
Water Act 20 Years Later. Washington, DC, Island Press.
CENR (1997). Integrating the Nation's environmental
monitoring and research networks and programs: A
proposed framework. National Science and Technology
Council, Committee on Environmental and Natural
Resources, The Environmental Monitoring Team.
Washington, DC, 100pp.
lmp://wvw.whitchouse.gov/WH/EOP/OSTP/html/OST
P Home.htinl.
Gilbert, R. 0. (1987). Statistical Methods for
Environmental Pollution. New York, Van Nostrand
Reinhold.
Helsel, D. R. (1995). Design of a relational water-quality
assessment program. 1995 Proceedings of the Biometrics
Section. Washington, DC, American Statistical
Association: 60-67.
ITFM (Intergovernmental Task Force on Monitoring
Water Quality) (1995). The strategy for improving water-
quality monitoring in the United States: Final report of
the Intergovernmental Task Force on Monitoring Water
Quality. Washington, DC, U.S. Geological Survey: 25.
Messer, J. J., R. A. Linthurst, et al. (1991). "An EPA
program for monitoring ecological status and trends."
Environmental Monitoring and Assessment 17: 67-78.
NRC (1995). Review of EPA's Environmental
Monitoring and Assessment Program: Overall Evaluation.
Washington, D.C., National Research Council, National
Academy Press.
Olsen, A. R., D. L. Stevens, Jr., et al. (1998).
"Application of global grids in environmental sampling."
Computing Science and Statistics 30: 279-284.
Overton, W. S., D. L. Stevens, et al. (1990). Design
Report for EMAP, Environmental Monitoring and
Assessment Program. Washington, D.C., U:S.
Environmental Protection Agency, Office of Research
and Development.
Paulsen, S. G., R. M. Hughes, and D. P. Larsen. (1998)
"Critical elements in describing and understanding our
nation's aquatic resources." J American Water Resources
Association 34: 995-1005. "
-------
Sanders, T. G., R. C. Ward, et al. (1983). Design of
Networks for Monitoring Water Quality. Littleton,
Colorado, Water Resource Publication.
Schreuder, H. T., T. G. Grcgoire, et al. (1993). Sampling
methods for multircsource forest inventory. New York,
John Wiley & Sons.
Stevens, D. L, Jr. (1997). "Variable density grid-based
sampling designs for continuous spatial populations."
Environmetrics 8: 167-195.
Stevens, D. L., Jr. and A. R. Olsen (1999). "Spatially
t,jtricted surveys over time for aquatic resources."
Journal of Agricultural. Biological, and Environmental
Statistics: (in press).
Stevens, D. L, Jr. and N. S. Urquhart (1999). "Response
designs and support regions in sampling continuous
domains." Environments (in press)
Thompson, S. (1992). Sampling. New York, John Wiley
& Sons.
Thornton, K. W., D. E. Hyatt, et al. (1993).
Environmental Monitoring and Assessment Program
Guide. Research Triangle Park, NC, U.S. Environmental
Protection Agency, Office of Research and Development,
Environmental Monitoring and Assessment Program,
EMAP Research and Assessment Center: 35.
USEPA (1995). Guidelines for preparation of the 1996
State Water Quality Assessments (305(b) Reports).
Washington, DC, U.S. Environmental Protection Agency,
Office of Water.
USEPA (1997). Environmental Monitoring and
Assessment Program (EMAP) Research Strategy.
Washington, DC, U.S. Environmental Protection Agency,
Office of Research and Development: 16.
White, D., A. J. Kimerling, et al. (1992). "Cartographic
and geometric components of a global sampling design
for environmental monitoring." Cartography and
Geographic Information Systems 19(1): 5-22.
------- |