Uncertainty analysis in hydrological modeling has become an essential step in the scientific interpretation of model results and a useful tool to support decision making. Among many uncertainty sources in the modeling practice, uncertainties in precipitation estimation play an important role since it is the main driving force for other hydrological processes. The present study demonstrates a statistical method for generating radar rainfall realizations that account for the uncertainties in radar-based quantitative precipitation estimation (QPE). The random sampling technique used to generate stochastic uncertainty fields is based on sequential Gaussian simulation. The hydrological impact of the uncertainties in radar QPE is analyzed by propagating the rainfall ensemble through a distributed and integrated water resources model. The study shows that the uncertainty of the simulated stream discharge depends on the intensity of the rainfall input signal. The coefficient of variation is calculated for simulated stream discharge and groundwater recharge at subcatchments with various sizes. The results reveal strong scale dependency showing higher variations of hydrological uncertainties at smaller catchments, especially for catchment areas smaller than 50 km 2. The uncertainties from precipitation input are generally amplified in the hydrological model. This effect is less obvious for groundwater recharge but rather substantial for stream discharge, where the coefficient of variation increases by a factor of three.
- Programme Area 2: Water Resources