Home > Research > Publications & Outputs > Threshold modeling of nonstationary extremes

Links

Text available via DOI:

View graph of relations

Threshold modeling of nonstationary extremes

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNChapter

Published
Close
Publication date2016
Host publicationExtreme Value Modeling and Risk Analysis: Methods and Applications
PublisherCRC Press
Pages87-108
Number of pages22
ISBN (electronic)9780429161193
ISBN (print)9781498701310
<mark>Original language</mark>English

Abstract

It is common for extremes of a variable to be nonstationary, varying systemati cally with covariate values. We consider the incorporation of covariate effects into threshold-based extreme value models, using parametric and nonparametric regres sion functions. We use quantile regression to set a covariate-dependent threshold. As an example we model storm peak significant wave heights as a function of storm direction, season, and a climate index. © 2016 by Taylor & Francis Group, LLC.