United* States
    Environmental -Protect i
   1 Agency       1J
                              V,
                    "  J/G?      *MH2
                                    1
                              Mwfa
                 ' *j,%


                                       £Z
                           strategy
environmental monitoring and assessment program

-------

-------
                                   EPA 620/R-02/002

                                        July 2002
        Research  Strategy

 Environmental Monitoring and
      Assessment Program
        U.S. Environmental Protection Agency
        Office of Research and Development
National Health and Environmental Effects Research Laboratory
         Research Triangle Park, NC 27711

-------

-------
                                     Notice
The United States Environmental Protection Agency through its Office of Research and
Development produced this research strategy. It has been subjected to the Agency's peer and
administrative review and has been approved for publication as an EPA document. Mention of
trade names or commercial products does not constitute endorsement or recommendation for use.
                                                                                  in

-------
IV

-------
Dear Reader,

       The Environmental Protection Agency's Office of Research and Development's National
Health and Environmental Effects Research Laboratory (NHEERL) is pleased to present this
research strategy on the Environmental Monitoring and Assessment Program (EMAP). In
assessing environmental risk and determining restoration priorities, current environmental
conditions must be known and rates of change must be measurable. Because of EPA's
responsibilities under the Clean Water Act, this program within the Office of Research and
Development has focused on improving monitoring and assessment methodologies for aquatic
ecosystems and their associated landscapes. EMAP has focused on developing indicators and
unbiased statistical design frameworks to assess the status and trends of aquatic ecosystems at
local, state, regional, and national scales. As is EMAP's primary mission, the goal of this
Strategy is the development of sound scientific approaches to determine the health of the  nation's
aquatic ecosystems and the stressors most closely associated with impairment.

       EMAP efforts ensure that comprehensive and comparable methods are being used at a
national level, allowing meaningful assessments and the first regional comparisons of aquatic
ecosystem conditions across the entire U.S. These results will significantly  improve the quality
of performance-based reporting to Congress and will better inform EPA national and regional
decisions on priority issues and areas.

       State managers and technical staff frequently struggle to balance local information needs
with federal reporting requirements.  EMAP will continue to work with State partners to develop
cost-effective monitoring methodology to aid in decision-making.  Results to date from EMAP
approach applications in more than 30 States show  cost-savings while producing full-coverage
condition estimates. Often these cost-savings are used to address priority issues also identified
through the EMAP process.

       Finally, EMAP's approach and associated indicators serve the Agency and the public by
contributing to scientifically based reports such as EPA's upcoming state of the environment
report and the Heinz Center's "The State of the Nation's Ecosystems" report. EMAP's efforts
help to fill important information needs at both national and at local levels.  EMAP information
will improve our ability assess our progress in environmental protection and provide valuable
information for decision makers and the public.  For further information, please go to
www.epa.gov/emap or contact Mike McDonald, NHEERL's EMAP Program Manager, at
mcdonald.michael@epa.gov.
                                         Since
                                         >aul Gilman, Ph.D.
                                         Assistant Administrator
                                         Office of Research and Development
                                         U.S. Environmental Protection Agency

-------
                                Peer  Review

      Peer review is an important component of research strategy development.  The peer
review history for this research strategy follows.
ORD Science Council

             Final Review Date:  October 1, 2001

             Lead Reviewers:    Lee Mulkey, Lead Reviewer, NRMRL
                                Hugh McKinnon, NRMRL
                                Jay Messer, NERL
                                Michael Slimak, NCEA
                                Hal Zenick, NHEERL
                                Barbara Levinson, NCER

External Peer Review:            December 14-15, 2000, Research Triangle Park, NC

External Peer Review Panel Members:

      James J. Alberts, Ph.D., Research Marine Scientist, Marine Institute, University of
      Georgia, Sapelo Island, GA.

      Brian P. Bledsoe, Ph.D., Assistant Professor, Department of Civil Engineering, Colorado
      State University, Fort Collins, CO.

      Richard C. Lathrop, Ph.D.,  Wisconsin Department of Natural Resources, Madison, WI.

      Steven W. Seagle, Ph.D., Associate Professor, Center for Environmental Science,
      Appalachian Laboratory, University of Maryland, Frostburg, MD.

      Jianguo Wu, Ph.D., Associate Professor, Department of Life Sciences, Arizona State
      University, Phoenix, AZ.

      Linda J. Young, Ph.D., Professor, Department of Biometry, University of Nebraska,
      Lincoln, NB.

Peer Review Coordinator:

      Robert E. Menzer, Ph.D., Senior Scientist, National Center for Environmental Research,
      Office of Research and Development, U.S. EPA.
VI

-------
                         EPA Authors

Authors

     Michael E. McDonald, EMAP, NHEERL, ORD
     Steven Paulsen, WED, NHEERL, ORD
     Roger Blair, WED, NHEERL, ORD
     Joseph Dlugosz, MED, NHEERL, ORD
     Stephen Hale, AED, NHEERL, ORD
     Steven Hedtke, NHEERL, ORD
     Daniel Heggem, ERL, NERL, ORD
     Laura Jackson, EMAP, NHEERL, ORD
     K. Bruce Jones, ERL, NERL, ORD
     Barbara Levinson, NCER, ORD
     Anthony Olsen, WED, NHEERL, ORD
     John Stoddard, WED, NHEERL, ORD
     Kevin Summers, GED, NHEERL, ORD
     Oilman Veith, NHEERL, ORD
                     Acknowledgments

     The external peer review was administered by ORD's National Center for
Environmental Research (NCER).  The authors would like to thank Dr. Robert E.
Menzer, NCER, who organized and chaired the review. The authors would also
like to thank NCER Director Peter Preuss and National Risk Management
Laboratory Associate Director Lee Mulkey for their help in the administration of
the peer review. The authors would also like to thank all the EMAP scientists that
have contributed to various drafts and reviews of the Strategy.
                                                               Vll

-------
Vlll

-------
                                                                              Table of Contents
                           Table of Contents
SECTION I.    INTRODUCTION to the EMAP RESEARCH STRATEGY	   1-1

SECTION II.   EMAP RESEARCH STRATEGY	   II-l

                       Science Barriers	   II-l

                       Implementation	   II-3

                       Framework for a National Design	   II-3
                              Statistical Design	   II-3
                              Current Status of Design	   II-5
                                      Sampling  Site Selection	  II-5
                                      Classification for Strata	  II-6
                              Ecological Indicators	  II-7
                                      Current Status of Ecological Indicators	   II-7
                                             Streams	  II-8
                                             Estuaries	  II-8
                              Variability	   II-8
                                      Current Status of Variability Research	   II-9
                              Reference Conditions	   II-9
                                      Current Status of Reference Condition Research	   II-9
                              Index Sites	   II-9
                                      Current Status of Index Site Research	   11-10
                              Key Research Issues for Developing a National Design Framework..  II-10
                                      Survey Design	  11-10
                                             Classification	  11-11
                                      Indicators	  11-11
                                      Variability	   11-11
                                      Reference Conditions	   11-11
                                      Index Sites	   11-12

                       Landscapes	   11-12
                              Current Status of Landscape Research	   11-12
                                      Landscape Indicators	   11-12
                              Key Landscape Research Issues	   11-13

                       Strategic Geographic Studies	   11-13
                              Current Status of Geographic  Studies	   11-14
                              Key Geographic Research Issues	   11-14
                                      Western Pilot	   11-14
                                             Estuaries	   11-14
                                             Streams and Rivers	   11-15
                                             Landscapes	  11-16
                                      Regional Environmental Monitoring and
                                         Assessment Program (R-EMAP)	    11-16

                       National Resource Assessments	   11-16
                              National Coastal Assessment	   11-17
                              National Stream Assessment	   11-17
                              Large and Great Rivers	   11-17
                                    EMAP Research Strategy                                  ix

-------
Table of Contents
                      Information Management	    11-17
                              Data Management	    11-17
                              Information/Technology Transfer	    11-18

                      Program Management	    11-18
                              Quality Assurance	    11-18
                              Measures of Success	   11-18

APPENDIX I - Needs for Improved Environmental Monitoring	    AI-1

APPENDIX II - Resource Sampling	    AII-1
                              Lake Sampling	    AII-1
                              Stream Sampling	    AII-1
                              Estuarine Sampling	    AII-2
                              Wetland Sampling	    AII-3
                              Great Lakes Sampling	    AII-3

APPENDIX III - Indicators	     AIII-1
                              Indicator Development	    AIII-1
                              Indicator Measurement	   AIII-1
                              Indicator Responsiveness	    AIII-1
                              Indicator Variability	    AIII-1
                              Indicator Demonstration	   AIII-1

APPENDIX IV - Components of Variance in Indicators	    AIV-1
                              Population Variance	    AIV-1
                              Year Variance	    AIV-1
                              Resource-year Interaction	    AIV-2
                              Index Variance	    AIV-2
                              Estimates of Variance Components	   AIV-3
                              Effect of Variance on Trend Detection	   AIV-4
                              Sensitivity of Trend Detection to Variance Components	AIV-5
APPENDIX V - Index Sites	    AV-1
                              National Park Intensive Monitoring Network (PRIMENet)	   AV-1

APPENDIX VI - Mid-Atlantic Integrated Assessment (MAIA)	   AVI-1
                              Estuaries	   AVI-1
                              Wadeable Streams	   AVI-2
                              Landscapes	   AVI-3
                              EMAP Integration in MAIA	   AVI-4

APPENDIX VII - Regional Environmental Monitoring and Assessment Program (R-EMAP)	   AVII-1
                              Example of Recent R-EMAP Results	   AVII-1

APPENDIX VIII - Information Management	    AVHI-1
                              EMAP Data Directory	    AVIII-1
                              EMAP Data Catalog	    AVIII-1
                              EMAP World Wide Web Site	   AVIII-1
                              EMAP Data Links	    AVIII-2

REFERENCES	     R-l
                                    EMAP Research Strategy

-------
                                                                                     List of Figures
                                 List of  Figures
Figure 1.        Science and implementation barriers to a nationally comprehensive monitoring
                        framework for EPA high priority aquatic resources	
Figure 2.        EMAP programmatic organization	
Figures.        Timeline for major EMAP research accomplishments	
Figure 4.        The EMAP base grid for the United States	
Figure 5.        Chronically and episodically acidic Northeast lakes	
Figure 6.        Agriculture on slopes greater than 3% in the Mid-Atlantic	
Figure 7.        Tiered monitoring approach used by EMAP and the CENR Strategy	
Figure 8.        EMAP ecological provinces used for classification of estuaries in the continental U.S.
Figure 9.        Stream length sampled to yield 90% of total species present	
Figure 10.       Response of an indicator to stressor gradient	
Figure 11.       Signal to noise ratio for stream physical habitat indicators	
Figure 12.       Stressors associated with impaired stream quality in Mid-Atlantic Highland streams....
Figure 13.       Hypothetical time series for Secchi transparency in five different lakes	
Figure 14.       Design options  considered	
Figure 15.       Power for detecting trends using the design options	
Figure 16.       Power for trend detection with varying levels of a2yeai	
Figure 17.       Standard error of estimated status compared among designs	
Figure 18.       Power to detect trends in lake trophic condition with different indicators	
Figure 19.       Geographic scope of MAIA	
Figure 20.       EMAP estuarine sampling locations in MAIA, 1990-1993	
Figure 21.       The 1997 stream/river probability sample sites	
Figure 22.       MAIA land cover and Mid-Atlantic Highlands area used in the USGS-EPA
                        monitoring study	
Figure 23.       Distribution of probability-based stream sampling sites in Nebraska, 1997-1998	
Figure 24.       Percentages of streams in Nebraska R-EMAP study meeting designated uses for
                        aquatic life	
Figure 25.       Significant change detected in water clarity in Nebraska using EMAP approach	
II-2
II-2
II-5
II-6
11-10
11-13
AI-2
AII-2
AIII-2
AIII-3
AIII-3
AIII-4
AIV-2
AIV-7
AIV-8
AIV-8
AIV-9
AIV-9
AVI-1
AVI-3
AVI-4

AVI-4
AVII-2

AVII-3
AVII-3
                                      EMAP Research Strategy
  XI

-------
                                                                            List of Tables
                              List of Tables
Table 1.        EMAP projected research accomplishments FYOO - FY07	   II-4
Table 2.        Core EMAP coastal indicators	   11-15
Table 3.        EMAP surface water core indicators	   11-16
Table 4.        Levels of biological organization to consider in indicator development	   AIII-2
xii                               EMAP Research Strategy

-------
                                                                                Executive Summary
Executive  Summary
The U.S. Environmental Protection Agency's
(EPA's) Environmental Monitoring and Assessment
Program (EMAP) is a long-term research effort to
enable status and trend assessments of aquatic
ecosystems across the U.S. with a known statistical
confidence. Initiated in the late 1980's within the
Office of Research and Development (ORD), EMAP
has addressed the condition of estuaries, streams and
lakes in selected geographic regions, as well as
having examined the surrounding landscapes in
which these resources occur.  EMAP is now
progressing towards national demonstrations of
monitoring science in these and other aquatic
resources.  This strategy forms the basis for the
research needed to establish the condition of the
nation's resources, as a necessary first step in the
Agency's overall strategy for environmental
protection and restoration.

EMAP's Goals:
        Develop the science needed for a state-
based statistical monitoring framework to
determine condition, and detect trends in
condition, for all the Nation's aquatic ecosystems.

        Transfer this technology in a useable form
to states, tribes, and regions.

        Have this approach adopted and
implemented by the states, tribes, and regions.

        • EMAP has developed approaches for:
        state-based statistical monitoring designs for
        streams and estuaries; aggregation of state
        data to the national level; and providing
        scientifically defensible measures  of changes
        in ecosystem condition in support of the
        Government Performance and Results Act
        (GPRA).

The EMAP statistical (or probability) sampling
design provides the framework for unbiased,
representative monitoring for condition of an aquatic
resource with a known confidence level.  These
estimates can be aggregated from the local to the
national level. The Mid-Atlantic Integrated
Assessment (Region III) served as the proof-of-
concept for this approach. Monitoring with this
approach through time allows statistical detection of
change (and subsequently trends) in condition.

        • EMAP has developed and tested
        ecological indicators as integrators of
        stressors and estimators of aquatic condition.
An ecological indicator describes the condition of the
ecosystem (e.g., blood pressure is an indicator that
helps describe the condition of a person), and reflects
an ecosystem's biological, chemical, or physical
attributes. EMAP primarily uses biological indicators
to integrate all the different stressors acting on an
ecosystem.

        • EMAP uses large-scale geographic
        projects for removing scientific barriers, and
        for demonstrating our approach to States,
        Tribes and EPA Regions.

These geographic studies typically include one or
more major national or regional coverages, and ten
smaller sub-regional studies (Regional EMAP or R-
EMAP).  For these large-scale studies we partner
with EPA Regions, States, and Tribes in an effort to
build regional assessments from the bottom up,
aggregating local data into broader regional
assessments. The smaller R-EMAP  studies, in each
of the ten EPA Regions, allow us to  test our
approaches in diverse geographic areas, engage
additional states, and transfer our technologies, while
helping address local problems.

        • EMAP is well integrated with the Office
        of Research and Development's (ORD)
        Science to Achieve Results (STAR) Grants
        Program, and derives maximum benefit from
        academic research.

Scientific uncertainties continue to exist in a number
of areas related to probabilistic monitoring of
ecological condition (e.g., ecological indicators,
ecosystem classification, reference conditions). By
integrating our EMAP research plans with the STAR
Grants Program, we can more effectively use
academic researchers to help answer many of the
                                      EMAP Research Strategy
                                            Xlll

-------
Executive Summary
current and future scientific questions necessary for
developing comprehensive national aquatic
ecosystem monitoring frameworks.

        • EMAP's Western Pilot is a major strategic
        geographic study to remove scientific
        barriers to a national stream assessment.

The Western Pilot (encompassing 12 western states in
EPA's Regions VIII, IX and X) is a scientific
evaluation of the generalizability of our stream
designs and indicators to a vast area of the country
with high ecological variability. This work will
remove the remaining scientific barriers to a national
framework for monitoring the condition of the
nation's streams.

        • EMAP's National Coastal Assessment is a
        national demonstration of our approach in
        estuaries.

We are demonstrating the framework for a consistent,
state-based, probabilistic monitoring framework in
the 24 marine coastal states and Puerto Rico, This
effort will produce the first national assessment of the
condition of the U.S. marine estuaries, and will be the
baseline for future measures.

        • EMAP has a well developed information
        management (IM) system for public data
        accessibility and is integrated into EPA's
        STORET database for long-term storage.

IM is an integral part of our large-scale research and
demonstration programs, and is guided by the need
for sharing environmental monitoring data. Our IM
group regularly revises and updates the EMAP public
home page on the web, providing access to: our Data
Directory, for ready identification of data of interest;
EMAP data and metadata; and our bibliography of all
EMAP-related publications (>1000+).  The EMAP
IM group is archiving our data in EPA's STORET
database.  All of our EMAP data for condition
assessments typically becomes publicly accessible
within two years of collection after all our sampling
              and data quality assurance and quality control
              procedures have been satisfied.

                      • Future plans for EMAP include: removing
                      the scientific barriers to establishing the
                      ecological condition of great and large rivers
                      (as receiving waters for streams), and
                      demonstrating a national stream monitoring
                      framework.

              Sampling designs and indicators do not currently
              exist to characterize the condition of large or great
              rivers in a scientifically defensible manner.  These
              large rivers are unique resources and are difficult to
              sample.  However, they are critical receiving waters
              for the nation's streams.  An understanding of their
              condition is an important step in dealing with
              problems associated with Total Maximum Daily
              Loads (TMDL's). Additionally, to ensure that a
              nationally consistent stream monitoring framework is
              adopted by the states, design and technological
              advice may be required for other large geographic
              areas of the country.

              With the Western Pilot, we are removing the
              scientific uncertainties that prevent state-to-national
              assessments of stream condition. EMAP has already
              established the scientific basis for monitoring
              estuarine condition nationally, and is currently
              engaged in a national demonstration  Sampling
              through time using our EMAP approach will allow
              significant trends in stream and estuarine conditions
              to be identified at the local, state, regional, and
              national levels.  An unbiased assessment of the
              condition of our streams and estuaries will provide
              the scientific basis for better-informed public
              decisions regarding these resources, and will provide
              baselines for measuring EPA's progress in improving
              environmental condition in support of the
              Government Performance and Results Act.  Future
              EMAP work will provide the science necessary for
              determining the condition of other critical aquatic
              resources and will be necessary for the full
              implementation of GPRA.
xiv
EMAP Research Strategy

-------
                                                                                 Introduction
I.
Introduction  to
the   EMAP
Research
Strategy
Calls for improvements in environmental monitoring
date back to the late 1970s. Along with the National
Academy of Sciences and the White House Office of
Science and Technology Policy, EPA has recognized
the critical need for nationally-consistent,
comprehensive, and scientifically-defensible
monitoring to detect environmental status and trends.


Because of EPA's statutory responsibilities under the
Clean Water Act, and the ability of inland surface
waters and estuaries to reflect total watershed
condition, EPA has committed to advancing the state
of the science in monitoring and assessing the
condition of the nation's aquatic ecosystems.  An
historical overview of reports and initiatives that have
identified the need for an improved national
monitoring framework or made recommendations for
improving environmental monitoring can be found in
Appendix I.


EPA's focus on improving monitoring and
assessment methodology derived from the need to
characterize and assess environmental risk (U.S. EPA
1996). This goal required scientific advances in the
way aquatic ecosystems and their associated
landscapes are measured, modeled, maintained, and
restored; these issues form the basis of the Ecological
Research Strategy of EPA's Office of Research and
Development (U.S. EPA 1998).


A critical component of this Strategy is ORD's
Environmental Monitoring and Assessment Program
(EMAP). EMAP focuses on developing indicators,
and unbiased statistical design frameworks that allow
the condition of aquatic ecosystems to be assessed at
local, state, regional, and national scales. The current
condition of the nation's aquatic ecosystems, and the
stressors most closely associated with impaired
condition, are key assessment activities.  Developing
sound scientific approaches for these activities has
been and continues to be EMAP's primary mission.


To assess environmental risk, the current condition of
the environment must be known and an estimate of
the rate of change in that condition must be available.
Through a probability-based sampling design, the
EMAP approach provides a statistically-valid basis
for determining aquatic ecological condition. The
EMAP approach also exemplifies monitoring for
results; when implemented over time, it can provide
quantifiable estimates of the environmental benefits
derived from the Agency's protection and restoration
strategies in support of full implementation of the
Congressionally mandated Government Performance
and Results Act (GPRA). Using the EMAP
approach, ORD hopes to reduce data gaps identified
by the Government Accounting Office (2000),
develop new hypotheses for testing  cause-and-effect
relationships in ecosystems, and provide scientifically
defensible assessments of change and trends.


