Remote Sensing and GIS-Based Lithological Discrimination in the Precambrian Shield of Kebbi State

Detailed lithological mapping of the Precambrian basement in Kebbi State using conventional field methods is constrained by limited outcrop in vegetated and agricultural terrains, making remote sensing and image processing techniques valuable supplements to field investigation for regional geological mapping programmes. This study investigates the capability of multispectral and hyperspectral remote sensing data for lithological discrimination in the Precambrian shield of Kebbi State. Advanced Spaceborne Thermal Emission and Reflection Radiometer multispectral imagery, Sentinel-2 data, and resampled ASTER thermal infrared data were processed to generate image composites, band ratio images, and principal component analysis decorrelation stretch outputs optimised for lithological discrimination. Spectral angle mapper classification and support vector machine supervised classification algorithms were applied using field-verified training sites representing gneiss, granite, migmatite, schist, and quartzite units. Classification accuracy was evaluated using a stratified random test sample yielding overall accuracy and kappa coefficients. The study compared the performance of different image datasets and classification approaches, identifying ASTER thermal infrared band combinations as most discriminative for granitoid versus mafic basement units. Resulting lithological maps were validated against Geological Survey of Nigeria base maps. Identified improvements in lithological delineation accuracy relative to existing maps are documented and the updated map is submitted to the Nigerian Geological Survey Agency. Keywords: remote sensing, lithological mapping, Precambrian, Kebbi State, ASTER.