United States
Environmental Protection
Agency
Office of Research and
Development
Washington, D.C. 20460
EPA/600/R-00/001
January 2000
A National Assessment
of Landscape Change
and Impacts to Aquatic
Resources
A 10-year Research Strategy
for the Landscape Sciences
Program
-------
EPA/600/R-00/001
January 2000
A 10-year Strategic Plan for
the Landscape Sciences Program
by
K. Bruce Jones1
Llewellyn R. Williams1
Ann M. Pitchford1
E. Terrence Slonecker1
James D. Wickham1
Robert V. O'Neill2
Donald Garofalo1
William G. Kepner1
1 United States Environmental Protection Agency
Office of Research and Development
National Exposure Research Laboratory
Environmental Sciences Division
2 Oak Ridge National Laboratory
Oak Ridge, Tennessee
Interagency Agreement No. DW89938037
-------
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development (ORD), partially funded and collaborated in the research described here. It has
been peer reviewed by the EPA and approved for publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation by the EPA for use.
11
-------
We are grateful to the Megan H. Mehaffey, Deborah J. Chaloud, Donald W. Ebert,
Maliha S. Nash, Anne C. Neale, Curtis M. Edmonds, Timothy G. Wade, Daniel T. Heggem,
David F. Bradford, Nita G. Tallent-Halsell, George T. Flatman, Michael P. Bristow, Robert D.
Schonbrod, Clayton E. Lake, Phillip A. Arberg, David B. Jennings, David J. Williams, Mary
Lacerte, S. Taylor Jarnagin, Rick W. Kutz, Elizabeth R. Smith, and Ross Lunetta of our
Landscape Sciences Program, Joe Walker of CSIRO-Australia, and Dennis Yankee and Roger
Tankersley of the Tennessee Valley Authority for helping us establish the vision for a national
landscape assessment. We thank Gil Veith, Steve Paulsen, Michael McDonald, Tom Barnwell,
Jay Messer, Tom DeMoss, and Joel Scheraga of the EPA for their support of landscape research
in EPA. We thank the EMAP Surface Waters, Near Coastal, and Information Management
groups for providing numerous ideas and comments on landscape research over the years. We
are grateful to Frank Golley (University of Georgia), Monica Turner (University of Wisconsin at
Madison), Ingrid Burke (Colorado State University), Ed Martinko (Kansas University), John
Jensen (University of South Carolina), Jingle Wu (Arizona State University), and John Lyon and
Susan Norton (EPA) for review of this plan. We thank our colleagues June Thourmosguard,
James Vogelmann, and Tom Loveland of the USGS EROS Data Center for making regional-
scale, ecological assessments possible. Finally, we thank Rick Linthurst (EPA) for his
unwavering support of the Landscape Sciences program, for none of what is proposed in this plan
would have been possible without Rick's vision, leadership, and encouragement.
in
-------
Notice ii
Acknowledgements iii
List of Tables vi
List of Figures vii
Executive Summary 1
Introduction 1
Approach 2
Priority Research 2
Section I. Purpose, Context, and Organization of this Plan 4
Section II. Introduction 6
A. Background 6
B. Landscape Approach 7
C. Goals and Objectives 8
D. Scope of the Program 9
E. Statement of Capabilities 9
Section III. High Priority Research 11
A. Spatial Data Acquisition, Assembly, and Accuracy Assessment 11
Spatial Data Acquisition and Assembly 11
North American Landscape Characterization (NALC) 12
Multi-Resolution Land Characteristics Consortium (MRLC) 12
Other Satellite Sensors 13
Aerial Photographs 13
National Technical Means 14
Other Spatial Data 14
Accuracy Assessment 15
B. New Remote Sensing Methods Development 16
Digital Photogrammetry 16
High-Spatial-Resolution Satellite Remote Sensing 17
Microwave Remote Sensing 17
Thermal Infrared Remote Sensing 17
High Spectral Resolution Remote Sensing 17
C. Change Detection 18
D. Quantification of Landscape Indicators Relative to Condition of Aquatic Resources 20
E. Assessment Methods Research and Development 23
IV
-------
Section IV. Strategic Approach for Implementing Research and the National Assessment 25
A. Implementation Themes 25
Stakeholder Needs 25
Highlighting Assumptions, Critical Research Steps, and Interdependencies 26
Setting Priorities 26
Developing and Testing Approaches and Methods 27
Assessing the Method Performance for Different Geographic Areas 27
Considering ORD's High Priority Geographic Areas in the Selection of Study Areas 28
Managing Information for Use by Collaborating Scientists, Clients, and the Public 28
Collaborating within the EPA and with Academia, Federal Agencies, and Others 29
Providing Customer Services and Products 29
B. Research Plans 30
C. Time Line for Research and Assessment Implementation 30
Section V. Measures of Success 32
Section VI. References Cited 33
Appendix I. List of Research and Assessment Activities by Geographic Areas by Fiscal Year (FY)40
Mid-Atlantic Region 40
Northeastern Region 41
Southeastern Region 42
Mid-West Region 42
Upper Mid-West Region 43
Pacific Northwest Region 44
Pacific Southwest Region 44
Northern Great Plains Region 45
Inter-Mountain Region 45
Southwestern Region 46
Rocky Mountain Region 47
South Central Region 47
National Assessment 48
Appendix II. Relationships to EPA Goals and Objectives 49
Appendix III. Anticipated Outputs of the Research 51
Appendix IV. Acronyms 52
-------
•. -'.'. : es
Table 1. Research Topics Areas and Goals
Table 2. MRLC Land Cover Classes
Table 3. New Remote Sensing Systems
Table 4. Landscape Sciences Program Responsibility Matrix
Table 5. Existing and Potential Collaborators Within ORD
Table 6. Ongoing Collaboration Organized by Topic
Table 7. List of Clients and Customers
....54
....55
....56
....58
....59
....60
....61
VI
-------
.•:'"
Figure 1. General conceptual model of landscape change and sustainability of
environmental attributes values by society 62
Figure 2. Comparison of multispectral, hyperspectral, and ultraspectral signatures 63
Figure 3. Unique reference spectra of minerals as derived from NASA's AVIRIS
instrument 64
Figure 4. General implementation strategy for landscape research and assessments
proposed in the Plan 65
vn
-------
Introduction
This 10-year strategic plan (the Plan) describes the rationale and approach for research
activities proposed by the Landscape Ecology and Landscape Characterization Branches
(Landscape Sciences Program) of the National Exposure Research Laboratory (NERL). The 10-
year goal of the Landscape Sciences Program is to conduct a national assessment of landscape
change between the early 1970s and the early 2000s at relatively fine spatial scales (60-meter
resolution) and to evaluate the consequences of observed change on aquatic resources, including
streams and estuaries. An emphasis on aquatic resource endpoints is based on legislative
mandates and the important role that the EPA plays in monitoring and protecting these resources.
However, the research strategy recognizes the importance of the terrestrial characteristics and
processes in determining the condition of aquatic resources; therefore, many of the indicators
being developed by the Landscape Sciences Program are terrestrial-based.
Five priority research and development areas have been identified to achieve this 10-year
goal: (1) acquisition and assembly of spatial databases on the environment that permit generation
of landscape indicators (as defined by O'Neill et al. 1988, 1997; Jones et al. 1996; 1997) for four
dates (early 1970s, mid-1980s, early 1990s, and early 2000s), (2) development of new remote
sensing methods to improve the quality of spatial databases and to measure landscape indicators
of stress (e.g., grazing, mining, urbanization, other land uses), (3) development of methods that
permit an analysis of landscape change (Change Detection) based on different satellite sensors,
including Landsat Multi-Spectral Scanner (Landsat MSS) and Landsat 4, 5, and 7 Thematic
Mapper (Landsat TM) imagery, (4) development and selection of landscape indicators based on
the degree to which they explain variability in aquatic resource conditions, and (5) development
of landscape assessment approaches and watershed models that permit an analysis of multiple
landscape indicators. The projects will yield information needed to understand how the activities
of humans disturb the quality of our Nation's waters and the habitats they provide. The Plan
describes the "landscapes approach" as we will apply it to better understand both the causes and
consequences of landscape change.
The Plan is organized in six sections and leads the reader through the sequence of events
from the identification of the research needs, the development of appropriate approaches, the
gathering and analysis of the critical data to close information gaps, to the validated indicators,
models, and databases that are the major research products of the Landscape Sciences Program.
The objectives of the program are to: (1) develop new remote sensing data collection and
processing techniques that measure the extent and magnitude of watershed-level stressors,
including grazing, mining, and different agricultural practices, that will improve future landscape
change estimates, (2) quantify relationships between measures of landscape attributes (landscape
indicators) and parameters representing aquatic resource conditions, and determine of how these
relationships vary within and among regions of the U.S., (3) compile a national comprehensive
landscape-change database representing dates from the early 1970s, the mid 1980s, the early
1990s, and the early 2000s, and make these data available to the public on the Internet, (4)
develop methods to analyze changes in landscape indicators between the early 1970s and the
early 2000s and evaluate the consequences of observed change on our Nation's aquatic resources,
(5)
-------
demonstrate how the landscape sciences can contribute to the assessment of the condition of our
Nation's resources, and (6) provide tools and guidance to the Federal, State, and other resource
managers and the public such that they may confidently apply landscape science techniques to
ecological assessments in various regions of the country. The relationships of these outcomes to
higher order goals and objectives of the EPA and Congress, and to specific environmental laws,
are described in Appendix II.
Approach
Successful implementation of this Plan requires close coordination among three
organizational units in the NERL: the Landscape Ecology Branch, in Las Vegas, Nevada, the
Landscape Characterization Branch in Research Triangle Park, North Carolina, and the
Environmental Photographic Interpretation Center, in Reston, Virginia. It also requires
coordination with the Agency's Environmental Monitoring and Assessment (EMAP), Regional
Vulnerability Assessment (ReVA), Global Change, and other EPA Programs. The Plan is
designed to facilitate that coordination, to provide guidance to the staff involved, and to help
Agency managers and the Congress understand our long-term research agenda and the resources
needed to accomplish it.
The "landscape approach" (Jones et al. 1996, 1997, O'Neill et al. 1997) relies on the
measurement of the spatial pattern of ecological characteristics relative to the condition of aquatic
resources and associated processes. In this approach, metrics of spatial pattern are used as
surrogates or indicators (landscape indicators) of aquatic resource conditions. Moreover, changes
in landscape indicator values can be used to evaluate the consequences of landscape change on
aquatic resources. However, we currently lack sufficient knowledge to interpret many of the
landscape metrics with regards to aquatic resource conditions and, therefore, have emphasized
this kind of research in the plan.
The Plan includes examples of how the approach can be used to address problems related
to broad geographic areas (e.g., vulnerability of streams to nutrient inputs). Because certain data,
such as those derived from satellites, can be composited over a large area, landscape analysis can
be performed at different spatial levels such as states or parks, watersheds or ecoregions.
Improved resolution of more recent data allows landscape indicators to identify both the nature
and the location of problems. And the ability to use data that have already been gathered, often
for very different purposes, can make the approach highly cost-effective. To implement such an
approach nationally, however, requires the development of landscape indicators that capture
important aspects of landscape change and an ability to relate the indicators to aquatic resource
conditions.
-------
Priority Research
We have focused our proposed research program on evaluating those aspects of
landscape change that pose the greatest ecological risk to aquatic resources, including (1)
hydrologic alteration, (2) habitat conversion, (3) turbidity/sedimentation, (4) habitat
fragmentation, (5) pesticides, and (6) nutrients (EPA 1999). The resulting Program is consistent
with an EPA NERL strategic goal, ". . . to facilitate environmental management and decision
making by providing the scientific tools needed to estimate (and ultimately reduce) risks posed by
exposures . . . [and] for protecting ecosystems at the complex levels at which ecological processes
actually operate."
Implementing the Program involves a variety of research activities, focused in five broad
and overarching areas summarized in Table 1. In addition to these five broad areas, the Program
will also address three more focused research areas of high priority to the EPA: (1) methods to
assess stream vulnerability to pesticides and toxic substances, (2) methods to target areas that are
most like to have high values for Total Maximum Daily Loads (TMDLs), and (3) methods to
evaluate status and trends of riparian ecosystems. A more complete listing of planned research
activities ~ with time lines and critical milestones over the 10-year life of the Plan — is found in
Appendix I.
Common to all the research and assessment activities will be a process to (1) understand
stakeholder needs, (2) develop and test methods and approaches, (3) assess method performance
for different geographic areas (including Agency priority study areas), (4) collaborate with other
EPA and external organizations to integrate results across the major research areas, (5) highlight
assumptions, critical steps, and interdependencies, (5) provide user-friendly products (indicators,
models, guidelines, and assessments); (6) make data and information generated from the program
readily available to participating scientists, clients, and the public, and (7) identify future research
needs and opportunities. The Plan presents an extensive list of current and planned research
collaborations that leverage limited resources to address this aggressive research agenda.
The success of this Program will be evaluated in terms of our ability to satisfy our various
clients with useful, timely products and knowledge that enhance their ability to measure, monitor,
and assess those changes in landscape level conditions that impact negatively upon the quality of
our Nation's aquatic resources. Our clients include the Congress (annual measures under the
Government Performance Results Act), other Agency programmatic and regional offices, other
Federal and State staff, the external scientific community and the general public. Specific
products are described in the Plan or are identified in Appendix I.
-------
Ml < , i > ' !i S
This Plan presents the rationale and a blueprint for those research activities to be
conducted over the next 10 years by the Landscape Sciences Program of the U.S. EPA's National
Exposure Research Laboratory (NERL) needed to achieve a national assessment of landscape
change between the early 1970s and the early 2000s and to assess the consequences of those
changes on aquatic resources. The Plan describes a strategic approach for implementing key
research activities that will ultimately lead to the national assessment. It describes the linkage
between our 10-year goal and the goals and objectives described in the Strategic Plan for the
Office of Research and Development (ORD, EPA 1997) and the ORD's Ecological Research
Strategy (EPA 1998a, see Appendix II), both of which embrace a focus on an ecological risk
assessment paradigm and a shift away from "at the pipe" monitoring to more integrated, multiple
scale environmental monitoring and assessment (EPA 1988, NRC 1997). It also reflects recent
recommendations from a peer review of the Landscape Sciences Program (Golley et al. 1997).
The Plan presents the goals, objectives, and priorities of the Program and projects the outputs and
outcomes to be realized through successful completion of the research objectives outlined herein.
The Plan describes how multi-disciplinary research projects in five critical research areas
will address critical gaps in the science and provide the tools (indicators, databases and models)
and the process knowledge required to assess changes in landscape characteristics and properties
that adversely affect the quality of our Nation's waters and the ecological habitats they provide.
The Plan will define and explain the "landscapes approach" and how it can be applied to
determine the status and trends in both ecological resources and the key components that alter
those resources. The information that will be generated from research activities in this Plan will
reduce future uncertainty in forecasting changes in our Nation's resources that result from the
activities of humans.
The proposed research activities are placed in the context of the broader ecological risk
assessment paradigm (Graham et al. 1991, Hunsaker et al. 1990, EPA 1998b) under which the
ORD has organized its major ecological research goals and objectives (EPA 1998a). It is our
belief that the science of landscape ecology and related disciplines are integral to the assessment
of the vulnerability and sustainability of ecosystem processes and functions. The combination of
landscape ecology with advanced technology such as remote sensing, geographic information
systems, and computer science is widely recognized as an effective approach to assess the
potential impacts of complex natural and anthropogenic forces on the structure and function of
ecological resources at various temporal and geographical scales.
The Plan provides a rationale for the research areas selected based upon the 10-year
project goal, national goals developed under the Government Performance Results Act (GPRA
1993), Agency priorities, and peer input from the scientific community. It is designed so that the
reader can follow the logical sequence from recognition and prioritization of the research needs
and information gaps to development of hypothesis-driven research approaches to the acquisition,
processing, interpretation and reporting of the critical data required to close that gap or fill that
need. Products developed in the process will be made available to both Agency users and the
broader scientific community as tools for reducing the uncertainty of our future estimates and
projections of the conditions of our ecological resources.
