Metadata - U.S.A. Coastal Change Analysis Program (CCAP)
GEODATASET NAME: GLSCCAP
IDENTIFICATION INFORMATION
Description:
NOAA's Coastal Change Analysis Program (C-CAP) land cover
geodataset, clipped to the Great Lakes Basin study area
boundary.
Abstract:
C-CAP includes the classification of 2000 Landsat 7 data to
produce a land cover product intended to improve the
understanding of coastal uplands and wetlands, and their
linkages with the distribution, abundance, and health of
living marine resources.
Data Type:
Grid
Data Originator:
National Oceanic and Atmospheric Administration
Coastal Services Center
Charleston, SC
(843)740-1210
csc@csc.noaa.gov
Data Processors:
Rick Van Remortel & Ed Evanson
Lockheed Martin Environmental Services
1050 E. Flamingo Road, Suite E120
Las Vegas, NV 89119
(702)897-3295
rvanremo@lmepo.com
Data Provider:
Ricardo Lopez, Ph.D.
U.S. Environmental Protection Agency
National Exposure Research Laboratory
P.O. Box 93478
Las Vegas, NV 89193-3478
(702)798-2394
lopez.ricardo@epa.gov
Keywords:
coastal, land cover classification data, C-CAP, wetland
Version:
N/A
Status:
Interim
Revision Number:
0
Series Name:
Online Link (URL):
http://www.csc.noaa.gov/crs
Time Period of Content:
Use Constraints:
Data set is not for use in litigation. While efforts have been
made to ensure that these data are accurate and reliable within
the state of the art, NOAA, cannot assume liability for any
damages, or misrepresentations, caused by any inaccuracies in the
data, or as a result of the data to be used on a particular
system. NOAA makes no warranty, expressed or implied, nor does
the fact of distribution constitute such a warranty. Acknowledgment
of NOAA CSC would be appreciated in products derived from these data.
Purpose:
GLB data browser
Date of metadata entry/update:
11/18/2004
No Publication Information Available
No File Security Information Available
DATA QUALITY INFORMATION
Cloud Cover:
Not applicable
Software:
ESRI ArcGIS ArcInfo 8.2 Workstation
Operating System:
Microsoft WindowsXP
Path Name:
d:\solec\gds\glsccap
Logical Consistency Report:
Not presently available
Completeness Report:
Not presently available
Horizontal Positional Accuracy:
Not presently available
Vertical Positional Accuracy:
Not presently available
Attribute Accuracy:
Not presently available
Procedures:
Individual C-CAP datasets for the year 2000 were assembled by state
and clipped to the study area boundary. Substitution of NLCD data
(reclassed to corresponding C-CAP classes) was performed for a few
very small HUC-data gaps present for the states of WI, IL, IN, OH,
and NY. A change legend was prepared with summary statistics
documenting the substitution activity, and random adjacent subsample
areas were used to perform QA checks regarding the substitution.
Reviews Applied to Data
Lockheed Martin Environmental Services internal review
Related Spatial Data Files:
All geodatasets with gls prefix.
Other References Cited:
Dobson, J. et. al., NOAA Coastal Change Analysis Program(C-CAP):
Guidance for Regional Implementation, NOAA Technical Report
NMFS 123, U.S. Department of Commerce, April 1995.
Notes:
Update Frequency:
As needed
SPATIAL REFERENCE INFORMATION
Description of Grid glsccap
Cell Size = 30.000 Data Type: Integer
Number of Rows = 28143 Number of Values = 15
Number of Columns = 49972 Attribute Data (bytes) = 48
BOUNDARY STATISTICS
Xmin = 213015.000 Minimum Value = 2.000
Xmax = 1712175.000 Maximum Value = 19.000
Ymin = 1990395.000 Mean = 7.044
Ymax = 2834685.000 Standard Deviation = 3.878
COORDINATE SYSTEM DESCRIPTION
Projection ALBERS
Datum NAD83
Units METERS Spheroid GRS1980
Parameters:
1st standard parallel 29 30 0.000
2nd standard parallel 45 30 0.000
central meridian -96 0 0.000
latitude of projection's origin 23 0 0.000
false easting (meters) 0.00000
false northing (meters) 0.00000
ENTITY AND ATTRIBUTE INFORMATION
Annotation Name:
ATTRIBUTE LISTING FOR: glsccap.vat
COLUMN ITEM NAME WIDTH OUTPUT TYPE N.DEC ALTERNATE NAME INDEXED?
