Evaluation of the Capability of Landsat MSS Data for Mapping Landforms in Arid Regions: A Case Study in the Centre of Iran

This research was performed to evaluate the potentials of Landsat MSS data for mapping land features in arid zones of southeastern Esfahan, Iran. Databases of the area were formed using all available relevant maps and reports which were supported by fieldwork. A supervised image classification approach was used and thirty-two training areas were applied. Separability of the spectral classes was examined using feature space plots of imagery data and self-classification of training areas. The accuracy of the classification was examined by using test and random pixels. The results show the potential of Landsat data for the discrimination of landforms and zones of the playas. Sand deposits (deflated sands and Barchans) expressed different spectral reflectance which could be due to mineralogy of these features. Soil classes differing in moisture content and salinity located on the soil line and limestone classes located along the soil line. The applied imagery data disable to discriminate Barchans from Bare soil I and Andesitic fans from Grey limestone. After image classification the spectral classes were merged to form landforms. The main landforms were mountains, piedmonts, sand dunes, valleys and playas. The results indicate that integration of field observation and supervised classification can compensate for the lack of detailed topographic maps in some areas.