A Censored Point Process Model for Extreme Volcanic Eruptions

摘要

The magnitude of a volcanic eruption is an essential component of risk assessment in volcano-sensitive regions. Extreme value models are a natural candidate for modelling such phenomena, and a point process representation for extreme value behaviour provides a convenient inferential framework. However, direct application to databases of volcanic events is complicated by an under-recording of historical events. This is complicated further by the fact that small events appear to have a greater tendency to go unreported relative to large events. In this article we suggest modifying the standard point process model for extremes with a parametric component that models the under-reporting mechanism.