Applying COTREC-derived rainfall forecasts to the rainfall-runoff model PDM—estimating error sources

The proposed contribution focuses on short-term rainfall forecasts for hydrological applications. The radar echo tracking algorithm COTREC (COntinuity of TREC vectors), using pattern recognition, provides information about movement and change in intensity of the precipitation. The rainfall forecast derived with COTREC will be used as input to the rainfall-runoff model PDM (Probability Distributed Model). To evaluate the quality of the rainfall forecast a parameter scheme able to distinguish between errors in the position and the intensity of the predicted precipitation is used. While investigating the potential utility of COTREC for flood forecasting we will analyse the dependence of flood forecasts on the quality of the position and intensity rainfall forecasts separately. To investigate the importance of refined intensity forecasts a comparison with the U.K. Met. Office GANDOLF system will be carried out for selected convective storms.