Geothermal potential assessment using Integrated GIS and machine learning approaches

Growing global demand for clean and reliable energy highlights the need for sustainable alternatives that can also contribute to water scarcity mitigation in semi-arid regions. Geothermal energy provides a constant, low-carbon resource capable of enhancing long-term energy security and environmental resilience. This study develops an integrated, data-driven assessment of geothermal potential in northern Tunisia, applying a multidisciplinary framework that combines Geographic Information Systems (GIS), Remote Sensing (RS), Multi-Criteria Decision-Making (MCDM), Principal Component Analysis (PCA), and Kohonen Self-Organizing Map (SOM) clustering. Twelve thematic layers including geomorphological, geological, hydrogeological, structural, and climatic were analyzed, with weighted indices assigned to key factors, such as lithology, aquifer depth, fault and lineament density, slope, precipitation, and land surface temperature. Results show that approximately 48% of the region exhibits favorable to highly favorable geothermal conditions, with 22% classified as very good and 26% as good. The highest-potential zones align with the Tell Atlas structural corridor, characterized by intense faulting, permeable lithologies, deep aquifer systems, and active hydrothermal circulation, while low-potential zones occur in compact, poorly fractured aquifer units. The integrated approach offers a robust and reproducible method for mapping geothermal resources and supporting sustainable energy development in northern Tunisia.