Premium
Using a hidden Markov model to analyse extreme rainfall events in Central‐East Sardinia
Author(s) -
Betrò Bruno,
Bodini Antonella,
Cossu Q. Antonio
Publication year - 2008
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.944
Subject(s) - hidden markov model , weibull distribution , precipitation , environmental science , climatology , meteorology , extreme value theory , statistics , geography , mathematics , computer science , geology , artificial intelligence
The aim of this paper is to investigate the capability of a hidden Markov model (HMM) in identifying possible recurrent patterns in the occurrence of extreme events over a small area of Central‐East Sardinia (Italy), for the purposes of hydrogeological risk prevention. In this area, three intense rainfall events (defined as more than 40 mm in a day) per year are expected during the period September–January. They often assume the characteristics of an extreme event, here defined as more than 100 mm of precipitation in a day. The main interest of modelling daily rainfall data by a HMM lies in the underlying correspondence between the hidden states and the concept of discrete weather states. To model the precipitation at each station, given the weather state, we suggest using a mixture of a point mass at zero and of Weibull distributions, these latter showing a better fit to the considered data than the more usual Gamma and Exponential distributions. Copyright © 2008 John Wiley & Sons, Ltd.