Independence of rainfalls through the rainy season and the implications for the estimation of rainfall probabilities

✍ Scribed by Dennett, M. D. ;Rodgers, J. A. ;Stern, R. D.

• Publisher: Wiley (John Wiley & Sons)
• Year: 1983
• Weight: 615 KB
• Volume: 3
• Category: Article
• ISSN: 2314-6214
• DOI: 10.1002/joc.3370030405

✦ Synopsis

Abstract

A detailed study of more than 50 years of daily rainfall records from three sites in seasonally arid West Africa revealed no dependence between rainfall at different times within the rainy season. Years were grouped on the basis of the date of the start of the rains and rainfall models fitted separately to early, medium and late start groups. Any differences between the groups in subsequent rainfall pattern bordered on statistical significance and were too small to be of practical significance. Thus, the best estimates of the probabilities of rainfall events are obtained by considering all years of the rainfall record. As rainfalls at different times of the season can be assumed to be independent, it is straightforward to calculate conditional distributions, for example, the probability distribution of the length of the rains given the actual starting date. The calculation and value of using conditional distributions are discussed.