-------
This plan is designed to serve multiple purposes. It will function as the principal
guidance document for the staff of the Landscape Ecology and Landscape Characterization
Branches in planning their specific research activities. It will help managers of the National
Exposure Research Laboratory and of the ORD set resource priorities and identify opportunities
for in-house collaboration. It will also help top Agency administrators and the Congress
understand the long-term ecological research agenda and the resources required to accomplish
that agenda. Finally, it will be made publicly available via the Internet to describe the nature,
rationale, and direction of Agency research and to invite collaborative research efforts toward a
common purpose. By providing information on how we plan and prioritize our research and by
providing specific goals, objectives, products, and timetables, we also are providing yardsticks by
which the progress of our research can be assessed by all of our stakeholders.
The plan will not include detailed descriptions of individual research components, such
as sets of hypotheses to be tested, detailed study designs, and resources allocated to individual
projects. Rather, those details will be included in individual research plans for specific projects
(see Appendix I for a list of projects).
The Landscape Sciences Strategic Plan includes an executive summary, six sections, and
three appendices:
• Section I introduces the Plan, its purpose, objectives, context, and organization;
• Section II introduces the need for regional- and national-scale landscape
assessments, the overall goal and objectives of the Program, the Program's relationship
to broader EPA goals and objectives, and the scope of the program;
Section III identifies and describes the five priority research areas within the
Program;
Section IV provides detail on implementation strategies, program assumptions,
project priorities, partnering with potential collaborators, a description of anticipated
products, and a discussion of how the research activities relate to other current and
anticipated initiatives;
Section V describes the measures of success;
Section VI provides a list of references cited in the Plan.
-------
]]
A. Background
There is a growing interest among Federal agencies, States, and the public to evaluate
environmental conditions at community, watershed, regional, and national scales. Numerous
consortia have been formed to address broader-scale environmental problems. These include the
Great Lakes Program, the Chesapeake Bay Program, the Gulf of Mexico Program, and the Grand
Canyon Trust, to name a few. In some cases, interagency teams have been formed to evaluate
status and trends in environmental conditions, e.g., the Pacific Northwest Initiative, the Interior
Columbia Basin Assessment, and the Mid-Atlantic Integrated Assessment. Finally, the
Committee on the Environment and Natural Resources (CENR), chaired by the Vice President of
the United States, has initiated a national assessment of the environment.
Each of these multi-agency organizations is faced with the daunting task of acquiring and
analyzing data that are of sufficient spatial and temporal similarity to permit regional and national
scale assessments. Moreover, because most existing data have been collected by different
agencies for different reasons, they are generally incompatible. Some agencies have established
their own monitoring programs specifically to address regional and national scale issues. For
example, the Environmental Monitoring and Assessment Program (EMAP) was designed to
assess the status and trends of ecological resources at regional and national scales (Messer et al.
1991). But these programs are often limited to a specific set of questions and ecological
resources (e.g., timber resources, impacts of agriculture, water quality of streams), and, therefore,
they are not generally comparable.
The relatively high cost of collecting environmental data has limited the implementation
of regional- and national-scale monitoring programs. However, a new set of land cover databases
being developed by the Multi-resolution Land Characteristics Consortium (MRLC), a historical
landscape database developed by the North American Landscape Characterization (NALC)
Program, and development of landscape pattern indicators from spatial data (O'Neill et al. 1997),
offers an unprecedented opportunity to assess ecological resource conditions at community,
watershed, regional, and national scales over the next five to 10 years. In 1996, a regional-scale
land cover database was developed for the five-state area of the United States Mid-Atlantic
Region, and this database, along with other regional landscape coverages (e.g., topography, soils,
road networks, stream networks, and human population density), was used to assess landscape
conditions across the entire region down to a scale of 30 meters (Jones et al. 1997). The
assessment used a set of landscape indicators (O'Neill et al. 1988, 1997) to evaluate the spatial
patterns of human-induced stresses and the spatial arrangement of forest, forest-edge, and riparian
habitats as they influence forest habitat suitability and aquatic resources. Advances in computer
technology and geographic information systems (GIS) have made it possible to calculate
landscape metrics over large areas (e.g., regions) at relatively fine scales.
Costs of spatial data derived from satellites and other sources are relatively low when
compared to similar spatial coverages acquired from finer-scale monitoring studies. Additionally,
spatial data lend themselves to analysis in a Geographic Information System or GIS (Ball 1994).
Because certain data, such as those derived from satellites, can be combined (in a mosaic) for a
-------
large area, landscape analysis can be performed on different spatial units, including natural units
such as watersheds and ecoregions and political units such as parks, counties, and states. This
makes it possible to report landscape statistics at a number of scales for many different types of
units and to determine cross-scale relationships between landscape composition and pattern,
fundamental ecological processes, and ecological goods and services. Finally, because they can
be generated from data that cover an entire region at scales as fine as 30 meters (as opposed to a
sample that covers only a small portion of the total area in a region), landscape indicators can
provide information both on the magnitude of problems and where the problems exist. As a
result, this information can be used to develop multi-scale plans to reduce vulnerability or risk
and prioritize activities to restore ecological function, condition, and sustainability.
Despite these new data and the development of landscape pattern indicators, the current
state-of-science precludes a national assessment of landscape change and the consequences of
observed change on aquatic resources. Significant scientific gaps exist in:
• our ability to detect and evaluate change in important stressors at watershed
scales, including mining, urbanization, and agricultural land uses,
• our ability to detect landscape change between the early 1970s and the early
2000s using different types of remote sensing imagery (e.g., Landsat Multi-spectral
Scanner (MSS) and Thematic Mapper (TM)),
• our knowledge of quantitative relationships between landscape metrics and
aquatic resource condition parameters within and among regions, and
• our ability to assess consequences of landscape change based on multi-indicator
data sets.
To eliminate the gaps in implementing a national landscape assessment, we propose
research in each of these areas (see Section III).
B. Landscape Approach
The landscape monitoring and assessment approach involves the analysis of the spatial
patterns of ecological characteristics such as soils, topography, climate, vegetation, land use, and
drainage pathways as they relate to processes affecting ecological and hydrological conditions on
areas ranging in size from small watersheds (a few hundred hectares) to entire basins (several
million hectares) (Forman 1990, O'Neill et al. 1994, Kepner et al. 1995, Jones et al. 1997.).
Landscape metrics can be used as indicators (landscape indicators) of the condition of ecological
resources because changes in landscape pattern result in changes to ecological processes that
sustain such resources (Turner 1989, Forman 1990, Figure 1). For example, losses of forested
riparian habitats in a watershed reduce that watershed's ability to filter sediment from overland
surface flows (Karr and Schlosser 1978, Lowrance et al. 1985, Peterjohn and Correll 1984). This
results in reduced water quality in the streams of that watershed. Similarly, clear-cutting of
forests on areas with steep slopes results in soil loss which in turn can decrease the likelihood of
sustaining harvestable forests on those areas and can increase sediment loadings to streams (Lee
1989, Franklin 1992).
-------
Five characteristics distinguish a landscape monitoring and assessment approach from most
traditional field or site-based monitoring programs:
(1) it involves analysis of spatial patterns of ecological characteristics (e.g., forests
near streams) to interpret ecological conditions,
(2) it applies the concept from the field of landscape ecology that changes in
landscape patterns result in changes in fundamental ecological processes, including
fluxes in energy, biota, materials and nutrients, and water,
(3) it applies the concept of ecological hierarchy theory that changes in landscape
pattern and ecological processes at relatively broad scales (e.g., river basins) constrain
and influence the condition of imbedded ecological resources (e.g., stream segments),
(4) it uses digital maps or coverages of biophysical (e.g., soils, vegetation,
topography, and climate) and human use (e.g., land use, population distribution)
characteristics to analyze and interpret landscape patterns relative to ecological
condition, and
(5) it includes humans as part of the environment.
C. Goals and Objectives
The Landscapes Sciences Program seeks to establish landscape ecology and related
disciplines as integral to the assessment of the condition, vulnerability and sustainability of
ecosystem processes and functions. The 10-year goal of the Program is to assess landscape
changes between the early 1970s and the early 2000s and to evaluate the consequences of
observed change on aquatic resources nationally. The assessment of landscape change
consequences is focused on aquatic resources because the EPA has a primary responsibility in
assuring their protection and restoration. However, the landscape approach evaluates many
aspects of the terrestrial environment because these attributes are intricately linked to ecological
and hydrological processes that influence aquatic resource conditions, as predicted from
ecological hierarchy theory (O'Neill et al. 1986). The objectives of the Program are to:
Develop new remote sensing data collection and processing techniques that
measure the extent and magnitude of watershed-level stressors, including grazing,
mining, and different agricultural practices, and that will improve future landscape
change estimates,
Quantify relationships between measures of landscape attributes (landscape
indicators) and parameters representing aquatic resource conditions, and determine how
these relationships vary within and among regions of the U.S.,
• Compile a national comprehensive landscape-change database representing dates
from the early 1970s, the mid 1980s, the early 1990s, and the early 2000s, and make it
available to the public on the Internet,
Develop methods to analyze changes in landscape indicators between the early
1970s and the early 2000s and evaluate the consequences of observed change on our
Nation's aquatic resources,
Demonstrate how the landscape sciences can contribute to the assessment of the
condition of our Nation's resources, and
-------
• Provide the tools and guidance to the Federal, State, other resource managers,
and the public so that they may confidently apply landscape science techniques
to ecological assessments in various regions of the country.
These goals are related to higher goals established by Congress, the EPA, and ORD.
Appendix II relates various aspects of this Strategic Plan to higher-order agency and government
objectives.
D. Scope of the Program
The 10-year goal of the Program is national in scope and ambitious. It seeks to provide
an unprecedented assessment of landscape change across the US at scales relevant to
environmental decision makers and the public. It builds upon ideas and concepts highlighted in
the Program's 1994 National Research Plan (O'Neill et al. 1994) and the Mid-Atlantic Landscape
Indicator Development Plan (Kepner et al. 1995).
At this time, we do not anticipate having the resources to assess landscape change across
the State of Alaska between the early 1970s and the early 2000s. However, we will explore the
possibility of assessing landscape change from the early 1980s to the early 2000s using data from
the Advanced Very High Resolution Radiometer (AVHRR), which has a 1-kilometer resolution.
The Plan is the culmination of extensive efforts to define those areas of scientific
uncertainty upon which the resources of this Program can have the greatest impact. The Program
sought (and continues to seek) peer input from international leaders in ecological risk assessment
and risk characterization in developing its proposed research agenda. The resulting Program is
mainstream with respect to one of three long-term strategic goals of the National Exposure
Research Laboratory, i.e., ". . . to facilitate environmental management and decision making by
providing the scientific tools needed to estimate (and ultimately reduce) risks to ecosystems
posed by exposures. . . . [and] for protecting ecosystems at the complex levels at which ecological
processes actually operate."
Findings of a formal expert panel peer review of the Landscapes Sciences Program
(Golley et al. 1997) helped to sharpen the research focus and to identify areas for disinvestment
and areas for intensification. Recognition of key research thrust areas as Congressional priorities
reinforced our confidence that the research program goals and products proposed are broadly
supported, politically as well as scientifically. One such product, A Landscape Atlas of the Mid-
Atlantic Region, released in 1997, received wide acclaim and provided additional support for the
approaches proposed in this Plan.
Based on the overall goal of the Program (see previous discussion) and the current state-
of-the-science (see later discussions), we identified five broad and three more specific areas of
research as the foci of the Landscape Program. These are discussed in greater detail in Section
III.
-------
E. Statement of Capabilities
The Landscape Sciences Program consists of approximately 40 full-time scientists and
technical experts that are assigned to two research branches in the Environmental Sciences
Division: the Landscape Characterization Branch which is located at Research Triangle Park,
North Carolina, and the Landscape Ecology Branch located at Las Vegas, Nevada, and Reston,
Virginia. The Reston, Virginia office, also know as EPIC, is co-located with the USGS in its
National Headquarters. Co-location with the USGS provides access to a number of USGS
scientists working on national-scale assessments. The two branches have a wide range of
scientific expertise including: (1) aquatic and terrestrial biology, (2) biochemistry, (3)
biogeography, (4) quantitative ecology, (5) landscape ecology, (6) hydrology, (7) remote sensing
applications, (8) remote sensing engineering, (9) geographic information systems, (10)
atmospheric sciences, (11) multi-variate statistics, and (12) spatial statistics. The staff have
produced numerous publications and have organized national and international symposia.
Scientists are supported by state-of-the-art computing, including multiple workstation,
UNIX and NT networks, and a wide range of remote sensing, GIS, and modeling software. Basic
remote sensing and GIS support also are provided through a contract. Four scientists have a Q
security clearance, which permits them to access National Technical Means (NTM) data.
The Landscape Sciences Program collaborates with scientists from a number of other
Federal agencies, including the USGS Earth Resources Observation System (EROS) Data Center,
the DOE Oak Ridge National Laboratory, the USGS Biological Resources Division (two
locations), the US Forest Service, the USDA Agricultural Research Service, the Tennessee Valley
Authority, and the USGS Water Resources Division. Additionally, the Landscape Sciences
program collaborates with scientists from other ORD Laboratories.
10
-------
We will focus our research and development in five general areas that currently prevent
implementation of a national landscape change assessment (e.g., gaps in the science). These
include (1) spatial data acquisition, assembly, and accuracy assessment, (2) development of new
remote sensing methods, (3) landscape change detection, (4) quantification of the degree to which
existing and new landscape indicators explain variation in aquatic resource conditions, and (5)
development of multi-indicator assessment techniques. Progress in all of these research areas will
be needed to implement and complete the national assessment. Additionally, we will evaluate
landscape indicators and watershed models for their applicability to three other types of
assessments of high priority to the EPA ~ (1) estimates of pesticide and other toxic chemical risks
to aquatic resources, (2) identification of aquatic resources at risk to high levels of nutrient and
sediment loading (TMDLs), and (3) determination of status and trends of riparian ecosystems.
A. Spatial Data Acquisition, Assembly, and Accuracy Assessment
Spatial Data Acquisition and Assembly
At the core of a landscape approach to monitoring ecological status and trends is the
basic need for quantifiable and consistent spatial data of biophysical characteristics (e.g., land use
and land cover) on national, regional and local scales. These data permit the calculation of
landscape indicators which we then intend to compare to aquatic resource conditions to build and
execute spatially-distributed models over regional scales (see Sections D and E). Multiple date
landscape data permit calculation of landscape indicators at different time intervals, and changes
in these values can then be interpreted with regards to potential changes in aquatic resource
conditions (see later discussion).
Land use and land cover data are most often derived from some type of overhead,
remotely sensed imagery such as aerial photographs and digital, satellite, remote-sensing data.
Various classifications of land use and land cover are derived from imagery based on manual
interpretations or a variety of digital processing techniques, depending on the application.
Remote sensing data and resources also provide valuable environmental monitoring services such
as change detection, topographic analysis, various types of mapping, indicator development and
analytical support to environmental regulatory programs.
This program will use existing and emerging remote sensing technology and resources as
the primary source of data for landscape analyses. Therefore, the goal of this part of the program
is to assemble a set of primary landscape data that permits an evaluation of landscape status and
change nationally. The principal sources of remote sensing and spatial data needed for landscape
assessments are listed below:
11
-------
North American Landscape Characterization (NALC)
The NALC project is a cooperative effort between the National Aeronautics and Space
Administration (NASA), the EPA, and the U.S. Geological Survey (USGS) to make historical
satellite-imagery data available to the widest possible user community for scientific research and
general public interest. The objectives of the NALC project are to (1) develop standardized
remotely sensed data sets and standard analysis methods in support of investigations of changes
in land cover, (2) to develop inventories of terrestrial carbon stocks, (3) to assess carbon cycling
dynamics, and (4) to map terrestrial sources of greenhouse gas (CO, CO2, CH4, and N2O)
emissions. NALC is a component of the NASA Landsat Pathfinder Program used to study global
change. The NALC database consists of Landsat MSS scenes from three decades (1970s, 80s,
and 90s), referred to as a "Landsat Triplicate" and digital elevation model data. The goal is to
produce these data sets for 536 Landsat World Reference System-2 path/row locations (scenes).