Because of the importance and innovative nature of
the research undertaken by EMAP, it has undergone
extensive scientific review. EMAP has had more
than 25 separate peer reviews of individual program
components (over the last 10 years) and a program-
wide review by a National Research Council (NRC)
panel. The EPA Science Advisory Board has also
reviewed several aspects of EMAP, paying particular
attention to the development of indicators and the
integration and assessment activities within the
program. Most recently (1998), EMAP was reviewed
jointly by the American Statistical Association and
the Ecological Society of America.  These reviews
have shaped the current focus of EMAP, as detailed
                                   EMAP Research Strategy
                                          1-1

-------
Introduction
in this strategy, and affirmed EMAP as a productive,
high-quality scientific research program.


In its 10-year history, EMAP has substantially
advanced the scientific basis for monitoring
ecosystem condition. Selected EMAP
accomplishments for this time include:
    More than 1000 scientific publications (for a
    complete bibliography, see
    http://www.epa.gov/eniap/htinl/piibs'):
    27 scientific symposia and workshops sponsored;
    The U.S. Forest Service adopting the EMAP
    approach for their Forest Health Monitoring
    (FHM) Program;
•   Demonstrating the feasibility of using an EMAP
    approach for regional monitoring in the mid-
    Atlantic (through the Mid-Atlantic Integrated
    Assessment or MAIA);
    Producing the first unbiased assessments of
    regional stream condition (U.S. EPA 903/R-
    00/015), estuarine condition (U.S. EPA 600/R-
    98-004, U.S. EPA 600-R-98-147),  and landscape
    condition (U.S. EPA 600/R-97-130);
•   EPA's Office of Water recommendation that
    probability surveys be incorporated into state
    monitoring programs (in the guidance to the
    States for reporting to Congress under Section
    305(b) of the Clean Water Act);
    Completion of the first systematic remotely-
    sensed land cover characterization for the entire
    lower 48 states; and
    More than 30 states currently using or testing the
    EMAP monitoring framework to determine the
    condition of one or more of their aquatic
    resources.


EMAP's current scientific effort centers on removing
the scientific barriers to application of our approach
in diverse regional environments. Next, we
demonstrate these approaches and technologies in
national assessments of resource condition, and lastly
we transfer this capacity to States, Tribes, and
Regions. Ongoing efforts include:
    Adapting EMAP indicators  and design to the  12
    western U.S. states in EPA Regions VIII, IX, and
    X (Western Pilot);
    Working with the 24 U.S. coastal states
    (including Alaska and Hawaii) and Puerto  Rico,
    to collect and analyze data from a core set  of
    indicators for the first national assessment  of
                 estuarine condition (National Coastal
                 Assessment);
                 Supporting one or more studies in each of the ten
                 EPA Regions to demonstrate the utility of the
                 EMAP approach in addressing specific regional
                 issues (this is Regional EMAP or R-EMAP); and
                 Sponsoring a strategic array of academic
                 research grants in emerging and cross-
                 disciplinary fields through ORD's Science To
                 Achieve Results (STAR) Grants Program.
                 Among other topics, these grants are exploring
                 scalar relationships, watershed processes,
                 resource classification and reference gradients,
                 and are intended to result in new indicators and
                 design adaptations for EMAP.


             Future plans for EMAP involve research and
             technology transfer to enable periodic national
             assessments of all aquatic ecosystems. Priority near-
             term efforts include:
                 Developing the methodology and partnerships
                 required for assessing condition of large and
                 Great rivers;
                 Continuing regional studies to familiarize staff of
                 EPA Regions and States with effective solutions
                 to their monitoring needs through R-EMAP; and
                 Continuing sponsorship of academic research
                 grants and partnerships through STAR Grants to
                 augment internal expertise in indicator
                 development, statistical designs and analyses,
                 and other monitoring and assessment issues.


             The intent  of these various research activities is to
             develop and transfer the necessary science  to meet
             monitoring and assessment needs for aquatic
             ecosystems across a range of decision makers and
             other user groups.  At the national level, EMAP's
             efforts towards comprehensive and comparable
             methods will enable, for the first time, meaningful
             assessments and regional comparisons of aquatic
             ecosystem condition across the entire U.S.  These
             results will significantly improve the quality of
             mandated reports to Congress, and will better inform
             EPA and other federal decisions.
             EPA Regions will benefit as well from consistent and
             comparable environmental data as a result of the
             EMAP approach. Regional decision-makers must
             also prioritize protection activities across multiple
             States and environmental media, and often seek to
1-2
EMAP Research Strategy

-------
                                                                                            Introduction
develop unbiased State of the Region reports for their
stakeholders.
At the State level, managers and technical staff
frequently struggle to balance local information needs
with federal reporting requirements. The goal that
EMAP seeks to achieve with State partners is cost-
effective monitoring methodology that
simultaneously serves both levels of decision-making.
Results to date from applications of EMAP tools in
more than 30 States indicate that these States are
improving full-coverage condition estimates while
saving money.  Often these cost-savings are used to
address priority issues identified through the EMAP
process.


Finally, EMAP intends to serve Tribes, local
communities, and the general public by facilitating a
technically-sound "environmental condition report."
This information will provide the larger context for
evaluating local conditions, and provide trends to
gauge protection needs and effectiveness.
                                        EMAP Research Strategy
                                               1-3

-------

-------
                                                                   EMAP Research Strategy
EMAP  Research
Strategy
 Our Scientific Goal: Develop the
 science needed for a State-based
 statistical monitoring framework to
 determine condition, and to detect
 trends in condition, for all the
 nation's aquatic ecosystems.

 A scientifically-rigorous determination of the
 condition of an aquatic resource is fundamental to all
 subsequent research and modeling questions.
 Environmental risk characterization is predicated on a
 knowledge of condition and the rate at which that
 condition is changing (U.S. EPA 1996). Thus, the
 scientific goal of EMAP has remained virtually
 unchanged since its beginning, although the
 approaches have evolved (see Appendix 1).

 The current EMAP approach is to remove the
 scientific uncertainties associated with developing
 State-based statistical monitoring designs for aquatic
 resources in support of Section 305(b) of the Clean
 Water Act. The power of this approach is that the
 designs provide: the framework for an unbiased,
 representative assessment of the condition of these
 resources with a known confidence level; aggregation
 of State data to the national level; and scientifically
 defensible measures of changes in condition in support
 of the Government Performance and Results Act
 (GPRA).

 Our Implementation Goal:
 Transfer the EMAP science and
 technology to the States, Tribes,  and
 EPA Regions, and have our
 approach adopted for long-term use.
The States are Congressionally mandated to assess
the condition of their waters. Currently, the EMAP
approach is the only statistically-valid means of
assessing the condition of all waters of a State. It is
a change from historic practices of fixed-site
monitoring.  Therefore, the increase in information
and the cost-effectiveness of this approach must be
demonstrated to Regions, States, and Tribes to gain
their acceptance.

To gain acceptance of the EMAP approach, we
partner with EPA's Regions, States, and Tribes in
large- and small-scale geographic demonstrations.
Using our sampling designs and collecting samples
with our partners, we aggregate local data into
broader state, regional, and even national condition
assessments. These demonstration efforts typically
serve both to remove scientific uncertainties by
allowing us to test our approaches in diverse
geographic areas, and to show how our scientific
approaches can be effectively used in support of
solutions to local, state and regional problems.

Within the five-year time frame of this strategy,
EMAP intends to meet its scientific goal with
respect to estuaries and streams, and be well on the
way to meeting the implementation goal for these
two resources.  However, in order to do this,
EMAP must continue to overcome the scientific
barriers that would prevent development of
national frameworks, and we must effectively
demonstrate and transfer the technology we are
developing (Figure 1).

SCIENCE BARRIERS

EMAP, an ORD-wide research program (Figure
2), maximizes and integrates intramural and
extramural scientific capabilities in pursuit of
                                   EMAP Research Strategy
                                       II-1

-------
EMAP Research Strategy
       Figure 1.  Science and implementation barriers to a nationally comprehensive

       monitoring framework for EPA high priority resources.
       (0
                      Science Barriers
 Figure 2. EMAP Programmatic Organization
NHEERL
ADE
	


EMAP
ORE ADVEORYCOMM 2TEEE


EM AP
....
i
G ulf ED

...
i
AtlanticED

1
1 ±i-C onthent
ED

1
fl estemED


NCER
ADE


NERL
ADE


STAR
ESD
	


i
EERD
NCEA
___
ADE


....


3 tsbalClm ate



NRMRL
_ _ .
ADE


Remedia tbn
I D esJgn Fiam ew oik    I
I Landscape         I
  G eograph±3 Studies
I
  JM ATA SitBgrat±;n w ,fceg 3|—

            r^
  JW estem Pibtw/Reg 8 ,9,w[~~

            i
                                                      -0-
                                                       ©
                            ©

Y Y Y ~.
I Design | (JJ


I lidicato^ | Vif
I B±5ci±BrH/&ef.Cc
^sn 	 s 	 0 	 x 	
X
I lifom atbn M an a gem ent  I
  I    Uto
                            ©
11-2
                         EMAP Research Strategy

-------
                                                                           EMAP Research Strategy
our goals. By aligning our scientific expertise and
linking it with the academic community (through
EPA's Science To Achieve Results (STAR) Grants
Program), we have established the critical research
capabilities necessary to develop a national
monitoring design framework for estuaries and
streams over the coming years (Table 1).

EMAP's research focus includes:

        establishing the statistical variability of
        EMAP indicators when used in aquatic
        ecosystems in diverse ecological areas of
        the country (ecoregions);

        establishing the sensitivity of indicators
        and design to change and trend detection
        in the condition of aquatic ecosystems;
        and

        developing indicators and designs that will
        allow a nationally comprehensive
        monitoring framework for other EPA
        high-priority aquatic resources (e.g.,
        large/Great Rivers, wetlands).

Other important research areas amenable to
academic research through the STAR Grants
Program include:

        developing new and better indicators of
        condition in various aquatic resources;

        establishing chemical and biotic index
        sites for estimates of seasonal variability
        to link to the EMAP design;

        developing new and better classification
        schemes for the various aquatic
        ecosystems and the landscapes in which
        they occur; and

        Linking remotely-sensed landscape
        indicators to the condition of aquatic
        ecosystems.

IMPLEMENTATION

For the EMAP approach to be successfully
implemented nationally, the States and Tribes must
accept it and incorporate it into their common
monitoring practices. Therefore, EMAP must:

        Demonstrate the efficacy and utility of the
        EMAP approach broadly throughout the
        U.S.;
        Build state and tribal capacity to develop
        and analyze statistical monitoring designs;

        Build state and tribal capacity for the use
        of geographic information systems (GIS)
        to estimate land-use change, and for
        improving statistical design; and

        Improve information storage and transfer
        to optimize the available information from
        an EMAP-like monitoring approach.

A time line has been developed for accomplishing
the major components (including field work,
analysis, and demonstrations) for reaching our goal
of a national monitoring framework for estuaries
and streams by 2007 (Figure 3). The allocation of
specific tasks related to demonstration and
implementation in this time period is given in
Table 1.

FRAMEWORK FOR A NATIONAL DESIGN

To determine the condition of a national aquatic
resource, we must either census this resource
across the entire U.S., or we must be able to make
unbiased statements on the overall condition of the
resource from selected samples. The first  approach
is currently not feasible for aquatic ecosystems.
The latter approach is feasible, but complex. The
validity of the latter approach depends on  statistical
inference from the  sampling design, and
subsequent analysis, to produce regionally
representative information. To do this requires:  a
design for sampling across diverse ecosystems;
indicators that are sensitive to stress and can be
measured to determine condition of the resource at
a wide range of sampling locations; and reference
conditions to act as benchmarks against which to
measure the indicators. These are fundamental
research areas for EMAP.

Statistical Design  - Probability-based sampling
within a statistical survey design (Cochran 1977)
provides the only unbiased estimate of the
condition of an aquatic resource over a large
geographic area from a small number of samples.
The principle characteristics of a probabilistic
                                       EMAP Research Strategy
                                            II-3

-------
EMAP Research Strategy
Table 1. EMAP Projected Research Accomplishments E - primarily EMAP Researchers
FYOO-FY07 S - primarily STAR Grants Research at Academic Institutions
E/S - combination of EMAP and STAR Research

E
>
u_
S
t
E
E
t.
E
E
Monitoring Framework for
Streams in Western Pilot
developed
Remotely Sensed Land
Cover for Western U.S.
(MRLC) produced
Designs for National
Coastal Monitoring
developed
Draft State of U.S.
Estuaries Report
produced
Monitoring Protocols
developed for condition
of Regional Aquatic
Resources
Integrated Ecosystem
Assessment Protocols
developed for MA1A
Biocriteria Monitoring
Protocols developed for
Aquatic Systems in
Western U.S.
National Land Cover Map
for U .S . produced for land
cover/land use change
(MRLC)
Draft State of U.S. Near-
shore Coastal
Ecosystems Report
produced
Report on Trends in
Select Estuarine
Ecosystems produced
Models for Integrating
Survey and Remote
Sensing applications
developed
Scaling Protocols for
Multi-tier Design
developed
State of Western Streams
Report produced
Ecosystem Classification
Protocols developed
Design Framework for
Integrated Monitoring in
the West developed
Additional/new Ecological
Indicators for
environmental condition
developed
National Landscape
Changes Assessed
State of the Western
Surface Waters Report
produced
Design for National
Estuarine and Stream
Monitoring Framework
established
Design Framework
Survey
Design
E

E

E

E



E/S
S

S
E/S



E/S
Indicators
E

E

E/S

E/S



E/S


S
S
S


E/S
Reference
Conditions
E

E

E/S

E/S






E/S
E/S


E/S
E/S
Index
Sftes






S



S
S


S


S
S
Landscape
Indicators










E/S



E/S
E/S
E
E/S
E/S
Land Cover
Mapping

E





E



E/S

E/S
E/S

E

E/S
Geographic Studies
I.1AIA
Integration




E/S
E



E









Western
Pilot
E
E
E
E
E/S

E/S

E



E

E/S


E
E/S
Coastal
Initjatrae


E
E
E/S

E/S

E
E




E/S



E/S
R-
EMftP




E

E








E

E
E
Information
Management
Data
E


E
E
E
E

E
E


E

E


E
E
Info



E
E
E


E
E
E

E

E


E
E
II-4
EMAP Research Strategy

-------
                                                                            EMAP Research Strategy
    YEAR
                •00
                        •01
                                 •02
                                         •03
                                                 •04
                                                          •05
                                                                  •06
                                                                          •07
  Aquatic Ecosystem
 Estuaries    w P
 Near-Shore
 Coastal
                                     W
                                                                    Coastal
                                  W P
                                           Natbnal
Stream s

Reservoirs
             w p
             w P
 Large Rivers  wp

 Lakes

 W eOands

 R-EMAP    	
N ationalM onitaDring
Fram ewoik for
Stzeam s and Estuardes
                                                                      Ii]and Surfece W aters forW estem U S .
 Landcover   MRLC

 Analysis     w P
                                                           Natbnal
                                                                       *wjf - fieli sam pUng ends

                                                                       WP -Western Pibt

                                                                     MRLC -Natbnalram ote]y sensed ]and cover

                                                                   Natbnal -Natbnw jde Assessm ent
  Figure 3.  Timeline for major EMAP research components, including field work and analysis.
         By 2007 we intend to have a national monitoring framework for streams and estuaries.
design are: (1) the population being sampled is
unambiguously described; (2) every element in the
population has the opportunity to be sampled with
a known probability; and (3) sample selection is
carried out by a random process. This approach
allows statistical confidence levels to  be placed on
the estimates and provides the potential to detect
changes and trends in condition with repeated
sampling.

        Current Status of Design - EMAP uses
a probabilistic survey design to select sampling
sites for an unbiased representation of the condition
of aquatic resources over large areas.  Our design
specifies the information to be collected and at
what locations. The validity of the inference rests
on the design, and subsequent analysis, to produce
regionally representative information.

        Sampling Site Selection - In attempting to
describe the condition of large geographic areas,
samples should be distributed throughout the study
area to be  maximally representative.  EMAP's
                                                      design accomplishes this by taking samples at
                                                      regular intervals from a random start (a systematic
                                                      random design). The systematic element spreads
                                                      out the sampling locations geographically, but still
                                                      ensures that each element has an equal chance of
                                                      being selected.

                                                      Grids are used to add systematic elements to the
                                                      EMAP design (Figure 4). The grid is positioned
                                                      randomly on the map of the target area, and sample
                                                      locations from within each grid cell are selected
                                                      randomly.  The grid ensures spatial separation of
                                                      randomly selected sampling units (systematic
                                                      random sample). Within this design approach there
                                                      is some additional and important flexibility.  There
                                                      is the potential to divide the entire target population
                                                      into any number of sub-populations (or strata) of
                                                      interest.  Subsequent random sampling within these
                                                      strata allows statistical inferences to be made about
                                                      each sub-population. Each of these strata can have
                                                      a different level of sampling effort depending on
                                                      the inference to be made, but the weighting due to
                                                      the differential effort must be accounted for in the
                                        EMAP Research Strategy
                                                                                                  II-5

-------
EMAP Research Strategy
analysis. As an example, stratified sampling could
be used in a regional stream survey to enhance
sampling effort in a watershed of special interest so
that its condition could be compared with the larger
regional area.  Simple random sampling of the
region would not likely provide sufficient samples
in the watershed to reliably estimate the
watershed's condition.

        Classification for Strata - EMAP's
primary interest in classification is to better define
 Figure 4. The EMAP base grid overlaid on the
 United States. There are about 12,600 points in
 the conterminous U.S. with approximately 27
 km between points in each direction. A fixed
 position that represents a permanent location
 for the base grid is established, and the
 sampling points to be used by EMAP are
 generated by a slight random shift of the entire
 grid from this base location.  (Full descriptions
 and rationale for the cartography and geometry
 of the grid are given in White et al. 1991).

different strata of aquatic systems for which similar
expectations exist. This allows for improved
information about each of the strata and can
improve statistical estimates of condition.

At the coarsest level, EMAP divides aquatic
resources  into different water body or system types,
such as lakes, streams, estuaries and wetlands.  Our
rationale is that the biological, chemical and
physical characteristics for these systems are
fundamentally different from one another and we
expect different indicators and different designs
will be necessary. Subsequently, we use a second
level of strata, ecoregions, to capture regional
differences in water bodies. Ecoregions are areas
                which have generally similar ecosystem
                characteristics (geology, physiography, vegetation,
                climate, soils, land use, wildlife, and hydrology,
                see Omernik 1995).  We have compiled ecoregion
                maps for use in strata development that are based
                on the patterns and the composition of biotic and
                abiotic characteristics (Wiken 1986; Omernik
                1987,  1995).

                The lowest level of strata in the EMAP design
                allows us to distinguish among different "habitat
                types" within an aquatic resource in a specific
                geographic region (see Appendix II).  For example,
                portions of estuaries with mud-silt substrate will
                have much different ecological characteristics than
                portions of estuaries with sandy substrates. It is
                within this lowest stratum that our sampling is
                done.  Within the lowest stratum, our sampling
                design takes one of two very different forms
                depending on whether the aquatic ecosystem to be
                sampled is discrete or extensive.  A discrete
                resource consists of distinct natural units, such as
                small to medium-sized lakes. Our population
                inferences for a discrete resource are based on
                numbers of units that possess a measured property
                (e.g., 10% of the lakes are acidic). Extensive
                resources, on the other hand, extend over large
                regions in a more or less continuous and connected
                fashion (e.g., rivers), and do not have distinct
                natural units.  Our population inferences here are
                based on the length or area of the resource. The
                distinctions between discrete and extensive are not
                always clear, and in some cases a resource may be
                viewed as both at different times (e.g., discrete -
                estuaries in a region; extensive - sample units
                within an estuary). Each aquatic resource has its
                own unique sampling characteristics.  The
                approach used depends on the nature of the
                resource and the available information for that
                resource.

                We use an intensification of our EMAP grid to
                sample discrete aquatic resources. If the units of
                the resource have appreciable area relative to the
                grid density, then the grid can be used directly by
                selecting those units in which one or more grid
                points fall (e.g., estuaries in a state). With this
                method, the probability that a unit gets into the
                sample (its inclusion probability) is proportional to
                the unit's area. The inference to the entire
                population is then in terms of area (e.g., 8% of the
                bottom water area in the mid-Atlantic estuaries is
                severely hypoxic in summer). Alternatively, a unit
                of a discrete resource can be treated as a point in
                space.  In this case, there must be a  rule identifying
                a point that is uniquely associated with each
II-6
EMAP Research Strategy

-------
                                                                           EMAP Research Strategy
resource sample unit. The point can be defined and
located based on properties of the resource unit,
regardless of the location of the grid. For example,
the center point of lakes could be used. This
method of sampling is appropriate for inference in
terms of numbers of units in a particular condition
(e.g., 7% of Northeastern lakes are chronically
acidic).