1 VALUE 4 10 B - Indexed
5 COUNT 4 10 B - -
9 CCAP_CLASS 40 40 C - -
METADATA REFERENCE SECTION
FGDC Content Standards for Digital Geospatial Metadata
FGDC Standards Version 6/98 / metadata.aml ver. 1.3 5/21/99
SUPPLEMENTAL METADATA (where available)
Record VALUE COUNT CCAP_CLASS
1 2 6471551 High_intensity_developed
2 3 14692433 Low_intensity_developed
3 4 85905980 Cultivated_land
4 5 40522278 Grassland
5 6 73492517 Deciduous_forest
6 7 17589896 Evergreen_forest
7 8 18928400 Mixed_forest
8 9 14008649 Scrub/shrub
9 10 32861767 Palustrine_forested_wetland
10 11 13719974 Palustrine_scrub/shrub_wetland
11 12 6947191 Palustrine_emergent_wetland
12 16 82553 Unconsolidated_shore
13 17 1564930 Bare_land
14 18 23649727 Water
15 19 591821 Palustrine_aquatic_bed
Process_Description:
This data is the 2000-era (or late-date) classification of the Great Lakes.
It consists of a series of partial Landsat 7 Thematic Mapper scenes which
were analyzed according to the Coastal Change Analysis Program (C-CAP)
protocol to determine land cover. The data were field validated and
subsequently mosaicked to produce a land cover inventory for
a portion of the US Great Lakes Coastal Zone.
This classification is to fulfill the final delivery of the NOAA
C-CAP. The dataset was created by Earth Satellite Corporation.
This version of the classification is a final version. The study
area is the US Great Lakes Coastal Region. This product is one
of a series of 7 states covering a portion of that region.
The process history of this file is as follows: First, Landsat 7
data covering the study area was ortho-rectified at EROS Data
Center. Next the datasets were assessed for spatial (horizontal)
accuracy by Earth Satellite Corporation. Scenes with an offset
greater than 1 pixel were not accepted. The data was then subset
by the boundaries of the study area. Cloud cover was removed and
in some cases was replaced by same-era data using a statistical
prediction technique with the software Cubist. The individual
scenes were then classified. Unsupervised classification was used
to create a signature file: 233 classes, 20 iterations, 0.999
convergence threshold, 3X3 skip factor. The signature file was
then run through a Supervised Classification process: Maximum
Likelihood, no non-parametric rule. The resulting clusters were
labeled using the Earth Satellite Corporation developed
addition to Imagine 8.5 called GeoTools. In GeoTools the "Summary"
function was performed using the field-collected data and
classified scene overlap to label the clusters of the image.
Subsequent small corrections were made to the data by renaming a
few clusters and, in some cases, AOIs were used.
Then the classified data were stratified by NWI data to classify the
wetland classes. In some scenes where the NWI overlayed a pixel, the
pixel was changed to the wet version of its own class (i.e. deciduous
changed to palustrine forest). In other scenes the entire NWI area plus
other wet areas missed by NWI were classified separately. The human
developed categories were refined by using TIGER2000 files (rasterized)
to stratify a clustered file of 233 unlabeled classes. Then clusters
were labeled for High Intensity Developed or Low Intensity Developed
within the 30 meter rasterized TIGER2000 stratification. This process
was also done using a 90 meter version of the rasterized TIGER2000 data
only where AOI's were used to stratify out larger cities. Then a water
model was used to refine the water category. The water model uses image
algebra to determine water boundaries and gives several options for
accepting water pixels. Also a model was developed to refine the
cultivated category by distinguishing cultivated pixels from grassland
pixels better. This model utilized different dates of same-year imagery
to use greenness cycle as a criteria for classification. The model
calculated an NDVI for each date which were then subtracted from each
other and threshold of change applied. Sometimes the threshold was
manually determined, and sometimes it was automated using 1 standard
deviation on either side of 0. Cluster-busting and screen digitizing
were used extensively in the production of these scenes. Also comments
from the consultant and NOAA CSC staff on the rough and provisional
classifications (the previous versions) were used to find and correct
problem areas.
