A geobotanical approach to the tropical rain forest environment of the Carajás Mineral Province (Amazon Region, Brazil), based on digital TM-Landsat and DEM data

Abstract

Digital TM Landsat images integrated with elevation model variables were used to evaluate the potentiality of geobotany for lithological unit discrimination in the tropical rain forest environment of the Carajas Mineral Province, Brazilian Amazon region. The data set was analysed through digital image processing techniques (enhancements and non-supervised classification). The investigation has shown that the Up-Land rain forest vegetation in the area is mainly controlled by elevation and slope which reflect variations in the geology. Botanical verification has also indicated that the physiognomy (density and stratification), would be the most important vegetation attribute which influences the remote sensing responses. The research has provided useful information for the geological model of the area. Thus, by understanding the relationships among vegetation, terrain descriptors and geology, geobotanical remote sensing provides an additional tool for geological exploration in this kind of environment.