Each scene is approximately 185 x 185 km (Lillesand and Kieffer 1994). Satellite data used in
the NALC project include (whenever possible) Landsat MSS data from the years of 1973, 1986,
and 1991, plus or minus one year. The Landsat MSS data is the principal source of satellite data
used in the NALC program for several reasons, including the 20-year Landsat MSS digital data
archive and the relatively low data costs as compared to other sources of satellite imagery.
The NALC represents an invaluable source of historical information on landscape
condition and change because it provides low-cost, consistently processed, temporal Landsat data
for major portions of North America. However, as much as 15% of the scenes from the early
1970s are of insufficient quality to produce reliable landscape indicator estimates. Therefore, a
primary task is to replace poor quality scenes with higher quality scenes from the USGS's Earth
Resources Observation System (EROS) Data Center archive.
A significant challenge to this project is the generation of high quality, land cover data
from which landscape indicators can be generated for the early 1970s, the mid-1980s, and the
early 1990s. NALC data acquired from EROS Data Center are spatially registered, and most of
the triplicates are terrain corrected but are not classified into land cover. Additionally, each
triplicate set (total of 536) comes in a separate file and, therefore, must be electronically joined or
stitched together into a mosaic in order to analyze conditions across the region. A major task of
this project will be to take the NALC data and produce digital land cover maps for the three
periods. We will use historical aerial photography and NTM imagery (see discussion in Section
B) to calibrate and assess the accuracy of the derived land cover databases. We will test the
effectiveness and accuracy of our image classification through a series of small studies in
different regions of the country. This approach is necessary because variations in climate,
topography, and vegetation among the regions present different types of classification problems.
For example, clouds and atmospheric haze in the eastern U.S. can make classification difficult,
whereas low overall greenness, low soil moisture, and shadows (from high topographic relief) can
make differentiation between desert land cover types challenging.
Multi-Resolution Land Characteristics Consortium (MRLC)
As a result of the increasing recognition of the need for current land cover data, several
federal agencies formed the MRLC in 1995. The MRLC is a cooperative effort for the purpose of
sharing the costs, data, and effort associated with the development of national land cover
databases. The participating agencies include the EPA, USGS, the National Oceanic and
Atmospheric Administration (NOAA), and the U.S. Department of Agriculture Forest Service
(USFS). The goal of the consortium is to produce seamless, conterminous national land cover
data using a consistent classification, and, to the extent possible, quantifiable accuracy
assessments of that data.
12
-------
The MRLC vision is to facilitate the development of a national, multi-resolution, land cover
database from Landsat TM satellite imagery and field data.
MRLC is producing two data sets that will be major sources of data for our program.
First, the MRLC has effected a common acquisition and preprocessing of Landsat TM data for
the entire conterminous United States. This pre-processed data, all in the 1991-1994 time frame,
is available to all participating agencies through the USGS EROS Data Center. Second, MRLC is
now in the process of creating a consistent, national land cover data set from the 1991-94 Landsat
TM imagery. Completion of the full national data set is scheduled for 1999.
The MRLC classification consists of 23 land cover classes, nationally (Table 2). This
level of classifications should permit an analysis of the impact of general types of land use,
including agriculture and urbanization, on aquatic resources within watersheds. However, other
sources of imagery will be needed to evaluate the impact of specific agricultural and urban land
uses on aquatic resources (see below).
The MRLC program is currently planning to repeat the national Landsat TM data
purchase and land cover database development in the year 2000. Although still in the planning
stage, this second collection of pre-processed satellite imagery and derived land cover data would
represent a consistent and unprecedented national data set for multi-temporal landscape analysis
for the early 1990s and the early 2000s.
Other Satellite Sensors
Data from other satellite systems will be acquired and analyzed on an individual project
basis for landscape analyses of specific watersheds (e.g., Sections B and D) or for site-specific or
technical-support projects. These data, along with aerial photography (see below) also will be
used to identify the geographic distribution of more detailed features of the landscape that present
potential risks to aquatic resources, including grazing and timber harvest, agricultural land uses,
and urban land uses (e.g., golf courses), that the MRLC classification cannot identify. Numerous
remote-sensing systems are currently available or planned for deployment in the near future that
will provide additional spectral, spatial and temporal resolution data sets for landscape analysis.
Aerial Photographs
Archives of historical aerial photographs represent one of the most comprehensive
sources of landscape information available. High-resolution aerial photographs from as early as
the 1930's are routinely available for most parts of the United States and can be used for a variety
of analytical purposes including change detection and landscape analysis. Such photos, because
of their analog format and limited footprint of coverage, are best suited to smaller, site-specific
types of analyses. They do, however, provide an excellent complement to broad-area satellite
coverage and can be used as a source of data for calibration and quality control for satellite land
cover data.
Large archives of aerial photographs are available from a variety of Federal, State, local and
commercial sources including the U.S. Geological Survey, the U.S. Department of Agriculture,
U.S. Forest Service, and the National Archives and Records Administration. Additionally, the
Environmental Photographic Interpretation Center, a field station of the Environmental Sciences
Division, has over twenty years of experience in aerial photographic applications and maintains a
library of historical aerial photographic holdings.
13
-------
The primary use of historical aerial photography will be to calibrate and perform accuracy
assessments of digital land cover maps generated from NALC imagery (see below).
National Technical Means
The civilian agency use of data gathered by the U.S. intelligence infrastructure, termed
"National Technical Means" (NTM), has long been recognized as a vital and proper use of
nationally funded resources. Especially in terms of remote sensing technology (satellite and
aircraft imaging), NTM data are inexpensive and of high technical quality compared to
unclassified sources. The active use of classified information sources could provide the EPA
with unique archives of environmental data and state-of-the-art scientific capabilities.
NTM collection systems have proven their utility in a wide variety of environmental
applications. Projects completed to date include large-area site discovery and inventory, disaster
monitoring, coordination and assessment of environmental emergencies, natural-resource
inventories, pesticide-application activities, noise abatement and control as well as basic land use
and land cover mapping. DOD remote sensing technology provides the civil Federal community
with a variety of sensors that have great potential for fulfilling Agency needs for spatial
information. Although these sensors are classified and cannot be discussed outside a classified
arena, the derived cartographic output products derived from them are not; they have been used
successfully on hundreds of occasions in many of the major research and development programs
within the Agency.
The primary use of the NTM data will be to calibrate and conduct accuracy assessments
of the digital land cover maps derived from NALC imagery (see below).
Other Spatial Data
In addition to land cover data, other spatial data will be necessary to calculate landscape
indicators. These data will also be useful in improving the accuracy of NALC land cover maps
and in evaluating the spatial pattern of error. We will acquire national spatial-data coverages as
follows:
digital elevation models (30-meter resolution) from the U.S. Geological Survey
• national stream networks (River Reach File 3) from the U.S. EPA
soils (STATSGO and SSURGO) from the USDA Natural Resources
Conservation Service
human population or census data from the U.S. Department of Commerce
road distribution and density data from the U.S. Geological Survey
• geology from the U.S. Geological Survey
In some cases, it may be necessary to construct regional mosaics of these data as the data are
usually provided in geographic pieces.
14
-------
Accuracy Assessment
Quantifiable information on the thematic and spatial accuracy of land use and land cover
data derived from remotely sensed sources is required if these data are to be used to develop
larger landscape metrics and relationships. The development of viable accuracy assessment data
requires the comparison of the remote-sensing-derived land cover data with reference-test
information, often referred to as 'ground truth' data (Jensen 1986). The reference-test data is
usually developed with some of type of statistical sampling design and contains an inherent level
of accuracy that is known and significantly higher than the remote-sensing-derived land cover
data. Ideally, ground truth data is developed by actual field visits and in-situ sampling of the
landscape and its attributes. However, in practice this is rarely accomplished because: (1) the
field work is too time consuming and expensive, (2) field visits and sampling require prior
landowner approvals, and (3) significant amounts of time often elapse between the date of
imagery acquisition and the field work, creating temporal change inaccuracies in the reference
data set.
Because of these problems, a priority research area will be the development and implementation
of methods to document the accuracy of classified land cover and land characteristics databases,
such as MRLC and NALC (see previous discussion). This will require the development of
techniques that can quantify the accuracy and precision of derivative products such as land cover
surface maps. Part of the research will focus on image classification techniques; it will involve
the application of high-spatial-resolution stereo imagery that will be commercially available
beginning in 1998 and historical imagery from aerial photographic and NTM archives. Other
portions of this activity involve enhancement of the statistical rigor for determining uncertainty
and the development of sampling designs and sampling techniques for determining accuracy of
derivative classification products.
Accuracy assessments are included in the MRLC effort, and results will be made
available as a standard MRLC product. However, since the NALC program does not include
digital land cover maps, no accuracy assessment is available. We will use historical aerial
photography to calibrate digital land cover maps produced from NALC imagery of the 1970s,
1980s, and 1990s. We will evaluate different accuracy-assessment sampling designs and test the
application of historical imagery derived from aerial photography and NTM in the initial NALC
study areas (see previous discussion).
We will use a range of methods for error evaluation in classified imagery, including (1)
methods to assess the accuracy of individual date remote sensing classification maps and
landscape indicators (Congalton and Green 1999), and (2) methods to assess the accuracy of
change detection products (Khorram et al. 1999).
Advances in statistical ecology will be needed to take advantage of the new data sources
and estimates of accuracy (Gardner and Turner 1991). For example, existing statistical methods
are limited to dealing with data on a single spatial scale (Turner et al. 1991). Classification errors
are more likely to occur at the edge, rather than the center of a patch.
We know little about how to incorporate such a priori knowledge into analyses and assessments
(O'Neill et al. 1982). Holling (1992) has suggested that important impacts result simply from the
disruption of the scaled structure of landscape patterns, but we have no sensitive methods to
analyze those scales (Turner et al. 1991).
15
-------
These, and similar statistical problems, may be the limiting factors in our ability to analyze and
apply the new sources of data. We will evaluate the spatial pattern of error in classifying the land
cover in each of the NALC study areas and develop statistical approaches to assist us in
identifying such patterns.
B. New Remote Sensing Methods Development
The goals of this key research area are to: (1) enhance our interpretation of landscape
change from the early 1970s to the early 2000s, (2) improve our ability to detect the spatial
distribution and magnitude of stressors (e.g., surface mines and grazing) that cannot be readily
detected by conventional remote sensing approaches, and (3) improve future landscape analyses
through the use of new remote sensing imagery (e.g., Moderate Resolution Imaging Spectrometer
(MODIS)).
The science of remote sensing is currently undergoing a revolution in terms of both technical
capabilities and the number and quality of data sources. Advances in spatial and spectral
resolution, new platforms and continually improving digital analysis techniques are creating
significant new data sources for landscape-level analysis. Government remote-sensing programs,
such as NASA's Earth Science Enterprise (ESE), are currently developing a wide range of
aircraft and space-borne remote-sensing capabilities. In addition, recent commercialization of
space-borne, remote-sensing has resulted in plans by the commercial sector to develop several
high-resolution sensors that promise to provide global coverage at unprecedented levels of detail.
Table 3 shows a partial list of relevant new sensors and capabilities that will be available in the
near future. The new sensors may be especially important in estimating the magnitude and
distribution of landscape-level stressors that go undetected by Landsat sensors; these include
grazing, timber harvesting, mining, and certain agricultural activities (e.g., raising hogs and
chickens). These sensors and data sources are likely to play a significant role in the future of the
Landscape Sciences program, not only from the perspective of additional sources of data for the
development of the current suite of landscape indicators, but also from the perspective of new
scientific techniques that could enhance landscape analysis.
Data derived from these new sensors will not contribute directly to our national
assessment of 30-years of landscape change. However, they may play an important role in
identifying land use stresses on watersheds that affect landscape conditions observed during the
early 2000s samples. Moreover, the new data will provide a higher resolution baseline from
which future landscape change can be evaluated.
The following are brief summaries of (1) some of the major new capabilities and (2)
research we plan to conduct.
Digital Photogrammetry
Photogrammetry is the science of extracting reliable measurements from imagery and,
until recently, generally involved complex optical instruments and computationally intensive
techniques. The basis for nearly all precision mapping, photogrammetry was once limited to
aerial photographs acquired with precision mapping cameras. However, recently developed
digital techniques can be applied to a wide range of imagery, including that from satellite
systems, and can be used in a desktop computing environment. Products that may be relevant to
landscape analysis include vegetation-height measurements, precision measurements of slope,
volume, and fill, and the development of sub-meter digital elevation models. We will evaluate the
application of digital photogrammetry in each of the NALC study areas.
16
-------
High-Spatial-Resolution Satellite Remote Sensing
Several new sensors, such as IKONOS 2, which is scheduled for launch in 1999, will deliver
global multispectral imagery at the 3-meter level of spatial resolution (1 meter panchromatic).
This level of spatial detail is an order of magnitude better than previous satellite imaging
capabilities and could be significant in the development of landscape indicators of land use
stressors. We will evaluate whether this sensor can be applied effectively to estimate the spatial
distribution of such land use stressors as mining, urban sprawl, hog and chicken operations, and
specific cropping practices and the potential risks they pose to aquatic resources. For example, an
improved estimate of the spatial distribution of crop types should improve our ability to estimate
pesticide loadings to streams. Where possible, we will conduct these tests in areas already under
study (e.g., NALC studies). However, it may be necessary to add additional study areas to
evaluate certain stressors (e.g., grazing in the western U.S.).
Microwave Remote Sensing
Active microwave remote sensing, also known as radar, is available through a number of
operational satellites such as RADARSAT, ERS-1 and the Shuttle Imaging Radar (SIR)
experiments. These systems are active remote sensing instruments in that they operate
independently of reflected electromagnetic energy and transmit their own microwave energy
toward the earth's surface and record the reflected signal. For this reason, they have unique
almost all-weather, day/night capabilities. Because of the special wavelength characteristics in
this part of the EM spectrum, microwaves provide unique information about the landscape such
as soil- and leaf-moisture content, the dielectric properties of materials, and the physical structure
of the landscape. Also, microwave systems have been shown to increase land cover classification
accuracy when merged with traditional Landsat data (Haack and Slonecker 1994), especially
relative to wetlands that are covered by vegetation. Our primary goal will be to develop methods
to improve our estimate of wetlands, especially in areas where wetlands lie under extensive
canopy cover (e.g., the southeastern U.S.).
Thermal Infrared Remote Sensing
Although Thermal Infrared (TIR) remote sensing has existed for decades, new
technological developments and systems will create the potential for a new range of
environmental applications. TIR measures emissive energy in the 3-5 and 8-14 micron
wavelength range and is generally related to the heat budget of surface phenomena.
Until recently, TIR data were acquired mainly by sophisticated aircraft-based sensors and/or
satellite systems that collected data at spatial resolutions that were often inappropriate for
landscape-level analysis. New sensors such as MODIS and Landsat 7 will deliver TIR data that
can be used to add a thermal dimension to landscape classification and analysis, such as to
determine the role of urban heat islands in ecosystems processes.
High Spectral Resolution Remote Sensing
The most promising new technology in environmental remote sensing is the development
of hyperspectral and ultraspectral instruments and analytical techniques. Traditional
multispectral systems, such as the Landsat Thematic Mapper, capture data in broad spectral bands
selectively placed across the reflective spectrum. Hyper and ultra-spectral systems gather data in
hundreds and thousands of discrete, narrow, contiguous bands.
17
-------
As a result, landscape features can be analyzed using the same types of spectroscopic techniques
that are traditionally used in chemistry and astronomy. Figure 2 show a graphic comparison of
multi, hyper, and ultra-spectral data collection.
These spectral analysis techniques record the interaction of photons with material at the
molecular level and return unique spectral 'fingerprints' of the material being sensed. Figure 3
shows hyperspectral fingerprints of various minerals, as acquired from the NASA Advanced
Visible Infrared Imaging System (AVIRIS) instrument.