We use two different approaches for taking a
probability sample of an extensive aquatic
resource. Depending on the nature of the
ecosystem, our approach is to use area sampling
(e.g., sampling for rivers) or point sampling (e.g.,
sampling for estuaries). In area sampling, the
extensive resource is broken up into disjoint pieces,
much like a jigsaw puzzle. The extensive resource
is basically converted into a discrete resource, and
sample selection is from a random (within the
systematic, stratified random sampling design)
selection of these pieces.  The value we obtain
from the samples is then used to characterize the
entire piece. The point sampling approach locates
points at random within the extensive resource, and
sampling occurs at these points. With this
approach, the sample value is usually taken to
represent the value of the resource at that point.

Our EMAP designs provide the unbiased statistical
framework for where we sample, and define how
we extend our inference from our samples to the
entire population of interest.  However, within this
context, we must have indicators which establish
the condition of an aquatic ecosystem.

Ecological Indicators - An indicator is one (or
more) measure(s) or model(s) that describes the
condition of the system in question (e.g.,  blood
pressure is an indicator of human health).
Historically, chemical indicators have been used to
monitor the condition of aquatic systems, but these
chemical measures do not necessarily tell us
anything about ecological condition. Additionally,
our perturbations of aquatic environments have
expanded from chemical inputs to include
disruption of physical habitat, alteration of
hydrologic patterns, introduction of non-indigenous
biota and widespread alteration of the landscape.
All of these are stressors which may potentially
affect the condition of aquatic ecosystems.

Effective aquatic ecological indicators are central
to  determining the condition of our aquatic
resources. We have been able to develop or
modify a number of ecological indicators (see
Appendix III). The most successful have been
multi-metric indices formed by combining
biological indicators within a taxon (e.g., those
indices related to fishes - number of species
present, number of pollution-tolerant species, etc.)
found in aquatic ecosystems. Separate indices for
different taxa (from plants to fishes) are important
because they may be critical components of aquatic
ecosystems, and because they integrate various
natural and anthropogenic stressors into their
responses.  However, further interpretation of these
biological indicators requires additional
measurements of the fundamental and associated
components of the abiotic environment, including
physical measures of habitat (e.g., substrate type
and quality, depth) and chemical measures of the
ambient water quality (e.g., dissolved oxygen,
temperature, salinity, nutrients, and toxics).

        Current Status of Ecological Indicators
- Most of the current EMAP aquatic community or
assemblage indicators come from analysis of the
fish, benthic macroinvertebrate and plant
communities. Fish are of immense interest to the
public, and many  species are declining. Because
many species are relatively long-lived and mobile,
fish can be useful indicators of multi-year and
broad-scale environmental conditions.  Fish
assemblages contain species representing a variety
of trophic  levels (omnivores, herbivores,
invertivores, planktivores, and piscivores);  since
they tend to integrate changes in lower trophic
levels, they often reflect overall ecosystem
condition. Macroinvertebrates  (e.g., snails, aquatic
insect larvae, crayfish) are important intermediate
members of the food web.  They integrate changes
in the lower trophic  levels and, because they are
not as mobile as fish, they often can provide more
information on specific stressors. Plants are the
base of the food chain and can act as the primary
link between the chemical constituents in aquatic
ecosystems and the higher trophic levels. Changes
in their abundance or species composition can
greatly affect food availability at higher trophic
levels.

        Streams - EMAP uses fishes, benthic
macroinvertebrates,  and algae attached to rocks as
biological indicators of stream ecosystem
condition.  Stream fishes have been well studied
and have been shown to be suitable indicators
(Matthews and Heins 1987, McAllister et al. 1986,
Miller and Rolbison 1980, Minckley 1973,  Moyle
1976, Plafkin et al. 1989, Platts et al. 1983, Rankin
and Yoder 1991).  EMAP uses  an index of biotic
integrity (IBI, a multi-metric biological indicator,
similar to Karr et al. 1986) to evaluate the overall
                                        EMAP Research Strategy
                                            II-7

-------
EMAP Research Strategy
fish assemblage, which provides a measure of
biotic condition.

Stream benthic macroinvertebrates (a
heterogeneous assemblage of animals inhabiting
the stream bottom) have a long history of use as a
biomonitoring tool. Changes in the assemblage
structure (abundance and composition) and
function can indicate water resource status and
trends (Cummins and Klug 1979, Plafkin et al.
1989). Different stressors on stream ecosystems
have been studied and have been found to result in
different assemblages and/or functions (Armitage
1978, Hart and Fuller 1974, Hilsenhoff 1977,
Metcalfe 1989, Resh and Unzicker 1975). EMAP
currently uses a benthic invertebrate IB I for stream
macrobenthos.

Algae on stream rocks are the food for many
primary consumers, such as macroinvertebrates and
herbivorous fish. The algal component of stream
periphyton (a mixture of organisms attached to
substrates, including algae, bacteria,
microinvertebrates, and associated organic
materials) has been used extensively in the analysis
of water quality for several decades (Lange-
Bertalot 1979, Patrick 1968,  Stevenson and Lowe
1986, Watanabe et al. 1988). It has proven to be a
useful indicator for environmental assessments
because algal species have rapid reproduction rates
and short life cycles, and thus respond quickly to
perturbation (Stevenson et al. 1991).

        Estuaries - To establish estuarine
biological condition, we currently use benthic
invertebrates (e.g., oysters, shrimp) and fishes.
Benthic invertebrates are reliable indicators of
estuarine condition because they are sensitive to
water quality impairments and disturbance
(Rakocinski et al.  1997). We have developed a
benthic  index of estuarine condition (similar to the
stream invertebrate IB I) that  incorporates changes
in diversity, structure, and abundance of selected
estuarine benthic species (Engle et al. 1994, Engle
and Summers 1999). The EMAP estuarine benthic
index has been used successfully in estimating the
ecological condition of Atlantic coast estuaries
from Cape Cod, Massachusetts, to central Florida
(Paul et al. submitted, Paul et al. 1999, Van Dolah
et al.  1999), and estuaries in the Gulf of Mexico
(Macauley et al. 1999).  It is  currently being tested
nationally as part of the National Coastal
Assessment (see below).

Unlike streams, fishes in estuaries are typically
difficult to sample quantitatively. Because these
fishes are important commercially and
recreationally, we have chosen to use a generalized
fish health assessment in EMAP. This assessment
consists of examining fishes for gross pathological
disorders and pathologies of the spleen
(specifically, splenic macrophage aggregates).
High rates of gross pathological and/or spleen
abnormalities can be associated with environmental
contamination.

Having indicators that are sensitive to changes in
their biotic and abiotic surroundings, and
statistically rigorous monitoring designs, still does
not guarantee an accurate estimate of the condition
of an aquatic ecosystem. We must understand how
variable our ecosystems and indicators are, and
reduce this "noise" in the system in order to better
detect our condition "signal."

Variability - Variability, both natural and
anthropogenic, can reduce the accuracy of
estimates of aquatic ecosystem condition, perhaps
leading to erroneous conclusions.  Therefore, prior
to the development of a national monitoring design,
variability-induced statistical errors associated with
condition estimates  must be reduced.  To
accomplish this, the components of variability must
be calculated, and design alternatives evaluated.
Sources of variability may include: indicator
response at different spatial and temporal scales,
technician error,  and choices in allocating sampling
effort within and among population elements (e.g.,
the trade-off between more samples within a lake
or sampling more different lakes). The sources of
variability affecting condition estimation will also
affect change and trend detection capabilities for
aquatic ecosystems.

        Current Status of Variability Research
- EMAP has expended considerable research effort
to understand variability and how it impacts status
and trend detection under different design
considerations (Larsen et al. 1995). To date, we
have been able to:  (1) summarize variance
components for several indicators of aquatic
condition; (2) evaluate the influence of sources of
variability on trend estimation; (3) evaluate the
sensitivity of different sample survey design
options for trend detection; and (4) demonstrate the
ability of sample survey designs to detect trends in
indicators of condition, given the summarized
variance components. (See Appendix IV.)

Even as we reduce the variability in our indicators
and designs and develop more accurate regional
                                        EMAP Research Strategy

-------
                                                                            EMAP Research Strategy
estimates of our indicators, we must have some
reference against which we measure our current
conditions. We can measure relative changes and
trends without reference conditions, but to
determine the current quality and to guide
protection and restoration, reference conditions for
the various aquatic ecosystems are needed.

Reference Conditions - A reference condition
establishes the basis for making comparisons and
for detecting use impairment.  It should be
applicable to an individual water body, such as a
stream segment, but also to similar water bodies on
a regional scale. A critical dimension to the use of
biological indicators is the standard or "reference
condition" against which existing indicators can be
compared to determine impairment. When these
reference conditions become legally adopted as
water quality standards, they become biocriteria.
EPA (U.S. EPA 1998) requirements for biocriteria
include: (1) being scientifically sound, (2) being
protective of the most sensitive biota and habitats,
(3) being protective of healthy, natural aquatic
communities, (4) being protective of biological
integrity, (5) using specific aquatic  community
assemblage characteristics to assess attainment of
designated uses, (6) providing for non-degradation
of water resource quality, and (7) being defensible
in a court of law.
        Current Status of Reference Condition
Research - Our reference condition research has
identified and examined a number of approaches
ranging from best professional judgement to
selection from a frequency distribution based on
probability sampling (e.g., the condition of the sites
above the SOthpercentile). However, no explicit
comparison of approaches has been made.
Currently, we use a combination of best
professional judgmental and probability-based
selection for the development of reference
conditions. Judgement is used to select sites from
excellent to poor along a pre-defined condition
gradient; these are then sampled for the indicator of
interest. Indicator values from the reference sites
are plotted on a frequency distribution, along with
those derived from probability sampling in the
same area of interest. Best professional judgement
is again used to determine the percentile of the
probability sample associated with the lowest level
of excellent condition as defined by the reference
sites, and that becomes the nominal threshold for
excellent condition for the indicator within the area
of interest. In all cases, we provide the explicit
criteria by which we judged reference sites to be
classified along the condition gradient. The criteria
are sufficiently rigorous that the sites can be
classified consistently.

EMAP ecological indicators in a probability
sampling design can provide estimates of aquatic
ecosystem condition in a region where reference
conditions have been established.  However, the
EMAP approach cannot determine cause and
effect, nor can it currently provide estimates of the
seasonal variability in condition. This type of work
must be done at index sites.

Index Sites - Index sites are localized study areas
that offer opportunities for intensive research and
monitoring. They permit detailed research on
ecological processes and mechanisms, and can
serve as known points along a condition gradient.
Continual monitoring at these sites can also provide
estimates of seasonal variability and capture
catastrophic events.

There are numerous existing models for index site
networks. One of these is the National Science
Foundation's (NSF's) Long-Term Ecological
Research (LTER) network, which collects data at
sites across the U.S.  Sites have selected process
studies in common and researchers coordinate
some common measurements across sites. The
LTER sites have provided valuable insights into
processes across a broad range of ecosystems.
Another approach to index sites is to identify a set
of stressors to monitor across all sites within a
network.

         Current Status of Index Site Research -
EMAP researchers have used seasonal data from a
few intensive study sites to improve our
assessments of lakes affected by acidic deposition
in the northeastern U.S. In our regional probability
surveys of northeastern lakes, we determined the
proportion of acidic lakes during a summer index
period (Figure 5). However, because of the
number of lakes visited, it was not possible to re-
sample each of them during the  spring snowmelt
when episodic acidification occurs.

At a smaller number of intensively studied sites,
visited multiple times during the year, the lakes
which experienced episodic acidification during
spring snowmelt were determined. No regional
estimates of spring acidification could be
generated directly from these lakes, because they
were not a statistically random subset of all lakes.
By developing a classification system which
identified sensitive watershed and lake properties,
                                        EMAP Research Strategy
                                            II-9

-------
EMAP Research Strategy
 models relating spring episodes to summer
chemistry for each of these classes were
established. These models were then applied back
to the survey data, allowing the estimation of the
regional extent of episodic acidification (Figure 5).
            -80     0     SO     100    ISO     200
                Add Neutralizing Capacity (neq/1)

 Figure 5. Proportion of chronically and
 episodically acidic Northeastern lakes.
 Episodic proportion modeled from Long Term
 Monitoring (LTM) data.

Clearly, index sites are key to complementing
large-scale monitoring with continual, intensive
research that addresses causal mechanisms and
seasonal events.  In addition to NSF's LTER
network, the USDA Forest Service has instigated
Intensive Site Monitoring as part of the national
Forest Health Monitoring program, and the
National Oceanic and Atmospheric Administration
(NOAA) supports National Estuarine Research
Reserves for intensive studies. While EMAP
scientists have explored these and other index site
networks, the program has learned that it often
requires specialized data from index sites in order
to integrate them effectively with EMAP indicators
and assessment questions.

To address this need, EPA developed an additional
network of sites with the National Park Service
(see Appendix V).  This network focuses on UVB
radiation;  it is now part of the ORD Global Climate
Change Program. EPA's STAR Grants Program,
through funds provided by EMAP, has also
established a set  of coastal index research sites,
known as  CISNet (http://es.epa.gov/ncer), in
cooperation with NSF and NOAA. EMAP
scientists helped to frame the research questions for
the initial  proposal solicitations for CISNet sites
and participated in the relevancy review of the
resulting proposals. As the awarded research is
currently underway, ORD and grantee scientists are
just beginning to explore methods to utilize index-
site information to advance EMAP indicators and
design.
               Key Research Issues for Developing a National
               Design Framework - We have made a strategic
               decision to focus our EMAP researchers on the
               survey design and reference conditions specifically
               for inland surface waters and estuaries in support of
               a national monitoring framework.  We will use the
               EPA STAR program to solicit academic research
               on new indicators, designs, and index sites for
               other aquatic ecosystems (e.g., wetlands). STAR
               will also be instrumental in bringing academic
               research to bear on our ongoing EMAP research
               issues. This research will be incorporated into
               EMAP as the research results develop, and as
               EMAP moves more fully into establishing the
               condition of other aquatic ecosystems in the future.

                       Survey Design - For survey designs:  (1)
               our stream design must include all sizes of flowing
               waters and not just wadeable streams; (2) our
               estuarine design must be applicable to Pacific
               estuaries and must include estuaries beyond the
               three initially used (large, >260 km2; small, 2.6 -
               260 km2; and large tidal rivers (> 260 km2)); and
               (3) our design must allow for the detection of
               change and trends in the condition of streams and
               estuaries.

               Approach: EMAP scientists will conduct research
               to resolve these issues using our geographic study
               areas.  We will expand our core technical design
               group to a larger multi-divisional team approach to
               increase our capabilities.

                       Classification - Classification research is
               needed to provide consistent structural frameworks
               for different types of aquatic ecosystems.  This
               would enable states that share the same or similar
               ecoregions to collaborate on establishing
               monitoring programs and reference conditions.

               Approach: EPA's STAR program will play a major
               role in developing new and consistent classification
               schemes for the various aquatic ecosystems by
               engaging the academic research community.
               STAR will use EMAP-designated grant funds for
               open requests for assistance (RFAs) to obtain the
               best relevant scientific proposals in support of
               improved classifications.  There is also
               classification research being planned in ORD with
               respect to aquatic stressors.  EMAP will coordinate
               with this research effort in addition to the work
               done by the academic community through STAR
               funding.
11-10
EMAP Research Strategy

-------
                                                                            EMAP Research Strategy
        Indicators - EMAP's intramural research
on ecological indicators will focus on assessing and
reducing the variability of our current indicators
when applied in highly diverse new geographic
areas (e.g., Western Pilot, and on calibrating
indicators across geographic areas to facilitate
national data aggregation.  We will also begin
establishing more formalized procedures for
selection of reference conditions.

Approach: The development of new and novel
indicators for future use will be pursued through
the STAR Grants Program.  STAR will develop
RFAs to stimulate academic research on new
indicators (such as the current work on indicators
of wetland condition) using EMAP funds in their
extramural program. Through STAR RFAs, we
will expand the analysis of variability in existing
and proposed indicators. The value of each
indicator for describing current resource  condition
and for trend detection will be determined during
the course of the grant. STAR researchers will also
test hypothesized responses of indicators to
different anthropogenic and natural stresses  and
evaluate these hypotheses in field studies
conducted along known environmental disturbance
gradients. This will allow evaluation of the
optimum balance of indicators and indices from
different taxa.  As part of the grant completion
process, all indicators developed will be  screened
using EMAP's Evaluation Guidelines for
Ecological Indicators (Jackson et al. 2000).  Peer-
reviewed publications resulting from the funded
research will allow wider dissemination, testing,
and use of the  new indicators within the scientific
community. To ensure that the findings of their
grantees are relevant to EPA, STAR intends to
integrate them into Agency science through the
development of state-of-the-science documents and
workshops.

        Variability - Our variability research
issues are: (1)  extracting variance components for
our remaining indicators in all resource areas, and
evaluating trend detection capabilities; (2)
evaluating the variability associated with aquatic
ecosystems in different parts of the country; and (3)
evaluating different design options for trade-offs
between determining status and detecting trends.

Approach: This research will be conducted by
EMAP scientists using data from our large-scale
geographic studies and initiatives, and through
regional applications of EMAP.

        Reference Conditions - Defining
reference conditions for biological indicators to
allow quantifiable measurements of deviation is
critical to the EMAP approach and for the
derivation of biocriteria. These reference
conditions, and the approach to producing them,
must be broadly applicable over large, diverse
areas.

Approach: We intend to accomplish this primarily
through EMAP researchers, using our probability-
based sampling design to provide the  basis for
scientifically credible, regionally representative,
reference conditions. We will use the results from
a probabilistic monitoring design within the
resource to examine the distribution of conditions
for subsequent use in developing the criteria
necessary for determining reference conditions.
We will also examine improvements in reference
condition determination by integrating results from
index sites and our probabilistic monitoring.
Lastly, we will establish approaches for developing
reference conditions in impaired areas, where
reference communities are rare or no longer exist.

        Index  Sites - EMAP must determine the
proper level of integration between survey and
index sites in order to fully evaluate the condition
of aquatic ecosystems at multiple spatial and
temporal scales. A critical component of index-site
research is in developing criteria for reference
conditions.

Approach:  Index site research will be pursued
through EPA's sites, other agencies' sites, and
through sites associated with academic institutions
via the STAR program. For example, the impaired
portions of condition gradients may be explored at
urban LTER sites, and at urban index sites
maintained by U.S. Geological Survey's (USGS's)
National Water Quality Assessment (NAWQA)
program.

LANDSCAPES

Land-use activities within watersheds are the major
causes of non-point source pollution problems
affecting aquatic ecosystems.  By using satellite-
based remote-sensing techniques, a complete
census of the landscape and determination of the
different land-cover types (e.g., pine forest, grass
lands, pasture) in an area can be made.  Changes in
landscape composition and pattern can influence
water quality, the biological condition of streams,
and the risk of watersheds to flooding. If
relationships can be established between landscape
pattern indicators and aquatic ecosystem condition,
                                        EMAP Research Strategy
                                           11-11

-------
EMAP Research Strategy
then a comprehensive low-cost alternative to field
monitoring can be developed.

Current Status of Landscape Research - In 1993,
realizing the potential for remote sensing, EMAP
initiated a partnership with other federal programs
to deliver processed land-cover imagery from
across the conterminous U.S. with reduced cost to
each agency.  The other partners in the Multi-
Resolution Land Characteristics (MRLC,
http://www.epa.gov/mrlc) Consortium were: the
USGS Biological Resources Division's Gap
Analysis Program, the NOAA CoastWatch Change
Analysis Program, USGS NAWQA, and the USGS
EROS Data Center.

The first step towards developing a remote-sensing
capability for monitoring condition was the
production of a national land-cover data base with
30-meter resolution from MRLC data (completed
in Spring, 2000).  The data base was produced from
Landsat Thematic Mapper images taken during
1991-1993, and purchased by Consortium
members.  The national land-cover data base
consists of four components:  1) the land-cover
legend; 2) the spatial and digital format of the data
base; 3) the data layers contained in the national
land-cover data base;  and 4) the supporting
documentation. The MRLC consortium is
committed to an ongoing land-cover
characterization effort on a 10-year cycle. Work
has begun on a revised national land-cover data
base for the year 2000.

