6533b828fe1ef96bd1287ad9
RESEARCH PRODUCT
A non-homogeneous Poisson based model for daily rainfall data
Alberto LombardoAntonina Pirrottasubject
Stochastic modellingSettore SECS-S/02 - Statistica Per La Ricerca Sperimentale E TecnologicaGeodetic datumConfidence Region Daily Rainfall Data Linear Stochastic Differential Equation Poisson White Noise Probabilistic Engineer MechanicsImpulse (physics)Poisson distributionsymbols.namesakeNon homogeneousStatisticssymbolsProbability distributionSettore ICAR/08 - Scienza Delle CostruzioniRandom variableConfidence regionMathematicsdescription
In this paper we report some results of the application of a new stochastic model applied to rainfall daily data. The Poisson models, characterized only by the expected rate of events (impulse occurrences, that is the mean number of impulses per unit time) and the assigned probability distribution of the phenomenon magnitude, do not take into consideration the datum regarding the duration of the occurrences, that is fundamental from a hydrological point of view. In order to describe the phenomenon in a way more adherent to its physical nature, we propose a new model simple and manageable. This model takes into account another random variable, representing the duration of the rainfall due to the same occurrence. Estimated parameters of both models and related confidence regions are obtained.
| year | journal | country | edition | language |
|---|---|---|---|---|
| 2007-08-03 |