期刊名称：Science et changements planétaires / Sécheresse
印刷版ISSN：1147-7806
电子版ISSN：1777-5922
出版年度：2008
卷号：19
期号：4
页码：227-235
DOI：10.1684/sec.2008.0148
出版社：John Libbey Eurotext
摘要：Figures See all figures Authors Abdou Ali , Thierry Lebel , Abou Amani Institut de recherche pour le développement (IRD)/Centre régional AGRHYMET BP 5040 Niamey Niger, Laboratoire de météorologie dynamique École Polytechnique Route de Saclay 91128 Palaiseau cedex France, Institut de recherche pour le développement (IRD) Avenue de Maradi BP 11416 Niamey Niger, UNESCO - Accra Office PO BOX CT 4949 Nortei Ababio Street 32 Airport Residential Area Accra Ghana Key words: drought, probabilistic approach, rainfall index, Sahel, spatial variability DOI : 10.1684/sec.2008.0148 Page(s) : 227-35 Published in: 2008 The Standardized Precipitation Index (SPI) is commonly used to determine whether the Sahelian region can be considered as wet or dry for a given year. There are however several methodological points worth examining to assess the pertinence of using such a global index for characterising the rainy season in a region where the spatial variability of rainfall is large. Three such major points are addressed in this study: i) do the operational networks make it possible to compute interannual rainfall anomalies with reasonable accuracy? The answer is yes. ii) The Sahel cannot be considered as a homogeneous region in terms of interannual rainfall variability. As a consequence, is it pertinent to use a unique SPI for the whole Sahel? The answer is no: three sub-regions (West, Central, East) are identified and it is shown that, especially since 1990, the West and East SPI are not strongly correlated, meaning that computing a global SPI for the whole Sahel does not make sense every year. iii) Given this result, what usage can be made of the SPI? By studying the spatial variability of the SPI computed at the local scale of a 0.5°x 0.5° grid mesh, a probabilistic model is proposed making it possible to compute the expected proportion of the Sahel where the SPI is below or above any threshold for a given value of the regional index.
