March 2000 classified RADARSAT image depicting winter shorebird habitat in the Willamette Valley, OR

Metadata:

• Identification_Information
• Data_Quality_Information
• Spatial_Data_Organization_Information
• Spatial_Reference_Information
• Entity_and_Attribute_Information
• Distribution_Information
• Metadata_Reference_Information

Identification_Information:

Citation:

Citation_Information:

Originator: Taft, Oriane W., and Susan M. Haig

Publication_Date: 2004

Title:

March 2000 classified RADARSAT image depicting winter shorebird habitat in the Willamette Valley, OR

Geospatial_Data_Presentation_Form: raster digital data

Online_Linkage: <http://fresc.usgs.gov/ArcIMS/Website/radar>

Larger_Work_Citation:

Citation_Information:

Originator: Taft, O.W., S.M. Haig and C. Kiilsgaard

Publication_Date: 2004

Title:

Use of radar remote sensing (RADARSAT) to map winter wetland habitat for shorebirds in an agricultural landscape

Geospatial_Data_Presentation_Form: Journal article

Series_Information:

Series_Name: Environmental Management

Issue_Identification: 33: 750-763

Description:

Abstract:

This dataset is one of a collection of three land cover maps of the Willamette Valley of Oregon, USA, depicting 4 habitat classes: wet with <= 50% vegetation cover (class 1), wet with >50% vegetation (class 2), dry with <= 50% vegetation (class 3), dry with >50% vegetation (class 4). Land cover maps were derived by classifying signatures from C-band (5.6 cm wavelength), HH-polarized radar remote sensing (RADARSAT) data collected at a 38 degree incidence angle (8-m resolution). The data are in the form of an ERDAS Imagine image.

Purpose:

Data were collected to map the distribution of winter shorebird habitat on agricultural lands in the Willamette Valley of western Oregon.

Time_Period_of_Content:

Time_Period_Information:

Single_Date/Time:

Calendar_Date: 20020315

Currentness_Reference: Date of satellite data collection

Status:

Progress: Complete

Maintenance_and_Update_Frequency: None planned

Spatial_Domain:

Bounding_Coordinates:

West_Bounding_Coordinate: -123.536316

East_Bounding_Coordinate: -122.408327

North_Bounding_Coordinate: 45.541194

South_Bounding_Coordinate: 43.933582

Keywords:

Theme:

Theme_Keyword_Thesaurus: ISO19139 Topic Category

Theme_Keyword: imageryBaseMapsEarthCover

Theme:

Theme_Keyword_Thesaurus: Biocomplexity Thesaurus

Theme_Keyword: Agricultural land

Theme:

Theme_Keyword_Thesaurus: None

Theme_Keyword: RADARSAT

Theme_Keyword: agricultural landscape

Theme_Keyword: shorebirds

Place:

Place_Keyword_Thesaurus: None

Place_Keyword: Willamette Valley

Place_Keyword: Oregon

Place_Keyword: OR

Temporal:

Temporal_Keyword_Thesaurus: None

Temporal_Keyword: winter

Access_Constraints: No legal or policy restrictions to data access

Use_Constraints:

Data set originators must be cited if these data are used for analysis or display.

Point_of_Contact:

Contact_Information:

Contact_Person_Primary:

Contact_Person: Haig, Susan M.

Contact_Organization:

US Geological Survey Forest and Rangeland Ecosystem Science Center

Contact_Address:

Address_Type: mailing and physical address

Address: 3200 SW Jefferson Way

City: Corvallis

State_or_Province: OR

Postal_Code: 97331

Contact_Voice_Telephone: (541) 750-7482

Contact_Facsimile_Telephone: (541) 758-8806

Contact_Electronic_Mail_Address: susan_haig@usgs.gov

Data_Set_Credit:

Funding was provided by Oregon Watershed Enhancement Board and U.S. Fish and Wildlife Service. Analysis was conducted by Northwest Habitat Institute, Corvallis, OR.