Cultivated/Grassland differences were determined using several
techniques. Up to 3 dates of imagery exists for the same year of the
late-date. PCA was run on the different dates of the same year and
layerstacked to produce a six band image of PCA. This was then clustered
to produce a 233 cluster file. Then the cultivated and grassland
categories were masked from this file and re-classified. After having
performed a second field trip, the field collected points were used to
summarize the masked clusters with the ground truth points. This created
an automated Cultivated/Grassland distinction. This was then adjusted
and corrected. Filtering and Spatial Auto-correlation were performed on
the Cultivated and Grassland categories. A spatial majority filter was
applied to the cultivated category (class 4)and in some instances to
class 5. The filter was applied with a .85 or .75 majority threshold
within a 3x3 matirx. This reduced speckle within agricultural fields.
Also a spatial auto-correlation algorithm was applied to the land cover.
This algorithm uses low pass filters and recodes stray pixels to most
likely classes based on surrounding pixels.
Scrub/shrub and forest categories were confused sometimes. This was
alleviated by classifying the forest or scrub categories with the summer
imagery if available. Mixed Forest and Scrub/Shrub were the least
accurate classes.
After each scene was classified, a mosaicking algorithm was applied to
all the scenes in a team area to join the data into four team areas.
Also the seven states were mosaicked into one study region and then cut
out again by state to create the final state classifications. This
mosaic algorithm first creates a hierarchy of classes so that when the
overlay takes place, the preferred classes dominate. The hierarchy of
classes is as follows in descending order of dominance: class 1, 20, 19,
16, 17, 6, 8, 7, 9, 5, 4, 13, 14, 15, 10, 11, 12, 18, 22, 21, 3, 2. This
hierarchy was applied to the overlap between the scenes being mosaicked.
The hierarchy was applied automatically in most cases but in some areas
the discrepancy in the overlap was highlighted and changed. Thus the
overlap areas in the mosaicking process were assessed.
All processing was performed using Imagine 8.4 and 8.5 and Earth
Satellite Corporation developed additions, models, and eml scripts.
Attribute_Accuracy_Report:
A team of field investigators participated in field collection of
verification points in October 2001 and July 2002. Data validation teams
consisted of personnel from the NOAA Coastal Services Center. Each team
was equipped with a portable color laptop computer linked to a Global
Positioning System (GPS). The field laptop runs software that supports
the classified data as a raster background with the road network as a
vector overlay with a simultaneous display of live GPS coordinates.
Accuracy assessment points were generated with ERDAS Imagine software using
a stratified random sample in 3x3 pixel homogeneous windows. This data
collected was used to produce accuracy assessments for the Great Lakes
C-CAP data. Both windshield survey methods of collection and airplane
reconaissance were implemented to collect the accuracy assessment points.
NOAA implemented an accuracy assessment. The accuracy assessment plan
included the collection of field points. Only areas containing at least
3 x 3 contiguous pixel clumps were assessed. Transects were created and
random points were generated along those transects. The results have been
shared with EarthSat. The overall accuracy for the Great Lakes region is
91.4% correct. All of the states are also independently higher than the 85%
accurate required by NOAA C-CAP. Kappa coefficient was used to determine
the overall accuracy of 90.2%. The class accuracies were determined by
the producer's accuracy, or error of omission. These were supposed to be all
above 80% but three categories were below in the overall and in many states
individually: Mixed Forest, Scrub/Shrub, and Palustrine Scrub/Shrub. These
are the more subjective classes in that they have hard to define boundaries.
No fuzzy assessments were implemented, and an error matrix was created.
The overall accuracies by state are as follows: NY - 85.1%, PA - 94.3%, OH -
91.6%, IN - 92%, IL - 100%, WI - 96.1%, MN - 91.8%.
The early-date classification is derived from the re-classification of
the of the change areas of the late-date (c.2000)classification using the
early-date (c.1995) imagery. Therefore its accuracy is inherently tied to
the accuracy of the late-date classification. The late-date classification
was based primarily on the fieldwork collected. Accuracy tests of later
versions of this provisional early-date classification will involve logic
change analysis.
Pre-processing steps:
Each Landsat TM scene was geo-referenced by USGS EROS DATA CENTER. Then
EarthSat staff verified the scenes for spatial accuracy to within 1 pixel
consistently. The data was geo-referenced to Albers Conical Equal Area,
with a spheroid of GRS 1980, and Datum of WGS84. The data units is in
meters.
Ancillary Datasets:
Non-TM image datasets used are NWI, TIGER2000, and field-collected points.
