Temporal Analysis and Prediction of MalariaDynamics Using Meteorological Data in Southeastof Senegal

In Senegal, malaria is a seasonal disease predominantly concentrated in the southeast even before 2016. This study investigates the relationship between weekly malaria cases and meteorological factors in four districts of Senegal (Kédougou, Sal´emata, Saraya, and Dianké Makha). Using a Generalized Additive Model (GAM), three techniques were tested to identify key meteorological variables. Data on malaria cases were sourced from DHSI2 and meteorological data from NASA GIOVANNI. Principal component analysis (PCA) was used to identify the most explanatory variables, with climate regimes (clusters) determined via Hierarchical Ascending Classification (HAC). For positive and high values, the two lag weeks of the first component are significantly associated with a reduction in malaria risk. This suggests that a substantial increase in this variable, with a delay of two weeks, could have a protective effect against malaria. The two week lag of the second component appears to have a nonlinear relationship with malaria cases: moderate levels may be protective, while very high values are associated with an increased risk. Moderate to high, the third component 15 weeks before the observation period appears to have a moderate protective effect on malaria risk, with significant reductions in risk observed in the first deciles (50- 70%). However, very high rainfall during the 15-week delay increases the risk of malaria in higher deciles (80-100%), suggesting that excessive values of the third component may be detrimental, promoting mosquito proliferation and increasing malaria transmission. These findings emphasize the impact of meteorological variations on malaria transmission, helping to improve predictive models and intervention strategies