        Landscape Indicators - With a
comprehensive and internally-consistent spatial
data coverage (MRLC), indicators of status  and
changes in landscape composition and pattern can
be developed (e.g., Normalized Differential
Vegetation Index (changes in greenness), and
percent anthropogenic cover).  We expect spatial
correlations between a watershed's terrestrial
components and aquatic ecosystems.  Even if these
relationships are not always strong, it may be
possible to use a landscape screening approach
(through an assessment  of indicators) to  identify
those areas where on-the-ground sampling would
need to occur.  Streams  draining areas with
agriculture on slopes greater than 3% may be at
increased risk from sedimentation, and potentially
from non-point source pollution (Figure  6).  Areas
with high amounts of urban development or
agriculture (e.g., >60%  of the land surface) may be
sufficiently degraded so that variation in landscape
pattern has little influence on the area's aquatic
               ecosystem condition. Conversely, areas with very
               low human occupancy may possess significant
               resilience and variation in landscape pattern that
               these areas will also have little effect on aquatic
               ecological processes and conditions. It is the area
               with moderate levels of human occupancy where
               the pattern of use may have the greatest influence
               on aquatic ecological conditions. It may be
               possible to use remote sensing to identify and
               screen out areas that fit the two extremes, and to
               focus our ground sampling on those areas where
               pattern matters to aquatic ecological condition.
               Even this reduction in ground sampling effort could
               result in a large reduction in the  monitoring costs
               for aquatic systems.
               Figure 6. Occurrence of agriculture on slopes
               >3% in the Mid-Atlantic.

               Key Landscape Research Issues - A systematic
               approach to developing landscape indicators to
               assess aquatic ecosystem condition at regional
               scales is  a key research need, if remote monitoring
               is to be feasible. Landscape indicators research
               will focus on the degree to which landscape
               composition and pattern co-vary with water quality
               and stream biotic condition. This research will also
               help determine the role of riparian habitat in
               landscape-water interactions, and evaluate critical
               threshold values of landscape indicators with
               regard to the quality of aquatic resources.
11-12
EMAP Research Strategy

-------
                                                                           EMAP Research Strategy
Landscape indicators are only as good as the
remotely-sensed landscape data from which they
are built. Data problems can include the number of
samples (land-cover data sets may vary from a few
hundred to several million samples), the number of
attributes (e.g., land-cover classes), and scale of
analysis.  Interpretation of landscape indicators is
also influenced by sensitivity of individual
indicators to misclassification embedded within
land-cover and other primary spatial data.
Moreover, many landscape indicators are
calculated by overlaying different spatial coverages
(e.g., woody vegetation (from land-cover data) and
adjacent streams (from Digital Line Graph data)),
which may compound problems.  We must be able
to account for the  uncertainties introduced by these
components.

We also need to determine how our indicators and
their variability are affected by changing the scale
of measurement and the scale of assessment. We
must begin to evaluate the scales at which specific
types of anthropogenic stresses operate, and the
scales at which they can be detected.  Only by
understanding the differences in response to scale
issues among landscape indicators, and their
aquatic ecological analogs, can we evaluate and
integrate information generated at each tier (index
sites, geographic surveys, and remote sensing) of
our monitoring framework (see Appendix 1).

A final research challenge will be to develop
statistically valid procedures for testing landscape
change over time.  Parametric statistical procedures
require independent and similar experimental units,
which are unlikely to be obtained in regional-scale
investigations (Hurlbert 1984, Hargrove and
Pickering 1992). Neutral models (Gardner et al.
1987, Gardner and O'Neill 1991, O'Neill et al.
1992) may be used to specify null hypotheses as
benchmarks to test observed changes, but there is
no a priori neutral model which is appropriate  for
all tests that could be made.

Approach: Landscape indicator research will be
conducted both by EMAP scientists and extramural
academic researchers. The academic community
will be engaged in this research topic through our
extramural funds in the STAR program. We
envision three major focus areas for our intramural
and extramural research activities:  (1) indicators
of landscape condition that are linked to aquatic
ecological condition, (2) indicators of stress on
landscapes, and (3) effects of data properties on
landscape indicator interpretation.
STRATEGIC GEOGRAPHIC STUDIES

The EMAP geographic studies consist of one or
more major regional assessments and ten or more
sub-regional studies (R-EMAP). These geographic
foci are intended to advance monitoring science by
reducing the variability associated with describing
the status and trends in the condition of individual
ecological resources (e.g., estuaries, streams/rivers)
or land cover across the nation.

Current Status of Geographic Studies - In the
Mid-Atlantic Integrated Assessment (MAIA), we
focused on characterizing both the ecological
quality of the region and the important
environmental stressors at multiple scales.  This
resulted in the first "State of the Region"  baseline
assessments for individual types of ecological
systems (stream,  estuarine, and landscape). MAIA
served as the "proof of concept" for the EMAP
approach (see Appendix VI).

In MAIA, where the feasibility of the EMAP
approach has been extensively tested, the
ecosystems are relatively homogeneous and
representative primarily of the mid-Atlantic and the
southeastern U.S. For the EMAP approach to be
applicable to other regions of the country, we must
reduce the uncertainty associated with our
indicators and designs in more ecologically diverse
areas of the country.  We must also demonstrate
that the EMAP approach can be used by the States
to assess the condition of an aquatic ecosystem
nationally.

Key Geographic Research Issues - Reducing the
uncertainty and variability associated with applying
EMAP's design and indicators for estuaries,
streams, and landscapes to areas of the country that
differ dramatically from the mid-Atlantic is a key
EMAP research need.

Approach:  EMAP is focusing its major design and
monitoring research on the twelve conterminous
western states within EPA Regions VIII,  IX, and
X. This is the Western Pilot. Its vast extent
supports high ecological variability, and
ecosystems that differ dramatically from those in
MAIA.  We will  also continue to use the  smaller R-
EMAP studies in each of the  EPA Regions to test
our designs and indicators in other diverse
ecological areas of the country. In R-EMAP, we
can further reduce our statistical design
uncertainties and engage the Regions, States, and
Tribes in using the EMAP approach to address
                                       EMAP Research Strategy
                                           11-13

-------
EMAP Research Strategy
their local environmental problems (see Appendix
VII).

        Western Pilot - The EMAP Western Pilot
will be the largest comprehensive study conducted
by EPA on the ecological condition of the West.
The Western Pilot is a cooperative venture
involving 12 western states, tribes, universities, and
the western EPA Regional Offices.  Initiated in the
spring of 1999 as a five-year effort, the EMAP
Western Pilot will establish the condition of aquatic
ecosystems throughout the West, and identify
stressors associated with the degradation of these
resources.

EMAP's Western Pilot research effort includes
three core components: estuaries, surface waters
(streams and rivers) and landscapes. A probability-
based sampling approach will be used to monitor
the ecological condition of estuarine and surface
waters, and produce state and regional condition
assessments. The landscapes component will make
use of remotely-sensed imagery for a census of
land cover types, and produce a landscape atlas for
the western U.S.

        Estuaries - Our first year's effort (1999)
involved sampling the small estuaries in California,
Oregon and Washington in partnership with these
states, NOAA, and USGS.  In 2000, the Western
Pilot coastal component focused on the large
estuarine systems:  Puget Sound, WA; Columbia
River Estuary, OR; and San Francisco Bay, CA.  In
subsequent years, we will develop the techniques
and designs necessary for monitoring the Pacific
Coast wetlands and the near-shore environment.
"Near-shore" is used here to mean the shallow-
water ocean area near the coastline (approximately
from high tide to 5 km offshore).

Our hierarchical design for western coastal
estuarine resources includes a minimum of 50
locations in the small estuaries of Washington,
Oregon, and California (not including Puget Sound,
Columbia River Estuary, or San Francisco Bay) as
a base.  Our intent is to modify  existing state
programs as little as possible, but still meet our
probabilistic requirements, and  to provide estimates
of estuarine health and guidance for the
development of estuarine reference conditions for
use with biocriteria.

In 1999, areas of design and sampling
intensification were included because of specific
regional interests. Region IX selected Northern
               California coastal streams (Bodega Bay, CA, north
               to the California-Oregon line) and Region X
               selected Tillamook Bay as their desired areas of
               intensification. Nested designs were created for
               these areas that add approximately 30 sites in each
               of the respective sub-regions.

               Over the period of 1997-1999, Washington
               Department of the Environment (DOE) sampled
               300 locations within Puget Sound and its adjacent
               systems to characterize sediment contaminants,
               toxicity, and benthic communities. This effort used
               an EMAP-type probabilistic design and collected
               EMAP  indicators. The Washington DOE data will
               be incorporated into our analyses.

               NOAA had an intensification effort in 2000 that
               resulted in 200 sites located in San Francisco Bay.
               USGS/BRD's Biomonitoring of Environmental
               Status and Trends (BEST) program has also
               proposed intensifications to complement our
               Western Pilot; these would focus on western
               wildlife refuges and measure contaminant effects
               on birds, fish and mammals. While still
               preliminary, we will work to integrate any potential
               intensifications by the BEST Program into the
               Western Pilot.  Also in FY2000, we organized the
               EMAP  field effort that focused on San Francisco
               Bay, Columbia River and Puget Sound; this was
               conducted by the resource/protection agencies  of
               California, Oregon, and Washington, respectively.

               Estuarine conditions are typically assessed by
               EMAP  through the use of biological indicators
               such as benthic community  structure, fish
               community analysis, and the incidence of disease
               or other pathologies in fish. The  presence of
               stressors is evaluated by assessing water quality
               parameters, sediment contamination and toxicity,
               and the presence of contaminants in fish tissue.
               The core EMAP coastal indicators (Table 2) were
               developed in the estuaries of the Southeast, MAIA
               and the Gulf of Mexico, and will be tested in the
               Western Pilot.

                        Streams and Rivers - EMAP has not
               previously undertaken a large-scale sampling of
               streams and rivers in the West. In order to produce
               unbiased estimates of the ecological condition  of
               surface waters of the West, we must establish that
               our current design framework (stream survey
               design, biological indicators, and estimates of
               reference condition developed in MAIA) is
               effective. We will also include an assessment  of
               the potential importance of stressors (habitat
               modification, sedimentation, nutrients, temperature,
11-14
EMAP Research Strategy

-------
                                                                             EMAP Research Strategy
grazing, timber harvest, etc.) in the streams and
rivers of the western U.S.  EMAP, in conjunction
with the States, is sampling perennial streams in the
twelve western states over four years to produce
statewide estimates of condition.  In this period, a
total of 50 sites per state will be monitored.
Intensive study areas in each EPA Region will also
be sampled to address more specific issues of
impairment and reference conditions.  Our
sampling will build on existing state capacities in
implementing our probability sampling framework,
which will allow for the aggregation of ecological
condition information across the West.

Biological, physical habitat, and water chemistry
indicators will be used in our assessment of stream
condition in order to characterize the biological
communities, habitat attributes and levels of stress.
All of our state partners in the  Western Pilot have
agreed to a group of core stream indicators that will
be used (Table 3).  An additional set of research
indicators (riparian condition, continuous
temperature,  microbial organisms, sediment
toxicity, additional tissue contaminants, etc.) may
also be experimentally tested on a smaller scale in
focused areas. These indicators may be added
depending on local  importance and available funds.
Landscapes - Unlike the sampling required for
individual aquatic resources, landscape data can be
gathered "wall-to-wall." Through the use of
          remote-sensing techniques and the availability of
          the MRLC data, the entire western U.S. will be
          censussed. In the landscape component, we will
          assess spatial variability in landscape pattern and
          the degree to which landscape pattern influences
          the conditions of estuaries and inland surface
          waters.  If we can link watershed-level aquatic
          resource condition with landscape patterns, then it
          may be possible to assess conditions of aquatic
          resources  from remotely-sensed landscape data at
          many scales across the western US.

          The western U.S. presents a challenge to
          developing and interpreting landscape indicators.
          There are  limitations to processing the data for an
          area this large. There are also some unique stresses
          on western landscapes, including grazing and
          timber harvest, that do not result in changes in
          land-cover types, but rather in substantially altered
          states of land-cover conditions.  New remote-
          sensing data and analytical approaches will be
          needed to determine the extent and magnitude of
          these stresses.

          We intend to develop a regional-scale land-cover
          database for the western U.S. based on the MRLC
Table 2. Core EMAP coastal indicators
Water Column
Dissolved oxygen
Salinity, temperature, depth
PH
Nutrients
Chlorophyll
Sediments
Grain size
Total organic carbon
Sediment chemistry
Sediment toxicity
Biota
Benthic community structure & abundance
Fish community structure & abundance
Fish pathologies
Fish tissue residues
Submerged vegetation
Table 3. Core EMAP surface water indicators
                     Core
                                                                  Additional Indicators
 Conventional water quality parameters
 Fish assemblage
 Macroinvertebrate assemblage
 Periphyton assemblage
 Physical habitat structure
 Riparian vegetation
Fish tissue chemistry/toxics
Sediment metabolism
Sediment chemistry
Sediment toxicity
Water column toxicity
Amphibians
Bacteria
Biomarkers
Riparian birds
                                        EMAP Research Strategy
                                                      11-15

-------
EMAP Research Strategy
 data set. This database, along with other regional
landscape coverages (e.g., topography, soils, road
networks, stream networks, and human population
density), will be used to assess landscape
conditions across the entire region down to a
resolution of 30 m2. This assessment will use a set
of landscape indicators to evaluate the spatial
patterns of human-induced stresses and the spatial
arrangement of forest, forest edge, and riparian
habitats, as they influence aquatic resources.

The components necessary for a western landscape
assessment are: (1) spatial data acquisition,
assembly, and accuracy assessment; (2)
development of new remote-sensing methods to
detect watershed-level stresses; (3) landscape
indicator generation;  (4) quantify the degree to
which existing and new landscape indicators
explain variation in aquatic resource conditions;
and (5) development of multi-indicator assessment
techniques. During 1999-2002, aquatic and
landscape data will be assembled, landscape
indicators will be developed and calculated, and
aquatic resource-landscape indicator quantification
will begin. In 2002-2003 landscape indicators will
be quantified across the entire West to establish
potential risks to aquatic resources.

         R-EMAP - Our R-EMAP projects will
provide numerous small-scale ecological
monitoring research opportunities across the wide
biogeographic and political boundaries associated
with the ten EPA Regional Offices (see Appendix
VII). We typically sponsor one or two R-EMAP
projects annually in each Region, with a focus on
applications  of the EMAP approach to local
problems. EMAP will continue to use R-EMAP to
complement our larger geographic assessments,
and to provide us with additional opportunities to
develop and test indicators and designs across the
nation.

NATIONAL RESOURCE ASSESSMENTS

A key issue for EMAP is building the capacity for,
and demonstrating the feasibility of, using our
approach by the States and Tribes for
environmental condition assessments that can then
be aggregated to regional and national levels.

National Coastal Assessment - We are now
demonstrating that EMAP's integrated probabilistic
monitoring approach can be used to produce a
comprehensive assessment of the condition of the
nation's estuaries and near-shore coastal
               environments. This is EMAP's National Coastal
               Assessment.  During 2000-2001, all 24 marine
               coastal states (including Alaska and Hawaii) and
               Puerto Rico will be sampled to estimate the
               condition of their estuarine resources. A minimum
               of 50 sampling locations in each state have been
               established within EMAP's probabilistic sampling
               framework.  The results of the coastal component
               of the Western Pilot in 1999-2000 (see previous
               section) will be integrated into this national effort
               and constitute the assessment of the West Coast
               estuaries.

               At all selected sampling sites, measurements are
               focused on ecological and biological response
               variables (see Western Pilot - estuaries).  However,
               sufficient environmental stressor and habitat
               information is also collected in order to interpret
               our response variables.  A special effort is being
               conducted to test ecosystem-level response
               variables delineating function (e.g. system
               productivity,  nutrient cycling, and systems
               energetics) as measures of system condition (see
               NRC 1999).

               Partners  in the EMAP National Coastal Assessment
               are EPA Regions, the Office of Water, state
               resource/protection agencies in the 24 marine
               coastal states and Puerto Rico, USGS and NOAA,
               all of whom are participating in sampling during
               the late summer months of 2000 and 2001.  In
               2002, we will begin our assessment of the
               conditions of the near-shore marine coastal
               environment.  This effort will complement our
               work on the nation's estuaries and EPA's increased
               efforts in beach monitoring.

               National Stream Assessment - The Western Pilot
               is removing the major scientific barriers to the use
               of the EMAP approach for a national stream
               monitoring framework.  As we remove the
               uncertainties  associated with using our designs and
               indicators in the western U.S., we will be in a
               position to provide designs for all the conterminous
               states. Use of these monitoring frameworks by  the
               states will allow the collected data to be aggregated
               nationally into the first national stream assessment.

               Large and Great Rivers - Within the Western
               Pilot, EMAP has begun developing sampling
               designs and indicators for large and Great Rivers.
               Large/Great Rivers have been determined by
               EPA's Office of Water as one of the next major
               national resources that requires a monitoring design
               for establishing ecological condition.  Current and
11-16
EMAP Research Strategy

-------
                                                                          EMAP Research Strategy
future EMAP research will focus on developing
indicators and a monitoring design that will allow
the condition of large/Great Rivers to be
established. This resource is one of our most poorly
understood aquatic ecosystems. Our focus on
large/Great Rivers would also allow the integration
of the findings from a National Stream Assessment
with the condition of the ultimate estuarine
receiving waters through the National Coastal
Assessment.  Large and Great Rivers are the
crucial link between wadeable streams and
estuaries, and an understanding of their condition
will be requisite to developing rational total
maximum daily loads (TMDLs) for both upstream
and downstream waters.

EMAP plans to address the rivers of the
Mississippi basin (which drains approximately 41%
of the conterminous U.S.) to lay the groundwork
for exploring integrated watershed processes from
the upper Midwest to the Gulf of Mexico.
Potential partners with ORD in this effort, and who
have already participated in preliminary planning
discussions, include USGS (NAWQA, BRD, and
Long-Term Resource Monitoring Program), the
U.S. Army Corps of Engineers, EPA Regions IV -
VIII and EPA's Gulf of Mexico Program Office.

INFORMATION MANAGEMENT

As with any large environmental effort, an
effective information management program is
critical. There are two key components to EMAP
information management: data management and
information/technology transfer.

Data Management - The data management
component of EMAP has been a key feature of the
Program for many years (see Appendix VIII).  It
supports the programmatic and policy objectives,
and is an integral part of the landscapes,  surface
waters, and coastal groups. EMAP data are used
primarily by study participants to fulfill study
objectives—assessments of the environmental
conditions of a region or regions. In addition,
because these ecological data are collected under a
consistent design and with consistent methods over
broad regions, are of high-quality, and are well-
described, they have many potential uses beyond
the original study. These circumstances lead to
both EMAP's analytical databases that support the
statistical analyses (Hale and Buffum, 2000), and
general-use databases that are widely available to
secondary users (Hale et al. 1998).
EMAP's data management team will continue to:
provide a Data Directory so that data of interest can
be readily identified; provide access to EMAP data
and metadata (including data authorship
information); and revise and update the EMAP
home page for the Internet on a regular basis.
However,  our approach will be guided by
considerations for sharing national environmental
monitoring data that will allow national
environmental assessments to be conducted. An
element of this will be to ensure long-term data
archival through EPA's STORET system. We are
currently working with STORET data managers to
modify their system to accommodate EMAP data,
and those from States using the EMAP approach.

EMAP will continue to use the Internet for
distribution of data and metadata through our
existing EMAP World Wide Web (WWW) home
page (http://www.epa.gov/einap/). Our intent is to
have all EMAP data publicly accessible through
our EMAP web page within two years of
collection. To this end, each EMAP research group
will initially manage the data it collects and be
responsible for documentation and transfer to the
EMAP home page in accordance with our
standards and formats (see EMAP Information
Plan, Hale et al. 1999). This will provide for a
sustainable, consistent, and continuously updated
data system in support of our continuing research
on aquatic ecosystem condition assessments.  Our
metadata files are periodically uploaded to ORD's
Environmental Information Management System
(EIMS) for broader public exposure (see Appendix
VIII).

Information/Technology Transfer - The transfer
of our approaches and techniques generally occurs
by direct interaction with our partners (EPA
Regions, States, and Tribes) in our study areas.
However,  in some cases (e.g., remote sensing,
statistical design), the techniques or approaches are
sufficiently new or novel that our partners may
have limited expertise. In these cases, we will
establish work groups in each of the involved EPA
Regions, with an overall  steering committee that is
chaired by the EMAP technical lead in the subject.
EMAP will also train and support regional
representatives to help in the  immediate use of the
technology or techniques, and to further transfer
this capability to Regional staff. We have
previously established a GIS/remote sensing team
in EPA Regions VIII, IX, and X in conjunction
with the Western Pilot.  Similarly, we intend to
establish a design and analysis team in the near
future. In addition, we intend to enhance the
                                       EMAP Research Strategy
                                          11-17

-------
EMAP Research Strategy
usefulness of EMAP data to our partners and the
public by developing interactive approaches to
simplify interpretation of EMAP data through our
Internet site.

PROGRAM MANAGEMENT

EMAP is an ORD-wide program which receives
strategic guidance from all of the ORD National
Laboratories and Centers (Figure 2). These
laboratories and centers are also directly involved
in conducting the research associated with EMAP's
goals.

Administratively, EMAP resides in the National
Health and Environmental Effects Research
Laboratory.  However, the EMAP Director is
advised by a Steering Committee made up of the
Associate/Assistant Directors for Ecology from
each of ORD's National Laboratories and Centers.
The EMAP Director is responsible for developing
the detailed research directions and ensuring that
the research is implemented through the working
groups in the laboratories and centers.

