Characterizing Meadow Vegetation With Multitemporal Landsat Thematic Mapper Remote Sensing

摘要

Wet meadows are important biological components in the Blue Mountains of eastern Oregon. Many meadows in the Blue Mountains and elsewhere in the Western United States are in a state of change owing to grazing, mining, logging, road development, and other factors. This project evaluated the utility of remotely sensed data to characterize and monitor meadow vegetation for 32 meadows in the Wallowa-Whitman and Umatilla National Forests. The meadows had been previously classified into different plant community types based on the occurrence of specific indicator species of sedge (Carex spp). We analyzed the spectral signature of the sample meadows by using the Kauth-Thomas (tasseled-cap) transformation calculated for a temporal series of five Landsat thematic mapper (TM) images. The Landsat TM images were obtained for monthly intervals from April to September in 2000. We found that the sequence of Landsat TM scenes provided ample resolution to differentiate most of the plant communities examined. A larger sample size and additional field verification of the meadow vegetation data would have provided a more definitive evaluation of the methods. The multitemporal approach holds promise for monitoring change in wetland indicator species like the sedges, and for rapid characterization of grass, grasslike, and herbaceous vegetation over large areas. The technique also may be useful to detect the current location and spread of plants, especially those with distinctive vegetative or floral phenologies.