Citation: Kennaway, T., E. H. Helmer, M. A. Lefsky, T. A. Brandeis and K. R. Sherrill. 2008. Mapping land cover and estimating forest structure using satellite imagery and coarse resolution lidar in the Virgin Islands. Journal of Applied Remote Sensing 2:023551, DOI:10.1117/1.3063939.
Originator:
Department of Forest, Rangeland and Watershed Stewardship, College of Natural Resources - Colorado State University
Originator:
USDA Forest Service, International Institute of Tropical Forestry
Publication_Date: December 12, 2008
Title: IITF_JRS02_usbvi_landcov2.img
Geospatial_Data_Presentation_Form: raster digital data
Publication_Information:
Other_Citation_Details:
Citation: Kennaway, T., E. H. Helmer, M. A. Lefsky, T. A. Brandeis and K. R. Sherrill. 2008. Mapping land cover and estimating forest structure using satellite imagery and coarse resolution lidar in the Virgin Islands. Journal of Applied Remote Sensing 2:023551, DOI:10.1117/1.3063939.
(Freely downloadable at <http://www.treesearch.fs.fed.us/pubs/>).
This raster dataset represents land-cover and woody vegetation formation designations for the United States and British Virgin Islands centered around the year 2000. The United States and British Virgin Islands are composed of six major islands and over 40 small islands and cays. Located in the Caribbean's Leeward Island chain, the region is situated between the Caribbean Sea to the south and the Atlantic Ocean to the north. The major islands in the U.S. territory include St Croix, St. Thomas, and St. John in the southern region, while the dominant islands in the British territory include Tortola, Virgin Gorda and Anegada to the north. The center of the island region (18 20'N, 64 40'W) is approximately 30 km east of Puerto Rico and has a combined area of approximately 50,000 ha consisting of subdued to rugged topography with elevation ranging from just below sea level in some wetlands and mangrove swamps to over 500 m on Tortola.
Citation: Kennaway, T., E. H. Helmer, M. A. Lefsky, T. A. Brandeis and K. R. Sherrill. 2008. Mapping land cover and estimating forest structure using satellite imagery and coarse resolution lidar in the Virgin Islands. Journal of Applied Remote Sensing 2:023551, DOI:10.1117/1.3063939. (Freely downloadable at <http://www.treesearch.fs.fed.us/pubs/>).
This particular dataset (IITF_JRS02_usbvi_landcov2.img) is generalized from the original data to have a classification scheme for forest formations that is consistent with that of the maps developed for Puerto Rico, Culebra, Vieques, St Kitts, Nevis, St. Eustatius, Barbados and Grenada (Helmer et al. 2008, Carib. J. Sci. Vol 44). The more detailed scheme is available in the file IITF_JRS2_usbvi_landcov1.img.
Woody vegetation formations on the islands are subtropical and include drought deciduous xeric coastal forest and shrub with succulents, evergreen coastal shrubland, deciduous, evergreen and mixed forest and shrubland with succulents, semi-deciduous, and seasonal evergreen forests. Pasture and young leguminous secondary forest and forest shrub formations are present at lower elevations where natural and human caused disturbance including sheep and grazing has occurred.
Multi-part Landsat ETM+ scenes centered around the year 2000 that had been processed for cloud and cloud shadow elimination were classified using decision tree classification software. The classification techniques use decision tree software to classify 15m panchromatic sharpened Landsat satellite image bands and ancillary geospatial data. The approach also uses a technique to create image mosaics where alternate scene dates fill cloud and cloud shadowed areas in a year 2000 base scene after undergoing regression tree normalization. See Helmer and Ruefenacht 2005 for complete details.
A multivariate error matrix accompanies the dataset calculating the overall percentage of correctly classified pixels for each mapped class, including statistics for producer and user's accuracy, variance and the simple kappa coefficient. The overall classification accuracy of 29 land-cover and woody vegetation classes is 72 % overall with a kappa coefficient of agreement of 0.76 ± 0.01.
Purpose:
Map land-cover and forest formation for the United States and British Virgin Islands.
Supplemental_Information:
This dataset represents a year 2000 classification for the United States and British Virgin Islands as part of a Caribbean wide effort to develop detailed landuse/landcover products for the Caribbean region. The dataset was completed as part of the fullfillment of a Master's of Science degree byTodd Kennaway within the Department of Forest Rangeland and Watershed Stewardship (FRWS) under the College of Natural Resources at Colorado State University.