Quality Assurance - EMAP intramural research is
conducted by scientists within the organizational
divisions of ORD Laboratories and Centers. As
such, each research product is  subject to the quality
assurance (QA) procedures established by its
originating organization. Extramural scientists
conducting EMAP-sponsored research through
STAR grants are required to submit a quality
assurance plan prior to collecting data. ORD's
rigorous QA standards ensure that research results
intended for EMAP assessments have been
thoroughly verified.

Measures of Success - Since its inception in the
late 1980s, the tenet of EMAP has been
"monitoring for results." This was reinforced by the
enaction of the Government Performance and
Results Act (GPRA) in 1993.  EMAP's success is
predicated on improving the science behind
environmental decisions. Our progress will be
measured in four key areas:
                   West Virginia.  And, in Maryland, the
                   Department of Natural Resources (using
                   EMAP's design) developed the report "The
                   State of Maryland's Freshwater Streams"
                   (EPA/903/R-99/023), which is being used to
                   make planning decisions in support of the
                   Governor's Smart Growth Initiative.

                   Adoption of Methodology -
                   Implementation and routine use of EMAP's
                   statistical design and ecological indicators by
                   States and other monitoring agencies is an
                   unmistakable sign of acceptance.  To date,
                   more than 30 States have implemented or are
                   testing EMAP protocols.  In addition, the U.S.
                   Forest Service has adopted EMAP's rotating
                   panel design for its national survey programs.
                   Also, EPA's Office of Water has issued draft
                   guidance to the States on using probability
                   designs (EMAP approach) for monitoring the
                   condition of all waters within the State under
                   305(b) of the Clean Water Act.

                   National Data Aggregation  -
                   The National Coastal Assessment is currently
                   demonstrating that EMAP indicators can be
                   meaningfully aggregated and compared across
                   U.S. regions. Similar verification for streams
                   data is expected within the next few years.

                   Trends Detection -
                   The ability to detect changes  over time will
                   complete validation for EMAP's statistical
                   design. It will require long-term
                   implementation of EMAP methods, as
                   continued sampling over multiple years will be
                   necessary to analyze for trends. We have
                   begun looking at change detection between
                   sampling periods as a prelude to having
                   sufficient data to detect trends. We have been
                   able to detect trends at the basin-level through
                   R-EMAP, and we will be examining change
                   detection at the scale of estuarine provinces
                   through our National Coastal Assessment.
    Effect on Decision-Making -
    Most directly, we can examine the use of
    program results in environmental decisions and
    policies. Already, EMAP data on streams in
    the mid-Atlantic have provided the
    justification for EPA to require a full
    Environmental Impact Statement for proposed
    mountain-top removal for mineral extraction in
11-18
EMAP Research Strategy

-------
                                                                                   Appendix I
Appendix  I.
 Needs  for  Improved
 Environmental  Monitoring
The need for significant advances in the way EPA
and other federal agencies monitor the condition of
our environment has been, and continues to be,
recognized (National Research Council (NRC)
1977, U.S. General Accounting Office (GAO)
1981, 1986, 2000, USEPA 1987).


In 1988, the U.S. EPA Science Advisory Board's
(SAB's) report, Future Risk: Research Strategies
for the 1990s (U.S. EPA 1988), was the stimulus
for many changes in EPA research.  The report
concluded that EPA needed more research on
relating the effects of cumulative, regional, and
long-term anthropogenic disturbances to ecosystem
responses.  Increased research was also needed to
develop  ecological indicators and protocols for
monitoring, and to analyze and quantify uncertainty
in assessments resulting from monitoring data. The
goals of such research were improved detection of
ecosystem status and trends, and greater predictive
capability. The authors  recognized that great
benefit could be derived from the identification of
trends in environmental  quality before they begin
to cause serious ecological or human health
problems. They recommended that EPA take steps
to enhance its ability to anticipate environmental
problems before public fears are aroused, and
before costly, after-the-fact  clean-up actions are
required. They also recommended that EPA
broaden its data-gathering and assessment efforts.
Embodied in their recommendation was the
perspective that monitoring programs can be
valuable for their ability to paint a picture of
present conditions, and if continued, they can help
describe what has happened to the quality of an
ecosystem over time. Their recommendations
urged EPA to begin monitoring a far broader range
of environmental characteristics and contaminants
than it had in the past.


Toward these ends, the SAB recommended that
EPA undertake research on techniques that can be
used to help anticipate environmental problems and
make a more concerted effort to be aware of, and
interact with, the research efforts of other Federal
agencies concerned with these problems.  EPA was
urged to  evaluate environmental trends and assess
other predictors of potential environmental
problems before they become acute.


The SAB recommendations, the emerging vision of
ecological risk assessment within EPA, and the
importance of high quality monitoring information
in this risk assessment paradigm were responsible
for the creation of the Environmental Monitoring
and Assessment Program (EMAP) within EPA's
Office of Research and Development (ORD).
EMAP's challenge was to develop the tools
necessary for measuring the condition of many
types of ecological resources and the designs for
detecting both spatial and temporal trends (Messer
etal. 1991). EMAP used a tiered monitoring
approach (Figure 7) and focused on developing
indicators of ecological condition and new
monitoring designs for major classes of natural
resources such as surface waters, estuaries, forests,
and wetlands.


It became apparent that EMAP alone could not
develop the designs and indicators for all the
nation's resources.  Also, EMAP would not be able
to implement and maintain a national monitoring
program in the states without additional resources
and the partnership of tribal, state, and federal
agencies. Since States, Tribes, and EPA have
statutory responsibilities within the Clean Water
Act to monitor the surface waters in their  States
and in the country, respectively, and because these
waters integrate the atmospheric, landscape and
upstream inputs, aquatic ecosystems were chosen
as EMAP's focus.


Under the auspices of the White House Office of
Science and Technology Policy, the  Committee on
                                    EMAP Research Strategy
                                       AI-1

-------
Appendix I
Environment and Natural Resources (CENR)
formed the Environmental Monitoring Team in
1995. The Environmental Monitoring Team took
the crucial step of bringing federal agencies
together to shape a national framework for
integration and coordination of environmental
monitoring and related research (CENR 1996).
The framework calls for all environmental agencies
to merge efforts in forming a national monitoring
and research network which will link remote
sensing, regional surveys, and intensive, multi-
resource monitoring areas.  Also, this framework
unites the respective agencies in achieving a
common national goal of understanding and
managing our ecological systems for their sustained
use and enjoyment. Their framework was very
similar to EMAP's multi-tier monitoring approach
(Figure 7). The interagency nature of the
framework allowed EMAP to focus on priority
EPA needs associated with aquatic ecosystems.


Also in 1996, the National  Research Council
(NRC) convened the first National Forum on
Science and Technology Goals.  The NRC called
for a greater focus on monitoring in order to build
better understanding of our ecological systems.
They felt current ecological data and understanding
were inadequate to: 1) Detect, monitor, and
characterize environmental changes; 2) evaluate the
consequences of human activities; and 3) provide
an information base for sustainable management
(i.e., "no loss") of both natural and human-
designed ecological systems. Current programs did
not address these issues in a sufficiently coherent
and comprehensive manner on a national basis to
determine what actions were needed to achieve a
desired environmental quality; these programs also
could not detect changes between current and past
conditions or trends in condition for projection of
future ecological states. Indicators were needed to
measure the status of ecological systems, to gauge
the likelihood of meeting society's environmental
goals, and to anticipate problems resulting from
economic growth.  The NRC concluded that we
were spending substantial time and money to
collect data that were neither compete nor always
relevant to the  decisions that society needed to
make about land use, transportation, industrial
activity, agriculture, and other human activities. A
cost-effective approach to monitoring the condition
of the environment on a national level needed to be
developed.

Currently, all states monitor a  subset of their waters
               because it is cost-prohibitive to physically monitor
               all their waters. However, most of this monitoring
               does not allow for statistically valid assessments of
               water quality conditions in unmonitored waters
               (GAO 2000). These data gaps limit the state's and
               EPA's abilities to identify water quality problems
               and set priorities,  and to carry out key management
               and regulatory activities. These data gaps are
               particularly serious for non-point sources, which
               are widely accepted as contributing to the majority
               of the nation's water quality problems (GAO
               2000).
                 Increasing
                 Spatial
                 Resolution
Increasing Site
Characterization
                             National and Regional
                             Geographic Surveys
                         Landcover and Remote Sensing
               Figure 7.  Tiered monitoring approach used in
               EMAP and the CENR Strategy.
AI-2
EMAP Research Strategy

-------
APPENDIX  II.
Resource  Sampling
                                                                                      Appendix II
Lake Sampling - EMAP defines target lake
populations as standing bodies of water >1 ha, with
1,000 m2 of open water and a maximum depth >1 m,
(Larsen and Christie, 1993).  Lakes are treated as a
discrete resource, and a two-stage approach is used to
identify the unique probability-inclusion point for
each lake.  We use digitized databases (USGS Digital
Line Graphs (DLG) database and USEPA RF3 river
reach files) to identify lakes in the templates around
each grid point. These databases contain information
at the same scale as 1:100,000 topographic maps.
The lake coverages are extracted from the databases
and a unique label point is associated with each lake
by a geographic information system (GIS) for use as
the point for the lake. The DLG representation also
provides the surface area for each lake, and only
lakes with areas >1 ha are extracted.


A GIS is used to overlay the random templates on the
lake points, and a list of the lake points covered by
templates is extracted. The identified lakes are then
checked against several sources to remove any known
non-lakes. Some non-lakes will still remain, but
these will be detected during reconnaissance or the
field visit.  This list of lakes makes up the first-stage
sample.


The size distribution of the first-stage sample of U.S.
lakes is strongly skewed toward the small end.
Almost 25% of the lakes have areas <2 ha, 80% have
areas <10 ha, and almost 95% have areas <50 ha.
Equal probability  sampling of the first-stage sample
would result in very few samples in the ecologically
important moderate and larger-sized lakes.
Moreover, experience with pilot studies indicates that
many of the sites identified in the digitized database
will not meet the operational definition of a lake.
This problem is particularly acute for the sites
identified as small (<5 ha) lakes. In order to counter
these features of the population in the second-stage
framework, several size classes were defined and
target sample numbers were assigned for each size
class.
Stream Sampling - The stream resource does not fall
neatly into either the discrete or extensive category.
To deal with this, EMAP focuses on the population of
stream miles, and we characterize the population in
terms of the condition of miles of streams. Therefore,
we want a sampling method that samples a stream in
proportion to its length.  To accomplish this we view
streams as an extensive resource with length.  This
method has been used in MAIA, and it appears
suitable for use in other regions.


The sampling templates used in lake selection are
also used in stream selection. Stream traces are
identified on l:100,000-scale DLGs, and a GIS is
used to intersect these with the sampling templates.
Each stream segment within a template is identified
and its length determined. The endpoints of a
segment are defined as confluences, headwaters ends,
or intersections with a template edge. Sets of
connected segments of the same order are always
kept together in the sample selection process.  The
appropriate stream order is also determined for each
segment (e.g., headwater streams are 1st order).


As with lakes, some differential weighting by size is
necessary because of the predominance of lower-
order streams. The weighting factors that are used
are 1, 2, and 4 for orders 1, 2, and 3, respectively.
The weighting factors are chosen to produce
approximately equal sized samples  of 1st, 2nd, and
3rd order streams.


The sample selection proceeds in very much the same
manner as for lakes, with inclusion probability for a
segment proportional to its length times the weight
for its order. The total inclusion probability for each
template is calculated as the weighted sum of stream
lengths in the template, the templates are partitioned
into groups using a partitioning algorithm, and the
samples are selected. The same systematic selection
protocol is used; the partitions are randomized, the
templates are randomized within the partitions, and
the sets of connected segments are randomized within
                                     EMAP Research Strategy
                                          AII-1

-------
Appendix II
the templates.  The selection not only identifies the
stream segment to be sampled, but also identifies the
point on that segment where the sample is to be
located. This is accomplished by recording the
relative distance from the beginning of the segment to
the selected point on the segment.


Estuarine Sampling - EMAP uses estuarine
provinces for classification purposes. In the
continental United States these provinces are:
Acadian Province, Virginian Province,  Carolinian
Province, West Indian Province, Louisianian
Province, Californian Province, and Columbian
Province (Figure 8; Holland, 1990). We use a
classification scheme to subdivide estuaries within a
province into classes that have similar physical
features and are likely to respond to stressors in a
similar manner. The defined classes include (1)
large, continuously distributed estuaries (e.g.,
Chesapeake Bay, Long Island Sound), (2) large,
continuously distributed tidal rivers (e.g., Potomac,
Delaware, Hudson Rivers), and (3) small, discretely
distributed estuaries, bays, inlets, and tidal creeks and
rivers (e.g., Barnegat Bay, Indian River Bay,
Elizabeth River).
Within each estuarine class, elements of systematic
             and random sampling are used. Large, continuously
             distributed estuaries are sampled using a randomly
             placed systematic grid. A ninefold enhancement of
             the EMAP grid is used, with one-quarter of the points
             sampled each year.  Grid points are about 18 km apart
             in all directions, and the entire estuary is sampled.
             Large tidal rivers  are sampled along systematically
             spaced transects across the river channel. Transects
             are located about 25 km apart.  The starting point for
             the first transect is randomly selected between 0 and
             25 miles from the mouth of the river.  Two sampling
             points are located on each tidal river transect; one is
             randomly selected and one is an index sample,
             located in the deep channel of the river. Small,
             relatively discrete estuaries are sampled using a
             population approach. First, a list of all small
             estuaries is defined, then the estuaries to be sampled
             are randomly selected from the list without
             replacement. The list is geographically ordered from
             north to south, split into groups of four, and one unit
             is selected from each group. Two  sampling points are
             located in each small estuary that is sampled;  one is
             randomly selected and one is an index sample,
             located in the deep, depositional portion of the
             estuary.
                                          EMAP-Estuaries
                                    Biogeographical Provinces
                                     Continental United States
                   Columbi
                   California!
                                                                             cadian
                                                                      Carolinian

                                                                         est Indian
           Figure  8.  EMAP Ecological Provinces used for classification of estuaries in the
           continental U.S.
AII-2
EMAP Research Strategy

-------
                                                                                          Appendix II
Wetland Sampling - We have not yet settled on a
single approach to sampling wetlands. Our pilot
studies, including one on coastal salt marshes,
selected sample sites using an intensified grid.  We
are engaged in identifying suitable means of
locating, characterizing, and delimiting wetlands,
and in developing suitable indicators of wetland
condition both through intramural efforts and
extramurally through the STAR Grants Program.


Great Lakes Sampling - Hedtke et al. (1992)
defined the Great Lakes resource as the waters of
the Great Lakes and the sediments below them at
high water including: river mouths up to the
maximum extent of lake  influence; wetlands
contiguous to the lakes; and the connecting
channels, Lake St. Clair, and the upper portion of
the St. Lawrence Seaway.
The domain of the Great Lakes resource consists of
five regions within the Great Lakes basin
corresponding to the five Great Lakes: Superior,
Michigan, Huron, Erie, and Ontario.  Each of the
lakes is physically distinct. The connecting
channels tend to have some basic characteristics of
their upstream lakes but are physically unique.
EMAP is initially pursuing a monitoring program
that will provide  status and trends in the five Great
Lakes through intramural efforts.


The Great Lakes  resource can be split into four
classes for sampling purposes: nearshore waters,
offshore waters, harbors  and embayments, and
wetlands.  Because of the different spatial and
biological properties of each of the classes, the
sampling strategy will likely be different for each
of them. The nearshore resource class consists  of
waters adjacent to the shoreline and no more than
85 m deep, and the offshore class is the remaining
central portion of the lake.  The offshore and
nearshore  areas of the lakes will be determined
from GIS-based bathymetric maps  digitized on a 2-
km grid (except Lake Superior which is on a 4-km
grid).  The two open-water classes  are extensive
resources, and the EMAP grid can  generate a frame
for these classes. A threefold enhancement of the
grid would be used for the nearshore  class in order
to provide an adequate number of samples.  For
both open-water classes, sampling would take place
at the grid point.


Harbors and embayments would be treated as a
separate resource class for two reasons. First,
many embayments are formed by tributaries that
are a source of nutrients and contaminants, and thus
have differing water quality and higher variability.
Second, harbors are often the site of more intense
human activity, such as industry, shipping,
recreation, and dredging, and thus may be sources
of nutrients and contaminants to the rest of the
nearshore zone and to offshore waters. Harbors
and embayments do overlap.  For example, there
are harbors at the mouths of tributaries in Lake
Michigan such as Milwaukee Harbor, Benton
Harbor, and Green Bay. There are also
embayments that are not harbors and harbors that
do not have tributaries associated with them.
Rather than try to make distinctions between these
two types of subclasses, we would consider them
as one resource class. Definitions of these areas
will be physical (natural), based on structures such
as breakwalls and docks or easily recognizable land
features such as peninsulas or points.


The lateral extent of the harbor or embayment area
will be the same as the general nearshore area (i.e.,
85-m depth contour). If the harbor or embayment
includes a tributary, the area will include the
tributary mouth and upstream to the zone of lake
influence as defined by conductivity gradients. If
the tributary is dammed, the area would include the
first dam or zone of influence as applicable.  Those
areas  of the lakes that are often referred to as bays
(e.g.,  Saginaw Bay, Green Bay) and yet are
themselves large enough to contain smaller bays,
would be treated as part of the general nearshore
class. Minimum size for an embayment has yet to
be determined. Preliminarily we would use a
minimum area approach, rather than a minimum
distance across the mouth of a harbor or bay.


Two techniques could be used for sampling harbors
and embayments.  One approach treats them as a
discrete resource, where all harbors and
embayments would be identified by examining the
lake perimeter with the aid of a GIS.  A natural
ordering would be induced by the lake perimeter,
so a systematic sample of the ordered list of
harbors and embayments would have good spatial
coverage, and characterization would be in terms  of
numbers of harbors and embayments. The other
approach would be to characterize them as units,
which would require multiple samples per unit, or
carefully selected index samples.
                                       EMAP Research Strategy
                                          AII-3

-------
Appendix II
Wetlands in the Great Lakes will be sampled              above), once we have identified a consistent
similarly to our general sampling of wetlands (see          sampling approach.
AII-4                                  EMAP Research Strategy

-------
                                                                                       Appendix III.
APPENDIX
Indicators
Indicator Development - Indicators can be
developed for any level of biological organization
(Table 4). However, the structural and functional
aspects of the biological or ecological characteristic
to be used as an indicator must be appropriate to
the question being asked.


Indicator Measurement - To measure an indicator
requires a standard protocol, whether in the field or
the laboratory, and a documentation of the type and
amount of sampling necessary. An example might
be using the number offish species present in a
stream reach as an indicator of water quality.  A
protocol such as electroshocking could be defined
and then the length of stream reach necessary  for a
predetermined level of accuracy must be found.  In
fact, Reynolds et al. (in prep.) have determined the
length of stream required to do this in terms of
stream channel width (Figure 9).


Indicator Responsiveness - Evaluating the degree
to which a particular indicator actually responds to
various stressor gradients, or if a stressor indicator
responds to changes in the source, is an important
aspect of the indicator development process.
Without some knowledge of the shape of the
response curve and the variability associated with
the response, it is difficult to evaluate the utility of
any particular indicator. This process must begin
with an hypothesis about the expected direction of
response and then a quantification of the actual
response to  stressor gradient (Figure 10).


Indicator Variability - A description of the
components of variability which impact status and
trends estimation in monitoring is provided by
Larsen et al. (1995).  The important consideration
to EMAP is the extent to which variability in
measuring the indicator (noise) masks detection of
the "change of status" signal.  This may be natural
variability such as the real differences within a
body of water or real differences due to period
when the sampling site is visited. It also contains
variability due to crew errors and differences
among crews, the laboratory processing and other
extraneous components. One way to evaluate the
signal to noise for a particular indicator is to look at
the ratio of population variability relative to the
measured indicator's variability.  If the ratio is
small, then detecting signals will be difficult
(Figure 11).


Ultimately, sufficient information on indicator
variability is needed to determine the power of the
indicator to detect a trend. Power analysis for
detecting trends in lakes using Secchi transparency
and zooplankton species richness was shown in
Figure 11. In this example, if we have  a sample
size of 50 lakes per year and have a 2% per year
change in Secchi transparency, we would be able to
detect it after eight years with 90% power.  In
contrast, it would take 11 years to detect a similar
trend in zooplankton species richness, with
comparable power.


