Modeling the Effects of Climate Change on the Transmission of Malaria in the Province of Limpopo, South Africa

Abstract

The association between malaria cases and climatic factors has been examined in recent research using statistical and mathematical models. Several statistical models in the literature used a time-series approach based on the Box-Jenkins methodology without considering the type of data used. Frequently, cases of malaria were seen to be over-dispersed and might exhibit an autoregressive conditional heteroscedasticity impact. This paper provides a thorough analysis of the features of malaria cases, including autoregressive conditional heteroscedasticity, randomness, linearity, and over-dispersion, from all five districts in the province of Limpopo. Additionally, this study develops integer-valued generalized autoregressive conditional heteroskedasticity (INGARCH) and autoregressive integrated moving average (ARIMA) models to examine the dependence of climatic variables on the malaria case, identify the model that best predicts reported cases of malaria, and determine how climate variables affect malaria transmission in that geographic area. According to the findings, the risk of malaria transmission increases in all the districts found in the province of Limpopo at lags of 0, 1, 2, 3, and 4 months, depending on the maximum and minimal temperatures as well as the type and amount of rainfall. Using the mean absolute error (MAE) to compare the two models' forecast accuracy, it was found that the ARIMA model performed better in four districts, while the INGARCH model suited one district municipality better. The research validates the anticipated correlation between climatic factors and malaria transmission; nevertheless, the disparate outcomes of the two models in distinct regions warrant further investigation.