Modeling the role of rainfall patterns in seasonal malaria transmission

Abstract

Seasonal total precipitation is well known to affect malaria transmission because Anopheles mosquitoes depend on standing water for breeding habitat. However, the within-season temporal pattern of the rainfall influences persistence of standing water and thus rainfall patterns can also affect mosquito population dynamics in water-limited environments. Here, using a numerical simulation, I show that intraseasonal rainfall pattern accounts for 39% of the variance in simulated mosquito abundance in a Niger Sahel village where malaria is endemic but highly seasonal. I apply a field validated coupled hydrology and entomology model. Using synthetic rainfall time series generated using a stationary first-order Markov Chain model, I hold all variables except hourly rainfall constant, thus isolating the contribution of rainfall pattern to variance in mosquito abundance. I further show the utility of hydrology modeling using topography to assess precipitation effects by analyzing collected water. Time-integrated surface area of pools explains 70% of the variance in simulated mosquito abundance from a mechanistic model, and time-integrated surface area of pools persisting longer than 7 days explains 82% of the variance. Correlations using the hydrology model output explain more variance in mosquito abundance than the 60% from rainfall totals. I extend this analysis to investigate the impacts of this effect on malaria vector mosquito populations under climate shift scenarios, holding all climate variables except precipitation constant. In these scenarios, rainfall mean and variance change with climatic change, and the modeling approach evaluates the impact of non-stationarity in rainfall and the associated rainfall patterns on expected mosquito activity.