Regional Modelling of Extreme Storms Via Max‐Stable Processes

Abstract

Asymptotic models for extremes of random processes often form the basis for estimating the extremal behaviour of environmental phenomena. Most such phenomena have a spatial dimension, and the aim of this paper is to develop a procedure for modelling in continuous space the spatial dependence within extreme events. A principal objective in the analysis-as with other current research on extremes-is to base inference on as much of the available data as possible. The modelling procedures are justified on simulated data and subsequently applied to a series of rainfall data