Prioritising Geothermal Resilience Strategies in Hazard‐Prone Regions Using Geospatial Machine Learning and Fuzzy <scp>AHP</scp> – <scp>VIKOR</scp>

ABSTRACT Climate change is increasing extreme events and making energy resilience a key planning goal in climate‐sensitive regions. Geothermal energy can support resilience because it provides stable and dispatchable power and heat. However, geothermal deployment is often constrained by geological uncertainty, induced seismicity concerns, groundwater protection needs and local acceptance. These constraints are difficult to manage because geothermal decisions depend on multi‐source geoscientific evidence, including geophysical, geochemical and geospatial datasets, which are not easy to combine using simple decision tools. This study develops a hybrid AI–MCDA framework that links geospatial machine learning (ML) with fuzzy multi‐criteria analysis. ML is used to convert heterogeneous datasets into comparable, map‐based indicators, such as geothermal potential signals, landslide susceptibility and hydro‐environmental sensitivity indices. These indicators support fuzzy AHP to derive criteria and sub‐criteria weights from expert judgements under uncertainty. The study then applies fuzzy VIKOR to rank geothermal resilience strategies through a compromise solution based on overall performance and worst‐case regret. The results show that strategies focused on monitoring, evaluation and traffic‐light risk governance rank highest, followed by data integration and information‐sharing measures. Overall, the framework strengthens transparency, reduces decision uncertainty and supports practical strategy selection for geothermal development aligned with resource sustainability and energy resilience goals.