Spatial modeling of extreme snow depth

Abstract

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