A 81,000-ha rangeland area in Kenedy and Willacy Counties, Texas, was used to test LANDSAT-2 MSS data from October 17 and December 10, 1975 for inventorying rangeland and various other land-use categories.
Computer land-use classification percentages of land cover for each overpass were compared with photo-estimated percentages from a ground correlated 1:100,000 scale LANDSAT color composite print. We found a highly significant correlation (r = 0.977**) between the photo- and computer-estimated hectarages for the October LANDSAT-2 overpass. The correlation was not significant for the December overpass largely because about half of the most extensive rangeland category (mixed brush) was misclassified as grassland, probably because the woody species were dormant and freeze damage had weakened the herbaceous vegetation reflectance.
Computer estimates of level I land-use (rangeland, wetland, agricultural land, water, and barren land) hectarage from both overpasses resembled photo-estimated hectarages, indicating the feasibility of estimating level I land-use categories in either October or December. Computer estimates of level II land-use (grasslands, mixed brush, and live oak rangelands) hectarages agreed with photo-estimated hectarages only in October, indicating that living vegetation is needed to spectrally discriminate between level II rangeland categories.
Date of this Version