Large-scale flow patterns and their relation to summer lightning in Europe

This study is a first step in deciphering the role that climate change-induced changes in flow patterns play in the evolution of thunderstorm frequency. Here, we investigate the relationship between large-scale flow patterns and the (temporal and spatial) distribution of lightning in Europe, as recorded by the Met Office Arrival Time Difference Network (ATDnet). The seasonal cycle shows that the largest number of lightning days occurs in summer from May to August, the period we therefore focus on. The large-scale flow pattern is shown using the daily mean 500-hPa geopotentials from the ERA5 reanalysis data. The k-means cluster analysis is applied to the daily mean geopotential heights in the selected four-month period between 1950 and 2020. The algorithm produces 14 patterns. The distributions of lightning associated with the clusters show that lightning frequently occurs under synoptic quiet conditions or even below a ridge. In addition, the occurrence of lightning over Western Europe appears to be more dependent on the synoptic situation, where it is strongly associated with clusters that have a southerly flow in 500 hPa, compared to lightning over the Alpine region or Southeastern Europe. However, changes in the occurrence of synoptic-scale patterns cannot alone explain the changes in the lightning frequency over Europe detected by ARCHaMo (Figure 1). The second reason is that lightning is more dependent on large-scale synoptic-scale patterns in some regions than in others. Across Europe, lightning and lightning ingredients are more strongly tied to individual clusters over the western part than over the eastern part. Therefore, changes in lightning ingredients derived from ARCHaMo are expected to best match changes in lightning frequency derived from changes in clusters over western Europe.

Figure 1: Percent changes in lightning frequency from the average per decade for four months MJJA. Hatching indicates areas where trends are statistically significant. (a) Trend in ARCHaMo due to influence of flow patterns (1950-2020), model values are averaged over the 71 years, (b) Trend in ARCHaMo (1950-2020).