Trends of thunderstorm activity and relation to large-scale atmospheric conditions in western and central Europe

Thunderstorms associated with severe weather phenomena such as heavy rainfall, wind gusts, hail, tornados or lightning frequently cause considerable damage across the globe. In light of climate change, it is generally anticipated that the increase in thermal instability due to an increase in low-level temperature and moisture leads to more and more intense thunderstorms, thus increasing the risk from convective storms. However, previous studies have shown high annual and interannual variability of thunderstorm frequency in Europe related to large-scale atmospheric processes and mechanisms such as blocking or teleconnection patterns.

To investigate the relationship between large-scale atmospheric conditions and local-scale thunderstorms and their temporal changes, as well as, their long-term trends, we used cloudto-ground (CG) lightning measurements in western and central Europe from 2001 to 2021 (summer half-year; EUCLID network), representing a spatially and temporally homogeneous dataset. Surprisingly, trend analyses of this dataset show a large contiguous area with significantly decreasing trends from central France to Belgium and the Eifel region in Germany – for both lightning activity and thunderstorm days (defined as at least 5 lightning strokes on a 10 x 10 km² grid). In contrast, significant positive trends occur in northern Spain and some parts of the Balkans. However, in other regions of Europe, the observed trends are not significant due to high annual variability.

Further analyses suggest a relation between large-scale flow characteristics and thunderstorm activity in different regions of Europe. For example, based on odds ratio analyses, episodes with negative values of the North Atlantic Oscillation (NAO) Index are associated with decreasing thunderstorm activity in the above mentioned region (France to Germany). We observed that negative NAO phases, which have occurred more frequently over the last decade, may be an explanation for the negative trends in lightning activity and thunderstorm days.

In addition, we present a new algorithm for the temporal and spatial clustering of CG lightning based on ST-DBSCAN (Spatial-Temporal Density-Based Spatial Clustering of Applications with Noise). This algorithm also filters lightning in low-density regions and thus allows for the detection of contiguous areas of high lightning activity (clustered convective events, CCE). Initial analysis of the newly generated dataset shows that in recent years the number of CCEs with small spatial extent has increased, whereas the number of CCEs with large spatial extent has decreased.