Date of Award:

1-1-2004

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Wildland Resources

Advisor/Chair:

R. Douglas Ramsey

Abstract

Throughout the Intermountain West there has been a substantial reduction in both the quantity and quality of sagebrush ecosystems. To ass ist current range management objectives, numerous efforts have been made to classify and map sagebrush communities using remotely sensed data. However, the amount of deta il provided by these broad-scale mapping projects is often limited. This research evaluated the ability of a suite of airborne and satellite imagery to detect sagebrush community structural attributes, specifically percent canopy cover, live cover, density, size-vigor, and spatial arrangement of shrubs. Field data was collected at Camp Williams National Guard Training Facility near Bluffdale, Utah, within a Wyoming big sagebrush community. High-resolution color infrared (CIR) aerial photography, panchromatic, and multi-spectral satellite imagery, including data from Orb image, IKO OS, and Landsat ETM+, were used. Comparisons were made based on the inherent spatial and spectral properties of each image. In addition to the traditional pixel-based method for classifying imagery, a relatively new object-oriented approach to measure sagebrush cover was also explored.

Results indicate that the quantification of sagebrush cover can be done fairly accurately in mid-level canopy cover areas regardless of the imagery used. Confidence in the cover estimates did diminish slightly in areas where sagebrush cover was relatively sparse or extremely dense. Not all structural variables were quantifiable using the coarser imagery, due to constraints of spat ial resolution. In these instances the 0.3-meter CIR imagery was exemplified. The object-oriented approach enabled an automatic delineation of the range of variability within sagebrush stands and provided an interesting alternative to measuring sagebrush community structural attributes when compared with the more traditional pixel-based approach.

This research was intended to provide a resource for anyone working wi thin sagebrush ecosystems, including rangeland managers, wildlife biologists, or other remote sensors, specifically when decisions related to the appropriate selection of remotely sensed data for some intended management application is necessary. The evaluation of wildlife habitat for sagebrush-obligate species, the direction of fire management strategies and restoration efforts, and the ident ification of appropriate grazing areas are only a few of the potential applications of this work

Share

COinS