Event-Scale Characterization of Precipitation Observed at Two Sites in West Africa: Abidjan, Côte d’Ivoire and Nalohou, Benin

Abstract

The aim of this work is to separate the convective and stratiform components of tropical rainfall in Africa without first imposing a subjective criterion on the intensity threshold. Indeed, understanding the structure of precipitation observed in the localities of Abidjan and Nalohou as well as distinguishing between its convection and stratiform components, remains a challenge for the scientific community. Previous studies in this tropical region of Africa have typically relied on the use of a fixed rain intensity threshold to separate convective from stratiform precipitation. However, in our study the cluster approach named Functional Principal Component Analysis Clustering (FPCAC) was used to automatically separate convective and stratiform precipitation without imposing a critical rain intensity threshold. The rainfall data used in this study were collected at a 5-minute temporal resolution from the Abidjan (Côte d'Ivoire) and Nalohou (Benin) sites. To facilitate the application of our algorithm, we segmented our data to constitute samples of rainy events. The optimal number of clusters was determined using the silhouette method. As a result, four precipitation groups were identified at the coastal site of Abidjan, while the inland site of Nalohou exhibited only two groups. The grouping was based on the rain intensity profiles of individual events. We were able to assign a convective or stratiform nature to each group. At Abidjan, three groups (C1, C2, and C3) were classified as convective and one group (C4) as stratiform. At Nalohou, one group (C1) was convective and the other (C2) stratiform. Such results will make it possible to increase the quality of remotely sensed measurements (radars/satellite) in tropical regions. Furthermore, the results obtained are consistent with those of previous studies, supporting the robustness of our approach.