An Integrated Approach to Flood Risk Management in the Panchganga River Basin: A Comparative Analysis of AHP and AI-Based fuzzy logic Susceptibility Models

Abstract

This study presents a multi-criteria decision analysis (MCDA) model using the Analytic Hierarchy Process (AHP) to map flood susceptibility in the Panchganga River basin in western Maharashtra, India. The objective is to evaluate the AHP model's utility and performance by validating its output against historical ground conditions and by conducting a comparative analysis with data-driven Artificial Intelligence (AI) models. The methodology employed AHP to systematically assign weights to 13 flood-conditioning factors, including Rainfall, Elevation, and Land Use/Land Cover (LULC), based on expert judgment. The pairwise comparisons yielded a final Consistency Ratio (CR) of 9.2%, indicating acceptable logical coherence in the expert evaluations. The comparison showed that although AI models typically achieve higher predictive accuracy (e.g., AUC-ROC scores exceeding 0.90), the AHP approach provides better transparency and works especially well in environments with limited data. The report's conclusion suggests a hybrid strategy for sustainable flood mitigation in the Panchganga basin, utilizing AI for high-accuracy spatial modeling and AHP for transparent factor prioritization. The fuzzy model can successfully forecast flood severity levels and provide early warning signals, according to the results. The model spatially delineates flood-prone zones when paired with GIS, improving the region's resilience and readiness. The study comes to the conclusion that fuzzy logic provides a transparent, flexible, and computationally effective method for real-time flood forecasting in areas with limited data, such as Kolhapur.