Quantitative comparison of METEOSAT-based thunderstorm detektion and nowcastin with in-situ reports in the European Severe Weather Database (ESWD)

Abstract

Severe thunderstorms constitute a major weather hazard in Europe, with an estimated total damage of € 5-8 billion each year. Yet a pan-European database of severe weather reports in a homogeneous data format has become available only recently: the European Severe Weather Database (ESWD). We demonstrate the large potential of ESWD applications for storm detection and forecast evaluation purposes and complement an earlier case-based study. The analysis of all warm-season (JJA) severe weather days in Europe in 2008 corroborated our earlier findings. There is a good agreement between ESWD reports and Cb-TRAM detected thunderstorms, even though no exact correspondence between ESWD reports and Cb-TRAM cells is required (e.g. due to storm morphology). Correspondingly, a large portion of ESWD reports regarded as misses by our strict in/out-of-Cb-TRAM-polygon criterion were still located close to a Cb-TRAM cell. Quantitatively, only the probability of detection (POD) can be evaluated due to the different characteristics of the two data sources. The POD for storm detection was 0.24 on average, with maximum values up to 0.58. The respective analysis for the 30 and 60 minutes nowcasts yielded average POD values of 0.11 and 0.08, respectively, with maximum values of POD exceeding 0.4 on 3 days for the 30 minutes nowcast and on one day for the 60 minutes nowcast.