|
Activity Number:
|
382
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #305185 |
|
Title:
|
Multivariate Modeling of Wind Speed Maxima
|
|
Author(s):
|
Pál Rakonczai*+ and Nader Tajvidi and András Zempléni
|
|
Companies:
|
Lund Institute of Technology and Lund Institute of Technology and Eötvös Loránd University
|
|
Address:
|
Mathematical Statistics, Centre for Mathematical Sciences, Lund, , Sweden
|
|
Keywords:
|
multivariate extremes ; dependence function ; copulas ; compact predict regions ; wind speed maxima
|
|
Abstract:
|
This paper proposes and compares various approaches for fitting multivariate extreme value distributions in order to describe the behavior of simultaneously appearing wind speed maxima. When modeling spatial extremes one should not leave out of consideration that observations measured in closely located stations usually show stronger dependence. Thus, besides fitting univariate marginals, the knowledge of the dependence structure between the coordinates is also crucial. Here we assume extreme value distributions as marginals and estimate the dependence structures in different ways. After the necessary marginal transformations parametric and nonparametric Pickands-like dependence functions as well as copula functions have been used. All of these methods have been applied for wind time series of the last 5 decades measured in different cities of North Germany.
|
