Uncertain Tempering and Extreme Events in Volcanic Risk Assessment

• Activity Number: 346
• Type: Invited
• Date/Time: Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
• Sponsor: IMS
• Author(s): Robert L. Wolpert*+
• Companies: Duke University
• Keywords: extreme value ; tempered stable distribution ; Pareto ; risk ; model uncertainty

Abstract

Predictions of risk associated with rare but catastrophic natural events---such as volcanic pyroclastic flows---are often made on the basis of historical records of less extreme events, coupled with statistical and mathematical models through which common, non-extreme events lend evidence about rare and catastrophic ones. Risk predictions are uncertain for many reasons including insufficiency or variability of historical records, inherent stochasticity of the systems, and inadequacies or uncertainties in the models relating historical record to predictions of more extreme events than those that have been observed. We explore the uncertainty in model predictions of the risk of destruction due to volcanic pyroclastic flows that arise from uncertainty about features of the model (tempering of distributional tails) that are difficult to assess on the basis of historic records alone.