UNITED STATES
ENVIRONMENTAL PROTECTION
AGENCY
OFFICE OF RESEARCH AND
DEVELOPMENT
WASHINGTON, D.C. 20460
EPA/600/R-01/018
FEBRUARY 2001
AN ECOLOGICAL ASSESSMENT OF INVASIVE AND AGGRESSIVE PLANT
SPECIES IN COASTAL WETLANDS OF THE LAURENTIAN GREAT LAKES:
A COMBINED FIELD-BASED AND REMOTE-SENSING APPROACH
RESEARCH PLAN
-------
AN ECOLOGICAL ASSESSMENT OF INVASIVE AND AGGRESSIVE PLANT
SPECIES IN COASTAL WETLANDS OF THE LAURENTIAN GREAT LAKES:
A COMBINED FIELD-BASED AND REMOTE-SENSING APPROACH
RESEARCH PLAN
Principal Investigators:
Ricardo D. Lopez
and
Curtis M. Edmonds
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
Cover Images: (Top - left to right) reproductive plant parts of Typha spp., Phragmites australis
(Cav.) Steudel, and Lythrum salicaria L. (Bottom) September 5, 1992 Landsat Thematic Mapper
(TM) image of St. Clair River Delta [Michigan, USA and Ontario, Canada]. Three TM images
(left to right) using natural color band combination (R=3, G=2, B=l), land-water definition band
combination (R=4, G=5, B=3), and band combination that increases the detection of moisture in
vegetation and soils (R=6, G=4, B=2). Phragmites australis and Typha spp. are abundant
throughout the coastal wetlands of the St. Clair River Delta.
Notice: This research plan has undergone a peer review process, involving input from experts
within the U.S. Environmental Protection Agency (EPA) and from outside the U.S. EPA.
Mention of trade names or commercial products does not constitute endorsement or
recommendation by EPA for use.
-------
TABLE OF CONTENTS
PROJECT DESCRIPTION 1
PROJECT OBJECTIVES 6
TECHNICAL APPROACH 10
PRODUCTS 14
SCHEDULE, MILESTONES 15
PROJECT BUDGET AND JUSTIFICATION 16
LITERATURE CITED 17
PROJECT MANAGEMENT 19
-------
PROJECT DESCRIPTION
Purpose
The aquatic plant communities within coastal wetlands of the Laurentian Great Lakes are
among the most biologically diverse and productive systems of the world. Coastal wetlands have
been especially impacted by landscape conversion and have undergone a marked decline in plant
community biological diversity in the past. The loss of biological diversity in coastal wetland
plant communities coincided with an increase in the presence and patch-dominance of invasive
(i.e., non-native and opportunistic) and aggressive (i.e., native and opportunistic) plant species.
The loss of biological diversity, by definition, may be the result of the increased presence of
invasive and aggressive plant species, and other ecosystem research suggests that such invasive
and aggressive plant species may be the result of general ecosystem stress in coastal wetlands
(see "Theoretical Basis of Project"). Thus, such losses of biological diversity in the plant
communities of Great Lakes coastal wetlands may be related to changes in the frequency of
landscape disturbance within a wetland or on the edges of wetlands (e.g., road fragmentation of
wetland ecosystems, conversion of wetland ecosystems to agriculture, or wetland hydrology
alterations). Little is known about such ecological relationships in the Great Lakes, especially at
the lake-basin scale. The purpose of this study is to examine some of the landscape-scale
ecological relationships by quantifying the extent and pattern of invasive/aggressive plant species
and testing for substantive relationships with local landscape disturbance in the past. Remote
sensing technologies may offer unique capabilities to measure the extent of these invasive and
aggressive species over a large area. Our approach is to use ground-based vegetation sampling to
calibrate remote sensing data, to develop spectral signatures of invasive/aggressive species that
may then be used to address the ecological vulnerability of coastal wetlands. This study will
focus on coastal wetlands along the coastal regions of Lake Michigan, Lake Huron, Lake St.
Claire, and Lake Erie that represent a full range of disturbance conditions in the lake basins, but
may also include coastal areas of the other Great Lakes (Figure 1). The outcome of this study will
help managers throughout the Great Lakes region target vulnerable coastal wetlands in need of
restoration or protection, an important component of improving the water quality and ecological
integrity of the Great Lakes Ecosystem. This project will also produce a method that could be
used by environmental managers to monitor the progress/success of wetland rehabilitation and
restoration projects where measures are taken to control or eradicate aggressive plant species.
Rationale/Need
This project is the first step needed to identify the extent of wetlands being stressed by
invasive and aggressive plant species throughout the Great Lakes basin. Depending on the sensor
of choice, the rate of invasion might be detectable utilizing the techniques developed during this
study. If this project is successful in creating a remote sensing protocol capable of identifying the
extent of these species' occurrence, this protocol could also be used for inland wetlands, as well
as other coastal wetland locations. This project is supported by the management of the U.S. EPA
Great Lakes National Program Office, the U.S. EPA Region 5 Wetlands group, and the U.S. EPA
Region 5 Critical Ecosystems Team. These offices have indicated that staff wetland scientists
and ecologists would be available to support the field component of this project.
