Prospectivity mapping of carbonatite-associated iron oxide deposits using an integration process of ASTER and Sentinel-2A multispectral data

Abstract

ABSTRACT Carbonatites and associated rocks are commonly known to host precious commodities, including rare earth elements (REE), niobium, and phosphates. They form ring-shaped complexes frequently found as small occurrences in hard-to-reach areas, making their exploration a challenging task using conventional methods. Remote sensing has been increasingly regarded as a time- and cost-saving method to map and identify new potential zones, meeting the growing need for mineral resources. In this study, we use multispectral data by combining information from ASTER and Sentinel 2A sources to delineate and map carbonatite-associated REE- and Nb-enriched iron oxide deposits. The Moroccan Oulad Dlim massif, at the NW margin of the west African craton, where carbonatite occurrences are reported, was selected as a study area. A fuzzy gamma approach was adopted to combine information from several image processing techniques, including mineral indices, principal component analysis (PCA) and minimum noise fraction (MNF). A concentration-area (C-A) fractal model is applied together with the prediction-area (P-A) plot for objective anomalies classification and prediction ability assessment for each fuzzy integrated map. Subsequently, a prospective map that predicts and defines potential areas was produced through the integration of Sentinel- and ASTER-derived evidential maps. Validation was carried out via field surveys coupled with in-situ portable X-Ray Fluorescence analysis and laboratory mineralogical characterization. The Results were consistent with field data, reaching a prediction rate of 90%. They reveal the importance of the investigated structures as potential sites for REE, Nb, and phosphates. These outcomes are encouraging and provide evidence that data fusion (ASTER-Sentinel 2A) is an effective and helpful approach for geological exploration programmes.