Should We Have Expected the 2022 Hail Losses in France?
Juergen GrieserOctober 13, 2022
2022 has seen hailstorm records being broken in France, from the number of storms to hailstone size, and with it has come record-breaking hail losses.
Through to mid-August 2022, France Assureurs (FA), formerly the French Insurance Federation, registered nearly one million individual claims due to severe convective storms, and the industry is expected to pay around €4 billion in compensation.
Severe convective storm losses can result from a variety of perils such as hail, wind gusts, tornadoes, and local flooding, but with these record-breaking losses, let’s investigate hail.
The European Severe Weather Database (ESWD), from the European Severe Storms Laboratory (ESSL), collects reports of various perils. For hail, it reveals that the number of hail reports in France up to the end of August 2022 had exceeded the maximum of previous years by a factor of seven.
Other perils showed no increased frequency of reports or extreme intensities. This confirms that the major driver of this year’s high losses in France is hail.
We also learn from the ESWD that it was not a single event that caused the losses. Figure 1 shows the number of days with reported hailstones larger than a given diameter threshold during the months of May to August 2022 (red line), and for the same period in previous years since 2009 (blue lines). Regardless of which threshold we choose, none of the previous years had as many days with observed exceedances as in 2022.
The high hail losses resulted from an accumulation of large-sized hail over many days. This leads to high business interruption and post-loss amplification as confirmed in mid-September when publication Insurance Insider revealed that reinsurers are expecting the total insured hail losses in France to be above €6 billion.
How Unlikely Were These Hail Losses?
According to FA*, hail damage to property accounted for losses of about €2 billion. When we compare this to the average annual hail loss of just above €200 million registered by the FA in the last two decades, we see that the 2022 annual loss is a factor of 10 higher than the average loss. This begs the question: What return period does this represent?
In Figure 2, we draw the exceedance probability (EP) curve of the observed annual hail losses of recent decades (black line) and the EP curve of the RMS® Europe Severe Convective Storm HD Models (red line).
The observed and modeled curves are very well in line with each other, as represented in the figure. The observations suggest an exponential EP curve, so we fit an exponential distribution (blue line) to the observed annual hail losses in France as well.
This allows estimating the return period for 2022 observed hail losses based purely on observations over the last two decades (extrapolating the blue line in Figure 2,) and alternatively from the Europe Severe Convective Storm HD Models (red line in Figure 2). The RMS models are based on physical concepts and detailed hail probability modeling.
This leads to a heavy-tailed distribution of annual hail losses in France and a return period for the 2022 annual hail loss in the order of 425 years, based on an event set covering 50,000 years.
The pure statistical fit of an exponential distribution to the observations, although a good fit, leads to estimated return periods in the range of 82,000 years, which is more than 190 times higher than the RMS estimate.
While the return period from the Europe Severe Convective Storm HD Models is large, it is realistic given that the 2022 hail loss in France was about 10 times the average annual loss over the previous two decades. This indicates that in 2022 France experienced hail losses that happen very rarely. This expectation, however, is based on a hail model that assumes a stationary climate.
Is It Due to Climate Change?
Basic statistics teaches us that for many distributions, a change in the mean of a random variable affects more extreme quantiles greater than it affects less extreme ones. For climate change, this implies that the frequency of very rare events is impacted much more than the frequency of less rare events.
In other words, while no trend is visible in ordinary hail loss years as observed over the last two decades, we cannot exclude that climate change has increased the likelihood of the large losses experienced this year.
We know that this summer was exceptionally hot in the region. Plus, the very high sea surface temperatures of the Mediterranean contributed to providing the humidity necessary for hail-producing severe convective storms to develop.
While this is conceptual evidence that the 2022 high hail losses in France may be at least somewhat attributable to climate change, it is not proof. Due to the uncertain and potentially competing impact of climate change on the numerous factors critical to severe convective storm formation, there is no consensus as yet on how climate change may be affecting the frequency and severity of damaging convective weather events, including hail.
As such, an assumption of stationarity remains reasonable for now. If several exceptional years like 2022 occur in short succession, that opinion may well change.
Juergen heads a group of modelers within Model Development, who are responsible for building RMS Severe Convective Storm and Winterstorm models.
During his fifteen years at RMS, Juergen has worked on several RMS models, leading the development of the Japan typhoon surge model, and the development of the tropical-cyclone rain model, and has also contributed to the European Winterstorm model.
Before joining RMS, Juergen worked for the United Nations Food and Agriculture Organization (FAO) and is also a contributing author of the Intergovernmental Panel on Climate Change (IPCC) Fourth Assessment report, which won the Nobel Peace Prize in 2007.
Juergen studied meteorology in Frankfurt, Germany, and holds a PhD in geosciences with a thesis on climate change detection in observations.