This project was funded through a cooperative agreement with the USDA Forest Service International Institute of Tropical Forestry.
VI Agreement Number: 04-CA-11120101-047
Time_Period_of_Content:
Time_Period_Information:
Single_Date/Time:
Calendar_Date: 2000
Currentness_Reference: ground condition
Status:
Progress: In work (unpublished data)
Maintenance_and_Update_Frequency: As needed
Spatial_Domain:
Bounding_Coordinates:
West_Bounding_Coordinate: -65.121962
East_Bounding_Coordinate: -64.210360
North_Bounding_Coordinate: 18.781607
South_Bounding_Coordinate: 17.653213
Keywords:
Theme:
Theme_Keyword_Thesaurus: none
Theme_Keyword: landcover
Theme_Keyword: landuse
Theme_Keyword: woody vegetation
Theme_Keyword: classification
Theme_Keyword: Landsat
Theme_Keyword: decision tree
Theme_Keyword: forest type
Place:
Place_Keyword: US Virgin Islands
Place_Keyword: British Virgin Islands
Place_Keyword: St. Thomas
Place_Keyword: St. Croix
Place_Keyword: St. John
Place_Keyword: Tortola
Place_Keyword: Virgin Gorda
Place_Keyword: Anegada
Temporal:
Temporal_Keyword: 2000
Access_Constraints: No access constraints.
Use_Constraints: No use constraints.
Point_of_Contact:
Contact_Information:
Contact_Organization_Primary:
Contact_Organization:
USDA Forest Service, International Institute of Tropical Forestry
Browse_Graphic_File_Description: screen capture of classification
Browse_Graphic_File_Type: GIF
Data_Set_Credit:
The Department of Forestry Rangeland and Watershed Stewardship, College of Natural resources at Colorado State University in cooperation with the USDA Forest Service International Institute of Tropical Forestry.
VI Agreement Number: 04-CA-11120101-047
Security_Information:
Security_Classification: Restricted
Security_Handling_Description:
Restricted access- Consent by the USDA Forest Service, International Institute of Tropical Forestry required. Data is in draft status and cannot be released to other entities until its has been finalized and has passed peer review. Please contact E. Helmer, IITF for more information.
Native_Data_Set_Environment:
Microsoft Windows XP Version 5.1 (Build 2600) Service Pack 2; ESRI ArcCatalog 9.2.6.1500
The accuracy assessment for the 2000 classification yielded Kappa coefficient of agreement of 0.76 ± 0.01, which indicates a significant agreement between the reference and map classifiers. The overall accuracy is 72 %.
Completeness_Report:
This dataset as well as its counterparts is fully attributed and no additional updates or alterations will be needed for the present datum and spheroid. Updates and edits will be issued when necessary.
Lineage:
Process_Step:
Process_Description:
We classified multi-part Landsat ETM+ imagery scenes that were pan-sharpened for the year 2000. Cloudy parts of the reference dates for each scene were replaced by image data from cloud-free parts of other image dates. Scene parts for non-reference dates had been radiometrically matched to reference dates for each scene with regression tree models, and cloud-filled scenes were radiometrically matched to each other with histogram matching. The imagery scenes and extents included the following:
.
Path /Row Scene Mosaic
4 47 p004r047_20000327. Base
3 47 p003r047_20011102. Match
4 47 p004r47_19990917 Match
4 47 p004r47_20000802 Match
4 47 p004r47_20010125 Match
4 48 p004r48_20000327 Base
4 48 p004r48_20010125 Match
.
Ancillary data was combined with satellite imagery to create an island-wide predictor variable dataset for image classification. Topographic variables, derived from USGS 30 m digital elevation models (DEM) included elevation, slope, aspect, curvature and facet. Climatic variables included mean annual precipitation and temperature. Vector derived variables included distance to roads, distance to streams (ghuts) and distance to coastlines. Finally, band indices and ratios were created from the Landsat image bands to produce the normalized difference vegetation index (NDVI) and the 4/5 band ratio. The ancillary predictor data was spatially co-registered with the cloud free image mosaics for the two time periods and stacked with the native Landsat reflectance bands 1-5 and 7 at a spectral resolution of 15 meters, resulting in a 18 band image for each classification time period.
