Searching for the most extreme temperature events in recent history

Because they are rare, extreme weather events inevitably attract public and scientific attention. The most unusual events are regularly documented as part of routine climate monitoring by meteorological services, and put into the perspective of climate change by attribution studies through quantities such as a probability ratio. However, it is often recognized that (i) the selection of studied events is geographically uneven, and (ii) the definition of a given event, in particular its spatio-temporal scale, is subjective, which may impact the results.

In previous work, we proposed an objective method of event definition, consisting in the automatic selection of the spatio-temporal window maximizing the event rarity. Importantly, we showed that maximizing the event rarity does not bias attribution statements, in the sense that it does not systematically maximize (or minimize) the probability ratio. Here we present how this method can be used to compare several events occurring on different years, seasons, regions, etc. Our objective research procedure works both over time, which can be useful for routine climate monitoring, and over space, which can resolve the geographical selection bias of attribution studies. Ultimately, we provide a selection of the most extreme hot and cold events that have occurred worldwide in the recent past, among which are iconic heat waves such as that seen in 2021 in Canada or 2003 in Europe.