In this paper, after some considerations on the weak performances of the statistical inference models currently used, we propose two strategies to improve the estimation of extreme events of short duration. A first strategy aims to identify, using the meteorological information, features and climatic indexes that are strongly correlated with the characteristics of rainfalls. These indexes are used for searching homogeneous rainfall regions. The first results of a comprehensive meteoclimatic analysis have shown, for some limited areas of the Italian peninsula and for short rainfalls, the presence of a "super-homogeneity" that allows to hypothesize the annual mean of precipitation to be constant in the considered region. The second strategy, once admitted the existence of a super-homogeneity, is aimed at the definition of a single spatial distribution of extreme events of 1h rainfalls. In this way, it is possible to carry out the statistical inference using a single time series formed by annual maxima of daily rainfalls occurring all over the region. The frequency distribution thus obtained, allows to estimate the probability of non-exceedance F(X) of an event X on the entire region. The proposed procedure was applied to the Vibo Valentia flood on 3 July 2006.
Un modello regionale meteo-idrologico per la valutazione delle piogge brevi / A Meteo-Hydrological Regional Model to Evaluate Extreme Short Rainfall
Francesco Chiaravalloti
2014
Abstract
In this paper, after some considerations on the weak performances of the statistical inference models currently used, we propose two strategies to improve the estimation of extreme events of short duration. A first strategy aims to identify, using the meteorological information, features and climatic indexes that are strongly correlated with the characteristics of rainfalls. These indexes are used for searching homogeneous rainfall regions. The first results of a comprehensive meteoclimatic analysis have shown, for some limited areas of the Italian peninsula and for short rainfalls, the presence of a "super-homogeneity" that allows to hypothesize the annual mean of precipitation to be constant in the considered region. The second strategy, once admitted the existence of a super-homogeneity, is aimed at the definition of a single spatial distribution of extreme events of 1h rainfalls. In this way, it is possible to carry out the statistical inference using a single time series formed by annual maxima of daily rainfalls occurring all over the region. The frequency distribution thus obtained, allows to estimate the probability of non-exceedance F(X) of an event X on the entire region. The proposed procedure was applied to the Vibo Valentia flood on 3 July 2006.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
