Maxima of moving maxima of continuous functions

TL;DR

This paper introduces CM3 processes, a class of max-stable models for continuous phenomena, and proposes a method for fitting these models to data by estimating dependence length and profiles.

Contribution

It defines CM3 processes with continuous functions and develops a novel fitting procedure based on clustering and Hausdorff distance for extremal dependence modeling.

Findings

CM3 processes can model spatio-temporal extremes with simple spectral expressions.

The method effectively estimates the number of profiles and dependence length from data.

Properties like joint regular variation and mixing conditions are characterized for CM3.

Abstract

Maxima of moving maxima of continuous functions (CM3) are max-stable processes aimed at modeling extremes of continuous phenomena over time. They are defined as Smith and Weissman's M4 processes with continuous functions rather than vectors. After standardization of the margins of the observed process into unit-Fr\'echet, CM3 processes can model the remaining spatio-temporal dependence structure. CM3 processes have the property of joint regular variation. The spectral processes from this class admit particularly simple expressions. Furthermore, depending on the speed with which the parameter functions tend toward zero, CM3 processes fulfill the finite-cluster condition and the strong mixing condition. For instance, these three properties put together have implications for the expression of the extremal index. A method for fitting a CM3 to data is investigated. The first step is to estimate the length of the temporal dependence. Then, by selecting a suitable number of blocks of extremes of this length, clustering algorithms are used to estimate the total number of different profiles. The number of parameter functions to retrieve is equal to the product of these two numbers. They are estimated thanks to the output of the partitioning algorithms in the previous step. The full procedure only requires one parameter which is the range of variation allowed among the different profiles. The dissimilarity between the original CM3 and the estimated version is evaluated by means of the Hausdorff distance between the graphs of the parameter functions.