A Hierarchical Bayesian Extreme Value Theory Approach to Modeling Heat Waves

• Activity Number: 329
• Type: Topic Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics and the Environment
• Author(s): Benjamin Shaby*+
• Companies: University of California at Berkeley
• Keywords: extreme value theory ; heat waves ; hierarchical Bayesian ; latent variable

Abstract

Heat waves have the potential to cause widespread public health and economic damage. The European heat wave of 2003 and the Russian heat wave of 2010 are recent, calamitous, examples. Important risk management questions include whether heat waves are becoming longer, more severe, or more frequent. While they are clearly examples of a type of extreme weather event, heat waves have largely not been explored from an extreme value theory perspective. Here we take a hierarchical Bayesian approach to extreme value modeling of heat waves. A latent variable with dependence in space and time indicates membership in the heat wave state. Within a heat wave, temperatures are modeled using extreme value distributions, with extremal dependence across time accomplished through an extreme value Markov model.