Author(s): G. Capparelli; M. Giorgio; R. Greco; P. Versace
Linked Author(s): Roberto Greco
Keywords: No Keywords
Abstract: Effectiveness of floods and rainfall early warning systems could be improved by reliable quantitative rainfall prediction, which can allow gaining larger lead-times. To this aim, an empirical predictor of landslides mobilization function which uses hourly rainfall data collected in real time is proposed. The core of the predictive procedure consists in a DRIP model of rainfall height series external structure. Thus, rain storms and dry periods, constituting the external structure of rainfall time series, are modeled as an alternating renewal process. Three random variables, rain storm duration, rain storm total height and dry period duration, have been defined so to obtain the required form of association. The main advantages offered by the adopted model structure is that it enables to use in the analysis only the pieces of historical information which really affect the future behavior of a point rainfall series. The predictive procedure has been tested against the 11 years long hourly rainfall height series of the rain gauge of Lanzo, northern Italy, showing encouraging results.
Year: 2009