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：2005
卷号：16
期号：3
页码：189-198
出版社：John Libbey Eurotext
摘要：Figures See all figures Authors Abdoulaye Deme , Pierre de Felice , Alain Viltard Laboratoire de Météorologie Dynamique/IPSL, École Polytechnique, 91128 Palaiseau cedex Key words: Senegal, Rainfall, Indices Page(s) : 189-98 Published in: 2005 The aim of this paper is to develop daily rainfall prediction regression equations for Dakar (Senegal), for lead-times of 1 to 3 days. The daily rainfall is an average calculated by the Institut de recherche pour le développement (IRD) based on the rain gauges situated in the region of Dakar, for the years 1968-1990. To avoid all problems of trend, only the month of August is used because then the Intertropical Convergence Zone (ITCZ) lies at its most northerly position. Sixty-five indices are built using the National Center for Environmental Prediction-National Center for Atmospheric Research (NCEP-NCAR) reanalysis dataset. Several of these indices have been applied by meteorologists of the Sahel, France and the United States to predict rainfall, thunderstorm and tornado occurrences. The regression equations have a skill better than that provided by climatological forecasts, though large rainfall amounts (>8mm) are poorly predicted. Regression equations for predicting rainfall accumulations over 3 consecutive days are also developed. With these equations, the error rate in the forecast of small rainfall amounts is very small.
