Max-stable Models Research Articles

Space‒time max-stable models with spectral separability

AbstractNatural disasters may have considerable impact on society as well as on the (re-)insurance industry. Max-stable processes are ideally suited for the modelling of the spatial extent of such extreme events, but it is often assumed that there is no temporal dependence. Only a few papers have introduced spatiotemporal max-stable models, extending the Smith, Schlather and Brown‒Resnick spatial processes. These models suffer from two major drawbacks: time plays a similar role to space and the temporal dynamics are not explicit. In order to overcome these defects, we introduce spatiotemporal max-stable models where we partly decouple the influence of time and space in their spectral representations. We introduce both continuous- and discrete-time versions. We then consider particular Markovian cases with a max-autoregressive representation and discuss their properties. Finally, we briefly propose an inference methodology which is tested through a simulation study.

Conditioned limit laws for inverted max-stable processes

Max-stable processes are widely used to model spatial extremes. These processes exhibit asymptotic dependence meaning that the large values of the process can occur simultaneously over space. Recently, inverted max-stable processes have been proposed as an important new class for spatial extremes which are in the domain of attraction of a spatially independent max-stable process but instead they cover the broad class of asymptotic independence. To study the extreme values of such processes we use the conditioned approach to multivariate extremes that characterises the limiting distribution of appropriately normalised random vectors given that at least one of their components is large. The current statistical methods for the conditioned approach are based on a canonical parametric family of location and scale norming functions. We study broad classes of inverted max-stable processes containing processes linked to the widely studied max-stable models of Brown–Resnick and extremal- t , and identify conditions for the normalisations to either belong to the canonical family or not. Despite such differences at an asymptotic level, we show that at practical levels, the canonical model can approximate well the true conditional distributions.

Non-Stationary Dependence Structures for Spatial Extremes

Max-stable processes are natural models for spatial extremes because they provide suitable asymptotic approximations to the distribution of maxima of random fields. In the recent past, several parametric families of stationary max-stable models have been developed, and fitted to various types of data. However, a recurrent problem is the modeling of non-stationarity. In this paper, we develop non-stationary max-stable dependence structures in which covariates can be easily incorporated. Inference is performed using pairwise likelihoods, and its performance is assessed by an extensive simulation study based on a non-stationary locally isotropic extremal t model. Evidence that unknown parameters are well estimated is provided, and estimation of spatial return level curves is discussed. The methodology is demonstrated with temperature maxima recorded over a complex topography. Models are shown to satisfactorily capture extremal dependence.

Conditional independence among max-stable laws

Let X be a max-stable random vector with positive continuous density. It is proved that the conditional independence of any collection of disjoint subvectors of X given the remaining components implies their joint independence. We conclude that a broad class of tractable max-stable models cannot exhibit an interesting Markov structure.

Inferring Spatio-temporal Patterns in Extreme Snowfall in the French Alps Using Max-stable Processes

Our objective is to study the dependence structure of extreme snowfall in the French Alps. Snowfall amount measurements (3 days accumulation period) from 90 monitoring stations providing data spanning from 1958 to 2013 were projected at the same altitude level (1800m). Annual maxima of projected snowfall were modeled with Generalized Extreme Value (GEV) distributions and transformed in unit Fréchet in order to focus on the dependence structure only. The final goal is to evaluate the spatio-temporal evolution of this dependence structure and to compare the ability of different max-stable models to capture it. In addition to the most classical max-stable models (Smith and Schlather), max-stable processes made available more recently (extremal-t, Geometric Gaussian and Brown-Resnick) were also considered, taking into account spatial anisotropy. These latter models are found to be the most suitable in our case according to CLIC. Further cross-validation on joint exceedance probabilities does not clearly discriminate these max-stable models. Results show that measurement stations in the Northern and Central Alps are strongly dependent whereas stations from the extreme Southern Alps, influenced by Mediterranean effects, are more isolated. Furthermore, extremal dependence appears controlled by the orientation of mountains and valleys. Finally, using an estimation window moving in time, we highlight a recent decrease in extremal dependence at large distances.

