Abstract
Ensemble short-term rainfall-runoff prediction in urban river basin is presented. An ensemble rainfall prediction is built by perturbing initial condition of the extrapolation model. The five ensemble members are subsequently considered as uncertain input of the distributed hydrological model. GLUE method is used to calibrate the rainfall-runoff model and quantify the uncertainty. Four sensitive parameters of rainfall-runoff model are considered. The set of behavioral simulations develop an ensemble of flood prediction. This approach allows a cascading of uncertainty from rainfall prediction to the flood prediction. The methodology is demonstrated throughout case studies in Kofu urban river basin, Japan. Having demonstrated the plausible results, this approach could serve as a reliable and effective method for estimating the uncertainty range of short-term prediction of runoff dynamics for operational flood disaster prevention in urban area.