Urban flood susceptibility modelling of Guwahati in India using GIS-based AHP-MCDM for sustainable disaster risk management

Rapid, often unplanned urbanisation and intensified monsoonal regimes have significantly escalated the frequency and magnitude of urban inundation in the South Asian Indo-Gangetic-Brahmaputra plain. This study develops a granular, ward-level flood susceptibility assessment for the Guwahati Municipal Corporation (GMC), a strategic urban gateway in North-East India characterised by a complex mosaic of low-lying floodplains and Precambrian hillocks. Utilising an integrated Geographic Information System (GIS) and Multi-Criteria Decision Analysis (MCDA) framework, sixteen hydro-geomorphological and anthropogenic conditioning factors were synthesised via the Analytic Hierarchy Process (AHP). Statistical independence of the parameters was verified through a multicollinearity assessment, yielding Variance Inflation Factor (VIF) values below 3.636. The hierarchical weighting identified monsoonal rainfall (18.20%), elevation (13.11%), and slope (11.57%) as the dominant drivers of inundation. Spatial analysis reveals that approximately 47.35% of the GMC diagnostic expanse resides within ‘High’ or ‘Very High’ susceptibility zones, with the most critical vulnerabilities concentrated in the western and southern administrative wards (Wards 1, 2, 9, 23, and 24). The model's predictive reliability was empirically validated using a historical flood inventory (2014–2024) and Receiver Operating Characteristic (ROC) curves, achieving an Area Under the Curve (AUC) of 0.909 for the prediction dataset. These findings provide a robust scientific foundation for transitioning from reactive disaster response to proactive, evidence-based urban governance, emphasising the strategic conservation of natural detention basins and the implementation of ward-specific mitigation frameworks to align with Sustainable Development Goal 11.