Inferring Spatio-temporal Patterns in Extreme Snowfall in the French Alps Using Max-stable Processes

Our objective is to study the dependence structure of extreme snowfall in the French Alps. Snowfall amount measurements (3 days accumulation period) from 90 monitoring stations providing data spanning from 1958 to 2013 were projected at the same altitude level (1800 m). Annual maxima of projected snowfall were modeled with Generalized Extreme Value (GEV) distributions and transformed in unit Fréchet in order to focus on the dependence structure only. The final goal is to evaluate the spatio-temporal evolution of this dependence structure and to compare the ability of different max-stable models to capture it. In addition to the most classical max-stable models (Smith and Schlather), max-stable processes made available more recently (extremal-t, Geometric Gaussian and Brown-Resnick) were also considered, taking into account spatial anisotropy. These latter models are found to be the most suitable in our case according to CLIC. Further cross-validation on joint exceedance probabilities does not clearly discriminate these max-stable models. Results show that measurement stations in the Northern and Central Alps are strongly dependent whereas stations from the extreme Southern Alps, influenced by Mediterranean effects, are more isolated. Furthermore, extremal dependence appears controlled by the orientation of mountains and valleys. Finally, using an estimation window moving in time, we highlight a recent decrease in extremal dependence at large distances.

Incorporating spatial dependence in regional frequency analysis.

The efficiency of regional frequency analysis (RFA) is undermined by intersite dependence, which is usually ignored in parameter estimation. We propose a spatial index flood model where marginal generalized extreme value distributions are joined by an extreme-value copula characterized by a max-stable process for the spatial dependence. The parameters are estimated with a pairwise likelihood constructed from bivariate marginal generalized extreme value distributions. The estimators of model parameters and return levels can be more efficient than those from the traditional index flood model when the max-stable process fits the intersite dependence well. Through simulation, we compared the pairwise likelihood method with an L-moment method and an independence likelihood method under various spatial dependence models and dependence levels. The pairwise likelihood method was found to be the most efficient in mean squared error if the dependence model was correctly specified. When the dependence model was misspecified within the max-stable models, the pairwise likelihood method was still competitive relative to the other two methods. When the dependence model was not a max-stable model, the pairwise likelihood method led to serious bias in estimating the shape parameter and return levels, especially when the dependence was strong. In an illustration with annual maximum precipitation data from Switzerland, the pairwise likelihood method yielded remarkable reduction in the standard errors of return level estimates in comparison to the L-moment method.

CRPS M-estimation for max-stable models

Max-stable processes provide canonical models for the dependence of multivariate extremes. Inference with such models has been challenging due to the lack of tractable likelihoods which has motivated use of composite likelihood methods Padoan et al. (J. Amer. Stat. Assoc. 105(489):263–277, 2010). In contrast, the finite dimensional cumulative distribution functions (CDFs) appear natural to work with, and are readily available or can be approximated well. Motivated by this fact, in this work we develop an M-estimation framework for max-stable models based on the continuous ranked probability score (CRPS) of multivariate CDFs. We start by establishing conditions for the consistency and asymptotic normality of the CRPS-based estimators in a general context. We then implement them in the max-stable setting and provide readily computable expressions for their asymptotic covariance matrices. The resulting point and asymptotic confidence interval estimates are illustrated over popular simulated models. They enjoy accurate coverages and offer an alternative to composite likelihood based methods.

Tapered composite likelihood for spatial max-stable models

Spatial extreme value analysis is useful to environmental studies, in which extreme value phenomena are of interest and meaningful spatial patterns can be discerned. Max-stable process models are able to describe such phenomena. This class of models is asymptotically justified to characterize the spatial dependence among extremes. However, likelihood inference is challenging for such models because their corresponding joint likelihood is unavailable and only bivariate or trivariate distributions are known. In this paper, we propose a tapered composite likelihood approach by utilizing lower dimensional marginal likelihoods for inference on parameters of various max-stable process models. We consider a weighting strategy based on a “taper range” to exclude distant pairs or triples. The “optimal taper range” is selected to maximize various measures of the Godambe information associated with the tapered composite likelihood function. This method substantially reduces the computational cost and improves the efficiency over equally weighted composite likelihood estimators. We illustrate its utility with simulation experiments and an analysis of rainfall data in Switzerland.

