Extreme value theory provides a class of models for the behaviour of stochastic processes at extreme levels. Since volcanic eruptions are (at least in a non-scientific sense) extreme events, it might be hoped that there is some role for the extreme value models in the science of volcanology. In this article we explore such a possibility through a particular catalogue of extreme eruptions registered in the last two millennia. The analysis is based on a particular point process characterization of extremes: it takes into account that historical events in the dataset are less likely to have been recorded than recent events and that this effect seems especially pronounced for events of relatively low magnitude. Ignoring these aspects can lead to a biased estimate of extremal behaviour.
Processi di punto parametrici e non parametrici per la modellazione di eventi vulcanici estremi in presenza di censura.
Furlan, Claudia
2005
Abstract
Extreme value theory provides a class of models for the behaviour of stochastic processes at extreme levels. Since volcanic eruptions are (at least in a non-scientific sense) extreme events, it might be hoped that there is some role for the extreme value models in the science of volcanology. In this article we explore such a possibility through a particular catalogue of extreme eruptions registered in the last two millennia. The analysis is based on a particular point process characterization of extremes: it takes into account that historical events in the dataset are less likely to have been recorded than recent events and that this effect seems especially pronounced for events of relatively low magnitude. Ignoring these aspects can lead to a biased estimate of extremal behaviour.File | Dimensione | Formato | |
---|---|---|---|
2005_6.pdf
accesso aperto
Licenza:
Non specificato
Dimensione
1.23 MB
Formato
Adobe PDF
|
1.23 MB | Adobe PDF | Visualizza/Apri |
Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.