Application of advanced very high-resolution radiometer (AVHRR)-based vegetation health indices for modelling and predicting malaria in Northern Benin, West Africa.

BACKGROUND: Vegetation health (VH) is a powerful characteristic for forecasting malaria incidence in regions where the disease is prevalent. This study aims to determine how vegetation health affects the prevalence of malaria and create seasonal weather forecasts using NOAA/AVHRR environmental satellite data that can be substituted for malaria epidemic forecasts. METHODS: Weekly advanced very high-resolution radiometer (AVHRR) data were retrieved from the NOAA satellite website from 2009 to 2021. The monthly number of malaria cases was collected from the Ministry of Health of Benin from 2009 to 2021 and matched with AVHRR data. Pearson correlation was calculated to investigate the impact of vegetation health on malaria transmission. Ordinary least squares (OLS), support vector machine (SVM) and principal component regression (PCR) were applied to forecast the monthly number of cases of malaria in Northern Benin. A random sample of proposed models was used to assess accuracy and bias. RESULTS: Estimates place the annual percentage rise in malaria cases at 9.07% over 2009-2021 period. Moisture (VCI) for weeks 19-21 predicts 75% of the number of malaria cases in the month of the start of high mosquito activities. Soil temperature (TCI) and vegetation health index (VHI) predicted one month earlier than the start of mosquito activities through transmission, 78% of monthly malaria incidence. CONCLUSIONS: SVM model D is more effective than OLS model A in the prediction of malaria incidence in Northern Benin. These models are a very useful tool for stakeholders looking to lessen the impact of malaria in Benin.

Towards an intelligent malaria outbreak warning model based intelligent malaria outbreak warning in the northern part of Benin, West Africa.

BACKGROUND: Malaria is one of the major vector-borne diseases most sensitive to climatic change in West Africa. The prevention and reduction of malaria are very difficult in Benin due to poverty, economic insatiability and the non control of environmental determinants. This study aims to develop an intelligent outbreak malaria early warning model driven by monthly time series climatic variables in the northern part of Benin. METHODS: Climate data from nine rain gauge stations and malaria incidence data from 2009 to 2021 were extracted from the National Meteorological Agency (METEO) and the Ministry of Health of Benin, respectively. Projected relative humidity and temperature were obtained from the coordinated regional downscaling experiment (CORDEX) simulations of the Rossby Centre Regional Atmospheric regional climate model (RCA4). A structural equation model was employed to determine the effects of climatic variables on malaria incidence. We developed an intelligent malaria early warning model to predict the prevalence of malaria using machine learning by applying three machine learning algorithms, including linear regression (LiR), support vector machine (SVM), and negative binomial regression (NBiR). RESULTS: Two ecological factors such as factor 1 (related to average mean relative humidity, average maximum relative humidity, and average maximal temperature) and factor 2 (related to average minimal temperature) affect the incidence of malaria. Support vector machine regression is the best-performing algorithm, predicting 82% of malaria incidence in the northern part of Benin. The projection reveals an increase in malaria incidence under RCP4.5 and RCP8.5 over the studied period. CONCLUSION: These results reveal that the northern part of Benin is at high risk of malaria, and specific malaria control programs are urged to reduce the risk of malaria.