Numerical Weather Prediction systems lack skill in predicting convective events one day in advance. This project aims in improving the skill of prediction systems by top-notch statistical post-processing methods for the region of the Eastern Alps.
Major scientific questions to be answer are: On what spatial and temporal scales is convection predictible? Does predictability increase by taking benefit from pre-dominant flow patterns in the mountainious terrain?
For this purpose lightning serves as a proxy of convective events. The lightning data is received by state-of-the-art Lightning Detection Network established and maintained within the ALDIS project. The post-processing methods include both classic analog approaches and non-linear flexible regression methods.
Co-PI: Achim Zeileis (Department of Statistics, UIBK)
Industrial Partner: Gerhardt Diendorfer, Wolfgang Schulz (ALDIS)