Integrating Remote Sensing and HEC-RAS Hydraulic Modelling for Flood Inundation Mapping in a Lowland Peatland Watershed

Abstract

The Bukit Batu Watershed in Bengkalis Regency, Riau Province, is a lowland peatland area highly vulnerable to seasonal flooding caused by intense rainfall, river overflow, and extensive land-use change. This study aims to accurately map flood inundation by integrating remote sensing techniques and hydraulic modelling. Sentinel-1 SAR imagery was used to detect flooded areas through backscatter analysis, while Sentinel-2 optical data supported land-use and land-cover (LULC) classification using a Random Forest algorithm. DEMNAS elevation data were processed to obtain a refined digital elevation model that reflects the extremely flat peatland topography. Hydrodynamic simulations were conducted using HEC-RAS 1D/2D to model flood propagation, water-surface elevation, and inundation depth. Results indicate that land-use conversion, particularly the expansion of plantations and canal construction, has accelerated surface runoff and expanded flood-prone areas, especially in regions experiencing peat subsidence. The simulated flood extent shows strong agreement with SAR-derived inundation maps, achieving F1-scores of 0.71–0.79 and IoU values of 0.52–0.62. The integration of remote sensing and hydraulic modelling significantly improves flood mapping accuracy and provides a scientific basis for flood mitigation, watershed management, and spatial planning in the peatland areas of Bengkalis