-------
Both Great Lakes coastal wetlands and invasive species have been identified by managers
and scientists within the international Great Lakes basin as important indicators of the ecological
health of the Great Lakes basin. These indicators have been proposed at the international forum
of the State of the Lakes Ecosystem Conferences and are available at the following Internet web
sites: http://www.epa.gov/glnpo/solec/98 or http://www.cciw/ca/solec/intro.html.
E
300- 500m
200-OOOm
100-20001
0- 100m
Figure 1. Coastal wetlands for this study are along the coastal regions of Lake Michigan, Lake
Huron, Lake St. Claire, and Lake Erie, but may also include coastal areas of the other Great
Lakes (GLNPO 1995).
Theoretical Basis of Project
Landscape Ecology
Disturbance theory suggests that the intensity and duration of disturbance within an
ecosystem is a key factor in the loss of ecological integrity (Connell and Slatyer 1977, Rapport
1990, Keddy et al. 1993, Opdam et al. 1993). One of the potential mechanisms for the loss of
ecological integrity may be the decline in biological diversity of an ecosystem, through the
-------
invasion of opportunistic species (Odum 1985). The loss of plant biological diversity in coastal
wetlands of the Laurentian Great Lakes has been widely described as a result of increased
dominance of opportunistic plant species (e.g., Stuckey 1989).
Although not the only invasive and aggressive plant species present in the Great Lakes,
the three opportunistic emergent plant species to be studied (Lythrum salicaria L., Phragmites
australis (Cav.) Steudel, and Typha spp) are frequently observed throughout many of the Great
Lakes coastal wetlands. Lythrum salicaria (Figure 2) is a flowering perennial dicot that was
imported from Europe in the early 1800's for its colorful flowers (Voss 1985). Phragmites
australis (Figure 3) and Typha spp. (Figure 4) are flowing perennial monocots that are both
native to North America and often dominate the vegetation of coastal wetlands. Although there
are three species of Typha encountered throughout the Great Lakes, for the purposes of this study
they are similar enough to analyze as a single taxa. The three cattail species typically encountered
in the Great lakes are Typha latifolia L., Typha angustifolia L., and a hybrid of the two species
Typha x glauca Godron (Voss 1972). Lythrum salicaria reproduce by seed and stolons;
Phragmites australis reproduce by rhizome or stolon but also produce copious amounts of seed
that is predominantly sterile (Voss 1972); Typha spp. reproduce by rhizomatous growth and
massive seed production (ca. 200,000 per spike). Typha growth in a single season from a single
seed has been reported to produce a rhizome system 3 meters in diameter, with a total of 100
shoots (Voss 1972).
Figure 2. Typical patch of Lythrum salicaria in a coastal marsh (photo: The
Nature Conservancy).
-------
Figure 3. Typical patch of Phragmites australis in a coastal marsh.
Figure 4. Typical patch of Typha spp. in a coastal marsh.
-------
In general, Lythrum salicaria, Phragmites australis, and Typha spp. are resource
generalists that have life histories and physiological characteristics that enable them to rapidly
invade new areas and flourish under environmentally stressful conditions, where other plant
species can not. Lythrum salicaria, Phragmites australis, and Typha spp. are similar in that they
are often found to dominate wetland plant communities. Therefore, the landscape ecology
questions posed (in the Project Objectives section) will be tested in wetlands that are dominated
by large patches of each of the three species, keeping in mind that there are distinct biological
and ecological differences between the species. Expansion of the three plant species in coastal
wetlands may be a benchmark for observing the potential effects of landscape disturbance on the
biological diversity of wetlands because each species form large, relatively homogeneous patches
that typically reach sizes in the range of 1 to 50 hectares. The expanding populations within the
patches observed in the Great Lakes may be the result of increased opportunities for the
migration of individuals (or genets) from small initial populations to newly opened gaps in the
landscape.
Many of the opportunities for the expansion of these plant populations into landscape
gaps may be the result of increased land-cover fragmentation (Forman 1995), creating small
pockets of suitable invasive/aggressive plant habitat (e.g., roadside clearings, disturbed water
courses, and eroded shorelines). The patch dynamics (i.e., either increases or decreases in extent)
of these plant species in disturbed Great Lakes coastal areas may therefore be facilitated by the
extent of patch disturbance in the vicinity of coastal wetlands. Attempts to control these plant
species in Great Lakes coastal marshes (e.g., during the 1970's, 1980's, and 1990's) may have
bolstered the population resiliency by increasing the proportion of 'management resistant'
characteristics in the plant populations (Diamond 1974), acting to further select for invasive
genotypes at managed wetland sites.
Some specific studies of reed patches in other regions support the patch disturbance
hypothesis, but indicate that the level of disturbance may be an important factor in the process.
Die-back of Phragmites in relatively undisturbed temperate European regions, and expansion of
Phragmites in European areas of climatic extremes (van der Putten 1997) suggest that periodic
disturbance may increase the rate and extent of expansion, such as has occurred in some coastal
areas of the Laurentian Great Lakes. Periodic stress may actually allow for the formation of
relatively small, interconnected metapopulations, where gene flow between patches maintains the
genetic diversity that might otherwise decline (i.e., in relatively large inbred populations). When
such populations become unable to bridge the gaps between populations, at the advanced stages
of patch isolation, entire populations may become locally extinct (Opdam 1990).