Indicator Demonstration - During the final stages
of indicator development, potential users must be
given some sense of how well the indicator
performs and how it can be used. This  often
implies demonstrating the indicator in a regional
scale project such as a geographic initiative or R-
EMAP assessment and showing the results and
potential conclusions which may derive from using
the indicator (Figure 12). This is part of a
necessary "proof-of-concept" for the indicator to be
accepted into widespread use (see Jackson et al.
2000 for complete guidelines).
                                       EMAP Research Strategy
                                        AIII-1

-------
Appendix III.
Table 4. Levels of biological organization to consider during indicator development with examples of
structural and functional aspects of each level.
Structure
Heterozygosity
Condition
Anomalies/Deformities
Maximum Size
Tissue Contamination
Abundance
Age Class Distribution
Size Class Distribution
Relative Abundance
Richness -Native
Richness - Total
Evenness
Trophic Composition
Reproductive Composition
Habitat Guilds
Regional Diversity (gamma)
Homogeneity
Hot Spots
Patches
Patterns
Fragmentation/Recovery
Level of Organization
Gene
Individual
Population
Assemblage (Community)
Watershed or Landscape
Processes
Polyploidy Rate
Mutation Rate
Recombination Rate
Metabolic Rate
Growth Rate
Fecundity
Reproduction Rate
Growth Rate (of Population)
Death Rate
Evolution/Speciation
Competition/Predation
Disease/Parasitism
Mutualism
Recovery Rate
Water Delivery
Chemical Delivery (Native and Exotic)
Material Delivery (Sediment, Wood)
Energy Flow
Nutrient Cycles and Spiraling
Population Sources and Sinks
Fragmentation Rate/Recovery Rate
                       100
                                   20      40      60
                                      Stream Length
                      80
Figure 9. Electrofishing in small streams indicates that sampling a stream length equal to 40 channel
widths would result in about 90% the toal species present.
AIII-2
EMAP Research Strategy

-------
                                                                                 Appendix III.
                               100
                                80
                             of
                                60
                                            -1012
                                                 Habitat
                                       Human	


                Figure 10.  Example of  observed response of indicator to  stressor
                gradient.
                % Sand/Fine

        Fish
                 62


               60
35
                                                     12

                                                    11
                                              0.8

                                            0     10    20    30    40    50    60    70

                                                          Signal  : Noise
Figure 11. Signal to noise ratio for stream physical habitat indicators. If the ratio is small, then detecting
signals will be difficult. The first 3 indicators would be potentially useful.
                                     EMAP Research Strategy
                    AIII-3

-------
Appendix III.
Total Nitrogen
Riparian Habitat
Instream Habitat
Mine Drainage
Acidic Deposition
Fish Tissue Contamination
Total Phosphorus












zn





i

i


i

i



                                         10    20     30    40     50
                                     Percent of Stream Miles
  Figure 12.  Stressors associated with impaired stream water quality (as determined by our
  stream benthic invertebrate IBI) in the Mid-Atlantic Highlands.
AIII-4
EMAP Research Strategy

-------
                                                                               Appendix IV.
APPENDIX   IV.
Components  of
Variance  in
Indicators
The use of particular indicators in estimating status
and trends requires evaluation of the magnitude and
the effects of different components of variance.
The choices for minimizing variance components,
or their effects on estimates of status
and trends, depends upon their relative magnitudes.
Using indicators of lake trophic status as an
example, the components of variance that could
have important effects on status description and
trend detection are:
                                                   s l
                                                             +
    total    =    population   +   year
    variance     variance         variance
It is possible to separate some of the variance
components into deterministic and random parts to
the extent that exogenous covariates can be
identified. For example, some of the seasonal
variation in water quality variables in streams can
be associated with variation in stream flow. The
identifiable effects of the deterministic component
can be removed from the variance associated with
the covariate (Steel and Torrie,  1980). In the
following examination of variance, the
deterministic components are treated as if they
have already been accommodated, and what
remains must be handled in other ways.

Population Variance  (ci2lake)  - lake effects -
Population variance describes the  measured
differences among lakes in a regional population or
subpopulation during the index period. For
example, suppose there were 20,000 lakes in our
population of interest.  One hundred were picked at
random as the sample of lakes on which indicators
of condition were monitored in a particular year.
In the absence of other variance (described next),
+   interaction +
    effects
    variance
index
variance
  this snapshot expresses the status of the population
  during the yearly index window.  We use
  cumulative distribution functions (CDFs),
  histograms, or other population descriptors to
  represent a population's status. If no other forms of
  variation interfere with the sampling process, index
  snapshots derived from the randomly chosen
  sample of lakes would unambiguously express
  population variance.

  Year Variance (a2year)  - year effects - Year
  variance measures the amount by which all lakes in
  a population or subpopulation are high or low in a
  particular year (Figure 13 A). The condition of
  regional populations of lakes fluctuates around a
  central value in the absence of a trend. In the
  presence of a trend, this variance component
  measures the year-to-year variation from the trend
  line or curve. Regional trend detection capability is
  very sensitive to the relative magnitude of this
  component of variance. This population level
  variation is sometimes called a year effect, since it
  is a measure of the amount by which all lakes in a
                                    EMAP Research Strategy
                                      AIV-1

-------
Appendix IV.
                                                 Panel A
                             .— 10
                             £ 9
                             2 8 (
                            ;l 7
                           as
                                0 (	
                                                5678
                                                  Years

                                                Panel B
                                                                 10   11  12
                           !!
                             S  2
                             *~  o I
                                               S   6   ?
                                                 V»ars
                                                                 10  11   12
                                                      from the year-to-year differences confounds two
                                                      components of variance, a2lakei     and a2lnde.
 Figure 13.  A hypothetical example of time series for Secchi transparency data in five different lakes;
 approximately 2 percent per year trend was imposed.  This illustrates the concordant pattern if Year
 Variance is high (A) vs. the nonconcordant pattern if Interaction Variance is high (B). In Panel A, cr2 year
 = 2.25, with CT2lake.year = 0; in panel B, a2 year  = 0,  with CT2lake.year = 2.25.

particular year are above or below a long-term
central value or trend. This common pattern of
variation among lakes is caused by regional-scale
factors affecting the population in a consistent way,
such as regional-scale climatological conditions
(e.g., wet years or dry years). Magnuson et al.
(1990) refer to this variance component as temporal
coherence.

Resource-year Interaction (cj2lakeiyear) -
interaction effects - The condition of an
individual lake fluctuates from year to year around
its central value or around a trend for that lake
(Figure 13B). These fluctuations are responses to
local effects operating at the individual lake level
and are unrelated among lakes. This component of
variance  specifically describes that part of a lake's
year-to-year variation not already accounted for by
the year component. This interaction variance is a
natural feature of the populations under study and
could be  of interest in itself as an indicator of stress
or change in ecosystems. It can be estimated by
repeat visits to lakes across years. However, if
lakes are sampled each year without revisiting
during the index period, the variance estimated
                                                       (described next).  There is no way to separate these
                                                       without revisiting multiple lakes for several years.

                                                       Index Variance (a2lndex) - index effects - Index
                                                       variance is the aggregate variation seen by repeated
                                                       applications of a sampling protocol for the
                                                       indicator of interest during the index period. Index
                                                       variance can be broken down into several
                                                       components that can be estimated, some of which
                                                       are natural features. If index variation is
                                                       unacceptably high it is worth evaluating the
                                                       components of index variance to determine a
                                                       strategy for minimizing its effect. A useful strategy
                                                       is to estimate this by repeat visits during the index
                                                       period and evaluate its magnitude relative to other
                                                       variance components. If it is unacceptably high,
                                                       then one should determine which of the
                                                       subcomponents can be reduced. The following is a
                                                       partial list of components of index variance that
                                                       might be targets for reduction:

                                                               Temporal (or seasonal) variance and
                                                               trends within the index period describe the
AIV-2
                                        EMAP Research Strategy

-------
                                                                                Appendix IV.
variation and consistent temporal patterns
in the indicator of interest during the index
period. Identification and characterization
of variance and trends within the index
period could allow removal of the effects
by redefining the index period. For
example, if sampling during the rainy
season increases sample variance, then
this would be a poor choice for an index
period.

Measurement variance arises from the
measurements made on the samples
collected. Most standard operating
procedures (SOPs) target measurement
variance in documenting data quality.
This component includes variation
introduced anywhere in the sequence of
events, from the point of sample collection
to the final resting place of the
validated/verified data point in the
database. This might include
measurement error in reading the depth at
which a sample was taken or a
transcription error in recording the data.

Procedural variance expresses the
variance that arises from applying the
same approach within a single sampling
period.  For example, if the  index
sampling procedure calls for collecting
samples at various locations in a lake and
consistent location patterns can be
detected,  then their effect can be reduced.
Modification and standardization of
procedures can minimize this component,
if reduction in its magnitude is cost
effective.

Team-to-team variance comes from
differences introduced by different
sampling teams that use the same
procedure.  In regional-scale monitoring
programs, it is not feasible for a single
team to visit all lakes selected for
monitoring. However, the magnitude of
team-to-team variance can be reduced or
controlled by using as few crews as
possible.  Training, experience, and
execution of follow-up audits during
sampling are also effective at reducing
this variance component.
        Remaining index variance consists of all
        other unaccounted for variance (e.g., the
        indicator noise exhibited during the index
        period).

Neither a2lake*year nor a2year  are subject to
reduction by methodological improvements, as
might be possible with index variance. These
components of variance are natural features of the
ecosystems we study, and just as we characterize
the condition of lakes, we must also characterize
these components of variance. We must both
estimate their magnitudes and evaluate the resulting
influence on the precision of estimates for status
and sensitivity in detecting trends. If these
components are unacceptably high, the options for
improving precision are to increase the number of
lakes sampled (a2lakeiyear)  or the number of years
of monitoring (a2year) , or to identify covariates
that can be used to remove the effect of variance.
We sometimes combine  ci2lakeiyear   and ci2lndex
as residual variance and use:
Estimates of Variance Components - Evaluating
the potential sensitivity of monitoring designs, such
as EMAP, requires estimates of the variance
components. There are two options for obtaining
these estimates. We can acquire databases on
which to perform analyses of variance or use
published accounts.  An ideal database with which
to estimate variances would contain data derived
from a consistent sampling program, including
many lakes (more than 20 for effective variance
estimation) across many years (more than five
years), and encompassing large regions (generally
the size  of several states).  To our knowledge, such
a data set is not available. However, some state
water quality agencies have supported consistent
monitoring programs in which a common set of
indicators of trophic condition have been measured
across many years. We obtained databases from
Maine, Minnesota, New York, and Vermont
containing measurements on total phosphorus (TP),
chlorophyll-a (chl-a), and Secchi disk transparency
(SD) that could be used to estimate magnitudes of
the variance components.

Each of these states uses slightly different
approaches to monitoring the trophic condition of
lakes. Thus, it would be unwise to combine data
across states. Differences arose in the frequency of
                                EMAP Research Strategy
                                         AIV-3

-------
Appendix IV.
sampling, but all lakes were visited several times
during the open water season. The states were also
using slightly different methodologies for
collecting samples (sometimes methods differ even
within states).  However, we were able to create a
single number representing each water column
sample for each state.

These were the databases we used to isolate the
components of variance pertinent to evaluating
EMAP's trend detection capability. We selected a
set of lakes for which data were available over
many consecutive years (at least 4) and during
periods corresponding to a July-August index
period. From this we were able to estimate the
variance components by  the General Linear Model
procedure (SAS, 1989).

Effect of Variance on Trend Detection - Trend is
considered here as a consistent change in an
ecological attribute over time. At the site-specific
scale, this usually means the time series trajectory
of the particular attribute (e.g., population mean).

Trend Detection Models - We use two models to
show the effects of the variance components on
trend detection. The first is a regression model
based on simple linear regression techniques. The
role of the variance components is explicit in this
model.  This can be used for a quick assessment of
the effects of the different variance components,
but the model leaves out some of the technical
details. The second model is a more detailed
description of the variance components that affect
trend detection capability, including effects of
temporal autocorrelation. It was initially used to
evaluate several alternative monitoring designs for
estimating population trends for EMAP-like
designs (Urquhart et al. 1993).

Trend detection capability  can be described in
terms of the variance associated with the slope of a
trend line, Var(p). Var(p)  can be translated into a
95% confidence interval estimate, (approximately
+2 standard errors of the slope). Casting this as a
null hypothesis, we can test whether the slope = 0.
For the null hypothesis to be rejected, the slope
must be greater than two times its  standard error to
be detected at oc= 0 . 0 5 .  We could then be
reasonably sure of detecting trends which are of
greater magnitude.

The general strength of a trend is a function of the
variance in the attribute under study and the
               measure of time.  The ability to "see" the trend
               depends upon the magnitude of these two features
               (e.g., the greater the period of record the clearer the
               trend and the smaller the variance of the attribute
               around the trend line, the clearer the trend).

               The variance of the slope is calculated as the ratio
               of these  two components (e.g., Draper and Smith,
               1981; Snedecorand Cochran, 1967):
                              <7
                                                  (1)
               The numerator contains the variance of the attribute
               around the possible trend and the denominator
               contains a variance-like term associated with time
               (years). Fundamentally, the development of
               monitoring designs and their resultant sensitivity to
               trend detection depend on our ability to minimize
               the numerator and maximize the denominator.

               This basic framework is useful for evaluating
               choices in designing monitoring systems and for
               focusing our efforts on important sources of
               variance over which we might have some control.
               Expansions of the numerator of Equation (1) show
               how the effects of different variance components
               (both natural and measurement error) can be
               accommodated or reduced. This framework is a
               useful approach for evaluating choices about
               allocating sampling effort.

               Trend detection in a particular ecosystem attribute
               at a site across years requires one or more within-
               year measurements of the attribute taken across
               multiple years. Two general components of the
               attribute's variation are important: 1) the
               combination of within-year natural variability and
               measurement error (a2lndex) that produces variation
               around the central tendency for that year; and 2) the
               year-to-year differences in condition that might not
               be attributable to trend (ci2annual).  How well the
               yearly condition is specified depends on  the
               magnitude of variation and the number of
               measurements. Even if we estimate the yearly
               condition variation, a trend could arise due to the
               year-to-year differences in condition. Equation (1)
AIV-4
EMAP Research Strategy

-------
                                                                                         Appendix IV.
can be rewritten to incorporate these two
components of variance in the numerator:
                                            (2)
where r = # repeat visits/yr
Trend detection is a function of the magnitude of
the two variance components in the numerator.
From a design perspective, the effects of the index
variance term can be reduced by increasing the
number of measurements taken during the index
period (r) or by improving measurement techniques
(if measurement error accounts for a large part of
this variance component). This increases the
precision of estimating the yearly condition. Also,
the gain in precision declines with increasing r.
Not much can be done to control year-to-year
variance in the absence of identifiable covariates.
The only solution is to extend the monitoring
record by waiting for more years to pass.
Consequently, the limitations in our ability to
improve trend detection capability can be quickly
evaluated by using Equation (2) under different
variance scenarios  and lengths of record.
Incremental improvement becomes more and more
expensive as the required number of measurements
increases.

Two additional components of variance are
introduced when we consider multiple sites. One
component merely  expresses the true differences
among the sample of sites (a2lake - site-to-site
variation).  The second component is the common
year-to-year variance exhibited by all sites in the
sample (ci2year) (Figure 13A). The year-to-year
term in Equation (2) is also altered to reflect the
site*year interaction (the independent variation
sites exhibit) and the nonconcordant year-to-year
variation.  Equation (2) now becomes:
2
°/ote , ^2
, ' Vyear '
9
-y2 . index
year ^
I
 Vat(p) =
where / = # sites in the sample
              Equation (3) describes situations in which different
              sites are visited each year. The number of sites in
              the sample (/) improves the precision of
              estimating the yearly condition across sites.  A
              large variance component is introduced by the site-
              to-site differences. The effect of this variance term
              disappears if the monitoring design includes site
              revisits across years. If we assume revisits across
              years is the basic format, Equation (3) reduces to:
                                               index
                                            ~
                                               r
                                                       (4)
              The variance component models described here can
              be used effectively to evaluate alternative design
              protocols and estimate the potential payoff of
              various alternative sampling allocations. Equation
              (4) is especially useful, since all the important
              features related to trend detection capability are
              clearly expressed in terms of variance components
              (sample allocation both within and among years
              and length of the interval over which trend is to be
              detected). More elaborate models can be used for
              refined exploration of alternatives after initial
              scoping with Equation (4). This is especially true
              for power calculations and for exploration of the
              effects of temporal and spatial autocorrelation.

              Sensitivity of Trend Detection to Variance
              Components - Inspection of Equation (4) reveals
              the relative importance of year effects (a2year) and
              residual effects (ci2res - within the brackets) for
              describing the potential sensitivity of designs, such
              as EMAP's, for trend detection. The effect of
              a2year  is sensitive to the length of the monitoring
              interval (i.e., denominator of Equation 4), but not
              to the number of lakes visited or the number of
              revisits. Thus, it places a fundamental limit on
              trend detection capability that cannot be altered by
              the numbers of lakes visited or the number of
              revisits. Its effect can only be accommodated by
              extending the period of record. If its magnitude is
              large relative to a2res  expending additional effort
              minimizing the effect of  a2res  yields little.
(3)
                                        EMAP Research Strategy
                                                       AIV-5

-------
Appendix IV.
On the other hand, if a2res is large relative to year
effects, there are several options for increasing
trend detection sensitivity. One is to evaluate the
components of variance comprising a2lndex  to
determine the extent to which methodological
improvements will decrease its magnitude as this is
the only variance component subject to
methodological improvements. Methods evaluation
should always be a routine part of any QA
program, oriented toward cost-effective variance
reduction. Training, improved analytical
techniques, and refined sampling protocols all can
contribute to reduction of cj2lndex.

Another option is to evaluate the  allocation of lake
visits.  Both revisits to lakes during the index
period and visits to additional lakes can improve
trend detection capability.  Revisits to lakes (r)
reduces only a2lndex and its influence on trend
detection. Adding additional lakes reduces the
effect of both a2lndex  and a2lakeiyear. Therefore,
when resources are fixed, sampling additional lakes
rather than revisiting lakes is always an
improvement (or equal, if a2lakeiyear  =0). The
amount of benefit derived from adding lakes to the
sample is related to the relative magnitude of
CT2lake*year  -t- a2lndex, where the larger the ratio,
the greater the benefit of adding lakes. However, if
all lakes in a population or subpopulation can be
monitored in a year, then additional resources
should go toward revisits.

These concepts of a variance component
framework and the estimates of these components,
from the literature and our own sampling, have
allowed us to begin rigorously evaluating design
options for surveys. Urquhart et al. (1998) have
taken the lake variance estimates  for indicators like
Secchi transparency, chlorophyll  a, total
phosphorous and zooplankton species richness and
applied them to the common design options which
               have been used or proposed by various groups (see
               Figure 14).  Using the estimates of these lake
               variance components, we then evaluated the power
               of these design options to detect trends over time.
               All of the designs with an element of repeat visits
               (designs 1, 3, or 4) have similar power to detect a
               1-2% trend (Figure 15). Only design 2 which visits
               completely different systems every year stands out
               as a poor option. We also examined how they
               functioned with different levels of a2yeai (this has
               the largest impact on trend detection, Figure 16). It
               is clear from this analysis that decreasing a2yeai can
               have a dramatic effect on power for detecting
               trends.

               Given that one of the competing objectives of
               monitoring is also status estimation, we felt it was
               important to also compare designs for this.  We
               used the standard error (SE) of the estimated status
               as a comparison among designs (Figure 17). In this
               instance, a higher SE indicates a less precise
               estimation of status. Design  1, which revisits the
               same sites every year, has a significantly poorer
               ability to estimate status compared with the designs
               which visit a larger number of different sites. The
               other three designs converge  on one another over
               time. When balancing the need for status and trend
               estimation, it appears that either design option 3 or
               4 are preferable to design options 1 and 2.