Spectral remote sensing has the potential to deliver critical new capabilities to the
Landscape Sciences Program, such as vegetation species identification, air and water pollutant
identification, and improved monitoring, e.g., of pesticide fate and transport. However, this
technology is in its infancy, and significant research is required, especially in the area of
environmental applications. We will evaluate the application of AVIRIS in the NALC study
areas and in those areas selected indicator quantification studies (see later discussion). Emphasis
will be given to detecting gradients of stressors as well as to improving land cover maps.
Another priority need is likely to fall in the area of cross-scale analysis. The new
imagery will provide simultaneous measures on individual, population, community, ecosystem,
watershed, and regional levels. Our ability to integrate this multi-scaled data is in a primitive
state (Rastettler et al. 1991). Hierarchy theory (O'Neill et al. 1986) provides a general framework
- it indicates that finer scale data can provide mechanisms, and coarser scale data can provide the
constraints. But the theories need to be developed into practical techniques for routine analysis of
multi-scaled data.
The most advanced thinking in cross-scale analysis has occurred in the national
monitoring and the global-change programs. The EPA's Environmental Monitoring and
Assessment Program developed a hexagonal spatial grid that facilitates the extrapolation of plot-
scaled measurements to regional assessments in a statistically valid manner (Overton et al. 1990).
The Global-modeling program has developed methods for extrapolating primary productions on
individual plots to carbon exchange across a continent (King et al. 1989). But these efforts only
scratch the surface of the problem (Risser 1987). New approaches will take advantage of the
emerging vector and parallel processing computers (Casey and Jameson 1988) to link ecosystem
models across a spatial grid (e.g., Running et al. 1989, Burke et al 1990). Further breakthroughs
will be required to make most effective use of the multi-scaled data that the new sensing
technologies will provide.
C. Change Detection
Identification, development and refinement of the means to detect change in landscape
pattern and composition overtime is central to the scientific objectives of this strategic plan.
Landscape patterns, processes and trends are directly related to natural and anthropogenic uses of
land and to the subsequent changes in those uses over time. Several key landscape indicators of
ecological conditions, such as vegetation gain/loss, human landscape alteration patterns and forest
fragmentation require assessment of changing landscape conditions (Jones et al 1997). Accurate
delineation of landscape change is fundamental to landscape ecology in general and is required to
fully understand the interface between dynamic biophysical and anthropogenic systems (Jensen
1986).
18
-------
Detection of changes in landscape attributes involves the comparison of multi-temporal
data sets to discriminate areas of land use or land cover change. This is generally accomplished
by comparing thematic or spectral spatial data sets, usually in a GIS environment.
Landscape change is identified using the direct spatial comparison of either the thematic
landscapes attributes, in the case of land cover data, by differences in reflected electromagnetic
energy in the case of remote sensing data, or by the calculation and comparison of changes in
landscape indicators derived independently from each time period. A critical issue relative to
detecting change over the 30-year period is the comparability of imagery generated from the
different sensors. The early to mid-1970s Landsat MSS data, the mid-to-late 1980s Landsat MSS
data, the early 1990s Landsat TM data, and the early 2000s Landsat TM data represent different
instruments that in some cases were or will be on platforms with different orbits and that have
different band characteristics. Moreover, differences in data collection dates and within and
among-year variations in climate will make comparisons across dates even more difficult. A
major task of ours will be to develop and test approaches within and among regions so that
landscape change can be accurately detected. We will focus our research on landscape change in
the NALC study areas (see previous discussion).
There are several technical considerations that can affect the quality of landscape change
estimates derived using these approaches, and they must be taken into account in planning change
detection research (Lillesand and Kieffer 1994). These include the spatial, spectral, radiometric
and temporal resolution of the remote sensing systems used. Several environmental
considerations, such as growing season, atmospheric conditions, and tidal stage can also be
critical. Major differences in any of these parameters can create erroneous positive or negative
change estimates and generally result in very noisy data. Additionally, the majority of change
detection research concentrates on relatively small, site-specific applications. Attempting to
develop change detection data on regional or national scales becomes more complicated because
of the need for temporally and spatially consistent data over large landscape units.
Change detection methods that overcome inherent spatial and temporal limitations of base data
are needed. Potential new methods include data development from new and improved sensors,
band differencing and ratioing techniques, data-set calibration by high-resolution sensors and
statistical development techniques such as those demonstrated in Jones et al (1997), pages 67-71.
Methods from Landscape Ecology must be adapted to distinguish natural changes, such as those
related to fire (irregular edges set by topography, dry sites, fuel-rich stands), from those related to
harvesting (straight edges, mature stands) or to development (straight edges, low topography)
(Krummel et al. 1987).
As demonstrated by the contributions of percolation theory (Gardner et al. 1987), fractal theory
(Milne 1988), and lacunarity theory (Plotnick et al. 1993), there is a need for land use change
theory to produce neutral models (Dunn et al. 1991). The neutral model provides a "null
hypothesis" that can then be tested statistically against the noisy change data (Gardner et al.
1989). This approach helps distinguish random errors from significant trends of change over a
landscape or region.
Landscape change theory has largely developed in the field of economic geography
where it predicts how to maximize utility by the spatial distribution of economic activities on the
landscape (Sklar and Costanza 1991). Recent approaches have used Von Thunen theory to
predict change as a function of distance to the nearest developing urban center (Wickham et al.
1999).
19
-------
Additionally, we will evaluate a sampling strategy to assess changes in landscape
indicators by ecoregion. An initial test of a landscape sampling design will be made in the mid-
Atlantic Region.
This region possesses enough variability in land cover types and human uses of the landscape to
test the sensitivity of the sampling design. Moreover, it will be the first region with wall-to-wall
land cover change estimates from the early 1970s to the early 1990s. The goal is to see whether
or not variability in landscape change can be reduced via a classification scheme and whether
change can be estimated through a sample-based approach. Additional evaluation studies will
likely be needed in topographically diverse areas of the western U.S. The sample-based approach
is viewed as a viable back up to the wall-to-wall assessment proposed in this project, especially if
wall-to-wall coverages of land cover data are not available.
D. Quantification of Landscape Indicators Relative to Condition of Aquatic
Resources
Landscape indicators are defined as a measure, an index of measures, or a model that
describes the condition of an ecosystem or one of its critical components (Hunsaker et al. 1990).
An indicator may reflect biological, chemical or physical attributes of ecological condition. The
primary uses of an indicator are to characterize current status and to track or predict significant
change. A landscape indicator may also be used to identify major ecosystem stress (Jones et al.
1997).
As our primary goal is to evaluate landscape change and aquatic-resource conditions, our
research will focus on landscape indicators that relate to changes in hydrological and ecological
processes.
Assessing the consequences of landscape change on aquatic resources requires a thorough
understanding of the degree to which aquatic resource conditions vary with landscape pattern.
Hierarchy theory (O'Neill et al. 1986) suggests that individual stream segments, wetlands, and
estuaries should be constrained by landscape composition and pattern within their watershed
areas. A number of studies have shown the strong relationships of water quality, water quantity,
and runoff to landscape characteristics. Water quality of streams, wetlands, and estuaries is
dependent on the landscape's ability to collect and purify water. A decrease in natural vegetation
indicates a potential for future water quality problems (Walker et al. 1993, Hunsaker and Levine
1995). Many studies have shown that the land uses within a watershed can account for much of
the variability in stream water quality (Omernik 1987, Hunsaker et al. 1992, Charbonneau and
Kondolf 1993, Roth et al. 1996). Agriculture on slopes greater than 3 percent, for example,
increases the risk of erosion (Wischmeier and Smith 1978). Empirical studies have established
the significant causal relationship between watershed characteristics and nutrient and sediment
loads (Levine et al. 1993).
A drastic change in vegetation cover, such as clear cutting in the Pacific northwest, can produce
90% more runoff than in watersheds not altered by human practices (Franklin 1992). The linkage
between intact riparian areas and high water quality is well established (Karr and Schlosser 1978,
Lowrance et al. 1985). Riparian habitat functions as a "sponge," greatly reducing nutrient and
sediment runoff into streams (Peterjohn and Correll 1984).
The percentage and location of natural land cover influences the amount of energy that is
available to move water and materials (Hunsaker and Levine 1995). Forested watersheds
dissipate energy associated with rainfall, whereas watersheds with bare ground and anthropogenic
cover are less able to dissipate that energy (Franklin 1992). The percentage of the watershed
surface that is impermeable, due to urban and road surfaces, also influences the volume of water
20
-------
that runs off and therefore the amount of sediment that can be moved (Arnold and Gibbons 1996).
Watersheds with highly credible soils tend to have greater potential for loss of soil to streams
than watersheds with non-erodible soils.
Moreover, intense precipitation events may more easily exceed the reduced energy threshold and
move large amounts of sediment across a degraded watershed (Junk et al. 1989, Sparks 1995). It
is during such events that human-induced landscape changes have their greatest negative impact.
The anoxia or hypoxia problem in the Gulf of Mexico results from extensive agriculture, loss of
riparian zones, and channelization throughout the Mississippi River Basin that permits mass
movements of sediment into the Gulf during extreme precipitation events, as witnessed in 1993
(Goolsby et al. 1999).
Growth of human populations often results in increasing water use and consumption.
Historical responses to these needs have often been the construction of dams and diversion of
water from streams and rivers. Dams create barriers to fish migration and often disrupt natural
hydrological patterns resulting in fundamental changes in aquatic habitats. Dams, therefore, can
decrease both the size and diversity of habitats and stream network connectivity. This, in turn,
generally increases the likelihood that whole fish populations will be lost due to a combination of
the higher extinction rates associated with smaller habitat sizes and the decreased probability of
recolonization.
Differential responses offish and benthic communities to landscape changes may reflect major
differences in life histories between these two types of organisms (Schlosser 1990, Poff and Allan
1995). Because of larger home ranges and migrational requirements, fish populations may
respond to changes over an entire network of streams, including barriers to migration, and to
overall habitat suitability. Conversely, benthic populations are more likely to respond to
landscape and habitat changes in local sites or individual stream segments. Research is needed to
test this hypothesis.
Considerable progress has been made in the development and application of landscape
indicators in measuring landscape conditions within watersheds that affect aquatic resources
(Hunsaker and Levine 1995, Jones et al. 1996, 1997, O'Neill et al. 1997). These indicators
include simple metrics generated from a single spatial database (e.g., the percentage of the
watershed surface with crop land calculated from a land cover database), metrics generated by
combining or overlaying spatial data (e.g., the proportion of stream miles in a watershed with
forested riparian cover calculated by intersecting spatial coverages of streams and land cover), or
metrics generated through the implementation of relatively simple, spatially distributed models
(e.g., estimation of soil loss with models incorporating the Universal Soil Loss Equation). The
following summarizes more recent advances in the application of landscape indicators:
1. quantification of landscape composition and pattern from remotely sensed and
other spatial data (U.S. EPA 1994, Riitters et al. 1995, Jones et al. 1996, O'Neill et
al. 1988, 1997),
2. (statistical independence of landscape metrics (Riitters et al. 1995),
3. influence of grain and pixel size in determining landscape composition and pattern
(Wickham and Riitters 1995),
4. influence of sample and assessment-unit size on landscape-indicator performance
(Hunsaker et al. 1994, Wickham and Riitters 1995, O'Neill et al. 1996),
21
-------
5. scale relationships between landscape composition and pattern in different
environmental settings (O'Neill et al. 1991, Rastetter et al. 1991),
6. sensitivity of landscape indicators to human-use gradients (Wickham et al. 1997a),
7. sensitivity of landscape indicators to misclassification of remotely sensed data
(Wickham et al. 1997a),
8. relationships between environmental stressors and landscape pattern (Krummel et
al. 1987, O'Neill et al. 1988, Mclntyre 1995, Wickham et al. 1997b),
9. analysis of landscape indicators to assess relative watershed conditions at the
regional scale (Jones et al. 1997, Wickham et al. 1999a and b),
10. relationships between landscape pattern and aquatic resources (Hunsaker et al.
1992, 1995, Charbonneau and Kondolf 1993, Walker et al. 1993, Poff and Allan
1995, Frenzel and Swanson 1996, Roth et al. 1996),
11. spatial GIS models (spatially distributed models) that relate landscape
characteristics to: water balance (Storck et al. 1998), nutrient loadings to surface
waters (Soranno et al. 1996,Young et al. 1996), erosion (Mitasova et al. 1996,
Zhang et al. 1996), soil chemistry (Burke et al. 1990), and nitrogen mineralization
(Fan et al. 1998) at watershed to regional scales,
12. biogeographic models that explicitly consider landscape composition and pattern
(Wiens 1985, Wiens et al. 1993, 1997, Schumaker 1996, Riitters et al. 1997).
The outstanding progress in all of these research areas, new spatial databases covering
large areas at relatively fine scales (e.g., 30 meters), and new computing capabilities make us
optimistic that the remaining critical gaps can be filled in a timely manner. However, despite
these advances, we lack a comprehensive understanding of how landscape pattern influences
aquatic resources and associated processes. For example, although we are aware of landscape
factors that influence water quality, recent papers (e.g., Roth et al. 1996, Weller et al. 1996)
suggest that the importance of landscape features may change in different environmental settings,
or when moving from one spatial scale to another. Differences in biophysical attributes of
watersheds (e.g., size, gradient, precipitation, soils, geology) may explain why these studies had
different results. Moreover, we lack knowledge of how these relationships vary among different
regions of the country. For example, highly variable precipitation and topographical patterns
make the western US hydrologically and ecologically distinctive from the eastern US, and these
differences are likely to result in different spatial and temporal relationships between landscape
conditions and aquatic resources. To fill some of these important gaps, our research will attempt
to:
1. establish quantitative relationships between various indicators of landscape pattern
and water quality, discharge, and biota in different biophysical settings (e.g.,
watershed size, gradient, geologic setting, climate regime);
2. determine how the importance of different landscape patterns in explaining water
quality, discharge, and biological condition varies within and among different regions
of the US; and
3. determine which stressors contribute most to observed landscape condition.
We propose to establish several paired watershed studies in regions of the country where
relationships between landscape pattern and aquatic resources and associated ecological and
hydrological processes are likely to be different. We will use existing spatial databases of
22
-------
landscape attributes and biophysical characteristics; national coverages of these data at relatively
fine scales (30-60 meters) should be available within the next two to five years. We will use
existing stream and estuary samples, primarily from the EMAP Surface Waters and USGS
National Water Quality Assessment (NAWQA) programs, to evaluate nutrients, sedimentation,
and indicators of stream biological conditions.
We will "pair" each set of aquatic and landscape indicator data related to the site's watershed
area. Using a range of multi-variate statistics, we will determine which landscape indicators
explain the greatest variability in aquatic site conditions.
From these results, we will develop empirical models to be used in the assessment process. In
conducting these studies in different locations across the US, we will consider how different
temporal and spatial scales of the data might affect our results. The results will also be critical in
deciding which landscape indicators to use in the assessment process, in parameterizing
watershed models to assess the consequences of landscape change on water resources, and in the
development of landscape indices (statistical combination and weighting of landscape indicators).
E. Assessment Methods Research and Development
We have two primary assessment goals related to the project: (1) an assessment of how
landscapes have changed from the early 1970s to the early 2000s across the U.S., and (2) an
assessment of the consequences of observed change on aquatic resources based on quantification
studies (see previous discussion). The primary objective of the assessment is to identify and
prioritize those areas that have undergone changes in landscape pattern that are likely to pose
significant risks to water quality and biota in streams and estuaries. The assessment will depend
on progress made in all of the high priority research areas discussed in this plan, and will require
close coordination between basic data gathering and processing, multi-watershed research, and
implementation of the resulting products. To achieve the overall assessment objectives, project
components must be carefully sequenced within each region (Figure 4).
