SPECTRAL DIFFERENTIATION OF SURFACE SOILS AND SOIL PROPERTIES - IS ITPOSSIBLE FROM SPACE PLATFORMS

Spectral radiance data from Landsat's Thematic Mapper (TM) sensor were used to evaluate its effectiveness in differentiating among surface s oils of similar properties and to identify the TM spectral band combin ation useful in estimating selected soil properties. Correlation, regr ession, and discriminant analyses were used in analyzing the data. The overall accuracy of the TM sensor in differentiating the surface soil s was 97.2%. The order of importance of the TM bands that contributed the most in differentiating the soils was TM 6, TM 4, TM 5, TM 2, TM 7 , TM 1, and TM 3, respectively. Significant correlations were found am ong the spectral radiance data and soil variables studied; however, th e amount of variance explained was quite low. The prediction equation for estimating soil properties accounted for less than 40% of the vari ability in the data. The low percentages were attributed to atmospheri c particles such as moisture, CO2, dust, etc. that affect the electrom agnetic energy sensed from airborne and space platforms. It was conclu ded that even though the TM sensor is adequate for differentiating amo ng similar soil types from the same soil order, the spatial resolution of the data (30 m) is still too coarse to be effectively used in gene rating equations for predicting soil properties such as clay percentag es, iron oxide, and organic matter content.