Automatic Generation of Winter Storm Warnings

It is a task of MeteoSwiss to warn the public and authorities of upcoming extreme weather, in particular of windstorms. The current warning system is based on fixed geographical warning regions which are manually classified into levels of danger by weather forecasters. This technical report describes how to automatize the generation of geographically variable warnings for extreme weather events. An algorithm is designed and implemented to automatically produce and classify eventbased warning polygons. The inputs to the algorithm are the maximum wind gusts computed by the COSMO2-E numerical weather prediction model. The data is processed with the image processing technique Morphological Filtering to reduce the heterogeneity in the wind speed data and to obtain warning polygons. Evaluated on past winter storms, the algorithm’s probability of detection has improved by 0.11 to 0.48 while the false alarm ratio has increased by 0.1 to 0.85, compared to the preprocessed COSMO2-E wind speed output. Therefore, the implemented algorithm can form contiguous warning polygons while preserving the COSMO2-E model’s accuracy. The algorithm can easily be tuned to generate warnings for other numerical weather predictions.