Native_Data_Set_Environment:

Microsoft Windows XP Version 5.1 (Build 2600) Service Pack 2; ESRI ArcCatalog 9.0.0.535

Data_Quality_Information:

Attribute_Accuracy:

Attribute_Accuracy_Report:

We assessed accuracy of final maps using the remaining ground reference sites not used for training (see methods); this resulted in 274 verification sites for December, 283 for January, and 284 for March. For each site, we compared its predicted class from final maps to the site?s true class and constructed an error matrix of these comparisons for each date. The error assessment characterized the accuracy of each map, with commission error or "user?s accuracy" (proportion of sites assigned to a class that are correctly assigned) and omission error or "producer?s accuracy" (the proportion of sites of a certain class that are correctly assigned) included. For each map, we calculated overall accuracy as the number of correctly classified sites divided by the total number of testing sites.

Logical_Consistency_Report: Not applicable

Completeness_Report:

In mapping shorebird habitat on agricultural lands, we focused on identifying flooded fields with exposed soil (less than 50% cover); these conditions are preferred by Dunlin and ecological allies (species in the family Scolopacidae). We designed a classification scheme of four classes based on presence of water and vegetation: 1) wet (either impounded water or shallowly flooded) with <= 50% vegetation, 2) wet with >50% vegetation, 3) dry (without visible standing water) with <= 50% vegetation, 4) dry with >50% vegetation.

Positional_Accuracy:

Horizontal_Positional_Accuracy:

Horizontal_Positional_Accuracy_Report:

Before rectifying scenes, we used Sigma and Lee filters to reduce speckle noise (Rio et al. 2000). Among nine scenes, we were able to reduce coefficients of variation for noise from 0.387-0.399 at raw imagery to 0.213-0.223 at the second pass. We then rectified all December scenes to a 1996 Landsat Thematic Mapper reference image of the Valley. Among the three December scenes, we selected between 18 and 34 ground control points (GCPs) to calculate third-order polynomial transformations used to geo-register input images to the reference scene. Among the three scenes, registration was achieved with an error of 22-28 meters, approximately the size of a Landsat TM pixel (30 m). Scenes for January (using 41-52 GCPs) and March (38-45 GCPs) were geo-registered to the rectified December scenes, also using third-order polynomial transformations. Registration errors were lower among these six scenes, ranging from 4.8-16.7 meters. We used a cubic convolution resampling procedure to assign new coordinate grids to scenes. After geo-registration, the three scenes for each date were mosaicked into composite images used in the classification effort.

Lineage:

Source_Information:

Source_Citation:

Citation_Information:

Originator:

Pater, D.E., S. Bryce, T.D. Thorson, J. Kagan, C. Chappell, J. Omernik, S. Azevedo and A. J. Woods

Publication_Date: 1997

Title: Ecoregions of western Washington and Oregon

Geospatial_Data_Presentation_Form: vector digital data

Publication_Information:

Publication_Place: Corvallis, Oregon

Publisher: U.S. Environmental Protection Agency

Source_Scale_Denominator: 100000

Type_of_Source_Media: electronic file

Source_Time_Period_of_Content:

Time_Period_Information:

Single_Date/Time:

Calendar_Date: 1997

Source_Currentness_Reference: publication date

Source_Citation_Abbreviation: ecoregions

Source_Contribution:

The Willamette Valley Foothills ecoregion was selected and rasterized and used to mask this area from the radar imagery.

Source_Information:

Source_Citation:

Citation_Information:

Originator: Kiilsgaard, C.

Publication_Date: 1999

Title:

Oregon vegetation: mapping and classification of landscape-level cover types

Geospatial_Data_Presentation_Form: report

Publication_Information:

Publication_Place: Moscow, Idaho

Publisher: USGS

Other_Citation_Details: Final report, USGS-BRD GAP Analysis Program

Source_Time_Period_of_Content:

Time_Period_Information:

Single_Date/Time:

Calendar_Date: 1999

Source_Currentness_Reference: publication date

Source_Citation_Abbreviation: GAP data

Source_Contribution:

Each of the final classified mosaicked images was overlaid with urban, forest, and permanent open water (reservoirs, rivers, etc.) land cover from this data set.

