Integration of multisource data to support the identification of lateritic regolith in Eastern - Bahia, Northeastern Brazil

Abstract

This work used multi-source data integration techniques (gamma-spectrometry, magnetometry, SRTMaltimetry) to identify areas favorable for the occurrence of well-developed lateritic regolith in the eastern region of the state of Bahia. The variables observed to define target potential were the high eTh/K and eU/K ratios and altitudes (obtained from SRTM). Based on these parameters, Boolean and fuzzy logic were applied to produce favorability maps. The best results were obtained under the fuzzy model (γ = 0.7) with a hit accuracy in the areas for potential laterite occurrence of 97.4% and a kappa value of 0.5, consisting of 118 control points obtained through fieldwork. The cross-referencing of the fuzzy image (γ = 0.7) with the magnetometry, total gradient (ASA-Analytical Signal Amplitude) suggested the predominance of more felsic protoliths in the most favorable areas. The lateritic index [IL = (eTh*eU)/ K2)] was also applied, demonstrating good correlation with the areas determined by the Boolean and Fuzzy models. Mineralogical associations (clay minerals and/or iron oxides) of the target areas were estimated by the Crósta Technique in OLI/Landsat-8 images. The processing results were integrated in GIS environment, together with control points and data found in the bibliography (e.g. geological map, vertical electrical profile, drill-holes, occurrence of Fe and Al, lateritic crusts, geochemical anomalies). The observed correlation between the generated models and the direct data validates the effectiveness of the techniques used.