One key issue in the assessment phase of this strategy is the selection of appropriate
assessment units. We propose to use two general types of assessment units: (1) a watershed-level
unit to identify and demonstrate those biophysical attributes that mediate the response of the
aquatic resources to landscape change, and (2) a unit that relates to the aquatic resource itself
(e.g., a stream segment or individual estuary). Research described in the previous sections should
assist us in determining how aquatic resources respond to landscape indicators in different
biophysical settings. The aim of this research will be to develop a watershed classification that
characterizes different aquatic resource responses based on biophysical properties of the
environment (e.g., watershed size, topography, climate; see Maxwell et al. 1995).
Another potential way to approach this problem is to report out on risks to individual stream
segments and estuaries based on landscape indicator values for watershed or drainage areas for
each of these aquatic resources. Instead of coding watershed risks based on landscape change, we
would code the risk to each individual stream and estuary. To do so will require research and
development in computer data analysis, including watershed-level applications of GIS, primarily
because of the large number of watersheds and associated indicators that would have to be
generated for each stream segment or estuary. An assessment of individual stream segments and
estuaries is important because the EPA's Office of Water and individual Regional Offices need
robust and systematic approaches to identify those water bodies that are most likely to exceed
certain water quality standards.
Another key assessment issue is how to report results of multiple indicator sets relative to aquatic
resource conditions. At least four approaches are possible:
23
-------
1. report values for those landscape indicators that explain the highest proportion of
variation in aquatic resource condition (based on quantification studies), but do not
combine landscape indicators into an index;
2. apply multi-variate statistical approaches, such as cluster and principal components
analyses, to determine those watersheds and individual aquatic resources at greatest risk
(see Jones et al. 1997 and Wickham et al. 1999c);
3. develop empirical models from the quantification research (see previous section)
that combine landscape indicator values into an index or relative risk
characterization;
4. apply process models that utilize landscape indicators to estimate changes in
specific aquatic variables (see Fan et al. 1998 and Storck et al. 1998 for examples).
It is likely that a combination of all four of these assessment approaches will be used.
Again, results from the quantification studies should assist us in deciding which approaches to
use. For example, we may find that one or two landscape indicators explain most of the variation
in certain water quality parameters. In this case, we might want to report out on only these two
landscape indicators without combining them into a multi-variate model.
Another important scientific gap in our ability to assess relative risk is the inability to
deal with synergistic effects. One assessment of the mid-Atlantic States (Jones et al. 1997) used
the relative magnitude of stressors: the greater the number and magnitude of stressors, the greater
the relative risk. But this approach assumes that the stressors act independently (Hunsaker et al.
1990). What is needed is a mechanistic understanding of how the stressors interact to impact
aquatic resources. This is a major research area, but every incremental improvement in our
understanding can improve our ability to prioritize.
Risk assessment is basically a quantitative method (Barnthouse et al. 1982, O'Neill et al.
1982). The approach estimates the probability of an undesirable impact as a function of
uncertainties in measurements (O'Neill et al. 1981), natural variability (O'Neill 1979), model
assumptions (Bartell et al. 1983), etc. Although we can begin to rank impacts to water resources
using qualitative approaches (Graham et al. 1991) ultimately, we must move to quantitative
estimates of risk (Barnthouse et al. 1983). This is an important area of research, with long-term
implications for our ability to manage the Nation's water resources.
A number of ecosystem and hydrologic models could be applied to prioritizing impacts.
Unfortunately, few of these models have been tested across regional scales. As a result, we
cannot confidently move the models across the full range of biophysical conditions in the Nation
without first confirming their "transportability." Statistical models are largely limited to specific
watersheds and, in general, cannot be applied to watersheds in different biophysical settings.
Therefore, research is needed to determine the applicability of these models in assessing
landscape change and its consequences on water resources.
Each of these assessment issues will be evaluated in geographic areas where other
research activities have been initiated, including NALC, stressor gradient, and indicator
quantification studies.
24
-------
Our strategic approach will be to implement research and assessment activities in a
systematic manner within each region (see Figure 4). Significant progress in spatial data
assembly and in key research areas is needed to complete a national landscape assessment. Our
approach will be to implement the program on a "region-by-region" basis through a series of
focused studies. Where possible, we will conduct key research activities in the same geographic
areas. This will depend upon the degree to which certain study areas possess data and the
physical environment to address the range of research questions. Once the studies have been
completed, we will apply the methodologies to the regional spatial databases to conduct the
national assessment.
A proof of concept of our approach will be carried out in the mid-Atlantic Region of the
U.S. as this area will be the first to have all key aspects of the approach completed.
A. Implementation Themes
A number of themes will be central to all of the research and assessment activities. These
themes include:
understanding stakeholder needs,
highlighting assumptions, critical research steps, and interdependencies,
setting research and development priorities,
developing and testing a method or approach
assessing the method performance for different geographic areas,
considering ORD's high priority geographic areas in the selection of study areas,
managing information for use by collaborating scientists, clients, and the public,
collaborating within the EPA and with academia, federal agencies, and others, and
collaborating within the EPA and with academia, federal agencies, and others, and
providing customer services and products
25
-------
These themes are discussed further below.
Stakeholder Needs
To better understand stakeholder needs, we will seek the assistance of Regional and
Program Offices, and various laboratories and programs (e.g., EMAP) in ORD. Through an
ongoing process of workshops, informal and formal meetings, surveys, and working with existing
forums (such as the ORD Landscape Working Group) we will refine our understanding of their
priorities and long-term concerns. Data quality objectives and assessment questions will be
developed. We will solicit input and feedback on interim results through various mechanisms,
including Internet web pages and user conferences (discussed further under customer services).
Highlighting Assumptions, Critical Research Steps, and Inter dependencies
Within each research area, clearly articulating our assumptions, critical research steps,
and interdependencies will be essential to the communication and interpretation of our interim
products and results. For example, additional land use data for the year 2000 is critical to
performing a 30-year assessment, so the acquisition of these data is a high priority for the
Landscape Sciences Program. NALC data variability from the 1970's compared to the 1980's and
1990's may prevent some types of land cover and change analyses, so this potential problem will
need to be addressed early in the program's development through the focused NALC studies.
Transferring approaches or indicators from one region of the country to another may pose
unexpected problems. Finally, the transition from research mode to full-scale application mode
will require careful judgements about the adequacy of research results across multiple geographic
areas. The research community and other stakeholders will need to be involved in these decisions
in some capacity. Full-scale national production will depend on the outcome of these
interactions. These topics will be discussed in detail in the individual research plans and also will
be revisited as the program develops.
Setting Priorities
Implementation of a national landscape change assessment requires establishment of
priorities for data collection and research activities. Priority setting across the entire Landscape
Sciences Program will rest with the Environmental Sciences Division management, subject to
approval by NERL management. Research projects within the program will be sequenced using a
"fast-fail" approach so that a workable plan to achieve the national assessment is established as
early as possible in the program development.
There are a number of priorities relative to accomplishing the national assessment, but
some of the activities must be completed within the first three years of the project to determine
the feasibility of the approach. The two most important priorities relative to the approach are:
• determining whether landscape change can be estimated accurately across
different sensors in different regions of the US
determining whether landscape indicators explain variation in aquatic-resource
parameters in different regions of the US
26
-------
These priorities are reflected in the number and type of research activities planned in the
early years of the project (see Appendix I). If the results of either of these two research areas are
poor or inconsistent, then a national assessment of landscape change (from the 1970s to the early
2000s) and its consequences on aquatic resources may not be possible. However, if the results
suggest that the basic concept will work, then the next most important priority will be the
assembly of spatial databases needed to generate the indicators nationally. Since the MRLC will
be generating land cover for the early 1990s and 2000s, our emphasis will be on the generation of
early 1970s and mid-1980s digital land cover maps from the NALC database. To do so will
require replacement of poor quality 1970s scenes, classifying the imagery into land cover, and
assessing the accuracy of the land cover databases. If the classification of early 1970s and mid-
1980s NALC proves infeasible, we will limit the wall-to-wall assessment of landscape change
from the early 1990s to the early 2000s. Alternatively, we might deploy a sample-based approach
to assess landscape change between the early 1970s and the early 2000s (see early discussion);
such an approach might require far fewer images.
Priority must also be given to upgrading computing capability within the Landscape
Ecology and Landscape Characterization Branches. We will analyze and identify current and
anticipated needs for hardware and software to meet both our near-term priorities and those of the
10-year national assessment goal.
Table 4 summarizes the respective duties and responsibilities of the three organizational
units that comprise the Program and how they will be coordinated.
Developing and Testing Approaches and Methods
Each of the major research areas involves developing and testing a combination of
landscape ecology, landscape-characterization, watershed-modeling, and remote-sensing
approaches. While working within such diverse disciplines, many differences in approach are
expected. However, some commonalities are expected. Typical stages will include specification
or refinement of the parameters upon which the research is based; identification or development
of appropriate conceptual models related to the parameters of interest; determination of what
landscape scales and units are appropriate; selection of a procedure (or an indicator in the case of
indicator development) that is consistent with the parameters and conceptual model; selection of a
geographic area for initial testing; acquisition of data to test the approach; and extension of the
testing to additional geographic areas. In the process, quality assurance approaches and methods
for data analysis and interpretation will be developed and tested. As mentioned earlier, research
will be conducted in a limited number of geographic areas. The number of different study areas
established will depend on the degree to which multiple research objectives can be accomplished
in individual geographic areas.
Study areas will be selected based on: (1) landscape data availability, (2) aquatic resource
data availability, (3) stressor and biophysical spatial gradients, (3) importance to, and support
from EPA Regional Offices, and (4) importance to integrated programs of ORD (e.g., EMAP).
The third criterion is critical because EPA Regional Office collaboration is essential in achieving
a meaningful national assessment.
Assessing the Method Performance for Different Geographic Areas
27
-------
With the goal of a national assessment, validating each procedure or indicator for
multiple geographic regions becomes a high priority. We anticipate finding differences between
regions in the temporal and spatial scales at which biophysical and landscape characteristics
influence water resources. We also expect to find differences in the importance of landscape
pattern among regions. Therefore, multi-regional studies are needed.
While initial testing will be performed in an area selected for scientific considerations,
accessibility, or availability of data, a successful procedure or indicator must be re-evaluated for
application in other geographic areas. We expect that 10 or possibly 11 areas of the U.S.
including the Northeast, the Mid-Atlantic, the Southeast, the Midwest/Central Plains, the Upper
Midwest, the Upper Great Plains, the Pacific Northwest, the Pacific Southwest, the Southwest,
the Inter-Mountain West, and the South-central U.S. will be needed. These geographic regions
are consistent with broad-scale ecoregions and physiographic provinces for the lower 48 States.
Some procedures or indicators will require more or less testing to determine their geographic
transportability; the research will be planned accordingly (see Appendix I).
Integration issues arise when results from multiple research activities and diverse
geographic areas are combined to create a coherent national assessment. Data collected by one
method in one part of the U.S. may or may not be comparable to data collected by a similar
method at another location. Differences may develop in interpretive analysis methods due to
inherent differences from one ecosystem to another. Methods and sensors used historically may
not agree with methods and sensors in current use. Quality of remote sensing data and derived
results may differ from one region to another because of the availability of cloud-free days.
Timing of projects and data delivery will be a critical factor in planning for a smooth flow of
information so that synthesis and integration issues can be addressed properly. Because of the
importance of the integration issues, the difficulty in predicting specific problem areas so early in
the program, and the potential for "show-stoppers," this topic will be revisited frequently in the
dynamic planning of the program.
Considering ORD's High Priority Geographic Areas in the Selection of Study
Areas
When considering where to test initially and how to expand the testing process
geographically, certain options offer special advantages. Significant opportunities for sharing
existing data and leveraging sampling activities are available by working in one of the high
priority areas of interest to the ORD. These areas currently include the Pacific Northwest, Great
Lakes, Gulf of Mexico, and South Florida. In addition, the ORD Ecological Research Program
has designated the Mid-Atlantic region as a single, primary area of coordination for ORD with
other federal and state agencies. Finally near-laboratory sites, which serve as field laboratories,
have been designated and include the Savannah, Neuse, Lower Colorado, and Little Miami River
watersheds. These geographic areas offers different types of ecosystems, topography, climate,
and local issue priorities.
Managing Information for Use by Collaborating Scientists, Clients, and the
Public
The acquisition, manipulation, analysis, and dissemination of large databases associated
with this research program require the development of a comprehensive information management
28
-------
plan. The plan must address a number of key issues needed to support a program of this
magnitude, including:
(1) exchange of data, analysis programs, and information among participating
scientists located across the country;
(2) availability of data and information on results to the scientific community and the
public;
(3) documentation of analytical methodologies and data quality such that the
methods and results can be replicated;
(4) archiving of data and analysis programs and results.
The plan also must consider upgrades to the computer system and network that will be
needed to support the increasing scope of the project over the next 10 years. As a result of these
needs, we propose to develop and implement a comprehensive information management plan by
September 2000. This plan will be coordinated with the EPA Office of Information Resources
Management, as well as several collaborating organizations, including EMAP and the EROS Data
Center.
Collaborating within the EPA and with Academia, Federal Agencies, and
Others
Because of the size of this national assessment, intellectual and financial collaboration
both within and outside the EPA is essential. Within the EPA we will solicit partners who share
our research interests and priorities. Because the EPA Regional Offices will be key to the
implementation of the national goal, we train and assist Regional Office staff to compile
landscape data, generate landscape indicators, and conduct regional-scale assessments. Regional
Offices have already started to reorganize and pull together a critical mass of landscape and GIS
staff. The EMAP Western Pilot will be a test of the proposed joint ORD/Regional Office
implementation of the Program. Additionally, implementation of the Program will correspond to
national implementation of EMAP Surface Waters and Near Coastal Programs. Close
coordination with these two groups will be necessary to conduct landscape indicator/aquatic
resources quantification (see earlier discussions). Two additional ORD programs are available to
Regional Offices desiring ORD support: the Regional Applied Research Effort (RARE); and the
Regional Environmental Monitoring and Assessment Program (REMAP). Studies under these
programs are developed jointly by the Regional Office and ORD staff and typically last one to
two years. Such studies, when focused within one of the key research areas listed above, can help
us to test our concepts and research results in settings which have high priority to a Regional
Office. Additionally, EMAP provides EPA Program Offices (e.g., the Office of Water) an
opportunity to be involved. Part of the focused research is aimed directly at their needs;
leveraging ORD resources with Program Office funds is also an option. We anticipate that EPA's
Offices of Water; Prevention, Pesticides and Toxic Substances; Policy, Planning and Evaluation;
International Affairs; Administration and Resource Management; Solid Waste and Emergency
Response will benefit from this research and may be involved directly as partners and
collaborators. In the ORD, a number of programs within the National Research Laboratories and
Centers offer opportunities for collaboration, as listed in Table 5. In some cases, we have long-
standing cooperative efforts, while in other cases, we will be developing new relationships. The
STAR Grants Program is a competitive, peer-reviewed, investigator-initiated, EPA-funded,
research grants program to foster innovative and far-reaching scientific projects and facilitate
cooperation between EPA and the scientific community. Each year research categories are
designated to which potential grantees respond. We anticipate working with NCERQA to
29
-------
develop research opportunities that can complement the scientific discovery taking place within
the Agency.
Outside the EPA, opportunities for collaboration exist among federal and state agencies,
especially with the Departments of Interior, Commerce, Agriculture, and Energy, and NASA.
Some of these relationships are summarized in Table 6. Two existing collaborative efforts
illustrate these relationships. The MRLC was developed to provide the capability for broad-based
research on current and future conditions of the physical and biological resources of the United
States, and the NALC Project was developed by EPA and the EROS Data Center/USGS to
provide a standardized digital-data set of satellite images for the 1970s, 1980s and 1990s for the
contiguous 48 states and Mexico (see earlier discussions).
