Models for Temporal Clustering of Extreme Events With Applications to Mid‐Latitude Winter Cyclones

The occurrence of extreme events like heavy precipitation or storms at a certain location often shows a clustering behavior and is thus not described well by a Poisson process. We construct a general model for the inter‐exceedance times (IETs) in between extreme events which combines different candidate models for such behavior. One of them is formulated in terms of clusters of dependent events with exponential IETs in between clusters, while the other assumes independent events separated by heavy‐tailed IETs. We propose a modification of the Cramér‐von Mises distance for fitting the combined model. The resulting estimator turns out to be competitive with specialized estimators if the data stem from one of the two submodels. Our modeling approach thus allows us to distinguish these different data generating mechanisms without the need of a‐priori model selection. An application to mid‐latitude winter cyclones illustrates the usefulness of our work as the combination of the two mechanisms improves the descriptions of such occurrences at many places.