Flood Susceptibility Analysis Using MCDA-AHP and Geoinformatics: A Case Study of the Kosi River in Supaul District, Bihar, India

Floods are among the most recurrent natural hazards in northern India. The low-lying floodplains of the Kosi River are repeatedly inundated owing to high sediment loads, embankment fragility, and intense monsoonal rainfall. This study develops a district-scale flood susceptibility model for Supaul district, Bihar, by integrating the Analytic Hierarchy Process (AHP) with Multi-Criteria Decision Analysis (MCDA) in a Geographic Information System (GIS) environment. Ten thematic layers - slope, elevation, aspect, curvature, drainage density, distance from river, rainfall, soil texture, geology, and river-flow direction-were derived from SRTM DEM (30 m), IMD rainfall records (mm year⁻¹), and government soil and geology databases. Land use/land cover from the ESRI Global LULC product (10 m, 2021) was used only as an exposure layer; it was not reclassified and carries no weight in the analysis. Pairwise comparisons were performed by a panel of five domain experts (academic and government practitioners with at least seven years of experience in flood-hazard assessment and GIS), and consensus weights were derived using Saaty’s 1-9 scale. The final composite susceptibility map produced by weighted overlay analysis in ArcGIS Pro 3.0 shows that roughly 2,382.44 km² (≈97.6%) of the district is in medium-to-high susceptibility classes; the low susceptibility class occurs only in the slightly elevated terrain of about 56.21 km². High susceptibility areas are mainly located in the low-lying tracts near active and abandoned Kosi channels, while medium susceptibility areas are mainly located in the agricultural blocks of Tribeniganj and Pipra. The pairwise matrix resulted in λₘₐₓ = 10.132, CI = 0.015, and CR = 0.010 (< 0.10), which is in accordance with Saaty's criterion. The validation with Modified Normalized Difference Water Index (MNDWI) and NDWI flood extents derived from multi-temporal Landsat imagery (1975-2018) revealed good spatial consistency between flood susceptibility classes and historical inundation. Since the medium and high classes cover most of the district, the comparison with village-level flood records is presented as qualitative support rather than as proof of predictive skill. The findings can be used to support the United Nations Sustainable Development Goals (SDG 1, 11, 13 and 15). The evidence base has recommended that mitigation actions are prioritised, early-warning systems are strengthened, and infrastructure is designed to be climate resilient in Supaul district's eleven administrative blocks.