Generalized Pareto processes for simulating space-time extreme events: an application to precipitation reanalyses
- PDF / 2,545,566 Bytes
- 20 Pages / 595.276 x 790.866 pts Page_size
- 45 Downloads / 139 Views
ORIGINAL PAPER
Generalized Pareto processes for simulating space-time extreme events: an application to precipitation reanalyses F. Palacios-Rodrı´guez1
•
G. Toulemonde2 • J. Carreau3 • T. Opitz4
Accepted: 3 October 2020 Ó Springer-Verlag GmbH Germany, part of Springer Nature 2020
Abstract To better manage the risks of destructive natural disasters, impact models can be fed with simulations of extreme scenarios to study the sensitivity to temporal and spatial variability. We propose a semi-parametric stochastic framework that enables simulations of realistic spatio-temporal extreme fields using a moderate number of observed extreme space-time episodes to generate an unlimited number of extreme scenarios of any magnitude. Our framework draws sound theoretical justification from extreme value theory, building on generalized Pareto limit processes arising as limits for event magnitudes exceeding a high threshold. Specifically, we exploit asymptotic stability properties by decomposing extreme event episodes into a scalar magnitude variable (that is resampled), and an empirical profile process representing space-time variability. For illustration on hourly gridded precipitation data in Mediterranean France, we calculate various risk measures using extreme event simulations for yet unobserved magnitudes, and we highlight contrasted behavior for different definitions of the magnitude variable. Keywords Extreme-value theory Precipitation Risk analysis Space-time Pareto processes Stochastic simulation
1 Introduction Extreme events of geophysical processes such as precipitation extend over space and time, and they can entail devastating consequences for human societies and ecosystems. Flash floods in Southern France constitute highly destructive natural phenomena causing material damage and threatening human lives (Vinet et al. 2016), as for example during two relatively recent catastrophic flashflood events in the Gard department in September 2002 (Delrieu et al. 2005), and near Montpellier in October 2014 (Brunet et al. 2018). Since damage and costs of floods have & F. Palacios-Rodrı´guez [email protected] 1
Departamento de Estadı´stica e Investigacio´n Operativa, Facultad de Ciencias Matema´ticas, Universidad Complutense de Madrid, Plaza de Ciencias nu´mero 3, 28040 Madrid, Spain
2
IMAG, CNRS, Inria, Universite´ de Montpellier, Montpellier, France
3
HydroSciences Montpellier, CNRS/IRD, Universite´ de Montpellier, Montpellier, France
4
Biostatistics and Spatial Processes, INRAE, Avignon, France
been increasing over the last decades, the understanding of temporal and spatial variability of rainfall patterns generating such floods receives considerable attention from the authorities (European Environment Agency 2007). To help with this understanding, we develop a method to stochastically simulate realistic spatio-temporal extreme scenarios, which can be fed to impact models. Examples of impact models are urban flood models, such as the shallow water models of Guinot and Soares-Fraza˜o (2006) and Gu
Data Loading...