Source_Information:

Source_Citation:

Citation_Information:

Originator: Radarsat, International

Publication_Date: various

Title: RADARSAT Imagery

Geospatial_Data_Presentation_Form: remote-sensing image

Other_Citation_Details: 8 meter pixel size

Source_Time_Period_of_Content:

Time_Period_Information:

Multiple_Dates/Times:

Single_Date/Time:

Calendar_Date: 19991210

Single_Date/Time:

Calendar_Date: 20000128

Single_Date/Time:

Calendar_Date: 20000315

Source_Currentness_Reference: date of satellite data collection

Source_Citation_Abbreviation: radar image

Source_Contribution:

We acquired imagery on three selected days when the Canadian RADARSAT satellite passed over the Valley study area: 10 December 1999, 27 January 2000, and 15 March 2000. With these dates we aimed to capture the extent of shorebird wetland habitat at three different time periods: at the onset of early winter rainfall (early December), at the peak of the mid-winter rainy season (January), and during waning rain activity in late winter/early spring (March).

Process_Step:

Process_Description:

We acquired imagery on three selected days when the Canadian RADARSAT satellite passed over the Valley study area: 10 December 1999, 27 January 2000, and 15 March 2000. With these dates we aimed to capture the extent of shorebird wetland habitat at three different time periods: at the onset of early winter rainfall (early December), at the peak of the mid-winter rainy season (January), and during waning rain activity in late winter/early spring (March).

The RADARSAT satellite carries a C-band (5.6 cm wavelength) synthetic aperture radar (SAR) with HH (horizontally transmitted/horizontally received) wavelength polarization. All scenes were taken in fine 1 far (F1F) beam mode (8-meter pixel resolution; 38.78o incidence angle) on descending orbits at 14:18 hours. Scenes at F1F are approximately 50 x 50 km, thus it took three scenes to cover our focal study area (central and southern Valley) on each acquisition date. These images were aligned and acquired north-south so that each final map, once mosaicked, would continuously cover an area roughly 50 x 150 km (4500 km2), encompassing Valley lowlands from Eugene to Newberg, Oregon. We used ecoregions to mask the Valley Foothills subecoregion from radar imagery acquired in the 50 x 150 km swath. The size of the study area portrayed by final maps was 4209 km2.

Source_Used_Citation_Abbreviation: ecoregions

Source_Used_Citation_Abbreviation: radar image

Process_Date: 2000

Process_Contact:

Contact_Information:

Contact_Person_Primary:

Contact_Person: Kiilsgaard, Chris

Contact_Organization: Northwest Habitat Institute

Contact_Address:

Address_Type: mailing address

Address: PO Box 855

City: Corvallis

State_or_Province: OR

Postal_Code: 97339

Contact_Voice_Telephone: 541-757-3118

Contact_Facsimile_Telephone: 541-753-2440

Contact_Electronic_Mail_Address: chris@nwhi.org

Process_Step:

Process_Description:

We used both supervised and unsupervised classification techniques to associate spectral signature with habitat classes. Using our field data, we were able to distinguish a specific signature to supervise the classification of class 1 habitat only; we used unsupervised training to produce signatures for the remaining classes. For supervised classification of habitat in class 1, we delineated polygons around pixel clusters of homogeneous spectral returns within class 1 reference sites, and calculated mean, standard deviation, minimum and maximum digital values for all pixels within each polygon. We then used a maximum likelihood classification algorithm to delineate class 1 habitat within mosaicked scenes. To classify habitat in classes 2-4, we separated returns remaining in the class 1-masked scene into 50 distinct spectral clusters. We chose the number 50 to accommodate the variability of signatures we expected from the myriad of permutations of wet and dry soils with varying amount of vegetation. Using roughly half of our ground reference dataset (415 sites from December, 448 sites from January, and 308 sites from March) as training sites (randomly chosen), we then visually determined which of the three categories each of the 50 return clusters represented and recoded the 50 clusters into three remaining habitat classes. Finally, we overlaid each of the final classified mosaicked images with urban, forest, and permanent open water (reservoirs, rivers, etc.) land cover using an ancillary map developed from 1996 Landsat TM imagery (Kiilsgaard 1999).

