Informations et ressources scientifiques
sur le développement des zones arides et semi-arides

Accueil du site → Master → Portugal → Previsão sazonal de índices de seca com modelos de Markov

Universidade de Lisboa (2009)

Previsão sazonal de índices de seca com modelos de Markov

Sousa, João Miguel Barbosa Botelho de

Titre : Previsão sazonal de índices de seca com modelos de Markov

Auteur : Sousa, João Miguel Barbosa Botelho de

Université de soutenance : Universidade de Lisboa

Grade : Mestrado, Ciências Geofísicas (Meteorologia 2009

Drought is a recurrent phenomenon that has affected civilizations throughout history, with random occurrence and uncertain severity, due to outliers in rainfall. It affects the natural habitats, ecosystems and many social and economic sectors from agriculture, transport, urban water supply and modern industrial plants. Due to the random characteristics of the drought, one has considered a stochastic approach to study and characterize this phenomenon. This paper used a stochastic model of type Markov chains to the SPI index (Standard Precipitation Index) in order to identify, characterize and monitor the phenomenon under study. This model is also useful as a tool for forecasting and prediction of the drought. Given the known influence of atmospheric circulation patterns in large-scale seasonal rainfall, we have added some changes to the Markov chain model in order to increase its predictive ability, including information from these patterns by using indexes of global atmospheric circulation (GAC) associated with them. The calibration of these two Markov models, which differ by the inclusion or not of the GAC indexes (or meteorological predictors of large-scale), is done through time series of rainfall and time series of GAC indexes, to the site of study, the Alentejo region. These two models are then validated for time instants in the future and its predictive quality is measured


Version intégrale (1,49 Mb)

Page publiée le 23 octobre 2013, mise à jour le 25 août 2018