Flood Modelling, Mapping and Monitoring of Sparsely Gauged Catchments Using Remote Sensing Products

Efficient water resources management is highly dependent on the use of simulation modelling although the success of simulation modelling calls for the availability of sufficient hydrometeorological and other required data. Scarcity of hydrometeorological data limits the possibility of optimal water management using simulated information. Advances in remote sensing provide opportunities to access alternative data and offer new scopes for hydrological and hydraulic modelling of ungauged/ sparsely gauged catchments. This paper presents the use of global datasets in a cascade of hydrological and hydraulic models for flood mapping of the sparsely gauged Brahmaputra basin. TRMM rainfall estimates, corrected for bias, were used in calibration and validation of the hydrological model. Output of the hydrological model was used in a hydraulic model to produce flood maps. Flood inundation maps were compared with satellite imageries of flood events, and reasonably good match was observed. Uncertainty in hydrological models may arise from a number of sources among which the effect of an uncertain rating curve in calibrating the hydrologic model was assessed in this research. Fifty rating curves were sampled, and they were used to calibrate fifty hydrological models. The ensemble of runoff at different basin locations was used as upstream and lateral boundary condition of the hydraulic model to generate an ensemble of flood inundation maps. The study demonstrated that the ensemble of inundation map, compared to the inundation map produced with a single rating curve, provided more realistic flood map.