Sagebrush Ecosystem Mapping Using Landsat ETM Imagery
[Completed January 2003]
The objectives of this project were to research the habitat needs for sage grouse and devise a strategy for mapping sagebrush in a way that would be meaningful for sage grouse habitat management. The primary focus of this effort was to examine the potential for using remote sensing technology to map different categories of sagebrush in four study areas in northern Utah. The study areas were in Rich, Box Elder and Toole Counties. The classification method involved the use of Classification and Regression Trees (CART). Datasets for the classification effort included Landsat 7 imagery for the three dates (spring, summer, fall), and DEM derived datasets such as elevation, slope, aspect and landform. The efforts produced successful results distinguishing the four sagebrush ecosystems—artemsia wyomingensis ecosystems, artemsia vaseyana ecosystems, artemsia tridentata tridentata ecosystems, and artemsia nova ecosytems. Overall accuracies for each study area ranged from 60-70 percent, however accuracies for individual classes were higher. An effort to map sagebrush canopy cover densities was taken using regression trees. Results from the sagebrush density analysis were not as reliable as the land cover classification, and should be viewed as such.
USU RS/GIS Laboratory | Open Monday through Friday, 8am to 5pm | Tel: (435)797-2385 | Contact information