Diagnostic Study of Seasonal Prediction of Malaria: Case of Senegal, West Africa.

Beyond assessment and analysis of observed and simulated malaria parameters, this study is undertaken furthermore in the framework of predictability of malaria outbreaks in West Africa and Senegal in particular. Predictability of September-November malaria incidence is evaluated over Senegal, employing the Canonical Correlation Analysis (CCA), where the predictand is gridded malaria incidence hindcasts simulated by the Liverpool malaria Model (LMM) fed by daily data of the CPC Global Temperature and Africa Rainfall Climatology Version 2 (ARC2); and the predictor is the observed sea surface temperatures, i.e. the observed sea surface temperatures of the Extended Reconstructed Sea Surface Temperature version 4 (ERSSTv4) and the predicted sea surface temperature of the North American Multi-Model Ensemble (NMME´s Models). The analysis is based on three main ocean basins: the Tropical Pacific, the Gulf of Guinea and the Tropical North Atlantic. The selected ocean basins are well known to have a strong influence on the West African monsoon dynamics which in turn indirectly impacts other sectors including the seasonal outbreaks of vector-borne diseases such as malaria. The analysis is performed for the period 1982-2010. The paper highlight that there is an indirect relationship between malaria and SSTs. Good skills for malaria predictability for the region are found. The overall signal for tropical conditions is analyzed, adding new ideas about the Pacific-Atlantic link over the monsoon period. This encourages malaria prediction diagnostic to be extended to the whole Sahel region and the Gulf of Guinea.

This study is part of an ongoing work on the generation of improved seasonal forecasts to promote for the health sector over West Africa. This work is not yet fully implemented into operational practice, and it is therefore timely to evaluate their potential benefits, particularly after post processing with the CCA technique discussed.