Regional estimation of extreme daily rainfall in West Africa (Côte d'Ivoire)

Abstract

The extremes rainfalls are meteorological hazards that cause great damage and many casualties. Information on the occurrence of extreme rainfall  events is therefore important to protect people. However, the rainfall information is not available in many areas because of deficiencies or lack of data.  Regional frequency analysis is an alternative to provide robust information about quantile. In this perspective, the monthly and annual maximum rainfall  of 44 stations in the Côte d'Ivoire over the period 1942-2002 were analyzed. Four contiguous homogeneous climatic regions were determined using the  hierarchical cluster and validated by the test the homogeneity of Wiltshire. The analyses show that the series are adjusted by generalized Logistics distribution and generalized extreme values distribution. The differences between local and regional quantile of extreme rainfall are relatively low. It is of  the order of 5% in terms of bias and 15% relative mean squared error. This attests to the robustness of the regional approach for the estimation of  extreme rainfall events.