A novel integrated GIS-AI framework for optimal CSP plant site selection: a multi-criteria approach under climate change scenarios in Bushehr, Iran

Optimal site selection for Concentrated Solar Power (CSP) plants remains a significant challenge in renewable energy planning, particularly in regions facing climate change impacts. This study addresses this challenge by developing an integrated framework for CSP site selection in Bushehr Province, Iran. We employed Geographic Information Systems with fuzzy multi-criteria decision analysis to evaluate 18 criteria across six categories (climatic, topographic, infrastructural, environmental, socio-economic, and technical). The Analytic Hierarchy Process determined criteria weights, while climate projections were generated using the RegCM4 model under the RCP 4.5 scenario. Machine learning algorithms, specifically Random Forest and Convolutional Neural Networks, enhanced spatial prediction accuracy by 12.7 % compared to conventional methods. Results identified 5.37 % (approximately 1246 km2) of the province as highly suitable for CSP development, with these areas demonstrating positive economic returns under Monte Carlo simulation-based Cost-Benefit Analysis. This framework’s primary innovation lies in its seamless integration of spatial analysis, climate modeling, and artificial intelligence techniques to support sustainable energy planning in arid regions experiencing climate variability.