Hybrid Data-Driven and Enhanced AHP Framework for Flood Susceptibility Mapping

Flood susceptibility mapping plays a vital role in understanding and mitigating flood hazards, particularly in rapidly urbanizing regions where land-use and climate variability intensify runoff and exposure. Developing reliable susceptibility maps enables planners and decision-makers to enhance resilience, prioritize mitigation strategies, and design future-proof urban infrastructure. The Analytical Hierarchy Process (AHP) is widely applied in multi-criteria flood assessment as it provides a systematic framework to determine the relative importance of topographical and environmental factors affecting flood susceptibility. However, traditional AHP relies on expert judgment or values adopted from previous studies; these subjective weights vary across regions and reduce the accuracy and consistency of susceptibility zonation. The present study establishes a data-driven framework to improve AHP weight determination through machine learning and objective evaluation techniques. The coastal region along Jakarta Bay, Indonesia, which was severely impacted by the extreme flooding event of late December 2019 and early January 2020— caused by exceptionally intense monsoon rains and widespread surface runoff—was selected as the case study. Multiple geospatial layers were incorporated, including DEM, slope, curvature, aspect, TWI, TRI, SPI, STI, distance to river, NDVI, LULC, soil lithology, and rainfall frequency. Four complementary categories of methods were utilized to derive and refine AHP weights which include (1) probabilistic approaches (FR, WoE) and (2) statistical approaches (LR, GAM) and (3) objective weighting techniques (CV, Shannon Entropy, Entropy–CRITIC hybrid) and (4) machine-learning algorithms (RF, XGBoost, CatBoost, AdaBoost, SVM). The proposed hybrid framework enhances AHP objectivity through systematic integration of these methods which creates a solid base for flood susceptibility mapping in urban areas. The resulting susceptibility assessment show improved reliability, transparency, and spatial consistency, which enables planners to make evidence-based decisions for flood-risk management and long-term urban resilience development.