Continuous synthetic precipitation for Germany for design in urban hydrology

Results from a research project are presented, dealing with the analyses of precipitation models for the generation of stochastic rainfall for urban hydrology. Two selected rainfall models are presented and compared. The models are regionalised for Germany using data from about 800 rainfall stations. They are validated on a representative subset of 45 stations based on precipitation statistics and simulated discharge in sewer systems. The results show the superiority of synthetic rainfall compared to a practice scenario where the nearest neighbour station always serves as data provider for the target location. The best results are reached using a mixture of data from both stochastic rainfall models leading to a compensation of errors. Recommendations for the required time series length for practical design are given. The synthetic rainfall data are now available on a web platform for practical applications for any location in Germany.