Both datasets were rasterized to 30 meter pixels and used to mask the
classified layer. This stratified thematic layer was
then labeled for appropriate categories. The NWI was used to classify the
wetland categories and the TIGER2000 was used for the human developed
categories. The TIGER2000 data was also rasterized to 90 meter pixels in
order to include some of the low intensity developed with occurred more
than one pixel from the roads. These ancillary data were used in a visual
capacity as well as a rule-based model approach.
Field-Collected Data:
EarthSat's method of field point collection is the windshield survey method.
This means field crew drive around secondary roads in the study area
and record land cover at as many points as possible within the given amount
of time. The type information recorded at each location are as follows:
· Canopy cover
· Vegetation types by species (where applicable)
· Land Cover characterization
· Soils (if relevant)
· Special conditions and remarks
· Photography/video index number
· Date/time
· X, Y location (Z if relevant)
These data are then associated with the image by X, Y point and are
written to Arc/Info vector files. The above recorded information is
located in the attributes. Also digital photos were taken at certain points.
These points were determined by getting a mix of typical examples of a
feature as well as atypical. Points were taken for all classes of features.
To facilitate point collection EarthSat requested an additional module
from ERDAS to run in the IMAGINE software as an add-on to the
current GMID module. GMID only shows where on the image you are
located but does not allow attribute information to be recorded for a
point. The new module, RGMID, allows the field worker to select a pixel
from the image and record observations in text fields. Then these points
are exported to an Arc/Info vector file.
The first fieldwork expedition was performed by four teams. The
Great Lakes region study area was broken into four roughly even
fieldwork zones broken up by state, corresponding to the four teams.
The teams consisted of at least one EarthSat employee and at least one
NOAA employee. All teams had three workers. The first field trip
took place between October 15 and 26, 2001.
Each team collected over 1000 points per day. One person drove the
car while another navigated and pointed out features. The third team
member logged the points on the lap-top computer. Points were
selected that were larger than one pixel footprint, i.e. 3x3 pixels.
Each team collected about 20 digitally photographed points per
day as well. Point collection lasted at least eight hours per day.
The generalized route of collections within the study areas were
pre-determined using DeLorme Street Atlas USA. A proposed path
for collection was printed out by day before the teams went to the field.
These paths were for the most part followed but there were many
deviations in order to collect interesting or confusing points. The paths
were determined by length (about 5-7 hour trips per day), by how well
the entire study area was covered, and by how many different
feature types and physiographic regions could be covered.
The data and equipment required for the fieldwork is as follows:
Ancillary datasets:
· TIGER 2000
· NWI - mosaic into states
· State road map and Delorme state atlas www.delorme.com
Hardware:
· Lap-tops with IMAGINE and data
· 2 GARMINs and external antennae, redundant data cables
· Digital Cameras
· 4 backup devices
· 8 extra batteries
· 4 DC to AC adapters, and splitters
· Car fuses, flashlights, basic tools
· Mobile phones (if available)
· Calculator
· 2 sets of cable
· Compass
· Notebooks with complete field instructions and contact numbers
Imagery:
· Make note of interesting points beforehand
· Cluster Images to 233 clusters
· If time and early-date scenes available do CCA for binary change layer
· If time, rough classification of late date
The purpose of the first trip was to gather training points and digital
photographs and to understand the landscape before mapping intensely.
Also the points collected were rasterized, resampled to 3 by 3 pixels
and summarized with the thematic cluster layer to produce initial
classifications.
Training of field staff involved testing the systems, setting
up and breaking down the equipment, taking points while
driving local areas, and processing the collected files to make sure
the data was useable.
The second field trip occurred eight months later, toward the end of the
classifying of the late-date imagery. It was decided that there
was not enough understanding of the Grassland and Cultivated
categories and their differences in the imagery. The focus of the second
field trip was to understand this distinction and to better categorize
these features on the late-date classification. This field trip was deemed
necessary after discussions with NOAA at the May 7 meeting.
This field trip was not planned before that meeting. It took place the
last week of May for the New York State, and the first week of June
for the other states. The field trips lasted for about one week.
The plan for the second fieldwork involved five EarthSat employees.