Providing Customer Services and Products
Customer services and products will be an important factor in achieving the long term
goal of the Landscape Sciences Program. These will include the following:
• Peer-reviewed research plans, revised as needed and annually updated
appendices that list products, completed outputs and expected outputs,
• Journal articles that provide current research results to technical audiences,
Guidance and user manuals that provide systematically organized advice,
instruction, and recommendations to technical audiences, especially within the
Regional Offices and States,
• Data management via EPA's Environmental Information Management System
(EIMS) and associated databases to make data available to interested users inside and
outside the agency. These databases are expected to include NALC, MRLC-92,
MRLC-2000, and others,
Two Internet web sites, one for Remote Sensing Methods, and one for Landscape
Indicators to make research results and guidance available to technical audiences and
the general public,
• Computer Landscape Assessment Tools to assist Regional Offices in conducting
landscape analysis relative to aquatic resources,
Biannual User Conferences on Remote Sensing and Landscape Indicators, hosted
by the Landscape Science Program. These conferences will offer sessions on research
results, Regional and Program Office applications, and training. These will be the
primary mechanisms for our outreach efforts (methods and indicator guidance). They
will result in wide recognition and use of our research products and lead to consensus
on research needs and priorities, and
• Training of the Regional and Program Office staff upon request.
We will prioritize among these product types based on feedback from our stakeholders.
30
-------
B. Research Plans
A key component of the research strategy will be development and peer review of
individual research plans for specific research activities outlined in this plan. These plans will
follow the form of National Science Foundation research grant proposals, including a description
of the research area, a review of the state-of-science, statement of hypotheses and questions to be
addressed, a description of the study and analysis designs, and time lines, resource needs, and
backgrounds of principal and co-principal investigators. The research plans will provide details
on individual projects that were too voluminous to describe in this plan. Generally, these
research plans will cover a two to three-year study period. Some research activities, including
Regional Environmental Monitoring and Assessment Program (REMAP) projects that already
have research plans will not require a new plan. Research plans will not be required for
development of analysis tools and web sites.
C. Time Line for Research and Assessment Implementation
Appendix I provides a road map of activities and milestones to achieve the goal of the
program. These projects represent a diversity of programs funded by the EPA, including EMAP,
the Regional Environmental Monitoring and Assessment Program (REMAP), the Regional
Vulnerability Program (ReVA), and the Global Change Research Program (GCRP). Although
not listed, the EPA funds a number of external grants (Science to Achieve Results (STAR)
Grants) that will contribute to our knowledge of landscape change and how it affects aquatic
resources. Information on the EPA External Grants program can be found at
http://es.epa.gov/ncerqa/grants/.
We believe that the national landscape assessment can only be achieved by coordinating
and leveraging these individual programs. We anticipate that the road map will change
somewhat, especially in out years where initiation of specific activities is dependent on
significant research progress, completion of national-scale data sets, and funding. Appendix III
provides an example of milestone outputs.
31
-------
. •"
The measures of success for the Landscape Sciences Program include publication
productivity, client satisfaction, recognition by the science community, successful outside
evaluations, the successful completion of GPRA objectives and milestones, and the adoption of
our methods and approaches by others, ranging from county planners to other countries.
Productivity can be evaluated in terms of the quality and quantity of our publications and
databases when considered in relation to our assignments, budget, and staff. Our scientists
produce technical journal articles, databases, reports, and research plans. As this list suggests, the
wide range of activities conducted within the Program does not always lead to journal articles.
However, we consider journal articles highly desirable because of the peer review and recognition
inherent in the process. We are striving to increase the numbers of journal articles and databases
by focusing our efforts in this area, while continuing to produce the other products when needed.
Client satisfaction involves the quality, relevance, and timeliness of our products and support
(Table 7), as perceived by our many stakeholders. We provide our clients several different means
for giving feedback on our work, including meetings to discuss results and tear-off pages with
product questionnaires in the back of reports. Recognition by the science community can be
judged by the number of requests received for invited presentations, and consultation and
collaboration with our scientists, and citations of our papers in other scientific publications.
Outside evaluations take several forms, including technical program reviews conducted
by a panel of academics and leaders in similar programs from other federal agencies. Such
reviews, scheduled every three years, in accordance with ORD policy, consist of comprehensive
presentations to the review panel of our research results, work in progress, and plans for the
future. Program peer review results are summarized in a report by the reviewers. Our progress in
meeting these measures is tracked both formally and informally. Published journal articles are
listed for each Branch in monthly reports, as are major requests for consultation or collaboration.
Client satisfaction as indicated by tear sheets is tallied and examined for trends, and delivery of
key products is evaluated by follow-up visits or telephone conversations by senior managers.
This combination of measures provides a sound basis for the ongoing evaluation of the Landscape
Sciences Program.
32
-------
Arnold, C.L., and C.J. Gibbons. 1996. Impervious surface coverage: the emergence of a key
environmental indicator. J. American Planning Assoc. 62:243-258.
Ball, G.L. 1994. Ecosystem modeling with GIS. Environ. Man. 18:345-349.
Barnthouse, L. W., D. L. DeAngelis, R. H. Gardner, R. V. O'Neill, C. D. Powers, G. W. Suter, II,
and D. S. Vaughan. 1982. Methodology for environmental risk analysis. ORNL/TM-8167.
67pp.
Barnthouse, L. W., G. W. Suter, II, and R. V. O'Neill. 1983. Quantifying uncertainties in
ecological risk analysis, pp. 487-489. IN J. F. Bell and T. Atterbury (eds.), Renewable
Resource Inventories for Monitoring Changes and Trends. College of Forestry, Oregon State
University, Corvallis, Oregon.
Bartell, S. M., R. V. O'Neill, and R. H. Gardner. 1983. Aquatic ecosystem models for risk
assessment, pp. 123-127. IN W. K. Lauenroth, G. V. Skogerboe, and M. Plug (eds.),
Analysis of Ecological Systems: State-of-the-Art in Ecological Modeling. Elsevier
Scientific Publishing, New York.
Brady, W., D.R. Patton, and J. Paxson. 1985. The development of Southwestern riparian gallery
forests. US Forest Service Gen. Tech. Rpt. No. RM-120, p 39-43.
Burke, I. C., D. S. Schimel, C. M. Yonker, W. J. Parton, L. A. Joyce, and W.K. Lauenroth. 1990.
Regional modeling of grassland biogeochemistry using GIS. Landscape Ecology 4:45-54.
Casey, R. M. and D. A. Jameson. 1988. Parallel and vector processing in landscape dynamics.
Applied Mathematics and Computation 27:3-22.
Charbonneau, R., and G.M. Kondolf. 1993. Land use change in California, USA: nonpoint
source water quality impacts. Environ. Man. 17:453-460.
Congalton, R.G., and K. Green. 1999. Assessing the accuracy of remote sensing data: principles
and practices. Lewis Publ, Boca Raton, Florida. 137 p.
Dunn, C. P., D. M. Sharpe, G. R. Guntenspergen, F. Stearns, and Z. Yang. 1991. Methods for
analyzing temporal changes in landscape patterns. Pp 173-198. IN M. G. Turner and R. H.
Gardner (eds.) Quantitative Methods in Landscape Ecology. Springer-Verlag, NY.
EPA. 1988. Future risk: research strategies for the 1990s. EPA Science Advisory Board.
Washington, D.C.
EPA. 1994. Landscape monitoring and assessment research plan. EPA 620/R-94/009,
Washington, D.C.
EPA. 1997. 1997 update to ORD's Strategic Plan. EPA/600/R-97/015, Washington, D.C.
33
-------
EPA. 1998a. Ecological research strategy. EPA/600/R-98/086, Washington, D.C.
EPA. 1998b. Guidelines for ecological risk assessment. EPA/630/R-95/002F. Washington,
D.C.
EPA. 1998c. National Exposure Research Laboratory Research Strategy. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina. Draft.
EPA. 1998d. Federal Water Pollution Control Act. EPA-840-R-98-001, U.S.C.A. Sect. 1251 et
seq.; Public Law 100-202.
EPA. 1999. EXPOSITION: Integrated environmental decision-making in the 21st Century. EPA
Science Advisory Board, Washington, D.C.
Fan, W., J.C. Randolph, and J. L. Ehman. 1998. Regional estimation of nitrogen mineralization
in forest ecosystems using geographic information systems. Ecol. Appl. 8:734-747.
Forman, R.T.T. 1990. Ecologically sustainable landscapes: the role of spatial configuration. Pp.
261-278, in Forman, R.T.T., and I.S. Zomeveld (eds.), Changing landscapes: an ecological
perspective.
Franklin, J. F. 1992. Scientific basis for new perspectives in forests and streams, pp. 25-72. in
R. J. Naiman (ed.) Watershed Management. Springer-Verlag, NY.
Frenzel, S.A., and R.B. Swanson. 1996. Relations offish community composition to
environmental variables in streams of central Nebraska, USA. Environ. Man. 20:689-795.
Gardner, 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., R. V. O'Neill, M. G. Turner, and V. H. Dale. 1989. Quantifying scale dependent
effects with simple percolation models. Landscape Ecology 3:217-227.
Gardner, R. H., and M. G. Turner. 1991. Future directions in Quantitative Landscape Ecology.
Pp. 519-525. in M. G. Turner and R. H. Gardner (eds.) Quantitative Methods in Landscape
Ecology. Springer-Verlag, NY.
Golley, F.B., J. Jensen, J. MacMahon, C. Pease, and D. Slack. 1997. Peer review of the
Landscape Sciences Program in the Environmental Sciences Division, Las Vegas, Nevada
USA, December 6-8, 1996. 35 pp.
Goolsby, D.A., W.A. Battaglin, G.B. Lawrence, R.S. Artz, B.T. Aulenbach, RP. Hooper, D.R.
Kenney, and G.J. Stensland. 1999. Flux and sources of nutrients in the Mississippi-
Atchafalaya River Basin. NOAA/Topic 3 Report. Paper submitted to the White House
Office of Science and Technology Policy Committee on Environment and Natural Resources
- Hypoxia Work Group.
Government Performance Results Act (GPRA). 1993. Government Performance and Results Act
of 1993 (GPRA), Public Law 103-62, signed August 3, 1993.
34
-------
Graham, R.L., C.T. Hunsaker, R.V. O'Neill, and B.L. Jackson. 1991. Ecological risk assessment
at the regional scale. Ecol. Appl. 1:196-206.
Haack, B.N., and E.T. Slonecker 1994. Spaceborne radar and thematic mapper data fusion for
location villages in Sudan. Photogr. Eng. Remote Sens. 55:1253-1257.
Rolling, C. S. 1992. Cross-scale morphology, geometry, and dynamics of ecosystems.
Ecological Monographs 62:447-502.
Hunsaker, C. T., R L. Graham, G. W. Suter, II, R. V. O'Neill, L. W. Barthouse, and R. H.
Gardner. 1990. Assessing ecological risk on a regional scale. Environmental Management
14:325-332.
Hunsaker, C.T., D.A. Levine, S.P. Timmins, B.L. Jackson, and R.V. O'Neill. 1992. Landscape
characterization for assessing regional water quality. Pp. 997-1006, in D.H. McKenzie, D.E.
Hyatt, and V.J. McDonald (eds.), Ecological indicators. Elsevier Appl. Sci., New York.
Hunsaker, C.T., and D.A. Levine. 1995. Hierarchical approaches to the study of water quality in
rivers. BioScience 45:193-203.
Jensen, J.R. 1986. Digital Image Processing, 2nd Edition, Prentice-Hall, Englewood Cliffs, NJ.
Jones, K.B., J. Walker, K.H. Riitters, J.D. Wickham, and C. Nicoll. 1996. Indicators of
landscape integrity. Pp 155-168, in J. Walker and D.J. Reuter (eds.), Indicators of catchment
health: a technical perspective. CSIRO Publishing, Melbourne, Australia.
Jones, K.B., K.H. Riitters, J.D. Wickham, R.D. Tankersley, RV. O'Neill, D.J. Chaloud, E.R.
Smith, and A.C. Neale. 1997. An ecological assessment of the United States mid-Atlantic
Region: a landscape atlas. EPA/600/R-97/130.
Junk, W. J., P.B. Bailey, and R.E. Sparks. 1989. The flood pulse concept in river-floodplain
systems. Can. Spec. Publ. Fish. Aquat. Sci. 106:110-127.
Karr, J. R., and I. J. Schlosser. 1978. Water resources and the land-water interface. Science
201:229-233.
Kepner, W.G, K.B. Jones, D.J. Chaloud, J.D. Wickham, K.H. Riitters, and RV. O'Neill. 1995.
Mid-Atlantic landscape indicators project plan. EPA/620/R-95/003.
King, A. W., R. V. O'Neill, and D. L. DeAngelis. 1989. Using ecosystem models to predict
regional CO2 exchange between the atmosphere and the terrestrial biosphere. Global
Biogeochemical Cycles 3:337-361.
Khorram, S., G. Biging, N. Chrisman, D. Colby, R. Congalton, J. Dobson, R. Ferguson, M.
Goodchild, J. Jensen, and T. Mace. 1999. Accuracy assessment of land cover change
detection. American Soc. Photogr. Remote Sens. Monogr. Series #1, Bethesda, Maryland.
Krummel, J. R, R. H. Gardner, G. Sugihara, R. V. O'Neill, and P. R. Coleman. 1987. Landscape
patterns in a disturbed environment. Oikos 48:321-324.
-------
Levine, D.A., C.T. Hunsaker, S.P.Timins, and J.J. Beauchamp. 1993. A geographic information
system approach to modeling nutrient and sediment transport. Oak Ridge Nat. Lab, Report
No. 6736.
Lee, M.T. 1989. Soil erosion, sediment yield, and deposition in the Illinois River Basin. Pp. 718-
722, in S.Y.Y. Wand (ed.), Proceedings of the International Symposium on Sediment
Transport Modeling. American Soc. of Civil Eng., New York.
Lillesand, T.M., and R.W. Kieffer. 1994. Remote sensing and image interpretation. JohnWiley
& Sons, Inc. New York.
Lowrance, R. R., R. Leonard, and J. Sheridan. 1985. Managing riparian ecosystems to control
nonpoint pollution. Journal of Soil and Water Conservation 40:87-91.
Maxwell, J.R., C J Edwards, M.E. Jensen, S.J. Paustian, H. Parrott, and M. Donley. 1995. A
hierarchical framework of aquatic ecological units in North America. Gen. Tech. Rep. NC-
176, U.S. Dept. of Agr., Forest Serv., North-Central Forest Exp. Stat, St. Paul, Minnesota. 72
P-
Mclntyre, N.E. 1995. Effects of forest patch size on avian diversity. Landscape Ecology 10:85-
99.
Messer, J.J., R.A. Linthurst, and W.S. Overton. 1991. An EPA program for monitoring
ecological status and trends. Environ. Man. 17:67-78.
Milne, B. 1988. Measuring the fractal geometry of landscapes. Appl. Math, and Comput.
27:67-79.
Mitasova, H., J. Hofierka, M. Zlocha, and L.R. Iverson. 1996. Modeling topographic potential
for erosion and deposition using GIS. Int. J. Geogr. Inform. Sys. 10:629-641.
NRC (National Research Concil). 1997. Building a foundation for sound environmental
decisions. National Academy Press, Washington, B.C.
Omernik, J.M. 1987. Ecoregions of the United States: Map at a scale of 1:7,500,000. Suppl.
Annals American Assoc. Geogr. 77(1).
O'Neill, R. V. 1979. Natural variability as a source of error in model predictions, pp. 23-32 IN
G. S. Innis and R. V. O'Neill, Systems Analysis of Ecosystems. International Cooperative
Publishing House, Fairland, Maryland.
O'Neill, R. V., R. H. Gardner, F. O. Hoffman, and G. Schwarz. 1981. Parameter uncertainty and
estimated radiological dose to man from atmospheric 1311 releases: A Monte Carlo approach.
Health Phys. 40:760-764.
O'Neill, R. V., R. H. Gardner, L. W. Barnthouse, G. W. Suter, S. G. Hildebrand, and C. W.
Gehrs. 1982. Ecosystem risk analysis: A new methodology. J. Environ. Toxicol. Chem.
1:167-177.
O'Neill, R.V., D.L. DeAngelis, J.B. Waide, and T.F.H. Allen. 1986. A hierarchical concept of
ecosystems. Princeton University Press, Princeton, NJ.
36
-------
O'Neill, R.V., J.R. Krummel, R.H. Gardner, G. Sugihara, B. Jackson, D.L. DeAngelis, B.T.