Source_Used_Citation_Abbreviation: GAP data

Source_Used_Citation_Abbreviation: radar image

Process_Date: 2000

Process_Contact:

Contact_Information:

Contact_Person_Primary:

Contact_Person: Kiilsgaard, Chris

Contact_Organization: Northwest Habitat Institute

Contact_Address:

Address_Type: mailing address

Address: PO Box 855

City: Corvallis

State_or_Province: OR

Postal_Code: 97339

Contact_Voice_Telephone: 541-757-3118

Contact_Facsimile_Telephone: 541-753-2440

Contact_Electronic_Mail_Address: chris@nwhi.org

Spatial_Data_Organization_Information:

Direct_Spatial_Reference_Method: Raster

Raster_Object_Information:

Raster_Object_Type: Pixel

Row_Count: 22302

Column_Count: 11008

Vertical_Count: 1

Spatial_Reference_Information:

Horizontal_Coordinate_System_Definition:

Planar:

Map_Projection:

Map_Projection_Name: Transverse Mercator

Transverse_Mercator:

Scale_Factor_at_Central_Meridian: 0.9996

Longitude_of_Central_Meridian: -123

Latitude_of_Projection_Origin: 0

False_Easting: 500000

False_Northing: 0

Planar_Coordinate_Information:

Planar_Coordinate_Encoding_Method: Coordinate Pair

Coordinate_Representation:

Abscissa_Resolution: 8

Ordinate_Resolution: 8

Planar_Distance_Units: meters

Geodetic_Model:

Horizontal_Datum_Name: D_GRS_1980

Ellipsoid_Name: Geodetic Reference System 80

Semi-major_Axis: 6378137

Denominator_of_Flattening_Ratio: 298.257222

Entity_and_Attribute_Information:

Detailed_Description:

Entity_Type:

Entity_Type_Label: march 15 2000.img

Entity_Type_Definition: Erdas Imagine file

Entity_Type_Definition_Source: ERDAS

Attribute:

Attribute_Label: Blue

Attribute_Definition: meter number of blue in the RGB color model

Attribute_Definition_Source: RGB color model

Attribute_Domain_Values:

Range_Domain:

Range_Domain_Minimum: 0

Range_Domain_Maximum: 1

Attribute:

Attribute_Label: Class_names

Attribute_Definition: Name of habitat class

Attribute_Definition_Source: Dataset developer

Attribute_Domain_Values:

Enumerated_Domain:

Enumerated_Domain_Value: 0

Enumerated_Domain_Value_Definition: Unclassified area

Enumerated_Domain:

Enumerated_Domain_Value: 1

Enumerated_Domain_Value_Definition: Wet, unvegetated (<50%)

Enumerated_Domain:

Enumerated_Domain_Value: 2

Enumerated_Domain_Value_Definition: Wet, vegetated (>50%)

Enumerated_Domain:

Enumerated_Domain_Value: 3

Enumerated_Domain_Value_Definition: Dry, unvegetated (<50%)

Enumerated_Domain:

Enumerated_Domain_Value: 4

Enumerated_Domain_Value_Definition: Dry, vegetated (>50%)

Enumerated_Domain:

Enumerated_Domain_Value: 5

Enumerated_Domain_Value_Definition: Urban area

Enumerated_Domain:

Enumerated_Domain_Value: 6

Enumerated_Domain_Value_Definition: Open water

Enumerated_Domain:

Enumerated_Domain_Value: 7

Enumerated_Domain_Value_Definition: Forest

Attribute:

Attribute_Label: Count

Attribute_Definition: Number of pixels in class

Attribute_Definition_Source: ERDAS

Attribute_Domain_Values:

Range_Domain:

Range_Domain_Minimum: 0

Range_Domain_Maximum: Unknown Maximum

Attribute:

Attribute_Label: Green

Attribute_Definition: meter number for the amount of green in RBG color model

Attribute_Definition_Source: RGB color model

Attribute_Domain_Values:

Range_Domain:

Range_Domain_Minimum: 0

Range_Domain_Maximum: 1

Attribute:

Attribute_Label: ObjectID

Attribute_Definition: Internal feature number.