Four employees were broken into two teams of two and one worker
went alone. The teams were outfitted with the same equipment as
in the previous field trip. Two teams started at the same place and
worked their way in opposite directions. The starting place was
Chicago. Team A traveled from the starting location to the
destination of Cleveland, OH, and then back to Chicago. Team
B traveled from the starting location to Duluth, MN, and back to
Chicago. The length of this field trip was one week. Also, one
employee drove the agricultural areas of New York State. This
portion of the fieldwork was completed in five days, from
May 24 to May 28, 2002. Fields of hay, pasture, row crops, and
especially winter wheat were targeted during the collection of points.
The data was used in the re-classification of the cultivated and
grassland categories.
After the field points were collected they were rasterized and an
ArcView project was made to view the vector points. The
ArcView project contains hyperlinks to the digital photos so that
when the appropriate points are selected, a digital photo pops up
automatically. This was made for each team area. These field
point ArcView projects were then submitted to NOAA as a deliverable.
They were made for both field trips.
The vector field points were rasterized in order to summarize them
with the thematic clustered data to produce a rough classification.
After the points were rasterized they were increased in size from
one pixel to nine. This allowed for more pixels to enter into the
calculations in the summary.
Post-Processing Steps:
After each scene was classified, a mosaicking algorithm was applied to all
the scenes in a team area to join the data into four team areas. Also the
four team areas were mosaicked into one study region. This mosaic
algorithm first creates a hierarchy of classes so that when the overlay
takes place, the preferred classes dominate. The hierarchy of classes is
as follows in descending order of dominance: class 1, 20, 19, 16, 17, 6,
8, 7, 9, 5, 4, 13, 14, 15, 10, 11, 12, 18, 22, 21, 3, 2. This hierarchy
was applied to the overlap between the scenes being mosaicked. The
hierarchy was applied automatically in most cases but in some areas the
discrepancy in the overlap was highlighted and changed. Thus the overlap
areas in the mosaicking process were assessed.
The early-date classifications were additionally mosaicked to overlay the
final late-date classifications. This was done for each team area. Then
this file was used as a comparison to the late-date final classification
for each team. The comparison was made by performing a bivariate analysis.
This contains data for both dates of imagery and all change information.
Spatial Filters:
A spatial majority filter was applied to the cultivated category (class 4)
and in some instances to class 5. The filter was applied with a .85 or .75
majority threshold within a 3x3 matrix. This reduced speckle within
agricultural fields. Also a spatial auto-correlation algorithm was applied
to the land cover. This algorithm uses low pass filters and recodes stray
pixels to most likely classes based on surrounding pixels.
Logical_Consistency_Report:
Tests for logical consistency indicate that all row and column
positions in the selected latitude/longitude window contain data.
Conversion and integration with vector files indicates that all
positions are consistent with earth coordinates covering the same
area. Attribute files appear to be logically consistent.
Completeness_Report:
The classification scheme does not included all anticipated land
covers. There are no pixels representing class 21 (Aquatic Vegetation).
All pixels have been classified. The NOAA Coastal Change
Analysis Program (C-CAP): Guidance for Regional Implementation, NOAA
National Marine Fisheries Service Report 123, discusses the interagency
effort to develop the land cover classification scheme and defines all
categories.
Enumerated_Domain_Value: 1 Unclassified
Enumerated_Domain_Value_Definition:
This class contains no
data due to cloud conditions or data voids.
Enumerated_Domain_Value: 2 High Intensity Developed
Enumerated_Domain_Value_Definition:
Contains little or no vegetation. This subclass includes
heavily built-up urban centers as well as large
constructed surfaces in suburban and rural areas. Large
buildings (such as multiple family housing, hangars, and
large barns), interstate highways, and runways typically
fall into this subclass.
Enumerated_Domain_Value: 3 Low Intensity Developed
Enumerated_Domain_Value_Definition:
Contains substantial amounts of constructed surface mixed
with substantial amounts of vegetated surface. Small
buildings (such as single family housing, farm
outbuildings, and sheds), streets, roads, and cemeteries
with associated grasses and trees typically fall into this
subclass.
Enumerated_Domain_Value: 4 Cultivated Land
Enumerated_Domain_Value_Definition:
Includes herbaceous (cropland) and woody (e.g., orchards,
nurseries, and vineyards) cultivated lands.
Enumerated_Domain_Value: 5 Grassland
Enumerated_Domain_Value_Definition:
Dominated by naturally occurring grasses and non-grasses
(forbs) that are not fertilized, cut, tilled, or planted
regularly.