Statistical Modeling of Spatial Extremes

The areal modeling of the extremes of a natural process such as rainfall or temperature is important in environmental statistics; for example, understanding extreme areal rainfall is crucial in flood protection. This article reviews recent progress in the statistical modeling of spatial extremes, starting with sketches of the necessary elements of extreme value statistics and geostatistics. The main types of statistical models thus far proposed, based on latent variables, on copulas and on spatial max-stable processes, are described and then are compared by application to a data set on rainfall in Switzerland. Whereas latent variable modeling allows a better fit to marginal distributions, it fits the joint distributions of extremes poorly, so appropriately-chosen copula or max-stable models seem essential for successful spatial modeling of extremes.

Spatial modeling of extreme snow depth

The spatial modeling of extreme snow is important for adequate risk management in Alpine and high altitude countries. A natural approach to such modeling is through the theory of max-stable processes, an infinite-dimensional extension of multivariate extreme value theory. In this paper we describe the application of such processes in modeling the spatial dependence of extreme snow depth in Switzerland, based on data for the winters 1966--2008 at 101 stations. The models we propose rely on a climate transformation that allows us to account for the presence of climate regions and for directional effects, resulting from synoptic weather patterns. Estimation is performed through pairwise likelihood inference and the models are compared using penalized likelihood criteria. The max-stable models provide a much better fit to the joint behavior of the extremes than do independence or full dependence models.

Storm processes and stochastic geometry

This paper is devoted to a prototype of max-stable models called the storm process. At first its spatial distribution is given in association with different observation supports. Then the compatibility relationships between extremal coefficients at various supports are completely characterized. Particular attention is paid to the special case where the storms are indicator functions of Poisson polytopes. Explicit formulae are found for the extremal coefficients with finite or convex supports. A new algorithm for exactly simulating the Poisson storm process in continuous space is also provided. Overall, the storm process can be used as a benchmark for comparing the performances of several estimators of extremal coefficients, or for model selection.

Asymptotic models and inference for extremes of spatio-temporal data

Recently there has been a lot of effort to model extremes of spatially dependent data. These efforts seem to be divided into two distinct groups: the study of max-stable processes, together with the development of statistical models within this framework; the use of more pragmatic, flexible models using Bayesian hierarchical models (BHM) and simulation based inference techniques. Each modeling strategy has its strong and weak points. While max-stable models capture the local behavior of spatial extremes correctly, hierarchical models based on the conditional independence assumption, lack the asymptotic arguments the max-stable models enjoy. On the other hand, they are very flexible in allowing the introduction of physical plausibility into the model. When the objective of the data analysis is to estimate return levels or kriging of extreme values in space, capturing the correct dependence structure between the extremes is crucial and max-stable processes are better suited for these purposes. However when the primary interest is to explain the sources of variation in extreme events Bayesian hierarchical modeling is a very flexible tool due to the ease with which random effects are incorporated in the model. In this paper we model a data set on Portuguese wildfires to show the flexibility of BHM in incorporating spatial dependencies acting at different resolutions.

Modelling Dependence within Joint Tail Regions

Summary Standard approaches for modelling dependence within joint tail regions are based on extreme value methods which assume max-stability, a particular form of joint tail dependence. We develop joint tail models based on a broader class of dependence structure which provides a natural link between max-stable models and weaker forms of dependence including independence and negative association. This approach overcomes many of the problems that are encountered with standard methods and is the basis for a Poisson process representation that generalizes existing bivariate results. We apply the new techniques to simulated and environmental data, and demonstrate the marked advantage that the new approach offers for joint tail extrapolation.