Comparing the environments of large hail, tornadoes and non-severe convection in the United Kingdom

Destructive tornadoes and damaging hail are a rare but present hazard in the United Kingdom (UK), as illustrated by the November 2023 Jersey storm. However, improving our ability to understand and forecast these events requires moving beyond case studies and towards a climatological view of both severe and non-severe convection. Not only do severe convective environments vary between regions, but differences in the non-severe ‘baseline’ further modulate the most useful parameters for forecasting.

We present the first climatological investigation of severe hail environments in the UK, and the largest to date of tornado environments, using thundeR convective parameters derived from ERA5 reanalysis. We show that lightning flash data, often used to identify non-severe (null) cases, are poorly suited to the UK, on top of only being available for the past ~20 years. Instead we develop a novel method using ERA5 convective precipitation, inclusive of unelectrified but potentially tornado-producing storms common during the cool season. The resulting long study period (1950–2022) enables a larger sample size (~400 severe hail days and ~950 tornado days) than has previously been achieved in regions with temperate, strongly maritime climates and low severe hail incidence such as the UK.

UK severe hail and tornado environments are unlike typical European or American ones. Most notably, instability (i.e. CAPE) is much lower in the UK. Nonetheless, CAPE effectively identifies severe hail environments, because it is also much lower in non-severe environments. Conversely, shear parameters generally show reduced separation between non-severe and severe distributions. Exceptions to this, such as the link between very large hail and 1–6 km shear, may identify physical constraints on severe hazard occurrence. Many UK severe environments, particularly outside of summer, are of the ‘high-shear, low-CAPE’ type. However, this kind of environment also typifies cool-season, non-severe convection, increasing the forecasting challenge.

These results suggest that most forecasting rules of thumb and composite parameters cannot simply be re-applied to the UK. Instead, we seek ‘analogue’ regions with which comparison and forecasting experience are likely to be most useful. Candidates include other parts of Europe’s maritime periphery, from Portugal to the Netherlands and potentially to Fennoscandia and, further afield, the US Pacific Coast and South Australia.