Anna Kiriliouk (UNamur)

Title: Climate event attribution using multivariate peaks-over-thresholds modelling

Abstract: Quantifying the human influence on climate change and identifying potential causes is a highly relevant research area which is often referred to as detection and attribution. A common approach is to compare the probability of an extreme event in the factual world to the probability of an extreme event in a counterfactual world, i.e., a world that might have been if no humans would have existed. The event probabilities can be calculated using large scale climate model runs that simulate the evolution of the climate with and without anthropogenic forcings. The Fraction of Attributable Risk (FAR) is defined as the relative ratio of event probabilities in the factual and in the counterfactual world. Estimating the FAR will allow us to quantify the extent to which human activities have increased the risk of occurrence of an extreme event. We propose a model for the FAR based on the multivariate generalized Pareto distribution, i.e., the asymptotic distribution of suitably normalized exceedances over a high threshold. The model is used to quantify the increased risk of an extreme rainfall event in central Europe.