the efficiency of landsat tm and etm+ thermal data for extracting soil information in arid regions

a knowledge of soil surface conditions, especially desert crust, salt crust and desert varnish is useful for improving classification of remotely sensed data. desert crust can generate high levels of reflectance, similar to those areas with high salt concentration and non-saline soil. therefore, soil surface crusts might bias thematic remote sensing of soils. in this study, we evaluated the efficiency of the thematic mapper (tm) and enhanced thematic mapper plus (etm+) reflective and thermal bands in detecting crusted surfaces and soil salinity conditions. the study areas were ardakan, damghan, lut desert, qom, and abarkooh which are located in arid regions of iran. to assess the landsat tm ther-mal data for detecting land cover types, the following steps were taken: 1) determination of correlation coefficients between tm wavebands, 2) assessment of the relationship be-tween tm thermal and tm reflective bands on land cover types, 3) assessment of the rela-tionship between soil salinity and tm digital numbers (dn), 4) two dimensional feature space (fs) analysis of the training samples, 5) field sampling, 6) image classification and accuracy assessment, and 7) comparison of surface reflectance of different soil surface types. the results show that the trend of correlation coefficients of tm6 with reflective bands is completely different from the correlation between reflective bands. the behav-iour of the thermal band on gypsiferous soils is completely different from that on saline soils. moreover, with an increasing correlation between soil salinity and reflective bands, the correlation between soil salinity and the thermal band decreases. in image classifica-tion, the thermal band improved the separability of the crusted and gravely classes. therefore the tm/etm+ regions of the electro-magnetic spectrum have complementary capabilities for spectral separability of gravely and crusted surfaces. in general, selection of the tm/etm+ thermal band combination is an important step for classifying the re-mote sensing data and for securing class separability of gravely and crusted surfaces in arid regions. we also concluded that tm/etm+ thermal bands may contain information complementary to the tm/etm+ reflective bands and therefore this combination of the tm/etm+ thermal and reflective bands provide a viable method for soil salinity studies in arid regions.