Submitted to: Society for Range Management
Publication Type: Abstract Only
Publication Acceptance Date: February 13, 2000
Publication Date: February 13, 2000
The Landsat Multi-spectral Scanner (MSS) acquired data from Landsat 1 in 1972 and ended in the mid 1990's with data from Landsat 5. Images from the 1970's, 1980's and 1990's are available from the USGS EROS Data Center as the North American Land Characterization Dataset. Standard change detection techniques in remote sensing can not be used for semi-arid regions because of the high year-to-year variability in precipitation. We developed a technique using the Advanced Very High Resolution Radiometer (AVHRR) in which the Normalized Difference Vegetation Index (NDVI) is adjusted for grassland and sagebrush rangelands based on the amount and timing of rainfall. This technique was applied to 15 MSS scenes covering most of Wyoming. We found the vast majority of rangelands had either improved or remained the same over the twenty-year period. Areas that decreased were usually associated with wildfires. Furthermore, we found riparian zones (defined as a swath 10 pixels around rivers and creeks from a vector GIS coverage) had to be separated from the above analyses; NDVI of most riparian areas either increased or remained the same over the twenty-year period. Since NDVI is related to the amount of vegetation and not the species present, these data can not be used for assessment of condition and trend. However, these data can be used to prioritize field checks and serve as a baseline for more detailed studies with current and future sensors.