Abstract

Skewed multivariate distributions, such as skew-normal or skew-t distributions, are flexible
parametric models that are suitable for modeling data sets with complex dependence structures. The respective limiting extreme-value distributions can capture both symmetric and asymmetric tail dependence structures thus providing greater flexibility when modeling multivariate extremes data, including spatial extremes. In this talk, we consider different methods of constructing multivariate extreme-value distributions for spatial data based on skewed multivariate distributions. We study some of these skewed distributions that combine flexibility in the bulk as well as in the tails. We apply the proposed models to analyze the maximum wind speed data.