Statistical-dynamical methods for scale dependent model evaluation and short term precipitation forecasting (STAMPF)

 

Application PHASE 2

Summary

 

The aim of the project is a scale dependent evaluation of precipitation forecast of the DWD weather prediction models (LME/LMK) in relation to dynamical and statistical parameters as well as cloud properties. The Dynamic State Index (DSI), scaling exponents of distribution functions, information entropy, cloud type and coverage are the evaluation parameters. The forecast skill of the DSI as a threshold parameter for precipitation processes has to be explored further in the different temporal and spatial model resolution, including data of the vertical humidity profile. The evaluation focuses on interactions between synoptic and convective scales, which are often the reason for extreme precipitation events. Besides, the separation of convective and stratiform components will be improved by using higher resolved Meteosat-8 data. For the evaluation of the LME/LMK, numerical analyses of comparable scales will be performed for Central Europe (7 km), Germany (2.8 km) and especially for Berlin (1 km) in a temporal resolution of 5 minutes. The analysis scheme developed in our project will be improved by combining cloud statistics from Meteosat-8 and radar information. In addition the displacement of precipitation fields along trajectories will be calculated considering the stochastic structure of precipitation and clouds fields. In a further step the importance of turbulent fluxes for the intensity of precipitation processes has to be tested in the frame of COPS field experiments.