Landscape Change Detection
The inherent complexity of the numerous biotic and abiotic variables involved in the
process of landscape change has prompted research and development of specialized sensors and
processing techniques. These sensors and techniques have been necessary to gather sufficient
data to determine ecological relationships over large areas of the landscape. There are also many
examples of the use of historical maps and aerial photographs to assess patch fragmentation and
ecological characteristics of the landscape (e.g., MacArthur and Wilson 1967, Howard 1970).
Airborne data is also useful for determining the ecological effects of abiotic environmental
factors (e.g., Lyon and Drobney 1984, Lyon and Greene 1992). Recently, ecological studies have
5
-------
successfully utilized data from more sophisticated sensors (e.g., Color Video, Landsat
MultiSpectral Sensor and Thematic Mapper), automated image processing software/techniques,
and geographic information systems. These tools have been used to describe better the identity of
materials within the landscape, and to describe landscape change (Jones et al. 1997, Heggem et
al. 1999). However, most of these tools are new and the application of these sensors and
techniques to wetland detection and analysis has not been fully explored.
The spectral and spatial detection characteristics of the 5 sensors to be used in this study
are listed in Table 1. Sensor comparisons are generalized to allow for comparison between
sensors, and to highlight the information utilization trade-offs between sensors. For example,
among the sensors utilized in this study, the imaging spectrometer (AVIRIS) offers the greatest
spectral resolution and the Color Video data offers the greatest spatial resolution. The range of
spectral and spatial resolution among sensors illustrates a fundamental trade-off when
considering sensor design and use. The trade-offs between sensor capabilities is the basis for
ongoing research to determine the ideal combination of spatial, spectral, and radiometric
characteristics for each application of remote sensing technologies (Schott 1997). The data types
utilized in this study have been used to answer many landscape ecology questions, however
cross-sensor applications of these sensors have not been widely used for wetland ecosystem
detection or for landscape disturbance research, especially at the lake basin scale.
Table 1. Comparison of Some Relative Sensor
Characteristics (as utilized for this study)
Passive Sensor Type Spectral Resolution Spatial Resolution
CV Coarse Fine
AVIRIS Fine Fine to Moderate
MSS Coarse Coarse
TM Coarse Moderate
*RADAR is an active sensor that can transmit and receive signals through clouds, haze, smoke,
and darkness, and obtain high quality images of the Earth in all weather at any time.
PROJECT OBJECTIVES
Three fundamental wetland detection questions will be addressed by the study:
1. Can airborne color video (CV) data (Figure 5), airborne visible infrared imaging spectrometer
(AVIRIS) data (Figure 6) or other hyperspectral data, satellite thematic mapper (TM) data
(Figure 7), radio detection and ranging (RADAR) data (Figure 8), satellite multispectral
scanner (MSS) data (Figure 8), or other selected remote sensing data (e.g., MODIS, ASTER,
or CASI) be used to accurately identify patches of purple loosestrife (Lythrum salicaria),
giant reed grass (Phragmites australis), or cattails (Typha spp.) in the coastal wetlands of the
Great Lakes?
-------
2. Which sensor(s) is(are) most effective for identification of vegetation patches of the above
species?
3. What is the detection accuracy of each sensor (or sensor combination), using ground-based
plant community data as truth?
Four fundamental landscape ecology questions will be addressed by the study:
1. What is the range of landscape stress among coastal wetland study sites, as evidenced by
historical land-cover change in the wetland vicinity (early 1970's - 1990's)?
2. What are the ecological relationships between landscape stressor(s) and the extent of the
invasive/aggressive plant species in the coastal wetlands of the Great Lakes?
3. What Great Lakes coastal wetlands are the most threatened by invasive/aggressive plant
species and what is the areal extent of Lythrum salicaria, Phragmites australis, and Typha
spp. at these sites?
4. Can this project serve as the baseline for measuring the rate of expansion or die-back (i.e.,
increasing or decreasing patch size) of invasive/aggressive plant species?
Figure 5. Airborne color video (CV) data of coastal marsh, and upland woods.
-------
Figure 6. Airborne visible infrared imaging spectrometer (AVIRIS) data of a coastal area,
including marsh, agricultural, and urban areas.
Figure 7. Landsat thematic mapper (TM) data of a coastal area, including marsh,
wooded, agricultural, and urban areas.
-------
Figure 8. Radio detection and ranging (RADAR) data of a coastal area, including marsh
and developed areas.
Figure 9. Satellite multispectral scanner (MSS) data including marsh, wooded,
agricultural, and urban areas.
-------
TECHNICAL APPROACH
Overview
The approach of this study is to A) determine which, if any, of the three study-species are
identifiable using remote sensing data, and by which sensor(s); B) identify coastal wetlands (per
Cowardin et al. 1979) in the Great Lakes that are dominated by each of the invasive or aggressive
plant species; C) establish a group of 'gradient test sites' that range from high to low dominance
of each invasive or aggressive species, using a stratified random sampling method; D) determine
how historical land-cover change among the 'gradient test sites' is related to the presence and the
spatial extent of existing invasive or aggressive species patches. Note that, depending upon the
outcome of the preliminary sensor/species studies, all of the species and sensors may not be
required to complete step D.
