AbstractUsing Satellite Imagery to Map Irrigated Lands Track: Remote Sensing Imagery Author(s): Sharon Qi, David Litke, Alexandria Konduris This paper presents the techniques used to map irrigated lands in the 174,000-square-mile High Plains area using Landsat imagery. Eighty scenes from summer and early spring (nominal date 1990) were used. The area was subdivided into nine regions based on similar environmental characteristics. To remove spectral overlap, only the agricultural areas as defined by the National Land Cover data set were used for processing. The scenes were processed using a band-ratio method to create a brightness theme and enhance the vegetation signature. A threshold was selected that would be the brightness cutoff between irrigated and nonirrigated pixels. Overall, 81 percent of the pixels were classified correctly when compared to 1,000 square miles of ground-truth data. Comparisons were made with 1980 Landsat irrigated lands data to produce change maps. Sharon Qi U.S. Geological Survey Water Resources Box 25046, MS 415 Denver Federal Center Lakewood , CO 80225 USA Phone: 303-236-4882x310 Fax: 303-236-4912 E-mail: slqi@usgs.gov David Litke U.S. Geological Survey Water Resources Box 25046, MS 415 Denver Federal Center Lakewood , CO 80225 USA Phone: 303236-4882x242 Fax: 303-236-4912 E-mail: dwlitke@usgs.gov Alexandria Konduris U.S. Geological Survey Water Resources Box 25046, MS 415 Denver Federal Center Lakewood , CO 80225 |