A POINT STOCHASTIC MODEL FOR CONTINENTAL DROUGHTS / Modèle stochastique ponctuel pour les sécheresses continentales

Open Access
Publisher Website
Google Scholar
Abstract
A two-level point stochastic model is used to represent the daily rainfall occurrences in the midwestern United States. The primary level corresponds to the generating mechanism of the rainfall occurrences such as fronts and thunderstorm clouds. The secondary level corresponds to the actual rainfall occurrences. The persistence of the rainfall occurrences is basically due to the persistence of the rainfall generating mechanisms. The persistence of the actual rainfalls can be explained in terms of their clustering around the rainfall generating mechanism. Based on these physical facts a point stochastic cluster model of the Neyman-Scott type is constructed for the rainfall occurrences around the Great Lakes region of the North American continent. This two-level point stochastic model utilizes the meteorological characteristics of the rainfall generating mechanisms for the simulation of the probabilities of dry and wet sequences. The probabilities of drought lengths and the return period of rainfalls are derived from the Neyman-Scott cluster model for practical drought computations. It is shown that the model can very well preserve the marginal probability of drought lengths in the central North American continent.

This publication has 1 reference indexed in Scilit: