REMOTE SENSING AND GEOCHEMICAL INTEGRATION FOR MINERAL PROSPECTIVITY MAPPING IN COMPLEX GEOLOGICAL TERRAINS

Abstract

In this study the integration of remote sensing and geochemical methods for mineral prospectivity mapping in a complex geological terrain has been investigated. Spatially referenced observations, which incorporate remote sensing, geochemical, geological and structural data, were used in the research, with 300 observations used. Data from remote sensing were analyzed for lithological variations, hydrothermal alteration, Fe oxide, clay mineral response, and structural lineaments; and geochemical observations were assessed for elemental anomalies and multiple element associations. The evidence layers were standardised and unified in a GIS-based framework and mineral prospectivity classes were created using a Random Forest model. Of the 186 observations, 62.0% exhibited hydrothermal alteration signatures, 195, 65.0% were related to structural lineaments, and 168, 56.0% were related to geochemical anomalies. The 87% observations classified as high prospectivity, 29.0% were classified as very high prospectivity. For the integrated model, predictive evaluation results were: accuracy of 91%, sensitivity of 89%, specificity of 93%, and precision of 91%, F1-score of 90%, and ROC-AUC of 0.95. Of 115 mineral occurrences known, 90 (78.3%) were in the high and very-high prospectivity areas. Results showed that the integration of remotely sensed alteration, structural indicators and geochemical data showed greater potential to provide a more holistic approach to identification of prospective mineralization zones than any single data-set. It provided practical guidance for prioritizing exploration, field verification, geochemical sampling, geophysical survey, and drilling in geologically complex areas. More recently, studies have confirmed the value and feasibility of incorporating multiple data sources from geoscience to aid in targeting mineral exploration.