Remotely Sensed, Geophysical and Geochemical Data as Aids to Mineral Exploration in Bahia State, Brazil

This thesis aims to establish methodological procedures to implement Geographical Information System (GIS) techniques applied to mineral exploration, in particular gold associated with an Archaean greenstone belt located in a semi-arid climate: - the Contendas-Mirante greenstone belt which is located in the southeastern part of Bahia State in northeastern Brazil. To achieve this goal a systematic approach was formulated. It begins with the correlation between Brazil and Africa. The geographical position of adjacent age-provinces, the coincidence of structural trends, their continuity in a pre-drift reconstruction and the similarities in the geochronological patterns, all indicate their correlation. A particular Brazilian portion of the South America Platform, termed the Sao Francisco craton, (SFC) became stabilised in the late Proterozoic (1800 Ma). In this crustal block the Archaean rocks include medium-grade (Gaviao Block) and high-grade (Jequie Block) terrains which form the infrastructure for a number of volcano-sedimentary or sedimentary sequences. One of these volcano-sedimentary sequences, the Contendas-Mirante belt (CMB), is located along the junction between the Gaviao and Jequie Blocks. It is divided into Lower Unit (mainly volcanogenic) and Upper Unit (essentially clastic). In the CMB the plutonic granitoid associations exhibit geochemical characteristics compatible with a continent-continent collision. The available data set for this area includes topographic and geological maps, stream sediment survey, Landsat TM and radar data, and airborne magnetic and radiometric surveys. The most suitable way for dealing with this enormous amount of data leading to mineral exploration exploration, is via GIS. Gold deposits can be distinguished according to depositional and genetic models. In this work I concentrate on the depositional model which can be divided according to the structural setting and lithological associations. From the available data set, the use of the Landsat TM data in lithological mapping was restricted by the vegetation and soil cover. However, TM data provided useful information on regional structures. Radar data were valuable for structural information and as ancillary data combined with radiometric data for lithological discrimination. The stream sediment reconnaissance was interpreted via a raster model. Several sets of very well correlated elements were defined: Fe-Cu-Co-As, Fe-As-Cr-V-Ba-Cu, Fe-Ni-Cu, Fe-Mg-Ti-Cu, Fe-Cu-As-Pb and Fe-Ba-As-Pb. From these element associations anomalous areas were defined. The magnetic data were useful for both the selection of zones with high magnetic fields and the detection of magnetic lineaments. The radiometric data in both small and large scale were valuable for lithological discrimination. The selected and diversified GIS functions generated a final product, called data fusion, which is far more sophisticated than data integration. The four selected areas for gold, which represent 12% of the total study area, and due to the convergent geological factors, have good potential for gold