               Our other application of this framework is to
               compare indicators for their ability to detect trends.
               Secchi transparency is quite good for detecting
               trends, but the zooplankton species richness, which
               many would consider too variable, also reaches a
               similar power after only a few additional years
               (Figure 18).  The variance component estimates
               were derived from EMAP and State data for
               northeastern lakes.  These types of analyses allow
               us to compare different indicators, and if necessary,
               choose among them.
AIV-6
EMAP Research Strategy

-------
                                                                                       Appendix IV.
Panel
Size
TIME PERIODS (= YEARS)
1
2
3
4
5
6
7
8
9
10
11
12

DESIGN 1 = SAME SITES (e.g., lakes) EACH YEAR
1
60
X
X
X
X
X
X
X
X
X
X
X
X


DESIGN 2 = NEW SITES EACH YEAR
1
2
3
4
5
6
7
8
9
10
11
12

60
60
60
60
60
60
60
60
60
60
60
60

X













X













X













X













X













X













X













X













X













X













X













X















PANEL
SIZE
TIME PERIODS (= YEARS)
1
2
3
4
5
6
7
8
9
10
11
12

                   DESIGN 3 = AUGMENTED SERIALLY ALTERNATING
1
2
3
4
Cornrno
n
50
50
50
50
10

X



X


X


X



X

X




X
X

X



X


X


X



X

X




X
X

X



X


X


X



X

X




X
X







               DESIGN 4 = PARTIALLY AUGMENTED SERIALLY ALTERNATING
  4


  1A

  2A
 1C    5   X
Figure 14. Design options considered for evaluation.  1. same sites visited
every year, 2.  different sites visited every year, 3. rotating  panel with 4
different sets of sites in each of 4 years and then revisiting them plus a constant
set visited each year, and 4. a more complex variation of design 3.
                           EMAP Research Strategy
AIV-7

-------
Appendix IV.
         11J
         UL
            0.8-
            0.6
         fi
         uj 0.4
         I
         °-0.2
   DESIGNS 1, 3, &
                                                             DESIGN 2
                                               10
                                           YEARS
                           15
20
        Figure 15. Power for detecting trends using the four design options.  It is clear that
        designs with repeat sampling have significantly greater ability to detect trends.
         UJ
         U.
             0.8-
             0.6
         u  0.4 H
         O
         a.
             0.2
                 0               5              10
                                           YEARS
        Figure 16. Power for trend detection with varying levels of a:
                           15
                                                              year*
20
AIV-8
EMAP Research Strategy

-------
                                                                              Appendix IV.
STANDARD ERROR OF
ESTIMATED STATUS
o o
» ro i».
^ 	 DESIGN 1
^^^^^^^^^DESIGN2
DESIGNS 3 & 4
U 1 1 1 1
0 5 10 15 20
Years
  Figure 17.  Standard error of estimated status compared among designs. Design 1 is significantly
  poorer than the other 3 options.
Figure 18.  The power to detect a 2% peryear trend in Secchi transparency and zooplankton species
 •_
 0)

 o
&H
                    0
                        5
                        Years
10
                                                                   Secchi
                                                                   Transparency
                                                                   Zooplankton
                                                                   Richness
richness with a sample size of 50 lakes per year. Data were generated from the 1991-1994 EMAP lakes
study in New England.
                                   EMAP Research Strategy
                                                                        AIV-9

-------

-------
APPENDIX V.
Index  Sites
                                                                                      Appendix V.
National Park Intensive Monitoring Network
(PRIMENet) - The National Park system provides
the potential coverage of all terrestrial ecosystem
types, and many of the goals and objectives of their
Inventory & Monitoring (I&M) Program are
similar to those of EMAP. The National Park
Service (through their Air Monitoring Division and
their Inventory and Monitoring Program) co-
developed a 14 site terrestrial intensive
monitoring/research network with EMAP.  Both
agencies contributed funds and efforts toward this
development.  Other federal agencies have been
invited to participate in the longer term. In 1996,
EPA and NFS created a formal interagency
agreement to create the Demonstration of Intensive
Sites Project (now PRIMENet - Park Research and
Intensive Monitoring of Ecosystems Network).
This inter-agency effort between EPA/ORD and
DOI/NPS selected 14 parks in a demonstration of
an intensive site network of monitoring and
research at locations across the United States (see
http://www.epa.gov/uvnet). All 14 parks are
readily accessible, have a history of environmental
monitoring, and represent a broad and sometimes
unique spectrum of ecological communities.
Through this network, EMAP and the Park Service
are using park sites for monitoring of global-scale
environmental stressors (e.g., air deposition) and
locale-specific stressors (e.g., water-borne
contaminants), and to coordinate with cause-effect
research related to these environmental stressors.
The intent was to initiate a consistent air
monitoring program at each site followed by
consistent monitoring within other media.


Effects research was based on known stressors at
the sites.  For example, at the Everglades site we
have: examined the flux of materials and nutrients
from Everglades canals into Florida Bay; examined
the role of humic materials in the complexation and
transport of mercury through the canals;
investigated the effects of increased nitrogen and
phosphorus from the canals on primary and
secondary productivity in Florida Bay; and
investigated the cause of black band disease in
corals in the Florida Keys National Marine
Sanctuary. At the Great Smoky Mountain site,
opportunities existed to  validate forest stand/ozone
models using new forest micro-meteorological and
dosimetry equipment, and to initiate mechanistic
studies of atmospheric nitrogen deposition effects
on watersheds,  expanding on similar studies at the
Sequoia, Acadia,  and Rocky Mountain sites.  EPA
also proposed extramural and/or cooperative
research (with other CENR-member agencies) to
examine the effects of increased UV-B exposure:
on the reproductive success of amphibians and
reptiles (Big Bend, Everglades, Sequoia); on coral
community structure (Virgin Islands); and on
plankton community structure and productivity
(Everglades, Virgin Islands).


With the interagency agreement in place, EPA and
NFS developed a management structure for the
PRIMENet network to determine the monitoring
and research needs at the sites and across the sites,
and management  of the information to come from
these sites. In late 1996, EPA and NFS initiated
the PRIMENet Oversight Committee with
membership from both agencies, including EPA's
Offices of Air and Water.


The Oversight Committee created 14 site
committees, one at each National Park, to establish
the monitoring  and research needs at each site.
Each of these committees is comprised of three to
five members representing the park, academic
researchers involved in park research, and an EPA
Regional representative, when possible.  Each
committee developed a prioritized list of
monitoring and research needs for their site to
characterize the long-term exposure of parklands to
the various environmental stressors. These
stressors do not include human use of the parklands
and its facilities as a public land.
                                      EMAP Research Strategy
                                          AV-1

-------
Appendix V.
The initial phases of PRIMENet focused on
augmenting monitoring of stressors via
atmospheric pathways.  However, the intent is to
include additional monitoring indicators and issue-
based effects research at the NFS sites. To do this,
the partners and cooperators have worked together
to establish research and monitoring plans for the
14 intensive sites. The cooperating agencies formed
the Joint Intensive Monitoring Committee and Joint
Effects Research Committee to guide these studies.
The Joint Intensive Monitoring Committee will
coordinate all new monitoring, such as atmospheric
nitrogen deposition or water quality monitoring. In
addition, the Joint Effects Research Committee will
develop and implement cause-effect research
programs at appropriate sites to provide additional
index sites.
             Specific implementation activities included:  NFS
             establishing UV-B monitoring at 14 sites; setting
             up support infrastructure in the parks,
             implementing quality assurance/quality control
             procedures, developing a data management system,
             and organizing reporting methods. EPA provided
             UV-B sensors for each site, and  statistical analysis
             and support for the network.  EPA provides for
             information management and the long-term storage
             of data. In addition, EPA supports both intramural
             and extramural research at several of the proposed
             sites that will increase the value  of long-term
             records at these sites (e.g., ecosystem process
             research on mercury and nutrient stress in the
             Everglades (South Florida), and  a variety of
             process research in the Big Bend, Great Smoky
             Mountains, Virgin Islands, and Sequoia sites.
AV-2
EMAP Research Strategy

-------
                                                                               Appendix VI.
APPENDIX VI.
Mid-Atlantic
Integrated
Assessment  (MAIA)
The first regional scale geographic study in EMAP
was conducted in the mid-Atlantic area (EMAP's
Mid-Atlantic Integrated Assessment). The mid-
Atlantic region of the eastern United States is
defined by the land and near-coastal area that
includes all of Standard Federal Region III and
parts of Regions II and IV (Figure 19).  States
completely covered are: Pennsylvania, Maryland,
Delaware, Virginia, and West Virginia.  Also
included are parts of New Jersey, New York, and
North Carolina.  The communities in the mid-
Atlantic are diverse in size, type, values, economic
and cultural influences.  They include the fishing
and crabbing communities of Delaware, eastern
Maryland, Virginia and North Carolina; the farm
communities of central Pennsylvania and western
Maryland; the coal-mining communities of West
Virginia and western Pennsylvania; and the major
metropolitan areas of Baltimore, Washington, D.C.,
Philadelphia, and Norfolk.
  Figure 19.  Geographic scope of MAIA.
The assessment activities within MAIA focused
first on single resource assessments and
subsequently have begun to focus on integrated
assessments across multiple resources. The
assessments were based on the measurement of
ecological condition and the major stressors
associated with impaired condition in the region
using a probabilistic monitoring design.  EMAP
stressor indicators cannot determine cause and
effect, but can provide weight-of-evidence for the
relative magnitude of the various stressors
impacting that resource.  (See EPA reports: An
Ecological Assessment of the United States Mid-
Atlantic Region, EPA/600/R-97/130: Condition of
the Mid-Atlantic Estuaries, EPA 600-R-98-147;
Condition of the Streams of the Mid-Atlantic
Highlands, EPA/903/R-00/015).


Estuaries - EMAP conducted surveys of the
estuarine resources within the Virginian Province,
which includes the area designated as MAIA
between 1990-1993. The results of that effort
(Strobel et al.  1995) have been combined with
those from other agencies and programs in our
Condition of the Mid-Atlantic Estuaries Report
(EPA600-R-98-147). This report  included:


       integrated EMAP surveys, and
       Chesapeake Bay and Delaware Bay
       Programs;


       common indicators and common designs,
       to offer better coverage and comparability
       for MAIA estuaries;


       multiple indicators of biological resource
       condition to improve the  assessment of the
       condition of the estuaries;


       the need for additional stressor indicators,
       especially for eutrophication.
                                   EMAP Research Strategy
                                      AVI-1

-------
Appendix VI.
The other primary partners in MAIA's estuarine
research were: EPA Region III, Chesapeake Bay
Program, Delaware Bay Program, National
Oceanic and Atmospheric Administration,
Delaware, Maryland, Virginia, and North Carolina.


The EMAP study of the Virginian Province
considered estuaries an extensive resource (Stevens
1994) and used an appropriate probability-based
sample survey design.  The survey considered three
strata based on the subpopulations of interest, large
estuaries (> 260 km2), small estuaries (2.6 - 260
km2) and large tidal rivers (> 260 km2) (Figure 20).
            MAIA have resulted from a variety of human
            activities and primarily impact the timing, amount
            and path of water flow. Physical habitat alterations
            have become a concern more recently and include
            changes in habitat complexity, substrate size and
            embeddedness, bank stability, and riparian
            vegetation. Biological alterations have often been
            overlooked by water quality agencies as a potential
            source of problems in streams. These biological
            alterations include introduction of plants and
            animals, exotic species, and management practices,
            such as stocking and harvesting.  MAIA provided
            an opportunity for EMAP to examine these issues
            on a broad regional scale.
                                                    The objectives for the wadeable streams portion of
                                                    MAIA were to:
                                                            Estimate the total miles of wadeable
                                                            streams;


                                                            Determine the condition of wadeable
                                                            streams;


                                                            Establish factors associated with poor
                                                            conditions in these streams;
Figure 20.  EMAP estuarine sampling locations
in MAIA, 1990-1993. Sampling in 1997-1998
was modified to improve coverage of smaller
estuaries and link with efforts on larger systems,
such as Chesapeake and Delaware Bays.
                    Demonstrate the EMAP approach for
                    regional and subregional scale monitoring
                    of wadeable streams;


                    Demonstrate the use of biocriteria,
                    ecoregions, watersheds, and probability
                    surveys to address environmental issues
                    within a region; and
In MAIA we worked with our partners to
collectively arrive at a design which met multiple
objectives.  However, the probability basis for the
design was fundamental. Sampling locations and
site selection criteria were used to allow as many
existing state sampling sites to be included and
aggregated for quantitative regional statements as
possible. Where significant gaps existed, we added
sampling to fill the gap. For example, it was
necessary to increase the coverage to adequately
describe the Virginia, Delaware and Del Marva
Coastal Bays.  While the coastal bays did not
constitute a large percentage of the estuarine area,
they are an important resource which had not been
adequately described previously.


Wadeable Streams - Hydrologic alterations in
                    Develop the baseline for future change
                    and trend comparisons for wadeable
                    streams.


            We used a probabilistic sampling survey to
            characterize the wadeable streams of the mid-
            Atlantic.  To ensure characterization of the stream
            reaches in the region it was necessary to classify
            streams on the basis of their size, using stream
            order as a size surrogate.  Stream order is relatively
            easy to calculate for all stream segments. Any
            metric used as a size surrogate must be calculated
            for all segments or it cannot be used.  This
            precluded other size surrogates (watershed area,
            discharge, stream width) which may have been
            better indicators of stream size, because of the
            difficulty  in defining them for the entire population.
AVI-2
EMAP Research Strategy

-------
                                                                                         Appendix VI.
The survey sample for wadeable stream systems for
1997 is shown in Figure 21.


In MAIA we also linked our EMAP sample survey
data with data from the more deterministic,
temporally intensive effort of the USGS National
Water Quality Assessment Program (NAWQA,
Hirschetal. 1988). Through this joint effort we
are evaluating the links between our regional data
and the data from USGS index sites.  Conceptual
and empirical models are being developed to
describe links between spatial and temporal
bio/physical/ chemical data. From these models we
should be  able to assess the "representativeness" of
the "index" monitoring sites within NAWQA.  The
models should also allow us to spatially interpret
the loading models developed at NAWQA study
sites.  The linkage between EMAP sample survey
information and NAWQA temporal information,
will allow EMAP to include seasonal variability in
its assessments of condition and will provide
NAWQA with greater spatial capability for more
regional level assessments.  Our studies began in
three basins of the Mid-Atlantic Highlands, the
Potomac, Susquehanna, and James river basins
(Figure 22), and are reported in Mid-Atlantic
Highlands Streams Assessment (US EPA 2000).


The information generated in MAIA has helped
identify areas of concern and suggest high priority
stressors.  In addition, the study provided baseline
information for comparing ecological condition of
wadeable streams throughout the mid-Atlantic
region, (e.g., Mid-Atlantic Highland Streams
Assessment, EPA/903/R-00/015). With continued
monitoring through time, the States (i.e., Delaware,
Maryland, New York, North Carolina,
Pennsylvania, Virginia, and West Virginia) and
federal agencies (e.g., USEPA, USGS) will have
critical information on change, and then trends, in
the condition of these Mid-Atlantic streams.
Quantitative estimates of change (and trends)
would provide a measure of the efficacy of state
and federal environmental programs and policies in
reducing current environmental problems.


Landscapes - In MAIA we used measurements
derived from satellite imagery to develop a land
cover map for the Mid-Atlantic (Figure 22).  This,
combined with spatial data bases on biophysical
features (e.g., soils, elevation, human population
patterns), were used to produce an ecological
assessment of the impact of changing landscape
conditions. Using fine-scale  spatial resolution
(e.g., 30-90 meters) to census MAIA, we were able
to analyze and interpret environmental conditions
of the 125 watersheds in the mid-Atlantic region
based on 33 landscape indicators.  The results of
this study were published in the report An
Ecological Assessment of the United States Mid-
Atlantic Region (EPA/600/R-97/130).  This report
included:
        spatial patterns of agriculture and urban
        lands;


        proportions of forests, forest connectivity,
        and forests near streams (riparian zones);


        Condition estimates for watersheds; and


        watershed conditions around major
        metropolitan areas.


We are currently conducting multiple watershed
studies to determine quantitative relationships
between landscape metrics (e.g., riparian forest,
interior forest, and agricultural land cover) and in-
stream variables (e.g., stream total nitrogen
concentration, benthic invertebrate community
condition). From this, we will be able to interpret
the hydrological and ecological meaning of
landscape metrics relative to aquatic resource
condition.
EMAP Integration in MAIA - EMAP is
continuing intramural work in MAIA to develop
better approaches to integrating the various
estuarine, stream and landscape condition estimates
into an integrated assessment of the overall
environmental condition.  This  work is being done
in conjunction with all of ORD's Laboratories' and
Centers' ecology programs to develop better
techniques for assessing: overall environmental
condition, environmental risks and their trade-offs,
and ultimately, restoration and protection foci.
                                       EMAP Research Strategy
                                           AVI-3

-------
Appendix VI.
                     1997 MAIA Stream/River Sample
                 Figure 21. The 1997 stream/river probability sample sites. An
                 additional set of ~200 sites was sampled in 1999.
                   ["""""1 Hfcjh intensity urban HHH| Conferk-us forest  [ """"'] nay^pasj,_re  [   | Emergent wetland
                   I    | Low Intensity utoan |   | Deciduous forest       Row crops   jI Woooy yretland
                   ^_ Water          |~  j M&red forest     |   | Barren


                   Figure 22.  Mid-Atlantic land cover map. The Highlands
                   region (in the oval) served as the initial study area for our
                   joint USGS-EPA streams monitoring.
AVI-4
EMAP Research Strategy

-------
                                                                              Appendix VII.
 Appendix VII.  Regional
 Environmental
 Monitoring  and
 Assessment  Program
 (R-EMAP)
The Regional Environmental Monitoring and
Assessment Program (R-EMAP) is a partnership
between the EPA Regional Offices and ORD's
EMAP for improving monitoring as a tool for
regional decision makers and resource managers.
The goals for R-EMAP are to transfer EMAP's latest
scientific techniques for ecological monitoring to
EPA Regions, States, Tribes, and local decision-
makers.  EMAP works with the EPA Regional
Offices to identify and support projects meeting our
criteria, and of importance to the Regions. R-EMAP
provides numerous smaller scale ecological
monitoring opportunities across wide the
biogeographic and political boundaries of the EPA
Regions, and complements our larger geographic
assessments. R-EMAP projects also provide initial
opportunities to develop and test indicators across the
nation.
EMAP support for these projects includes:
contributing to development of the scientific design
of the projects; selection and evaluation of
appropriate indicators and methods for measurement;
application of information management approaches;
analysis and interpretation of data; and providing a
source of funding. It is expected that the
organizations involved with EPA's Regions and
EMAP could include States, Tribes, local
governments, and academic institutions. Because
much of the research will use approaches and
techniques that have not been traditionally employed
by monitoring programs, ORD's experience and
expertise in these new techniques will often be
required to  fully develop the proposals. This is
consistent with R-EMAP's goal of transferring these
new techniques into Regional, State, Tribal, and local
decision-making.


EPA Regions submit pre-proposals to the ORD
R-EMAP Work Group for possible funding. The
R-EMAP Work Group selects pre-proposals for
which full proposals will be requested. Each of the
selected proposals is adopted by one of ORD's
Ecological Divisions. These Divisions assume
responsibility for helping in the development of a
high quality full proposal.  Each R-EMAP full
proposal undergoes a rigorous external peer review,
prior to any funding decisions. The R-EMAP
projects undertaken and proposed can be viewed on
the R-EMAP website:
http://nraxp.nar.epa.gov/emap/html/remap/).


Example of Recent R-EMAP Results - In 1994 the
State of Nebraska (Region VII), started using a 5-yr
rotating basin design for its ambient surface water
monitoring program (a basin level monitoring that is
rotated to different basins in subsequent years, such
that all major river basins in the state are monitored
after a period of 5 years). This rotating basin
framework worked from an organizational standpoint,
but the state realized it was negatively biasing its
sampling by targeting its monitoring to waters with
suspected problems. With targeted monitoring, it was
impossible for the state to make an unbiased estimate
of the status and trends in condition of its streams.
As a result the state's assessment of waters meeting
designated use ("305(b) Report") tended to show a
disproportionately large percentage of the state's
monitored waters with water quality problems.


In 1997, the state decided to try an EMAP approach
within its rotating basin assessment through R-
EMAP. Using an EMAP sampling design, random
stream segments were selected within the basins of
interest (Figure 23).  The random selection of stream
segments  permitted unbiased estimates of conditon
of the stream resources with known confidence
intervals (Figure 24). Using water quality
information from this R-EMAP study with data from
a previous R-EMAP project (Region VII, 1994-
                                    EMAP Research Strategy
                                      AVE-1

-------
 Appendix VII.
1995), the state was able to detect a significant trend
towards improved water clarity in the river basins
sampled (Figure 25). We will continue to work with
the Regions in the development new R-EMAP
projects. Pre-proposals for projects for funding in
2000 include:
        Region 1.  Condition of New England's
        Wadeable Streams.
2.       Region 2.  NY/NJ Harbor Study, Cohansey-
        Maurice Watershed Assessment, and an
        Environmental Assessment of Barnegat Bay,
        NJ.
                                                    6.
                    Region 4.  Everglades Ecosystem
                    Assessment and Condition of Southeastern
                    Wadeable Streams.


                    Region 5.  An Ecological Assessment of
                    Invasive and Aggressive Plant Species in
                    Coastal Wetlands of the Laurentian Great
                    Lakes.


                    Region 7.  A Probabilistic Survey of Iowa
                    Stream Resources and Probability-based
                    Monitoring Design within Missouri's
                    Statewide Resource Assessment and
                    Monitoring Program.
3.       Region 3. Watershed-based Design frame
        for estimating Biotic Integrity of West
        Virginia Streams.
                    Region 8, 9, & 10.  Development of
                    Reference Conditions and Use of
                    Intensification Sites in the EMAP Western
                    Pilot Design.

