Séminaire
Extreme weather or climate events are often the result of large-scale, non-static phenomena. Understanding the underlying drivers can be extremely complex, and one should account for the fact that similar atmospheric configurations might not share the same physical origin nor evolve in a similar way.
Recent developments in extreme value theory try to take into account these complexities by using a dynamical systems framework which considers the entire evolution, or trajectory, of atmospheric flows. This approach provides powerful insights into the predictability afforded by a given atmospheric configuration. Climate extremes can be diagnosed in terms of rarely recurring segments of such trajectories, or dynamical extremes. However, it is important to note that the dynamical systems definition of climate extremes is not the same as the standard statistical definition of the maxima or minima of a given observable, even though the two quantities are closely interlinked.
In this talk I will illustrate a simple application of this framework to the North Atlantic atmospheric circulation, with a focus on extreme surface temperature and wind occurrences over Europe.