Sampling from a Max‐Stable Process Conditional on a Homogeneous Functional with an Application for Downscaling Climate Data. (27th September 2017)

• Record Type: Journal Article
• Authors: Oesting, Marco; Bel, Liliane; Lantuéjoul, Christian
• Is Part Of: Scandinavian journal of statistics. Volume 45:Number 2(2018:Jun.)
• Page Start: 382
• Page End: 404
• Publication Date: 2017-09-27
• Subjects: downscaling -- Markov chain Monte Carlo -- max‐linear model -- spatial extremes; Statistics -- Periodicals; 310
• DOI: 10.1111/sjos.12299
• Languages: English
• ISSNs: 0303-6898

Abstract

Abstract: Conditional simulation of max‐stable processes allows for the analysis of spatial extremes taking into account additional information provided by the conditions. Instead of observations at given sites as usually done, we consider a single condition given by a more general functional of the process as may occur in the context of climate models. As the problem turns out to be intractable analytically, we make use of Markov chain Monte Carlo methods to sample from the conditional distribution. Simulation studies indicate fast convergence of the Markov chains involved. In an application to precipitation data, the utility of the procedure as a tool to downscale climate data is demonstrated.