Intensity-based classification of North Atlantic and European extratropical cyclones

Most of the day-to-day variability in weather in Europe, including damaging events, is caused by extratropical cyclones (ETCs). ETCs are very different from one another and to more easily study their development, intensity, and structure, various ETC classification schemes have been proposed. Here, we propose an intensity-based scheme in which we first identify necessary ETC intensity measures to describe ETC intensity comprehensively from both dynamical and impact-relevant perspective, and then use them to produce an ETC classification.

ERA5 reanalysis data from 1979 to 2022 was used to track ETCs and compute their intensity measures in the extended winter season (October-March). A total of 7361 ETC tracks were identified in the North Atlantic and Europe. Eleven intensity measures were analysed including 850-hPa relative vorticity, mean sea level pressure, wind speeds at various levels, wind gust, wind footprint, precipitation, and storm severity index. Among the 11 intensity measures, relevant ones were identified by analysing their correlation with each other combined with a sparse principal component analysis (sPCA). The selected measures were used to classify the ETCs by performing a cluster analysis with Gaussian mixture modelling.

Based on the sPCA and relationships between the intensity measures, the set was reduced to 5 measures: 850-hPa relative vorticity, 850-hPa wind speed, wind footprint, precipitation, and storm severity index. Therefore, to describe ETC intensity comprehensively, one needs to use more than one or two intensity measures. The cluster analysis with these 5 measures as input produced 4 discernible clusters. Between these clusters ETCs differed in terms of their intensity, life cycle characteristics, and geographical location. Despite only 9 % of all ETCs belonging to the most intense cluster, it contained 17 out of 21 investigated impactful named storms, which demonstrates the relevance of the classification and its ability to identify potentially impactful ETCs.