Calibration of the remote sensing data will be accomplished by mapping the vegetation in
coastal wetland plant communities that are known a priori to contain patches of either purple
loosestrife, giant reed grass, or cattails. During vegetation sampling the field team will use a
Global Positioning System (GPS) to delimit the boundary of invasive/aggressive species
patch(es) within the wetland, and sample fixed points within the core of the patch(es) and on the
perimeter using a hand-held spectroradiometer and GPS. These field data will be used in the
laboratory to calibrate the AVIRIS (or other hyperspectral) data from over-flights scheduled for
the same time period. Within the patches dominated by the above species (and on the periphery
of each patch), standard cover estimates, and stem density measurements will be performed using
a quadrat method (Mueller-Domboise and Ellenberg 1974, Barbour 1987). Field data will also be
used to test the efficacy of using CV, AVIRIS, TM, MSS, and/or RADAR (or other selected)
sensor data for identifying the patches of invasive and aggressive plant species in coastal
wetlands.
Sensor accuracy assessments from the study will provide information about which
sensor(s) are suitable to detect the invasive/aggressive species, and will be used to determine
which type(s) of remote sensing data will be most useful to complete the site analyses. Ground-
based sampling, maps, and aerial photographs will be used to accuracy-assess wetland
characteristics determined from remote sensing data.
A stratified, random sampling design will be used to select wetlands for verification of
the remote sensing methods for detection of invasive/aggressive plant species and to attain the
objective of assessing the extent and degree of invasion into coastal wetlands of the Great Lakes.
Each wetland will be assessed for the presence of large homogeneous stands of the three study
plant species during summer sampling using information from the sensor comparison portion of
the study. The field site selection and field work will be coordinated with wetland ecologists in
the local area.
North American Landscape Characterization (NALC) data from the 1970's to the 1990's
(i.e., Landsat Multispectral Scanner data), along with TM data collected during the 1990's and
2000's will be used to develop land-cover maps. The land-cover maps will be used to quantify 30
years of land-cover change in the vicinity of the coastal wetlands studied. Multivariate statistics
(e.g., multiple regression and multiple analyses of variance) will be used to test for substantive
relationships between land-cover change and the invasive/aggressive plant characteristics at
wetland study sites (i.e., all appropriately sampled wetland study sites).
10
-------
Project Tasks
1. Locate coastal wetlands along the shore of the Laurentian Great Lakes in coordination with
local wetland ecologists.
2. Use a stratified, random sampling protocol to determine the spatial extent of homogeneous
stands of Lythrum salicaria, Phragmites australis, and/or Typha spp. among those wetlands
selected in Task 1, and ground-survey sites
3. Develop and test efficacy of utilizing the selected sensors to locate the patches of wetland
plants delineated in Task 2.
4. Determine and map the spatio-temporal land-cover change (1970's - 1990's) adjacent to the
wetlands selected in Task 2
5. Identify any landscape-ecological relationships between invasive/aggressive plant patch
configuration and land-cover change that has occurred (1970's - 1990's) in the vicinity of
coastal wetland study sites
Non-Exclusive List of Data Utilized in the Study
Airborne
Wetland Inventory Maps (i.e., Ohio, Michigan, Wisconsin, or National Wetland Inventories).
These digital maps are currently available and will be used to locate and verify locations of
coastal emergent wetlands.
Aerial Photographs. These data will vary by region and historical availability. Color infra-red
photographs as stereo-pairs will be the ideal selection criteria but final selection will depend
upon availability. These data are available from a variety of historical photograph sources and
will be obtained from the appropriate government offices and distributors. The ideal range of
scales are from approximately 1:1000 to 1:24,000, however scales up to 1:40,000 may be used if
finer scales are unavailable.
CV. These data are available from Positive Systems Inc., Whitefish, Montana, and will be
acquired from low altitude fixed wing aircraft. Digital image data are available as three visible
bands (RGB) and 1 near infra-red band, with sub-meter spatial resolution capability. This system
is based on the True Color Digital Camera System, developed by the Landscape Ecology Branch
and the U.S. Department of Agriculture (Everitt et al. 1996).
AVIRIS. These data are available from NASA/JPL , Pasadena, California, and may be acquired
from mid-altitude fixed wing aircraft. Digital image data are available in 224 bands in the visible
and infra-red regions of the spectrum, with a spatial resolution range of 3.5 to 20 meters.
Hyperspectral Mapper (HyMap). These data are available from the HyVista Corp., North Ryde,
Australia and may be acquired from low altitude fixed wing aircraft. The sensor is a commercial
system with 126 spectral bands at approximately 15 nm spectral resolution and spatial
resolutions of 3-10 m. The sensor operates in a 3-axis gyro stabilized platform to minimize
image distortion. Geolocation and image geocoding is achieved with GPS and an integrated
inertial monitoring unit.
11
-------
Satellite
Landsat TM. These data are available from EROS Data Center, Souix Falls, SD, and are acquired
from earth orbit. Digital image data are available in 3 visible bands (Red, Green, and Blue), 1
near infra-red band, and 2 mid-infra-red bands with a spatial resolution of 30 meters. There is
also 1 thermal band with a spatial resolution of 120 meters.