Attribute_Definition_Source: ESRI

Attribute_Domain_Values:

Unrepresentable_Domain:

Sequential unique whole numbers that are automatically generated.

Attribute:

Attribute_Label: Opacity

Attribute_Definition: Code identifying a pixel as a class set or part of background

Attribute_Definition_Source: Dataset developer

Attribute_Domain_Values:

Enumerated_Domain:

Enumerated_Domain_Value: 0

Enumerated_Domain_Value_Definition: background

Enumerated_Domain:

Enumerated_Domain_Value: 1

Enumerated_Domain_Value_Definition: class set

Attribute:

Attribute_Label: Red

Attribute_Definition: meter number for the amount of red in RGB color model

Attribute_Definition_Source: RGB color model

Attribute_Domain_Values:

Range_Domain:

Range_Domain_Minimum: 0

Range_Domain_Maximum: 1

Distribution_Information:

Distributor:

Contact_Information:

Contact_Organization_Primary:

Contact_Organization:

U.S. Geological Survey (USGS), Forest and Rangeland Ecosystems Science Center (FRESC) Rangeland Ecosystem Science Center (FRESC)

Contact_Person: Metadata Coordinator

Contact_Address:

Address_Type: Mailing and Physical Address

Address:

777 NW 9th St Suite 400

City: Corvallis

State_or_Province: Oregon

Postal_Code: 97330

Contact_Voice_Telephone: 541-750-1030

Contact_Facsimile_Telephone: 541-750-1069

Resource_Description: s:\projects\gis\workspace\taft\

Distribution_Liability:

The USGS Forest and Rangeland Ecosystem Science Center and its research partners make every effort to provide and maintain accurate, complete, usable, and timely information. However, some USGS data and information may, of necessity, be preliminary in nature. These data and information are provided with the understanding that they are not guaranteed to be correct or complete. Users are cautioned to consider carefully the provisional nature of these data and information before using them for decisions that concern personal or public safety or the conduct of business that involves substantial monetary or operational consequences. It is also strongly recommended that careful attention be paid to the contents of the metadata file associated with these data and information. Conclusions drawn from, or actions undertaken on the basis of, such data and information are the sole responsibility of the user.

Standard_Order_Process:

Digital_Form:

Digital_Transfer_Information:

Format_Name: Winzip

Format_Information_Content:

Raster digital data as an ERDAS Imagine image (.img) and metadata.

File_Decompression_Technique: Use Winzip or PKUnzip for decompression

Transfer_Size: 46.745

Digital_Transfer_Option:

Online_Option:

Computer_Contact_Information:

Network_Address:

Network_Resource_Name:

<http://fresc.usgs.gov/products/dataset/radarsat/wetlands_mar.zip>

Access_Instructions: Downloadable data

Fees: None

Ordering_Instructions:

Download from FRESC Website: <http://fresc.usgs.gov/products/dataset/radarsat/wetlands_mar.zip> or ArcIMS website <http://fresc.usgs.gov/ArcIMS/Website/radar>

Metadata_Reference_Information:

Metadata_Date: 20040830

Metadata_Review_Date: 20040903

Metadata_Contact:

Contact_Information:

Contact_Organization_Primary:

Contact_Organization:

U.S. Geological Survey (USGS), Forest and Rangeland Ecosystems Science Center (FRESC) Rangeland Ecosystem Science Center (FRESC)

Contact_Person: Metadata Coordinator

Contact_Address:

Address_Type: Mailing and Physical Address

Address:

777 NW 9th St Suite 400

City: Corvallis

State_or_Province: Oregon

Postal_Code: 97330

Contact_Voice_Telephone: 541-750-1030

Contact_Facsimile_Telephone: 541-750-1069

Metadata_Standard_Name:

FGDC Biological Data Profile of the Content Standard for Digital Geospatial Metadata

Metadata_Standard_Version: FGDC-STD-001.1-1999

Metadata_Time_Convention: local time

Metadata_Extensions:

Online_Linkage: <http://www.esri.com/metadata/esriprof80.html>

Profile_Name: ESRI Metadata Profile