Quantification of extremal dependence in spatial natural hazard footprints: independence of windstorm gust speeds and its impact on aggregate losses

• DOI: 10.5194/nhess-18-2933-2018
• Publication Date: 2018-11-08T10:49:27Z
• Authors: Laura C. Dawkins; David B. Stephenson
• Keywords: Spatial Dependence; Gumbel distribution; Tail dependence

Abstract. Natural hazards, such as European windstorms, have widespread effects that result in insured losses at multiple locations throughout a continent. Multivariate extreme-value statistical models for environmental phenomena must therefore accommodate very high dimensional spatial data, well correctly representing dependence the extremes to ensure accurate estimation of these losses. Ideally one would employ flexible model, able characterise all forms extremal dependence. However, are restricted few dozen dimensions, hence an priori diagnostic approach be used identify dominant form Here, we present various approaches exploring class hazard fields: tail dependency measures, copula fits, and conceptual loss distributions. These illustrated by application data set high-dimensional historical windstorm footprints (6103 maps 3-day maximum gust speeds 14 872 locations). We find there is little evidence asymptotic footprints. Furthermore, empirical properties shown reproduced using Gaussian copulas but not extremally dependent Gumbel copulas. It conjectured lack generic property turbulent flows. results open up possibility geostatistical process fast simulation fields.