In recent years considerable progress has been made in numerical weather prognosis. Special progress has been made in doing local forecasts up to five days of temperature, wind and atmospheric pressure and al so o f the weather det erminant flow s ystems. In contrast, the local prognosis of precipitation (liquid and ice phase) has not been improved. T his circumstance has lead to the DFG p rogram “Quantitative rainfall prognosis”. I t co vers broadly base d activities that ai m on the i mprovement o f t he knowledge on t he pr ocesses of r ainfall f ormation and t heir num erical pr ognosis. The main objective is to improve the routine prognosis of the German Weather Association (DWD). The program covers the modeling of microphysical processes as well as the description of essential meteorological conditions in different temporal and spatial scales. Especially, co nvective c loud sy stems that ar e often responsible for e xtreme r ainfall situations are studied. In a G ermany-wide monitoring campaign in the year 2007 comprehensive measurements are conducted. The gained information and data will be used to improve process description and to support model evaluation. The st udy at hand describes the anal ysis and pr ognosis of temporally (5 m in) and spatially (500 m) highly distributed rainfall data for the Berlin area. The data will be used in the frame of the EVA project of Kompetenzzentrum Wasser Berlin to analyse and evaluate the potential of online rainfall measurement and forecast to support the operation of wastewater pump stations.

In Berlin, Germany, the demand for enhanced protection of the environment and the growing economic pressure have led to an increased application of control concepts within the sewage system. A global control strategy to regulate the pumpage of the combined sewage system to the treatment plant was developed and evaluated in a theoretical study. The objective was to reduce CSO. In this paper an extension of the existing control algorithm by information from online rainfall measurement and radar nowcasting is described. The rainfall information is taken into account by two additive terms describing the predicted volume from rainfall runoff. On the basis of numerical simulation the potential of these two complementary forecast terms in the global control algorithm to further reduce CSO is evaluated. The investigations are based on long-time simulations that are conducted with the dynamic flow routing model InfoWorks for three subcatchments of the Berlin drainage system. The results show that at the current Berlin system a CSO reduction of only 0.8% is possible. The effect of the forecast terms is limited by operational constraints. Limits are set to both, the delivery from each individual pump station and the total pumpage to the treatment plant.