Milne, M.G. Turner, B. Zygmunt, S.W. Chistensen, V.H. Dale, and R.L. Graham. 1988.
Indices of landscape pattern. Landscape Ecol. 1:153-162.
O'Neill, R V., S. J. Turner, V. I. Cullinen, D. P. Coffin, T. Cook, W. Conley, J. Brunt, J.
M.Thomas, M. R. Conley, and J. Gosz. 1991. Multiple landscape scales: an intersite
comparison. Landscape Ecology 5:137-144.
O'Neill, R.V., K.B. Jones, K.H. Riitters, J.D. Wickham, and I.A. Goodman. 1994. Landscape
monitoring and assessment research plan. PEA-620/R-94-009. 53 pp.
O Neill, R.V., C.T. Hunsaker, S.P. Timmins, B.L. Jackson, K.B. Jones, K.H. Riitters, and
J.D.Wickham. 1996. Scale problems in reporting landscape pattern at the regional scale.
Landscape Ecology 11:169-180.
O'Neill, R.V., C.T. Hunsaker, K.B. Jones, K.H. Riitters, J.D. Wickham, P. Schwarz,
I.A.Goodman, B. Jackson, and W.S. Baillargeon. 1997. Monitoring environmental quality at
the landscape scale. BioScience 47:513-520.
Overton, W.S., D. White, and D.L. Stevens. 1990. Design report for EMAP (Environmental
Monitoring and Assessment Program). EPA/600/S3-91/051.
Peterjohn, W. T., and D. L. Correll 1984. Nutrient dynamics in an agricultural watershed:
observations on the role of a riparian forest. Ecology 65:1466-1475.
Plotnick, R.E., R. H. Gardner, and R. V. O'Neill. 1993. Lacunarity indices as measures of
landscape texture. Landscape Ecology 8:201-212.
Poff, N.L, and J.D. Allan. 1995. Functional organization of stream fish assemblages in relation
to hydrological variables. Ecology 76:606-627.
Rastetter, E. B., A. W. King, B. J. Cosby, G. M. Hornberger, R. V. O'Neill, and J. E. Hobbie.
1991. Aggregating fine-scale ecological knowledge to model coarser-scale attributes of
ecosystems. Ecological Applications 2:55-70.
Riitters, K.H, RV. O'Neill, C.T. Hunsaker, J.D. Wickham, D.H. Yankee, S.P. Timmins, K.B.
Jones, and B.L. Jackson. 1995. A factor analysis of landscape pattern and structure metrics.
Landscape Ecol. 10:23-29.
Riitters, K.H., R.V. O'Neill, and K.B. Jones. 1997. Assessing wildlife habitat at landscape
scales: a multi-scaled approach. Biological Conservation 81:191-202.
Risser, P. G. 1987. Landscape ecology: state of the art. Pp. 3-14. IN M. G. Turner (ed.)
Landscape Heterogeneity and Disturbance. Springer-Verlag, NY.
Roth, N.E., J.D. Allan, and D.L. Erickson. 1996. Landscape influences on stream biotic integrity
assessed at multiple scales. Landscape Ecology 11:141-156.
37
-------
Running, S. W., R. R. Nemani, D. L. Peterson, L. E. Band, D. F. Potts, L. L. Pierce, and M. A.
Spanner. 1989. Mapping regional forest evapotranspiration and photosynthesis by coupling
satellite data with ecosystem simulation. Ecology 70:1090-1101.
Schlosser, I.J. 1990. Environmental variation, life history attributes, and community structure
instream fishes: implications for environmental management and assessment. Environ. Man.
14:621-628.
Schumaker, N.H. 1996. Using landscape indices to predict habitat connectivity. Ecology
77:1210-1225.
Sklar, F. H., and R. Costanza. 1991. The development of dynamic spatial model in landscape
ecology: a review and prognosis. In M. G. Turner and R. H. Gardner (eds.) Quantitative
Methods in Landscape Ecology. Springer-Verlag, NY.
Soranno, P.A., S. L. Hubler, and S.R. Carpenter. 1996. Phosphorus loads to surface waters: a
simple model to account for spatial pattern of land use. Ecol. Appl. 6:865-878.
Sparks, R.E. 1995. Need for ecosystem management of large rivers and their floodplains.
BioScience 45:168-182.
Storck, P., L. Bowling, P. Wetherbee, and D. Lettenmaier. 1998. Application of a GIS-based
distributed hydrology model for prediction of forest harvest effects on peak stream flow in
the Pacific Northwest. Hydrol. Processes. 12:889-904.
Turner, M. G. 1989. Landscape ecology: the effect of pattern on process. Annu. Rev. Ecol. Syst,
20, 171-97.
Turner, S. J., R. V. O'Neill, W. Conley, M. R. Conley, and H. C. Humphries. 1991. Pattern and
Scale: statistics for landscape ecology. Pp. 17-49. In M. G. Turner and R. H. Gardner (eds.)
Quantitative Methods in Landscape Ecology. Springer-Verlag, NY.
U.S. EPA. 1994. Landscape monitoring and assessment research plan. U.S. EPA 620/R-94/009,
Office of Research and Development, Washington, B.C.
Walker, J., F. Bullen, and E.G. Williams. 1993. Ecohydrological changes in the Murray-Darling
Basin. I. The number of trees cleared over two centuries. J. Appl. Ecol. 30:265-273.
Weller, M.C., M.C. Watson, and D. Wang. 1996. Role of wetlands in reducing phosphorus
loading to surface water in eight watersheds in the Lake Champlain Basin. Environ. Man.
20:731-739.
Wickham, J.D., and K. H. Riitters. 1995. Sensitivity of landscape metrics to pixel size. Int.
J..Remote Sensing 16:3585-3594.
Wickham, J.D., RV. O'Neill, K.H. Riitters, T.G. Wade, and K.B. Jones. 1997a. Sensitivity of
landscape pattern metrics to land cover misclassification and environmental condition
gradients. Photogram-metric Engineering and Remote Sensing 63:397-402.
Wickham, J.D., J. Wu, and D.F. Bradford. 1997b. A conceptual framework for selecting and
analyzing stressor data to study species richness at large spatial scales. Environ. Man.
21:247-257.
38
-------
Wickham, J.D., R.V. O'Neill, and K.B. Jones. 1999a. A geography of ecosystem vulnerability.
Landscape Ecology, submitted.
Wickham, J.D., K.B. Jones, K.H. Riitters, T.G. Wade, and R.V. O'Neill. 1999b. Transitions in
forest fragmentation: implications for restoration opportunities at regional scales. Landscape
Ecology 14:137-145.
Wickham, J.D., K.B. Jones, K.H. Riitters, R.V. O'Neill, R.D. Tankersley, E.R. Smith, A.C.
Neale, and D.J. Chaloud. 1999c. An integrated environmental assessment of the US Mid-
Atlantic Region. Environ. Manag., in press.
Wiens, J.A. 1985. Vertebrate responses to environmental patchiness in arid and semi-arid
ecosystems. Pp. 169-193, in Pickett, S.T.A., and P.S. White (eds.), The ecology of natural
disturbance and patch dynamics. Academic Press, New York.
Wiens, J.A., N. C. Stenseth, B. Van Home, and R.A. Ims. 1993. Ecological mechanisms and
landscape ecology. Oikos 66:369-380.
Wiens, J.A., R.L. Schooley, and R.D. Weeks, Jr. 1997. Patchy landscapes and animal
movements: do beetles percolate? Oikos 78:257-264.
Wischmeier, W. H., and D. D. Smith. 1978. Predicting rainfall erosion loss: A guide to
conservation planning. Agricultural handbook 537. U. S. Department of Agriculture,
Washington, B.C.
Young, W.J., P.M. Marston, and J.R. Davis. 1996. Nutrient exports and land use in Australian
catchments. J. Environ. Man. 47:165-183.
Zhang, L., A.L. O'Neill, and S. Lacey. 1996. Modelling approaches to the prediction of soil
erosion in catchments. Environ. Software 11:123-133.
39
-------
i -f
;"." i
by
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Mid-Atlantic Region
Replacement of Poor Quality NALC Imagery
NALC Land Cover Classification
Change Detection Across Sensors
Landscape Indicator/Stream Correlations
Landscape Indicators of Pesticides and Toxics
New Remote Sensing Methods to Evaluate
Surface Mining Stress
New Remote Sensing Methods to Evaluate
Hog and Chicken Farms
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Landscape Indicator/Estuary Correlations
National Landscape Sampling Design Pilot
40+ Watersheds in Mid- Atlantic
Region
Entire Mid- Atlantic Region
Selected Watersheds in Mid- Atlantic
Region
600 Watersheds in Mid- Atlantic
Region
50 Watersheds in Mid- Atlantic Region
Mining Gradient in Mid- Atlantic
Region
Agricultural Gradient in Neuse River
Watershed
Urban to Rural Gradient in Mid-
Atlantic Region
20+ Large Watersheds in Mid-Atlantic
Region
5 Different Ecoregions in Mid-
Atlantic Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Mid-Atlantic Region, Continued
Landscape Change Consequences - Proof of
Concept (1970- 1990)
Watershed Classification to Characterize
Responses of Streams to Landscape
Characteristics
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Entire Mid- Atlantic Region
Entire Mid- Atlantic Region
Selected Watersheds in Mid- Atlantic
Region
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Northeastern Region
NALC and Thematic Mapper Land Cover
Classification
Change Detection Across Sensors
Landscape Indicator/Stream Correlations
Landscape Indicator/Estuary Correlations
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
New York City Watersheds
New York City Watersheds
New Jersey; New York City
Watersheds
New England Coast
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Southeastern Region
Landscape Indicator/Stream Correlations
Landscape Indicator/Stream Correlations
Landscape Indicator/Stream Correlations
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Landscape Indicator/Estuary Correlations
Validation of Remote Sensing Tech to Assess
Impervious Surface
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Savannah River Watershed
Alabama Watersheds
SAMAB Region
South Carolina
Alabama, Florida, South Carolina
SAMAB Region
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Mid-West Region
Landscape Indicator/Stream Correlations1
Landscape Indicators of Pesticides and Toxics
Replacement of Poor Quality NALC Imagery
NALC Land Cover Classification
Change Detection Across Sensors
Kansas, Nebraska, Missouri
Kansas, Nebraska
Kansas, Nebraska, Missouri
Kansas, Nebraska, Missouri
Selected Areas of Kansas, Nebraska,
Missouri
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Mid-West Region, Continued
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Nebraska, Kansas
Entire Region
Entire Region
X
X
X
X
X
X
X
Upper Mid-West Region
Landscape Indicator/Stream Correlations
Replace Poor Quality NALC Imagery2
NALC Land Cover Classification
Change Detection Across Sensors
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Landscape Indicators of Pesticides and Toxics
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Minnesota, Wisconsin
All Watersheds Feeding into the Great
Lakes
All Watersheds Feeding into the Great
Lakes
St. Claire Watershed (WS/MN)
Selected Watersheds in Wisconsin,
Illinois and Indiana
Selected Watersheds in Wisconsin,
Illinois and Indiana
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Pacific Northwest Region
Landscape Indicators/Stream Correlations3
Watershed Classification to Characterize
Responses of Streams to Landscape
Characteristics
Watershed Classification to Characterize
Stream Loadings Risk
Landscape Indicator/Estuary Correlations4
Landscape Indicators of Pesticides and Toxics
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Additional Landscape Indicator Studies (if
needed)
Validation of Remote Sensing Tech to Assess
Impervious Surfaces
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Northwest Oregon, Willamuth Basin,
Deschutes Basin
Columbia River Basin
Columbia River Basin
Selected Watersheds in PNW
Selected Watersheds in Columbia
River Basin
Selected Watersheds in the Columbia
River Basin
To Be Determined
Selected areas within Region
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Pacific Southwest Region
Landscape Indicator/Stream Correlations3
Landscape Indicators of Pesticides and Toxics
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Southern LA Basin
Central Valley, other areas in
California
Selected areas in California
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Pacific Southwest Region, Continued
Landscape Indicator/Estuary Correlations
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Coastal areas of California
Entire Region
Entire Region
X
X
X
X
X
X
X
X
Northern Great Plains Region
Landscape Indicator/Stream Correlations3
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Selected Watersheds in Upper
Missouri Basin
Selected Watersheds in the Upper
Missouri Basin
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
Inter-Mountain Region
Landscape Indicator/Stream Correlations1
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Watershed Classification to Characterize
Responses of Streams to Landscape
Characteristics
New Remote Sensing Method to Detect
Grazing Pressure
Humbolt River Basin
Selected areas of Nevada and Utah
Nevada, Utah
Nevada, Utah
X
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Inter-Mountain Region, Continued
Validation of Remote Sensing Method to
Detect Mining Stress
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Nevada, Utah
Entire Region
Entire Region
X
X
X
X
X
X
X
X
Southwestern Region
Change Detection Across Sensors
Hydrologic Model Development to Assess
Consequences of Landscape Change on
permanent and intermittent streams
New Remote Sensing Method to Detect
Grazing Pressure
Landscape Indicator/Stream Correlations
Additional Landscape Indicator Studies (if
needed)
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Watershed Classification to Characterize
Responses of Streams to Landscape
Characteristics
New Remote Sensing to Detect Riparian
Habitat and Hydrologic Conditions
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Upper San Pedro River Basin
Selected Watersheds in the Lower
Colorado River Basin
Grazing Pressure Gradients in
California, Nevada, Arizona, and Utah
Selected Watersheds in Arizona and
New Mexico
To Be Determined
Selected areas of California, Arizona
Arizona, New Mexico
Selected Watersheds in Arizona, New
Mexico, and Utah
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
Southwestern Region, Continued
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Entire Region
X
X
Rocky Mountain Region
Landscape Indicator/Stream Correlations3
Additional Landscape Indicator Studies (if
needed)
New Remote Sensing Approaches to Detect
Surface Mining Stress
New Remote Sensing Approaches to Detect
Grazing Stress
Watershed Classification to Characterize
Responses of Streams to Landscape
Characteristics
Landscape Assessment Approach to Target
Water Bodies at Risk to Loadings
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
Southwestern Portion of Rocky Mtns.
To Be Determined
Selected areas in Colorado
Selected gradients in Colorado,
Wyoming, and Montana
Colorado, Wyoming
Selected Areas on the Colorado
Plateau
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
South Central Region
Change Detection Across Sensors
Landscape Indicator/Stream Correlations
Landscape Indicator/Estuary Correlations
Tensas River Basin
Selected Watersheds in Texas
Watersheds Feeding into Louisiana
and Texas Estuaries
X
X
X
X
X
X
X
-------
Appendix I. Continued
Research Area or Activity
Geographic Area(s)
FY99
FYOO
FY01
FY02
FY03
FY04
FY05
FY06
FY07
FY08
South Central Region, Continued
Additional Landscape Indicator Studies - If
Needed
Validation of Remote Sensing Approach to
Detect Riparian Habitat Condition
Acquisition of All Primary Spatial Data
Needed to Generate Landscape Indicators
Across the Three Dates
Analysis of Landscape Change and Its
Consequences on Aquatic Resources
To be Determined
Selected Watersheds in Louisiana and
Texas
Entire Region
Entire Region
X
X
X
X
X
X
X
X
X
X
X
National Assessment
Compilation of Findings from Regional
Studies
Publish National Assessment
Lower 48 States4
Lower 48 States4
X
X
X
1 Regional Environmental Monitoring and Assessment Program Project
2 Work will include the Canadian Watersheds Feeding into the Great Lakes
3 Western EMAP Landscape Project
4 If funds are available, may also include Alaska and Hawaii
-------
•"• •:'. . "." '.
The Program is integral to EPA goals and objectives as articulated in the NERL Strategic Research
Plan (EPA 1998c), the Strategic Plan for the Office of Research and Development (ORD Strategic Plan,
EPA 1997) and the ORD Ecological Research Strategy (EPA 1998a) which respond, in turn, to national
goals, e.g., the Governmental Performance Results Act (GPRA 1993) or to specific legislative mandates,
e.g., Amendments to the Clean Water Act (CWA). The following are examples of specific goals and
objectives which this Program directly addresses or supports.
