Digital classification of contact metamorphic rocks in Extremadura, Spain, using Landsat thematic mapper data

Abstract

Landsat Thematic Mapper (TM) images of an extensively cultivated part of Extremadura, Spain, have been used to distinguish soil developed on contact metamorphic rocks in aureoles around late Hercynian granitic plutons from soil formed on stratigraphically equivalent Late Proterozoic slate and metagraywacke that has been regionally metamorphosed to the greenschist facies. Reflectance spectra of contact metamorphic soil have lower reflectance, especially in the 1.6 μm wavelength region, and weaker Al‐OH, Mg‐OH, and [Formula: see text] absorption features than do spectra of the slate‐metagraywacke soil. These spectral differences are attributed to highly absorbing carbonaceous material in the contact metamorphic soil that was subjected to high temperatures during emplacement of the plutons. These spectral reflectance differences are evident in a density‐sliced TM band 5 image, in color‐ratio composite images that incorporate TM 4 : 5, 4 : 3, and 3 : 1 ratios, and in principal‐component composite images. Digital classification of the numerous tilled, vegetation‐free fields was used to map the contact metamorphic soil in an August, 1984, TM scene of the Caceres study area. First, TM 4 : 3 was used to identify these fields. Then ranges of TM 5 and TM 3 : 1 values were determined for selected tilled fields within and outside the contact aureoles. Field evaluation of a classification map based on TM 5 plus TM 3 : 1 shows more extensive aureoles than published geologic maps and few misclassified areas. Similar results were achieved using TM 4 : 5 instead of TM 5. This approach was subsequently used to map two linear zones of contact metamorphic rocks in the San Nicolas mine area where only two small exposures of granite have been documented. Landsat Multispectral Scanner (MSS) images can also be used to map contact aureoles in the study areas, but extensive field evaluation is required because of more frequent misclassification.