Distributed model flow sensitivity to uncertainty in radar-rainfall input

The aim of the current research is to assess the sensitivity of the range of flow simulations produced by a distributed hydrological model, suitable for use in an operational environment, under various definitions of uncertainty in the radar rainfall estimates. Monte Carlo sensitivity analysis is used, sampling from various assumed distributions of mean areal precipitation error, and the variation in resulting ensembles of flow simulations is measured using a normalized range of 10% high minus 10% low flows. The results show that the average flow ensemble range over all events in a given sub-catchment possesses significant spatial-scale dependence for all assumed radar rainfall uncertainty characterizations.