ORD Strategic Plan (EPA 1997):
• To provide credible, state-of-the-art risk assessments, methods, models, and guidance
° Assessing the risks to ecosystems from non-chemical stressors (e.g., habitat loss and uvB
due to stratospheric ozone depletion)
° Developing and supporting the implementation of guidelines for assessing the ecological
impacts of environmental stressors
° Integrating scientific and technical information from ORD Laboratories and other
sources to provide a sound scientific base and technical support for Agency decisions
and policy
° Assuring adequate quality assurance for all research, testing, and applications
• To provide common-sense and cost-effective approaches for preventing and managing risks
° Developing diagnostic and characterization methods and protocols for use in determining
appropriate ecosystem restoration goals and requirements for specific sites, watersheds,
landscapes, and ecoregions
• To provide leadership ... in identifying emerging environmental issues, characterizing the
risks ... developing ways of preventing or reducing those risks
° Provide national and international leadership in risk assessment and its application for
risk reduction and risk management
° Conduct/sponsor workshops . . . reaching consensus on crucial research needs, and
defining the role of ORD and others in addressing those needs
• To develop scientifically sound approaches to assign and characterize risks to human health
and the environment
o Establish approaches to characterizing and understanding risks to ecosystems and, in
cooperation with other agencies, develop a national, multi-scale, integrated
environmental status and trends program
49
-------
o Characterizing national land cover/land use patterns and developing measures of
landscape condition at multiple scales for specific sites, watersheds, landscapes, and
ecoregions
ORD Ecological Research Strategy (EPA 1998a; Government Performance and Results Act of 1993).
• Goal 8 (Sound Science), Objective 1 — Provide the scientific understanding to measure, model,
maintain, or restore, at multiple scales, the integrity and sustainability of ecosystems now, and
in the future.
° Sl.l — By 2008, develop indicators, monitoring systems, and designs for measuring the
exposures of ecosystems to multiple stressors and the resultant response of ecosystems at
local, regional, and national scales.
° SI.2 — By 2008, develop models to understand, predict and assess the exposure and
response of ecosystems to multiple stressors at multiple scales.
• Develop a prototype multimedia, effects and exposure modeling framework for evaluating the
impact of watershed management practices, at multiple scales, on stream and estuarine
condition
• Develop advanced measurement, computing, modeling, and data management technologies,
and integrate them into an effective system for real-time delivery of multi-media, multi-scale,
multi-parameter information on environmental statue and risk
Clean Water Act Amendments (EPA 1998d):
• Provide improved, cost-effective mechanisms for the estimation of Total Mean Daily Loads
(TMDL's) of pollutants within watersheds
From the NERL's [draft, 8/28/98] Research Strategy (EPA 1998c), one of three strategic goals:
• "... facilitate environmental management and decision-making by providing the scientific
tools needed to estimate (and ultimately reduce) risks to ecosystems posed by exposures. . . .
Landscape, biological, and modeled indicators (used in an integrated manner at multiple
scales) can answer questions of importance for protecting ecosystems at the complex levels at
which ecological processes actually operate."
50
-------
11 f
F I- -
Web sites
° Remote sensing
° Landscape indicators
Quantification of landscape indicators in each major region (Reports & input to web)
Western region landscape assessments (western EMAP)
Quantification of landscape indicators in each major region (Reports & input to web)
Models to evaluate consequences of landscape change
° 3 to 4 regions of U.S.
National landscape assessment
Journal articles on indicator applications
Journal articles on new remote sensing application
National & regional databases on landscape @ 30-90 meter resolution
Remote sensing method to evaluate riparian habitat condition
Landscape indicator to evaluate vulnerability of water resources to pesticides & toxics
Models to evaluate vulnerability of water resources to pesticides & toxics
Landscape assessment methods to target high priority area for TMDL mitigation & 303 (d)
CWA designations
Remote sensing methods to evaluate surface/watershed impacts of surface mines
51
-------
," ""''
;• " .
AMI Advanced Monitoring Initiative
AVHRR Advanced Very High Resolution Radiometer
AVIRS Advanced Visible Infrared Imaging System
CENR Committee on the Environment and Natural Resources
CSIRO Commonwealth Scientific and Industrial Research Organization
CWA Clean Water Act
DOD Department of Defense
DOE/ORNL Department of Energy/Oak Ridge National Laboratory
EIMS Environmental Information Management System
EMAP Environmental Monitoring and Assessment Program
EPA Environmental Protection Agency
EPIC Environmental Photographic Interpretation Center
EROS Earth Resources Observation System
ESE Earth Science Enterprise
GCRP Global Change Research Program
GIS Geographic Information Systems
GPRA Government Performance Results Act
MODIS Moderate Resolution Imaging Spectrometer
MRLC Multi-resolution Land Characteristics Consortium
MSS Multi-spectral Scanner
NAWQA National Water Quality Assessment
NASA National Aeronautics and Space Administration
NCERQA National Center for Environmental Research and Quality Assurance
NERL National Exposure Research Laboratory
NHEERL National Health and Environmental Effect Research Laboratory
NOAA National Oceanic and Atmospheric Administration
52
-------
NRC National Research Council
NTM National Technical Means
OARM Office of Administration and Resources Management
ORD Office of Research and Development
OPPE Office of Policy, Planning and Evaluation
RARE Regional Applied Research Effort
REMAP Regional Environmental Monitoring and Assessment Program
ReVA Regional Vulnerability Assessment
SIR Shuttle Imaging Radar
STAR Science to Achieve Results
TIR Thermal Infrared
TM Thematic Mapper
TMDL Total Maximum Daily Loads
USDA-ARS United States Department of Agriculture-Agricultural Research Service
USFS United States Forest Service
USGS United States Geological Survey
USGS-BRD United States Geological Survey-Biological Research Division
USGS-WRD United States Geological Survey-Water Resources Division
UV Ultraviolet
53
-------
Table 1. Research Topic Areas and Goals
Research Topic Area
Change Detection
New Remote Sensing Methods
Spatial Data Acquisition and Accuracy
Assessment
Landscape Indicators of Water Resource
Vulnerability Assessing the
Consequences of Landscape Change
Multi-Indicator Assessment Approaches
Including Watershed Models
Purpose, Goal or Objective of Activity
Develop methods to monitor landscape change using
current remote sensing systems; perform change detection
analysis/interpretation; make findings public.
Capture new technologies for synoptic coverage and
enhanced resolution to support cross-scale analyses;
improve analysis of watershed-level stresses on aquatic
resources.
Exploit data sources and develop and implement methods
to document the accuracy of classified and other land
cover and land characteristics databases.
Develop quantitative linkages between a variety of existing
and new landscape indicators and aquatic resource
conditions.
Develop multi-indicator assessment techniques including
watershed models to prioritize vulnerable areas and
evaluate consequences of landscape change on water
resources.
54
-------
Table 2. MRLC Land Cover Classes
Class #
Class Name
Water
11
12
Open Water
Perennial Ice/Snow
Developed
21
22
23
Low Intensity Residential
High Intensity Residential
Commercial/Industrial/
Transportation
Barren
31
32
33
Bare Rock/Sand/Clay
Quarries/Strip Mines/Gravel Pits
Transitional
Vegetated; Natural Forested Upland
41
42
43
Deciduous Forest
Evergreen Forest
Mixed Forest
Class #
Class Name
Vegetated; Natural Shrubland
51
52
53
Deciduous Shrubland
Evergreen Shrubland
Mixed Shrubland
Vegetated; Non-Natural Woody
61
Orchards/Vineyards/Other
Herbaceous Upland
Natural/Semi-natural Vegetation
71
Grasslands/Herbaceous
Herbaceous Planted/Cultivated
81
82
83
84
85
Pasture/Hay
Row Crops
Small Grains
Fallow
Urban/Recreational Grasses
Wetlands
91
92
Woody
Emergent Herbaceous
55
-------
Table 3. New Remote Sensing Systems
Satellite
EOS AM-1
KOMSAT
CRSS-2
Landsat-7
IRS-P5, IRS-2A
CBERS 1,2
OrbView-3,4
IKONOS 2
NEMO
Warfighter
EO-1
ENVISTAT
IRS 1-D
QuickBird 1
Owners/Operators
Japan/U.S. Gov
Korean Gov
U.S. Commercial
U.S. Gov
Indian Gov
China/Brazil Gov
U.S. Commercial
U.S. Commercial
U.S. Gov
U.S. Gov
U.S. Gov
ESA
Indian Gov
U.S. Commercial
Proposed
Launch Date
1999
1998
1998, 1999
1999
1998, 1999
1998, 1999
1998, 1999
1999
2000
2000
1999
1999
1998
1999
Sensors
Multispectral
Panchromatic
Panchromatic
Panchromatic
Multispectral
Panchromatic
Multispectral
Hyperspectral
Panchromatic
Multispectral
Panchromatic
Hyperspectral
Panchromatic
Multispectral
Hyperspectral
Multispectral
Radar
Panchromatic
Multispectral
Panchromatic
Multispectral
Spatial
Resolution
(m)
15
10
10
15
30
1
4
8
1
4
5
30,60
1
4
8
30
30
10
20
1
4
Number of
Color Bands
14
7
200
4
210
4
80, 200
9
4
4
-------
Table 3. Continued
Satellite
QuickBird 2
Spot-5A
EOS PM-1
HRST
ARIES
EOS CHEM-1
EOS AM-2/Landsat-8
Spot-SB
Owners/Operators
U.S. Commercial
French Gov
U.S Gov
U.S Gov
Australia
U.S. Gov
U.S. Gov
French Gov
Proposed
Launch Date
1999
1999
2000
2000
2000
2002
2004
2004
Sensors
Panchromatic
Multispectral
Panchromatic
Multispectral
Hyperspectral
Hyperspectral
Panchromatic
Multispectral
Panchromatic
Multispectral
Spatial
Resolution
(m)
1
4
5
10
10
30
5
10
Number of
Color Bands
4
4
7
4
-------
Table 4. Landscape Sciences Program Responsibility Matrix
Responsibility
1 . Coordination of spatial data mosaic completion; MRLC;
NALC; complete data acquisition and processing
2. Landscape indicator development; indicator quantification1
3. New remote sensing systems capabilities; access; pilots
4. Exploitation of National Technical Means data/systems
5. Remote sensing technical support to EPA regions and program
offices
6. Change detection (NALC; and others)
7. Landscape indicator/model interface2
8. Land cover classification; operations and research and
development
9. Accuracy assessment of databases/land cover
10. Regional landscape assessments3
1 1 . Remote sensing support to research operations
12. Regional vulnerability assessments
LEB
CP
CP
CP
p
CP
p
CP
LCB
CP
CP
CP
p
CP
CP
EPIC
CP
CP
p
p
CP
s
CP
p
CP
Note: We anticipate some areas of overlap between the EPIC and the LCB in the area of
remote sensing research and development, but in order to maximize the use of our
currently small qualified research and development staff, it is appropriate and
necessary that some research and development be accomplished at both RTP and
Reston. This can best be accomplished with close communication between the
groups and in the spirit of coordination and cooperation.
Table Keys
P = Principal Role
S = Supporting Role
CP = Co-principal Role
1 We anticipate significant interaction and collaboration with EMAP Surface and Coastal Waters groups
2 We anticipate significant interaction and input from NERL-Athens
3 We anticipate significant interaction and collaboration with EPA Regional Offices
58
-------
Table 5. Existing and Potential Collaborators Within ORD
Collaborator
National Health and Environmental Effects
Research Laboratory (NHEERL)
NHEERL
National Risk Management Research
Laboratory
Global Climate Change Research Program
(OPPE)
National Center for Environmental Research
and Quality Assurance
NERL/Ecosystems Research Division-Athens
NERL/Atmospheric Modeling Division-RTP
NERL/Ecological Exposure Research
Division-Cincinnati
Project
Environmental Monitoring and Assessment
Program (EMAP)
Anthropogenic nitrogen [in streams]
Riparian Restoration Program
Global Change Research Program
Science to Achieve Results (STAR) Grants
Program
Modeling research
Climatological effects assessment
Biomarkers research
59
-------
Table 6. Ongoing Collaboration Organized by Topic
Topic
Remote sensing data acquisition and
classification
New Remote Sensing Methods
Development
Landscape characterization and
assessment research
Spatial data acquisition and accuracy
assessment
Change detection
Landscape indicators
Regional assessments
Regional to local scale assessments
Neuse River Basin data acquisition
Internal Partners
MRLC-OARM
AMI
STAR Grants Program -
NCERQA, EPA Regional
Offices
EPA Regional Offices
EMAP-NHEERL, EPA
Regional Offices
REMAP, RARE, EPA
Regional Offices
External Partners
USGS-EROS Data Center
NASA, USGS EROS Data
Center
University grantees
USGS-EROS Data Center,
NTM
NASA, USDA-ARS
DOE/ORNL, USGS-BRD,
USDA-ARS, CSIRO-Australia
USGS-WRD, USDA
State of North Carolina
60
-------
Table 7. List of Clients and Customers
Customer
EPA Regional Offices
EPA Offices of Water; Prevention, Pesticides and
Toxic Substances; Policy, Planning and
Evaluation; International Affairs
EPA Regional Offices and Office of Solid Waste
and Emergency Response
Academia
Public
Landscape Science Products Provided
Ecological assessments, e.g., Atlas; databases
Research and consultation; guidance documents
Remote sensing and aerial photography analyses
and reports for hazardous waste site evaluation
Databases for use in research
Data and analyses; consultation
61
-------
of
in of
i
r
Indicators
in and
Socio-Economic
Indicators
in
in
the Environment
Figure 1. General conceptual model of landscape change and sustailiability of environmental attributes values by society.
-------
Remote Sensing Theory
Wave
Length
Gamma
Rays
X-Rays
Ultra
Violet
0)
.a
'w
0)
Z
Medium and
Far Infrared
Microwave
VHFtoLF^
E E E E E E E
§3 1 s 5 2 S
Multispectral - Available
Number of Bands Several to Tens
Band Width - Wide Spectral
Resolution - Medium
Detects Solids and Liquids
Hyperspectral - Emerging
Number of Bands Hundreds
Band Width - Narrow Spectral
Resolution - High
Detects and Identifies
Solids and Liquids
Ultraspectral - R&D
Number of bands Thousands
Band Width - Very Narrow
Spectral Resolution - Very High
Detects and Identifies
Solids, Liquids, and Gases
Figure 2. Comparison of multispectral, hyperspectral, and ultraspectral signatures.
63
-------
CD
O
c
OJ
O
J)
•5
Of
0.4
0.2
0.0
i i i r
1
Iron Oxide and Hydroxide
(a) Hematite (GDS69,g)
(b) Goethite (WS222)
I I 1 i I I
10
15
20
25
Wavelength
(Mm)
Figure 3. Unique reflectance spectral of minerals as derived from NASA's AVIRIS instrument.
64
-------
Acquire and Process Primary
Landscape Data
Develop and Apply
"*" New Remote Sensing
Techniques and
Landscape Indicators
Determine Quantitative Relationships
Between Stream and Estuary and Landscape
Indicators (Within-Region, Watershed Studies)
i
Select Landscape Indicators to be
Used in the Models
Develop Multivariate, Landscape Indicator Models
Acquire and Assemble
Aquatic Resource Data
Test and Validate Models •*•
1
Select Subset of Aquatic Data
to Validate Models
Apply Landscape Models to Regional Data Sets (Wall-to-
I 1
1970s Landscape Conditions 1980s Landscape Conditions 1990s Landscape Conditions 2000s Landscape Conditions
Evaluate consequences of
Landscape Conditions
Evaluate consequences of
Landscape Conditions
Evaluate consequences of
Landscape Conditions
Evaluate consequences of
Landscape Conditions
Evaluate consequences of Landscape Change on Aquatic Resources -- 1970s to 2000s
Figure 4. General implementation strategy for landscape research and assessments proposed in the Plan.
The strategy will be implemented on a region by region basis.
------- |