betaFeedback

WorksMathematics

Spatial Extremes on Domains with Physical Barriers

Abstract

We propose a model for spatial extremes what realistically accommodates nonconvex domains. Most existing models assume that tail dependence between locations depends only on their separation. Even if extended to be anisotropic or nonstationary, they still cannot accommodate features between points that may mediate the dependence. This assumption limits their applicability in settings where physical barriers like mountains, coastlines, or islands may attenuate or even block dependence in extreme events. In this work, we use the idea of a visibility graph to modify an existing spatial random scale model. This results in a tail dependence model that naturally accommodates nonconvex spatial domains, where the nonconvex features are interpreted as physical barriers. Furthermore, the model inherits key properties from the scale mixture like the ability to represent asymptotic independence at long ranges and either asymptotic dependence or asymptotic independence at short ranges. We apply our model to two applications in which physical barriers feature prominently in determining the spatial structure of extreme events. The first is precipitation in the South Island of New Zealand, where the Southern Alps shelter the eastern region from the effects of atmospheric rivers which arrive from the west. The second is significant wave heights in the North Atlantic, where the Faroe Islands tend to block the effects of extratropical cyclones traveling into the North Sea.

Paper details

vlnc:W_6r4h9tut
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
arXiv nonexclusive license
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…