SUPPLEMENTARY MATERIAL: STABLE SUMS TO INFER HIGH RETURN LEVELS OF MULTIVARIATE RAINFALL TIME SERIES

First, we complement the case study of heavy rainfall in France by implementing Pareto-based methods using declustering techniques. Second, we develop on the asymptotic theory of the stable sums method. To prove Theorem 6.1, we give a more general statement and prove the multivariate central limit theory of regularly varying time series with unit (tail)-index.