Landsat MSS (e.g., NALC). These data are available from the joint EROS Data Center/LEB
North American Landscape Characterization (NALC) database that resides at the LEB facility.
These data were acquired from earth orbit and are processed to facilitate land-cover change
analyses (i.e., minimal cloud cover, co-registered decadal images). Digital image data are
available in 2 visible bands (Red and Green) and 2 near infra-red bands, and have an approximate
spatial resolution of 80 meters.
RADAR. These data will be obtained from the RADARS AT-1 orbiter. RADARSAT-1 is an
advanced Earth observation satellite project developed by Canada. RADARSAT-1 is equipped
with a Synthetic Aperture Radar (SAR). The SAR is a microwave instrument that can transmit
and receive signals through clouds, haze, smoke, and darkness, and obtain high quality images of
the Earth in all weather at any time. This provides an advantage over observation by aircraft and
optical satellites under sub-optimal weather conditions. Depending upon the operation mode,
RADAR data has a spatial resolution of 10 meters to 100 meters. Data from different sensors
may also be merged with RADAR bringing in the concept of multi-sensor data fusion. For
example, multi-spectral optical data can be combined with RADAR imagery. The optical data
(e.g., MSS) can provide spectral information for discriminating surface cover types and the
RADAR imagery can be used to detect landscape structural details (Jensen 1996, Howarth et al.
1997).
GPS. These data are available from uplink to global positioning satellites, using a Trimble Pro
XRS receiver/data-logger. Position data will be downloaded in the field to the GPS receiver/data-
logger, along with site identification information (GPS accuracy is sub-meter).
Field and Other Data Sources
Spectroradiometer. These data are available from field measurements using a FieldSpec Pro
Spectroradiometer, Analytical Spectral Devices, Inc., Boulder, CO. The Spectroradiometer will
be positioned above the plant canopy using a tripod system, and the position of the reading will
be logged automatically with the GPS (above).
Digital coverages. Digital map coverages will be obtained from a variety of sources and will
contain landscape data including (but not limited to): hydrology, soil chemistry, fertilizer
application, pesticide application, geology, topography, water chemistry, land cover, population,
roads, political boundaries, and biological characteristics.
12
-------
Gradient Analyses
A fundamental component of the study design is the selection of a sufficient number of
study sites to account for the full range of disturbance conditions (Green 1979, Karr and Chu
1997) that are present in coastal wetlands of the Great Lakes. An stratified, random sampling
design will be used to attain a gradient of potential wetland disturbance, ranging from least-
impacted sites to potentially impaired sites. An initial targeted sampling technique may be
necessary to initially stratify the gradient into coarse categories (e.g., regions that are likely to
have substantial land-cover change along coastlines and regions that are likely to have lesser
amounts of land-cover change along coastlines). Initial studies, prior knowledge of local experts,
existing data, and publication of prior work in Lake Michigan, Lake Huron, Lake St. Claire, and
Lake Erie indicate that coastal wetlands have sufficiently large stands of the study-species,
sufficient variability in patch size among sites, and sufficient variability of adjacent land-cover
characteristics to successfully perform the gradient analyses.
The land-cover gradient will be determined from the land-cover maps produced from
historical (1970's - 1990's) satellite imagery or other available land-cover data sets. Table 2 lists
the minimum Anderson (1976) "Level 1" land-cover class detail to be used for the analysis of the
land-cover gradients. Land-cover gradient(s) will be determined from the spatio-temporal
characteristics of land-cover change in the vicinity of each wetland site. Historical data can be
used to verify the duration of land-cover change in an area and to assess the rate of change since
the 1970's. The gradient(s) of land-cover change (the 'stressor' variable(s)) will then be
compared to the patch characteristics of the applicable plant species (i.e., the response variables).
The relationships between the stressor and response variables will be tested with multivariate
statistics (e.g., multiple regression or multiple analyses of variance) to determine ecologically
relevant and statistically significant relationships. The comparison of contemporary plant species
patch characteristics to temporally static land-cover change (i.e., variability among sites) will be
performed as well as an analysis of land-cover change from the 1970's to 1990's (e.g., rate of
change among sites). Each of these gradient analysis techniques solves the problem of having
only limited availability of plant species data at the field sites in the 1970's, 1980's, and the early
1990's.
Table 2. Anderson Land Cover Classes
Land-Cover Class Regional Examples
Urban or Built-Up Land Shopping center, parking areas, sub-urban area
Range Land Grassy 'old field'
Forest Land Forest
Water Pond, lake, stream
Agricultural Land Corn, soybean row crop
Barren Land Gravel pit, beach
Wetland Riverine wetland forest, coastal marsh
13
-------
Data Management
All of the study results, data sets (including metadata), and final project documents will be
made electronically available at the time of final report publication. These data will also be made
available on the Landscape Ecology Branch web site, and directly linked to the EMAP Web site.
These activities will comply with ORD guidance on federal data and metadata standards.