Enumerated_Domain_Value: 6 Deciduous Forest
Enumerated_Domain_Value_Definition:
Includes areas dominated by single stemmed, woody
vegetation unbranched 0.6 to 1 meter (2 to 3 feet) above
the ground and having a height greater than 6 meters (20
feet).
Enumerated_Domain_Value: 7 Evergreen Forest
Enumerated_Domain_Value_Definition:
Includes areas in which more than 67 percent of the trees
remain green throughout the year. Both coniferous and
broad-leaved evergreens are included in this category.
Enumerated_Domain_Value: 8 Mixed Forest
Enumerated_Domain_Value_Definition:
Contains all forested areas in which both evergreen and
deciduous trees are growing and neither predominate.
Enumerated_Domain_Value: 9 Scrub/Shrub
Enumerated_Domain_Value_Definition:
Areas dominated by woody vegetation less than 6 meters in
height. This class includes true shrubs,young trees, and
trees or shrubs that are small or stunted because of
environmental conditions.
Enumerated_Domain_Value: 10 Palustrine Forested Wetland
Enumerated_Domain_Value_Definition:
Includes all nontidal wetlands dominated by woody
vegetation greater than or equal to 6 meters in height,
and all such wetlands that occur in tidal areas in which
salinity due to ocean-derived salts is below 0.5 parts per
thousand (ppt).
Enumerated_Domain_Value: 11 Palustrine Scrub/Shrub Wetland
Enumerated_Domain_Value_Definition:
Includes all nontidal wetlands dominated by woody
vegetation less than or equal to 6 meters in height, and
all such wetlands that occur in tidal areas in which
salinity due to ocean-derived salts is below 0.5 ppt.
Enumerated_Domain_Value: 12 Palustrine Emergent Wetland
Enumerated_Domain_Value_Definition:
Includes all nontidal wetlands dominated by trees, shrubs,
persistent emergents, emergent mosses, or lichens, and all
such wetlands that occur in tidal areas in which salinity
due to ocean- derived salts is below 0.5 ppt.
Enumerated_Domain_Value: 13 Estuarine Forest Wetland
Enumerated_Domain_Value_Definition:
Includes all tidal wetlands dominated by woody vegetation
greater than or equal to 6 meters in height, and all such
wetlands that occur in tidal areas in which salinity due
to ocean-derived salts is above 0.5 parts per thousand
(ppt).
Enumerated_Domain_Value: 14 Estuarine Scrub/Shrub Wetland
Enumerated_Domain_Value_Definition:
Includes all tidal wetlands dominated by woody vegetation
less than or equal to 6 meters in height, and all such
wetlands that occur in tidal areas in which salinity due
to ocean-derived salts is above 0.5 ppt.
Enumerated_Domain_Value: 15 Estuarine Emergent
Enumerated_Domain_Value_Definition:
Characterized by erect, rooted, herbaceous hydrophytes
(excluding mosses and lichens) that are present for most
of the growing season in most years. Perennial plants
usually dominate these wetlands. All water regimes are
included except those that are subtidal and irregularly
exposed.
Enumerated_Domain_Value: 16 Unconsolidated Shore
Enumerated_Domain_Value_Definition:
Characterized by substrates lacking vegetation except for
pioneering plants that become established during brief
periods when growing conditions are favorable. Erosion and
deposition by waves and currents produce a number of
landforms, such as beaches, bars, and flats, all of which
are included in this class.
Enumerated_Domain_Value: 17 Bare Land
Enumerated_Domain_Value_Definition:
Composed of bare soil, rock, sand, silt, gravel, or other
earthen material with little or no vegetation.
Enumerated_Domain_Value: 18 Water
Enumerated_Domain_Value_Definition:
Includes all areas of open water with less than 30 percent
cover of trees, shrubs, persistent emergent plants,
emergent mosses, or lichens.
Enumerated_Domain_Value: 19 Palustrine Aquatic Bed
Enumerated_Domain_Value_Definition:
Includes wetlands and deepwater habitats dominated by
plants that grow principally on or below the surface of
the water for most of the growing season in most years.
|
![[logo] US EPA](../gif/logo_epaseal.gif){kind=logo}
Great Lakes Basin (GLB) Landscape Ecology Metric Browser (v2.0)
- You are here: EPA Home
- Landscape Ecology
- GLB Landscape Ecology Metric Browser (v2.0)
- Metrics Map
- Metadata - U.S.A. Coastal Change Analysis Program (CCAP)