                  I1"* -v^l. f— **""\ \x:r'TV'''~'     '""  : * ----- i "" -/^"T-^_™. I  /.-;' a T! 'rf~\ f'~^/>. x?
                   VSCAVV -"             f' *                  '

        Figure 23.  Distribution of probability-based stream sample sites on the Big/Little Blue,
        Loup, Niobrara, Republican and White-Hat River Basins in Nebraska in 1997-1998.
 AVE-2
EMAP Research Strategy

-------
                                                                                 Appendix VII.
                              1997-98 Nebraska Stream Data
                                 Aquatic Life Use Support
                                     (percent of streams)
                          n Full Support  n Partial Support   • Non-support
Figure 24. Percentages of Nebraska streams in the Big Blue, Loup, Niobrara, Republican
and White river basins supporting aquatic life use. These percentagesare based on a
draft IBI. All estimates are at the 90% confidence level and are +/-10%.
              
-------

-------
                                                                               Appendix VIM.
APPENDIX VIM.
Information
Management
Compiling, maintaining, and distributing the data
from EMAP has presented unique issues. The
information superhighway has stimulated new
information management possibilities as well as
general appreciation for the need to coordinate data
and network standards.  Early EMAP work
supported an information management program
that anticipated the expanding use of the Internet.
The program developed much of the software for
linking EMAP to the Internet. The Information
Management Workgroup was established to
maintain our data infrastructure and place a priority
on making the early EMAP field data available in
electronic format for analysis. The information
management priorities for EMAP (see Hale et al.
1999 for details) are: to make all the data
electronically available; to provide linkages or
flags to useful external data sets (as identified by
the EMAP researchers); to inventory environmental
monitoring data in the various management
programs in our study regions and elsewhere (an
extensive, interactive inventory has been developed
for MAIA (Jackson and Gant 1998,
WWW.epa.gov/monitor); and provide data
interpretation templates for end users.


EMAP Data Directory - The EMAP Data
Directory is the primary means for users to find out
what EMAP data are available and where the data
files can be found.  It contains all EMAP and
MAIA data set information.  The EMAP Data
Directory is based on the NASA Directory
Interchange Format (DIP), and is now FGDC
(Federal Geographic Data Committee) compliant.
Not all groups who will be making Directory
entries have access to Oracle. Thus, we will use
templates, including Web tools,  to help in acquiring
this information easily.


EMAP Data Catalog - EMAP policy states that
Data Catalog files must accompany all data files.
Therefore, no data files will be moved to the public
access EMAP home page without the metadata
files. This requirement allows us to provide
information about the data files so that the data can
be correctly interpreted and used. Also, we are
then interoperable with other federal agency data
catalogs (FGDC standards). The Data Catalog is
maintained with a WordPerfect template and
converted to HTML for the EMAP home page.


EMAP World Wide Web  Site - Primary
objectives of the EMAP home page
                       are to provide
information about the EMAP program, and to allow
access to the Directory of EMAP data sets, the
metadata files, and the EMAP data sets. Links are
provided from the EMAP home page to other
appropriate home pages (e.g., Global Change
Master Directory, STORET, NAWQA, and the
Chesapeake Bay Program).


EMAP home page contents include:


       Data Directory (HTML)


       Data and metadata files (ASCII, SAS,
       Oracle)


       Publications (WordPerfect, RTF, PDF)


       List of contacts (HTML)


       EMAP Bibliography (HTML)


•      EMAP Information: EMAP Research
                                   EMAP Research Strategy
                                      A VIII-1

-------
Appendix VIM.
        Strategy, EMAP Information Management
        Plan, EMAP Updates (WordPerfect, RTF,
        PDF, ASCII summaries)


•       Links to other environmental monitoring
        research sites (HTML)


        Geographic Reference Database for GIS
        coverages or links to the OEI (EPA's
        Office of Environmental Information) GIS
        library and the EROS Data Center
•

A future focus of EMAP's WWW site will be to
develop and provide the necessary algorithms for
key data analysis. This will allow agencies and the
public to routinely use EMAP data to answer
questions.  As the public and their management
agencies become increasingly sophisticated this
will become a greater component of EMAP
Information Management.


EMAP Data Links - Two other EPA information
management systems are directly linked to EMAP.
These are the Office of Water's STORET and
ORD's Environmental Information Management
System (EIMS).
             STORET - EPA Regions and many state agencies
             operate local copies of the STORET Oracle
             database that they use to house regional and local
             data. States will soon be able to choose to load
             their EMAP data into STORET. As STORET has
             been undergoing a recent revision, EMAP and
             STORET information managers have been working
             closely to enhance STORET's capabilities to store
             EMAP indicator and probabilistic survey data.
             This should be completed soon, and will allow
             EMAP data to be stored and accessed directly
             through EPA's STORET system.
             EIMS - In ORD the EIMS stores information about
             ORD (and other) data (e.g., what data exist, where
             they can be found, etc.). The EMAP Data
             Directory is periodically uploaded to the EIMS.
             EIMS makes the existence of EMAP data sets more
             widely known.  EIMS is also available  to use as a
             directory for the myriad of non-EMAP  data used in
             our geographic assessments. Any EPA Region,
             state, or ORD laboratory can make EIMS directory
             entries for these external (non-EMAP)  data sets and
             have them labeled as part of the larger study. The
             linkage between EMAP and EIMS is ongoing.
AVEI-2
EMAP Research Strategy

-------
                                                                                    References
REFERENCES
Armitage, P.D.  1978. Downstream changes in the composition, numbers, and biomass of bottom fauna in the
Tees below Cow Green Reservoir and an unregulated tributary, Maize Beck, in the first five years after
impoundment. Hydrobiologia 58:145-156.


Cochran, W.G.  1977. Sampling Techniques. 3rd edition. John Wiley & Sons.  New York. 428pp.


Committee on the Environment and Natural Resources.  1996. Integrating the Nation's Environmental
Monitoring and Research Networks and Programs: A Proposed Framework. White House National Science
and Technology Council. Washington, D.C.


Cummins, K.W. and M.J. Klug.  1979. Feeding ecology of stream invertebrates.  Ann. Rev. Ecol. Syst.  10:147-
172.


Draper, N. R. and H. Smith, Jr., 1981. Applied Regression Analysis (2nd Edition). John Wiley & Sons, New
York, New York, 709 pp.


Engle, V.D. and J.K. Summers. 1999. Refinement, validation, and application of a benthic condition index for
the Gulf of Mexico estuaries.  Estuaries 22:624-635.


Engle, V.D., J.K. Summers, and G.R. Gaston.  1994. A benthic index of environmental condition of the Gulf of
Mexico estuaries. Estuaries 17:372-384.


GAO (U.S. General Accounting Office), 1981. Better Monitoring Techniques Are Needed to Assess the
Quality of Rivers and Streams.  Volume 1. U.S. General Accounting Office, Washington, D.C.


GAO (U.S. General Accounting Office), 1986. The Nation's Water: Key Unanswered Questions About the
Quality of Rivers and Streams.  U.S. General Accounting Office, Washington, D.C.


GAO (U.S. General Accounting Office), 2000. Water Quality: Identification and Remediation of Polluted
Waters Impeded by Data Gaps. U.S. General Accounting Office, Washington, D.C.


Gardener, R.H., B.T. Milne, M.G. Turner, and R. V. O'Neill.  1987.  Neutral models for the analysis of broad-
scale landscape pattern. Landscape Ecology  1: 19-28.


Gardner, R.H. and R.V. O'Neill. 1991. Pattern, process and predictability: The use of neutral models for
landscape analysis  In: Turner, M.G. and R.H. Gardner (eds).  Quantitative Methods in Landscape Ecology:
The Analysis and Interpretation of Landscape Heterogeneity. Ecological Studies Series. Springer-Verlag. New
York.  pp. 289-307.


Hale, S.S., M.H. Hughs, J.F. Paul, R.S. Mcaskill, S.A. Rego, D.R. Bender, N.J. Dodge,  T.L. Richter, and J.L.
Copeland.  1998. Managing scientific data: The EMAP approach. Environmental Monitoring and Assessment


                                    EMAP Research Strategy                                 R-l

-------
References
51:429-440.


Hale, S., J. Rosen, D. Scott, J. Paul, andM. Hughs. 1999. EMAP Information Management Plan: 1998-2001.
EPA/600/R-99/001a.  U.S Environmental Protection Agency. Washington, D.C.


Hale, S.S. and H.W. Huffman.  2000. Designing environmental monitoring databases for statistical analyses.
Environmental Monitoring and Assessment 64:55-68.


Hargrove, W.W. and J. Pickering. 1992. Pseudoreplication: a sine qua non for regional ecology.  Landscape
Ecology 6: 251-258.


Hart, C.W., Jr., and S.L.H. Fuller. 1974. Pollution Ecology of Freshwater Invertebrates. Academic Press,
New York.


Hedtke, S., A. Pilli, D. Dolan, G. McRae, B. Goodno, R. Kreis, G. Warren, D. Swackhamer, and M. Henry.
1992. Environmental Monitoring and Assessment Program, Great Lakes Research Plan - Peer Review Draft.
Duluth, Minnesota: U.S. Environmental Protection Agency.


Hilsenhoff, W.L.  1977. Use of arthropods to evaluate water quality of streams.  Technical Bulletin 100.
Wisconsin Department of Natural Resources, Madison, WI.


Hirsch, R.M., Alley, W.M. and Wilber, W.G. 1988. Concepts for a National Water-Quality Assessment
Program. U.S. Geological Survey Circular 1021. U.S. Geological Survey. Denver, Colorado.


Holland, A.F., ed.  1990. Near Coastal Program Plan for 1990: Estuaries. EPA 600/4-90/033. U.S.
Environmental Protection Agency. Washington, D.C.


Hurlbert, S.H. 1984. Pseudoreplication and the  design of ecological field experiments. Ecological Monographs
54: 187-211.


Jackson, L.E. and M.P. Gant. An interactive, spatial inventory of environmental data in the Mid-Atlantic
region. In: Sandhu, S. et al. (eds). Monitor'ing Ecological Condition at Regional Scales. Proceedings of the
Third Environmental Monitoring and Assessment Program (EMAP) Symposium.  Kluwer Academic Publishers.
Boston, pp. 325-329.


Jackson, L., J. Kurtz, and W. Fisher, eds.  1999.  Evaluation Guidelines for Ecological Indicators.  EPA/620/R-
99/005. U.S. Environmental Protection Agency. Washington, D.C.


Karr, J.R., K.D. Fausch, P.L. Angermeier, P.R. Yant and L.J. Schlosser.  1986. Assessment of Biological
Integrity in Running Water:  A Method and its Rationale.  Special Publication 5.  Illinois Natural History
Survey. Champaign, IL.


Lange-Bertalot, H. 1979. Pollution tolerance as a criterion for water quality estimataion. Nova Hedwigia
64:285-304.


Larsen, D. P., and S. J. Christie, eds. 1993. EMAP-Surface Waters 1991 Pilot Report. EPA/620/R-93/003.
R-2                                   EMAP Research Strategy

-------
                                                                                        References
U.S. Environmental Protection Agency. Washington, D.C.


Larsen, D.P., N.S. Urquhart, and D.L. Kugler. 1995. Regional scale trend monitoring of indicators of trophic
condition of lakes.  Waters Resources Bulletin 31: 117-140.


Macauley, J.M., J.K. Summers, and V.D. Engle.  1999. Estimating the ecological condition of the estuaries of
the Gulf of Mexico. Environmental Monitoring and Assessment.  57(l):59-83.


Magnuson, J. I, B. J. Benson, and T. K. Kratz, 1990. Temporal coherence in the limnology of a suite of lakes in
Wisconsin, U.S.A. FreshwaterBiol. 23:145-159.


Matthews, W.J., and D.C. Heins. 1987. Community and Evolutionary Ecology of North American Stream
Fishes. University of Oklahoma Press, Norman.


McAllister, D.C., S.P. Platania, F.W. Schueler, M.E. Baldwin, and D.S. Lee. 1986. Ichthyofaunal patterns on a
geographic grid.  In: Hocutt, C.H. and E.O. Wiley, eds. The Zoogeography of North American Freshwater
Fishes. Wiley & Sons, New York.  P 17-5.


Messer, J.J., R.A. Linthurst, and W.S. Overton.  1991. An EPA program for monitoring ecological status and
trends. Environmental Monitoring  Assessment 17:67-78.


Metcalfe, J.L. 1989. Biological water quality assessment of running waters based on macroinvertebrate
communities: History and present status in Europe. Environ. Pollut. 60:101-139.


Miller, R. J., and H.W. Rolbison.  1980. The Fishes of Oklahoma. Oklahoma State University Press, Stillwater.


Minckley, W.L. 1973. Fishes of Arizona. Sims Printing, Phoenix.


Moyle, P.B.  1976. Inland Fishes of California. University of California Press, Berkeley.


National Research Council. 1977.  Environmental monitoring. Volume IV. National Academy of Sciences.
Washington, DC. 150 p.


National Research Council. 1999.  Ecological Indicators for the Nation. National Academy of Sciences.
Washington, DC. 153 p.


Omernik, J.M. 1987. Ecoregions of the conterminous United States (map supplement). Annals of the
Association of American Geographers  77:118-125.


Omernik, J.M. 1995. Ecoregions -  a framework for environmental management. In:  Davis, W.S., and Simon,
T.P.  eds.  Biological Assessment and Criteria - Tools for Water Resource Planning and Decision Making.
Lewis Publishers, Boca Raton, Florida,  p. 49-62.


O'Neill, R.V., R.H. Gardner and M.G. Turner. 1992. A hierarchical neutral model for landscape analysis.
Landscape Ecology  7:55-61.
                                       EMAP Research Strategy                                   R-3

-------
References
Patrick, R.  1968. The structure of diatom communities in similar ecological conditions.  Am. Nat.  102:173-
183.


Paul, J.F., J.H. Gentile, KJ. Scott, S.C. Schimmel, D.E. Campbell and R.W. Latimer. 1999. EMAP-Virginian
Province Four-Year Assessment (1990-93). EPA/620/R-99/004. U.S. Environmental Protection Agency,
Washington, D.C.


Paul, J.F., KJ. Scott, D.E. Campbell, J.H. Gentile, C.S. Strobel, R. Valente, S.B. Weisberg, A.F. Holland, and
J. A. Ranasinghe. (Submitted). Developing and applying a benthic index of estuarine condition for the Virginian
Biogeographic Province. Ecological Indicators.


Plafkin, J.L., M.T. Barbour, K.D. Proter, S.K. Gross, andR.M. Hughes. 1989. Rapid Bioassessment Protocols
for Use in Streams and Rivers: Benthic Macroinvertebrates and Fish. EPA 444/4-89/001. U.S. Environmental
Protection Agency, Washington, D.C. 172 pp.


Plaits, W.S., W.F. Megahan, and G.W. Minshall. 1983. Methods for Evaluating Stream, Riparian, and Biotic
Conditions.  General Technical Report INT-138.  U.S. Forest Service, Ogden, Utah.


Rakocinski, C.F., S.S. Brown, G.R. Gaston, R.W. Heard, W.W. Walker, and J.K. Summers.  1997.
Macrobenthic responses to natural and contaminant-related gradients in northern Gulf of Mexico estuaries.
Ecological Applications 7:1278-1298.


Rankin, E.T., and C.O. Yoder. 1991.  The Nature of Sampling Variability in the Index of Biotic Integrity (IBI)
in Ohio Streams. Ohio Environmental Protection Agency, Columbus, OH.


Resh, V.H., and J.D. Unzicker. 1975. Water quality monitoring and aquatic organisms: The importance of
species identification. J.Water Pollut. Control. Fed. 47:9-19.


Reynolds, L., S. Gregory, P. Kaufmann, A. Herlihy, and R. Hughes, (in prep).  Spatial sampling requirements
for electrofishing Willamette Valley and Cascade Mountain streams in Oregon, USA. Trans. Am. Fish. Soc.


SAS Institute Inc., 1989. SAS/STAT® User's Guide. Version 6, Fourth Edition, Volume 2, SAS Institute Inc.,
Gary, North Carolina, 846 pp.


Snedecor,G.W. And W.G. Cochran. 1967. Statistical Methods (Sixth Edition). Iowa State University Press,
Ames, Iowa.


Steel, R. G. D. and J. H. Torrie.  1980. Principles and Procedures of Statistic: A Biometrical Approach
(Second Edition). McGraw- Hill Book Company, New York, New York.


Stevens, D. L., Jr. 1994. Implementation of a national monitoring program. Journal Environ. Management
42:1-29.


Stevenson, R. J., and R.L. Lowe.  1986.  Sampling and interpretation of algal patterns for water quality
assessments. In: Isom, E.G. ed. Rationale for Sampling and Interpretation of Ecological Data in the
Assessment of Freshwater Ecosystems.  ASTM STP 894:118-149. American Society for Testing and Materials,
Philadelphia.


R-4                                   EMAP Research Strategy

-------
                                                                                        References
Stevenson, R.J., C.G. Peterson, D.B. Kirschtel, C.C. King, and N.C. Tuchman.  1991. Density-dependent
growth, ecological strategies, and effects of nutrients and shading on benthic diatom succession in streams. J.
Phycol. 27:59-69.


Strobel, C.I, D.J. Keith, H.W. Buffum and E.A. Petrocelli. 1995. Statistical Summary:  EMAP-Estuaries
Virginian Province -1990-1993.  EPA/620/R-94/026. U.S. Environmental Protection Agency.  Washington,
D.C.


Urquhart, N.S., W.S. Overton, and D.S. Birkes.  1993. Comparing sampling designs for monitoring ecological
status and trends: Impact of temporal patterns. In: Barnett, V., and K.F. Turkman, ed.  Statistics for the
Environment, John Wiley & Sons, Ltd., Sussex, England,  p.71-85.


Urquhart, N.S., S.G. Paulsen, and D.P. Larsen. 1998. Monitoring for policy-relevant regional trends over time.
Ecological Applications 8:246-257.


U.S. Environmental Protection Agency. 1987. Surface water monitoring: A framework for change. Office of
Water and Office of Policy, Planning, and Evaluation. U.S Environmental Protection Agency.  Washington,
D.C.


U.S. Environmental Protection Agency. 1988. Future Risk: Research Strategies for the 1990s.  Science
Advisory Board. SAB-EC-88-040. U.S Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1996. Strategic Plan for the Office of Research and Development.
EPA/600/R-96/059.  U.S Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1997. EPA Strategic Plan.  EPA/190-/R-97-002. U.S Environmental
Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1997. An Ecological Assessment of the United States Mid-Atlantic
Region. EPA/600/R-97/130. U.S Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1998. Biological Criteria: Technical Guidance for Streams and Small
Rivers. EPA-882-B-98-003. Office of Water.  U.S Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1998. Ecological Research Strategy.  EPA/600/R-98/086. U.S
Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 1998. Condition of the Mid-Atlantic Estuaries.  EPA 600-R-98-147.
U.S Environmental Protection Agency.  Washington, D.C.


U.S. Environmental Protection Agency. 1999. The Ecological Condition of Estuaries in the Gulf of Mexico.
EPA/620-R-98-004.  U.S Environmental Protection Agency. Washington, D.C.


U.S. Environmental Protection Agency. 2000. Mid-Atlantic Highlands Streams Assessment. EPA/903/R-
00/015. U.S Environmental Protection Agency. Washington, D.C.
                                      EMAP Research Strategy                                   R-5

-------
References
Van Dolah, R. F.,  J. L. Hyland, A. F. Holland, J. S. Rosen and T. R. Snoots.  1999.  A benthic index of
biological integrity for assessing habitat quality in estuaries of the southeastern USA. Marine Environmental
Research 48:269-283


Watanabe, T., K. Asai, and A. Houki.  1988. Numerical water quality monitoring of organic pollution using
diatom assemblages. In:  Proceedings of 9th Diatom Symposium, September 1986.  Frank Round Press, Bristol,
England. P 123-141.


White, D., A. Jon Kimmerling, and W. Scott Overton.  1991.  Cartographic and geometric components of a
global sampling design for environmental monitoring.  Cartography and Geographic Information Systems
19:5-22.


Wiken, E. 1986 Terrestrial Ecozones of Canada:  Ottawa. Environment Canada, Ecological Land
Classification Series no. 19. 26 pp.
R-6                                   EMAP Research Strategy

-------

-------
&EPA
      United States
      Environmental Protection
      Agency
Please make all necessary changes on the below label,
detach or copy, and return to the address in the upper
left-hand corner.

If you do not wish to receive these reports CHECK HERE Q ;
detach, or copy this cover, and return to the address in the
upper left-hand corner.
PRESORTED STANDARD
 POSTAGE &     PAID
         EPA
   PERMIT No. G-35
      Office of Research and Development
      National Health and Environmental
        Effects Research Laboratory
      Research Triangle Park, NC 27711

      Official Business
      Penalty for Private Use
      $300

      EPA 620/R-02/002
      July 2002
      www.epa. gov/e map/

-------