Quality Assurance/Quality Control
Accuracy of data and procedures is an important aspect of this research and efforts to ensure
the quality of the final products are incorporated into every step of the project. These quality
control procedures include verification of geographic coordinates and projection information on
map and imagery products. After data transfer from contributors outside the Landscape Ecology
Branch is completed, data will be recorded and routinely checked for accuracy and completeness.
Accuracy assessment of land-cover classifications will be performed at the Environmental
Photographic Interpretation Center in Reston, Virginia using high resolution aerial photographs
of the study areas and GPS data collected at field sites. A contingency table (i.e., a confusion
matrix) will be generated to describe errors of omission and errors of commission during
classification of the imagery. All field and laboratory equipment will be regularly calibrated per
the established laboratory schedule. This project plan has passed a peer review by experts in
remote sensing, geographic information systems, and landscape ecology, and a summary of their
comments and reconciliation of concerns is available from the Principal Investigators upon
request.
PRODUCTS
A. Project Report:
An Ecological Assessment of Invasive and Aggressive Plant Species in Coastal Wetlands of
the Laurentian Great Lakes: A Combined Field-Based and Remote-Sensing Approach
B. Refereed journal articles:
#1 Coastal wetlands detection in the Laurentian Great Lakes: A cross-sensor test
#2 Identification of invasive and aggressive plant species in coastal wetlands of the
Laurentian Great Lakes: A cross-sensor test
#3 The landscape-ecological relationships between land-cover change and invasive and
aggressive/invasive plant species in the coastal wetlands of the Laurentian Great Lakes
C. Images/data from all wetland study sites
D. Historical land-cover maps
14
-------
SCHEDULE, MILESTONES
Table 3. Project Milestones
Time Period Activities
2000 April - June Investigation of MSS (NALC) and TM scene availability and selection of
scenes
2000 July 13 Project proposal submitted for U.S. EPA and external peer review
2000 July - Sept Site reconnaissance with Region 5 scientists, GPS preliminary wetland sites,
preliminary vegetation field-sampling
2000 Oct. - Dec. Initial team member meetings to develop sampling protocols, site selection
2001 Jan. Research Plan published
2001 Feb. - April Order site maps and aerial photographs, order available satellite imagery,
schedule overflights, preliminary site reviews with team members, ensure
site access
2001 May - Sept. Field/remote-sensing data collection
2001 Oct. - Dec. Processing of summer 2001 remote sensing data (CV, AVIRIS, RADAR, or
other selected sensor data), analyses of vegetation data and study site
GPS/spectroradiometer data. Processing of NALC and TM data for land-
cover change analyses
2002 Jan. - April 2002 study site selection with team members, ensure site access, protocol
effectiveness team meetings
2002 May - Sept. Field/remote sensing data collection
2002 Oct. - Dec. Processing of summer 2002 remote sensing data, analyses of field data
2003 Jan. - June Sensor data validation assessment using ground-based and other data
2003 July - Dec. Analyses of patch characteristics, complete project report, develop and
prepare manuscripts for refereed journal articles
15
-------
PROJECT BUDGET AND JUSTIFICATION
Overview
The budgetary requirements to conduct this project are provided by the U.S. EPA Region
5 R-EMAP, with in-kind support from the Landscape Ecology Branch. The in-kind services
include selection and processing of the remote sensing data, GIS analysis of patch characteristics
of the wetlands, statistical analysis and ecological interpretation of the land-cover change data,
technical writing and editing, production of final maps, production of final images, and field
work that will be coordinated with local wetland ecologists.
Project Justification
Geographic and Remote Sensing Data and Services
Maps in digital or hard-copy format will be used to validate the location of wetland sites
and the identity of features in the landscape (e.g., roads) and will aid in the GPS mapping of the
sites. Digital map data will be used for producing relevant GIS coverages that, when combined
with the land-cover classifications, will aid in the description of the wetland site and the
surrounding landscape. Aerial photographs will be used for reconnaissance of potential field sites
before visiting site, while scouting a site, and during the land-cover classification portion of the
study. CV, MSS, TM, RADAR, and AVIRIS (or other selected) data will be used to locate
wetlands, determine boundaries of wetlands, and to measure patches of Lythrum salicaria,
Phragmites australis, and Typha spp. within wetlands of the Laurentian Great Lakes. MSS, TM,
and RADAR (or other selected sensor) data will also be used to produce land-cover maps for
change analyses.
Map Products
Land-cover maps from the 1970's through the 1990's will be used to develop land-cover
change images, which will subsequently be used to determine the relationships between plant
species patch configuration and local landscape change.
Field Sampling
Field measurement of the boundaries of each patch of plant species is necessary for
comparison to the remotely sensed data. Field measurements will primarily involve GPS
recording of the patch boundary, stem count sampling within the core area of the patch,
estimating percent cover in the core area of the patch, and spectrometry in the core area of the
patch. The spectrometry data are required to develop accurate spectral signatures from the
hyperspectral data.
16
-------
LITERATURE CITED
Anderson, J. R., E.E. Hardy, J.T. Roach, and R.E. Witmer, R. E. 1976. A Land Use and Land
Cover Classification System for Use with Remote Sensor Data (USGS Professional Paper 964).
U.S. Geological Survey, Reston, Virginia, USA..
