Rastreamento de unidades ricas em ferro no norte de Minas Gerais: modelamento de dados aerogeofísicos e de sensores remotos

• Ano de defesa: 2020
• Autor(a) principal: Eliane Voll
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; Brasil; IGC - DEPARTAMENTO DE GEOLOGIA; Programa de Pós-Graduação em Geologia; UFMG
• Palavras-chave em Português: Rochas ricas em ferro; Formação ferrífera; Modelagem espacial; Lógica Fuzzy; Grupo Macaúbas; Ferro; Minas Gerais; Modelagem de dados; Lógica difusa
• Link de acesso: http://hdl.handle.net/1843/34221; https://orcid.org/0000-0002-0490-7713

The thesis presents the main results of a scientific research that aimed to create an integrated model for tracking iron-rich rocks in northern Minas Gerais, a region with occurrences of ferruginous metadiamictites (Fe = 10-15 wt%) and diamictitic iron formation (Fe > 15 wt%) that are largely hidden by extensive Cenozoic tablelands (plateaus). The research extends to a case study, on a local scale, which aimed to test the model feasibility for different scales. In geological studies on regions lacking rock exposures and/or covered by dense vegetation, the use of indirect mapping methods is useful, including geophysical and remote sensing data with GIS-based spatial analysis, facilitating the treatment of spatial data and its integration for modeling, analysis, and interpretation of the studied phenomenon. The study region comprises the geomorphological-geological domains of the easternmost São Francisco Craton and the western Araçuaí Orogen. The eastern domain of the focused region corresponds to the large plateaus supported mainly by units rich in metadiamictite and quartzite of the Macaúbas Group, including the iron-rich metadiamictites (10-60 wt% Fe) of the Riacho Poções Member. However, the iron-rich rocks are largely hidden under covers on tablelands, showing only a few well-preserved outcrops from lateritization, a weathering-related process that can reach dozens of meters deep in the studied region. Along those sites, iron tends to be more concentrated in laterites developed on iron-rich rocks than in lateritic soils formed on other rocks. To track the iron-rich rocks, several products were generated based on magnetic and gammaespectrometric data from aerogeophysical surveys and Landsat 8 satellite images. To produce the final model, the selected data were combined using spatial analysis by knowledge-driven fuzzy logic methodology. Fieldwork data and compiled information, including data from ANM (National Mining Agency) and SAM (Sul Americana de Metais S.A.), were used for the development and validation of the model. Being coherent with previous knowledge on the region, the final model has proved to be efficient from regional to local scales, enhancing the accuracy to outline favorable areas for the occurrence of iron-rich rocks even in the large tableland domains.