Introduction
Pascal Yiou, Philippe Naveau, Mathieu Vrac (LSCE), Gabriele Hegerl (U Edimburg), Valerio Lucarini (KlimaCampus Hamburg) and Ted Shepherd (U Reading) are thrilled to organize a Fall school on climate extremes in Cargèse, in November 2015 (from 9 to 13th). We welcome graduate students, early career researchers (e.g. postdocs) from academic research institutions and private sector. A background/experience in mathematics/statistics and atmospheric sciences will be necessary.
Climate extremes, such as storms, floods, droughts, heatwaves and cold spells, are an important part of climate variability, although rare by essence. Their impacts on society and some ecosystems have fostered many research initiatives in the recent years. A few approaches have been proposed to study those events. Such approaches have been based on climate model analyses, statistical descriptions of the climate variables, and (simple) dynamical system analogies. Although there are formal connections between them, the time is ripe to share tools and concepts between communities and provoke the formalization of a new science of atmospheric extreme events (“extremology”).
A flyer can be downloaded from here.
The science of climate extremes faces three major challenges:

In statistics, the question of rarity of events (and scarcity of data) demands the development of specific methods, based on the theory of extreme value distributions (multivariate, in a non stationary context), in order to assess the probability distribution of events. A relatively new challenge has emerged when one wants to attribute a cause for singular events (rather than a distribution of events). This challenge also requires the development of ad hoc statistical tools.

In mathematics, the relations between the general behavior of an attractor and the extremes of the system are still being explored. A general question is to assess how climate change (as usually defined by an increase of mean temperature due to external forcings) translates into alterations of the climate attractor. New tools and theories are being developed to tackle this problem.

Atmospheric sciences implicitly link the previous two challenges by specifying the variables of underlying dynamical system. The atmospheric conditions leading to extreme events could change in the long term (due to internal or forced variability). In addition, expertise on climate model behavior (from various databases) is crucial to draw proper conclusions from statistical analyses.

A basic problem in the investigation of extremes in a meteoclimatic context is the definition of the spatial resolution of interest. It seems crucial to define protocols for comparing outputs of models featuring different resolution and for comparing models and observational data, going beyond simple interpolation schemes.
The school will propose a series of lectures by leading scientists in the fields of atmospheric sciences, statistics and mathematics. In the afternoons, the attendants will be divided into working groups in order to familiarize themselves with the tools and methodologies, and tackle a concrete project (e.g. the analysis of a recent climate extreme, the analysis of datasets…). The computer and network system of IESC will be used to tackle those projects. At the end of the school the participants will be invited to present the results. When appropriate, they will be encouraged (and helped) to write short papers.