Barbour, M.G., J.H. Burk, and W.D. Pitts. 1987. Terrestrial Plant Ecology.
Benjamin/Cummings, Menlo Park, California, USA.
Connell, J.H. and R.O. Slatyer. 1977. Mechanisms of succession in natural communities and
their role in community stability and organization. American Naturalist. 111(982):1119-1144.
Cowardin, L.M., V. Carter, E.G. Golet, and E.T. LaRoe. 1979. Classification of Wetlands and
Deepwater Habitats of the United States. U.S. Fish and Wildlife Service Publication FWS/OBS-
79/31. Washington, D.C., USA.
Diamond, J.M. 1974. Colonization of exploded volcanic islands by birds: The supertramp
strategy. Science. 184:803-806.
Everitt, J.H., D.E. Escobar, J.R. Noriega, M.R. Davis, and I. Cavazos. 1996. A True Digital
Imaging System For Remote Sensing Applications. U.S. Department of Agriculture, Weslaco,
Texas, USA.
Forman, R.T.T. 1995. Land Mosaics. Cambridge, New York, USA.
Gleason, H.A. and A. Cronquist. 1991. Manual of Vascular Plants of Northeastern United States
and Adjacent Canada. The New York Botanical Garden, Bronx, New York, USA.
Government of Canada and GLNPO (Great Lakes National Program Office). 1995. The Great
Lakes: An Environmental and Resource Atlas (3rd edition). Government of Canada
Toronto, Ontario, Canada and United States Environmental Protection Agency Great Lakes
National Program Office Chicago, Illinois, USA.
Green, R.H. 1979. Sampling Design and Statistical Methods for Environmental Biologists. J.
Wiley and Sons, New York, USA..
17
-------
Heggem, D.T., A.C. Neale, C.M. Edmonds, L.A. Bice, R.D. Van Remortel, and K.B. Jones.
1999. An Ecological Assessment of the Louisiana Tensas River Basin. EPA/600/R-99/016. U.S.
Environmental Protection Agency, Washington, D.C., USA.
Howard, J.A. 1970. Aerial Photo-Ecology. American Elsevier, New York, USA.
Howarth, P., J. Wang, J. Shang, and M.Y. Jollineau. 1997. Feasibility of Integrating Radar and
Optical Data for Wetland Mapping and Monitoring: a Case Study from Southern Ontario in
Proceeding of International Symposium: GEOMATICS IN THE ERA OF RADARS AT (May
25-30), Ottawa, Canada.
Jensen, J.R. 1996. Introductory Digital Image Processing. Prentice Hall, Upper Saddle River
New Jersey, USA.
Jones, K.B., K.H. Ritters, J.D. Wickham, R.D. Tankersley Jr., R.V. 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. U.S. Environmental Protection Agency,
Washington, D.C., USA.
Karr, J.R. and E.W. Chu. 1997. Biological Monitoring and Assessment: Using Multimetric
Indexes Effectively. EPA/235/R97/001. University of Washington, Seattle, Washington, USA.
Keddy, P.A., H.T. Lee, and 1C. Wisheu. 1993. Choosing indicators of ecosystem integrity:
Wetlands as a model system in (S. Woodley, J. Kay, and G. Francis eds.) Ecological Integrity
and the Management of Ecosystems. St. Lucie Press, Delray Beach, Florida, USA.
Lyon, J.G. and R.D. Drobney. 1984. Lake level effects as measured from aerial photos. Journal
of Surveying Engineering. 110(2): 103-111.
Lyon, J.G. and R.G. Greene. 1992. Use of aerial photographs to measure the historical areal
extent of Lake Erie coastal wetlands. Photogrammetric Engineering and Remote Sensing.
58:1355-1360.
MacArthur, R. and E.G. Wilson. 1967. The Theory of Island Bio geography. Princeton University
Press, Princeton, New Jersey, USA.
Mueller-Domboise, D. and H. Ellenberg. 1974. Aims and Methods of Vegetation Ecology. Wiley
and Sons, London, UK.
Odum, E.P. 1985. Trends expected in stressed ecosystems. Bioscience. 35(7):419-422.
Opdam, P. 1990. Understanding the ecology of populations in fragmented landscapes in Trans.
19th IUGB Congress. Trondheim, Norway.
18
-------
Opdam, P., R. V. Apeldoorn, A. Schotman, and J. Kalkhoven. 1993. Population responses to
landscape fragmentation in (C.C. Vos and P. Opdam eds.) Landscape Ecology of a Stressed
Environment. Chapman and Hall, London, UK.
Rapport, DJ. 1990. Challenges in the detection and diagnosis of pathological change in aquatic
ecosystems. Journal of Great Lakes Research. 16(4):609-618.
Schott, J.R. 1997. Remote Sensing: The Image Chain Approach. Oxford University, New York,
New York, USA.
Stuckey, R.L. 1989. Western Lake Erie Aquatic and Wetland Vascular-Plant Flora: Its Origin
and Change in Lake Erie Estuarine Systems: Issues, Resources, Status, and Management.
National Oceanographic and Atmospheric Administration, Washington, D.C., USA.
van der Putten, W.H. 1997. Die-back of Phrasmites australis in European wetlands: An
overview of the European Research Programme on Reed Die-back and Progression
(1993-1994). Aquatic Botany. 59:263-275.
