Integrated Geomorphic Approach for Floodplain Zonation Along an Anthropogenically Impacted Reach of the Narmada River, Central India

ABSTRACT We propose an integrated geomorphic framework for floodplain zoning along a 140 km stretch of the Narmada River, India (between Bargi Dam and Barmanghat). This framework utilizes Sentinel‐2 multispectral images (10 m), Copernicus Forest and Building‐Free DEM (FABDEM), and topographic profiles measured in the field using Differential GPS (DGPS). Following Central Water Commission (India) guidelines, we classified the floodplain into three zones. These include the No‐Go Zone (NGZ; 2–5‐year return period), Regulatory Zone (RZ; 25‐year return period), and Warning Zone (WZ; 100‐year return period). We delineated these zones entirely based on geomorphic features, topographic breaks, and riparian vegetation. The NGZ includes hydrologically connected depressions, swamps, and bank‐intact gullies with water‐tolerant riparian vegetation. The RZ comprises flood channels, large depressions, crevasse splays, and wetlands with lush woody vegetation. The WZ encompasses elevated terraces, paleochannels, and ravines with wide‐canopy, trunk‐root trees. We performed field validation through reconnaissance, local surveys, RTK‐DGPS surveys (< 2 cm vertical accuracy), and drone images (1–2 cm resolution). To assess this holistic geomorphological approach, a 1D HEC‐RAS model was set up for comparison along a 70‐km sub‐reach (Bhedaghat–Barmanghat). We merged the channel bathymetry measured using an Acoustic Doppler Current Profiler (ADCP) and FABDEM for the true representation of channel geometry into the model. We calibrated and validated the model using water‐level data recorded from 1977–2023 at a gauge station near Belkhedi. This model yielded an RMSE of 1.25 m and . Comparing the floodplain zonation obtained from geomorphic approach to the model, we found the value of Intersection‐over‐Union (IoU) for NGZ = 0.53, RZ = 0.42, and WZ = 0.31. The NGZ area demonstrates a consistently close agreement between the two approaches. However, flow regulation and anthropogenic modifications significantly decoupled the RZ and WZ in contemporary hydraulic inundation patterns. Hydraulic models underestimated the flood extents of the RZ and WZ due to dynamic terrain changes. Geomorphic analysis and local surveys enabled us to refine these flood zone boundaries by identifying land‐use impacts. These landform alterations lead to rapid flood pattern changes observed in the field survey. The geomorphic component does not rely on discharge data; we believe that the geomorphic framework directly applies to managing flood risk in ungauged or data‐scarce river corridors. The geomorphological zonation provides a scientific foundation of first‐order functional classification of floodplains for long‐term river corridor planning.