HR: 0800h
AN: B51C-0967 [Abstracts]
TI: Monitoring Rangeland Health With MODIS Vegetation Index Data
AU: * Brown, J F
EM: jfbrown@usgs.gov
AF: SAIC, EROS Data Center, 47914 252nd Street, Sioux Falls, SD 57198
United States
AB:
Rangelands cover approximately one third of the land area of the conterminous U.S. These lands supply much of the forage for
the U.S. cattle industry. Large area monitoring of these vast expanses of range has proved challenging since most of these
lands are in the western U.S., are relatively sparsely populated, and are not well covered by meteorological weather
stations. Improvements in the spatial and temporal precision of rangeland health information would be useful both for the
cattle industry and for scientific studies of soil erosion, water runoff, ecosystem health, and carbon cycling. Optical
multispectral remote sensing data from satellites are an objective source of synoptic, timely information for monitoring
rangeland health.
The objective of this study is to develop and evaluate a method for measuring and monitoring rangeland health over large
areas. In the past, data collected by the Advanced Very High Resolution Radiometer has proved useful for this purpose,
however the basic 1 km spatial resolution is not ideal when scaling up from ground observations. This study assesses MODIS
250 meter resolution vegetation index data for this purpose. MODIS data not only have finer spatial resolution and improved
geolocation, but they also exhibit enhanced vegetation sensitivity and minimized variations associated with external
atmospheric and non-atmospheric effects. Ground data collected over 51 sites in western South Dakota over four years are
used as training for regression tree models of range health. Range health maps for the growing season derived from the
models are presented and evaluated.
DE: 9350 North America
DE: 0400 Biogeosciences
DE: 0933 Remote sensing
SC: Biogeosciences [B]
MN: 2004 AGU Fall Meeting