Identification and Mapping of Sagebrush/Grass Successional Stages with Landsat Thematic Mapper Data at Yakima Training Center, Washington
Abstract
In the future, remote sensing technologies will become an increasingly important and valuable tool for military land managers in semiarid regions. These technologies, when combined with field samples, have the potential to accurately monitor rangeland trends from year to year with smaller monetary investments compared to field sampling exclusively. This research attempted to identify and map successional changes on semiarid rangelands at Yakima Training Center, WA, using remote sensing techniques by developing a model derived from analysis of dependent and independent variables chosen from field surveys of vegetation and geomorphic data, along with the interpretation of Landsat TM data. Preliminary results based on small data sets separated by elevation and slope direction showed both low and some reasonable R2 values, including some R2 near 0.70. The removal of elevation and slope direction and consideration of multicollinearity and outliers and influentials provided generally significant relationships among dependent and independent variables. Significant relationships between multiple dependent and independent variables were also identified using canonical correlation analysis. Variability among the releves, collection of field vegetation and soil data over the entire summer including many phenophases, and the correction of the raster radiance values for topography were assumed to be factors that may have reduced the predictive capabilities of the techniques investigated.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 2004
- Accession Number
- ADA422424
Entities
People
- Craig R. Leedy
- Paul T. Tueller
- Scott A. Tweddale
- Wei Gao
Organizations
- Engineer Research and Development Center