Process_Contact:
Contact_Information:
Contact_Person_Primary:
Contact_Person: Todd Kennaway
Contact_Organization: FRWS Colorado State University
Field reference surveys were performed during a reconnaissance visit in 2004 to collect training and reference data. The island wide field surveys and expert consultation with regional vegetation experts enabled us to discern land cover attributes and forest formation from 1 m IKONOS panchromatic sharpened imagery and 1m color DOQQs. Land-cover and forest formation were then identified in the satellite imagery based on a comparison to areas interpreted in reference imagery and the spatial location of field reference survey data. Forest formation was identifiable in both the reference imagery and the Landsat imagery by a number of factors including spectral components such as color, contrast, tone and texture as well as spatial indicators including geology and elevation. Difficulties distinguishing forest formation were encountered in transitional areas including semi-deciduous and seasonal evergreen forest. Field survey data collected meeting this criteria proved useful in the identification of uncertain forest transitions.
Process_Contact:
Contact_Information:
Contact_Person_Primary:
Contact_Person: Todd Kennaway
Contact_Organization: FRWS Colorado State University
The decision tree classifier See5 (www.rulequest.com) was used for landcover and forest formation classification of Landsat ETM+ imagery for the year 2000. Training data polygons for each land-cover and land-use class were converted to an ASCII point file providing spatial coordinates, class name and assigned class number. The classification was performed with a 10 trial adaptive boosting option to improve the overall accuracy by combining many decision trees into a single combined classifier. This process provides more predicted data to analyze before determining each final class (www.rulequest.com). The default global pruning option was also used to reduce the likelihood of over fitting the tree to the training data. The pruning process removes predicted parts of the decision tree with relatively high error rates and makes decisions as to the final class assignment.
Manual edit confused classes using 1 meter IKONOS multispectral imagery and 1 meter DOQQ reference data.
Manually digitize areas where remaining cloud and cloud shadow artifacts exist in image mosaics.
Process_Contact:
Contact_Information:
Contact_Person_Primary:
Contact_Person: Todd Kennaway
Contact_Organization: FRWS Colorado State University
A stratified random sample generated approximately 50 validation points for each landcover class. The 1 m IKONOS imagery (dates ranging from 2000 to 2003) and DOQQ (2004) data were used to verify the accuracy assessment points for each class in the classified images. Using high resolution imagery to validate the classification eliminated restrictions on point locations, permitting us to include points that would otherwise not be accessible because of topography, remoteness, or property ownership. An error matrix was produced for each mapped class that estimates the overall percentage of correctly classified pixels, producer and user's accuracies, and the kappa coefficient.
Process_Contact:
Contact_Information:
Contact_Person_Primary:
Contact_Person: Todd Kennaway
Contact_Organization: FRWS Colorado State University
The dataset was completed as part of the fullfillment of a Master's of Science degree by Todd Kennaway within the Department of Forest Rangeland and Watershed Stewardship (FRWS), College of Natural Resources, at Colorado State University under contract with USDA Forest Services International Institute of Tropical Forestry (IITF) (04-CA-111200101-047).
FRWS/IITF provides geographic data "as is". Although proper data development procedures have been followed in accordance with Federal and Industry standards, FRWS makes no guarantees or warranties, either expressed or implied, concerning the accuracy of any information contained or any other matter whatsoever, including, without limitation, the condition of the product, or its fitness for any particular purpose. The burden for determining fitness for use lies entirely with the user. See source data or conduct field surveys for the most accurate portrayal of any data layer. Although these data have been processed successfully on computers of FRWS/IITF, no warranty, expressed or implied, is made by FRWS/IITF regarding the use of these data on any other system, nor does the fact of distribution constitute or imply any such warranty.
In no event shall FRWS/IITF have any liability whatsoever for payment of any consequential, incidental, indirect, special, or tort damages of any kind, including, but not limited to, any loss of profits arising out of use of or reliance on the geographic data or arising out of the delivery, installation, operation, or support by FRWS/IITF.
Standard_Order_Process:
Digital_Form:
Digital_Transfer_Information:
Transfer_Size: 0.000
Ordering_Instructions: Contact distributor for ordering instructions.