Voss, E.G. 1972. Michigan Flora (Part I, Gymnosperms and Monocots). Cranbrook Institute of
Science and University of Michigan Herbarium, Ann Arbor, Michigan, USA.
Voss, E.G. 1985. Michigan Flora (Part II, Dicots, Saururaceae - Cornaceae). Cranbrook
Institute of Science and University of Michigan Herbarium, Ann Arbor, Michigan, USA.
PROJECT MANAGEMENT
The Landscape Ecology Branch
The Landscape Ecology Branch (LEB) is located on the campus of the University of
Nevada, Las Vegas and is under the U.S. EPA's Office of Research and Development,
Environmental Sciences Division. The Branch employs 32 professionals, including biologists,
research ecologists, engineers, statisticians, GIS specialists, remote sensing specialists, photo-
interpreters, data management specialists, graphic artists, and procurement specialists. The
Branch has robust hardware and software capabilities (e.g., ESRI-Arc/Info, ESRI-Arc View,
ERDAS-Imagine, RSI-ENVI) that enable the efficient processing and analysis of data. The
Branch employs 6 full-time Ph.D. researchers. The LEB possesses a modern laboratory and is
supported by the resources and facilities of the Environmental Sciences Division and the U.S.
EPA Environmental Photographic Interpretation Center in Reston, Virginia.
The LEB conducts research in the field of landscape ecology and related disciplines,
develops landscape assessment and characterization applications, and develops tools and
methods for solving regional environmental problems. Many of these projects involve the
analysis of ecosystem and watershed vulnerability to human-induced stresses. The primary goal
of the LEB is to develop tools for: 1) understanding and forecasting ecosystem trends; 2)
assessing the ability of an ecological resource to provide desired benefits; 3) anticipating
19
-------
emergency environmental problems; 4) monitoring and documenting progress in maintaining and
restoring ecosystems. LEB research includes developing ecologically-meaningful indicators of
landscape condition and trends related to endpoints of importance to the EPA; developing,
applying, and transferring tools for measurement, assessment, and prediction of the sustainability
and vulnerability of landscapes at multiple spatial and temporal scales; maintaining pace with
rapidly developing science and technology of remote sensing. Current LEB projects include
applied research in Arizona, Arkansas, California, Colorado, Delaware, Georgia, Idaho, Illinois,
Indiana, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Mississippi, Missouri, Nevada,
New Jersey, New Mexico, New York, North Carolina, North Dakota, Ohio, Oregon,
Pennsylvania, South Carolina, South Dakota, Tennessee, Vermont, Virginia, Washington, West
Virginia, and Wisconsin.
U.S. EPA Team Members and Contact Information
David W. Bolgrien, Ph.D.
U.S. EPA Environmental Effects Research Laboratory
Mid-Continent Ecology Division/ORD
6201 Congdon Boulevard
Duluth, MN 55804
tel (218) 529-5216
fax (218) 529-5003
bolgrien.david@epa.gov
Chad Cross, Ph.D.
Physical Scientist
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2148
fax (702) 798-2692
cross.chad@epa.gov
Donald Ebert
Ecologist
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2158
fax (702) 798-2692
ebert.donald @ epa.gov
20
-------
Curtis Edmonds
(Co-Principal Investigator)
Electronic Engineer
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2264
fax (702) 798-2692
edmonds.curtis @ epa.gov
Daniel Heggem
Research Environmental Scientist
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2278
fax (702) 798-2692
heggem.daniel@epa.gov
K. Bruce Jones, Ph.D.
Chief, Landscape Ecology Branch
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2671
fax (702) 798-2692
j ones .bruce @ epa. go v
Ricardo D. Lopez, Ph.D.
(Co-Principal Investigator)
Physical Scientist
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2394
fax (702) 798-2692
lopez.ricardo @ epa.gov
John G. Lyon, Ph.D.
Director, Environmental Sciences Division
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2525
fax (702) 798-2692
lyon.johng@epa.gov
21
-------
Anne C. Neale
(Project Officer)
Physical Scientist
U.S. EPA National Exposure Research Laboratory
Environmental Sciences Division/ORD
P.O. Box 93478
Las Vegas, NV 89193-3478
tel (702) 798-2347
fax (702) 798-2692
neale.anne@epa.gov
John Schneider, Ph.D.
(Great Lakes National Program Office)
G-17J, U.S. EPA REGION 5
77 West Jackson Boulevard
Chicago, IL 60604-3507
tel (312) 886-0880
Schneider .j ohn @ epa.gov
Terrence Slonecker
Research Environmental Scientist
U.S. EPA Environmental Photographic Interpretation
Center/ORD 555 National Center
12201 Sunrise Valley Drive Suite 2D-115
Reston, VA 20192
tel (703) 648-4289
fax (703) 648-4290
slonecker.t@epa.gov
David Williams
Research Physical Scientist
U.S. EPA Environmental Photographic Interpretation
Center/ORD 555 National Center
12201 Sunrise Valley Drive Suite 2D-115
Reston, VA 20192
tel (703) 648-4798
fax (703) 648-4290
williams.davidj@epa.gov
22
------- |