Spatial Modelling of Malaria Risk in Kwara State Using Remote Sensing Environmental Predictors

Malaria transmission in Nigeria is strongly modulated by environmental conditions that are observable through satellite remote sensing. This study develops a spatial risk model for malaria transmission in Kwara State using remote sensing-derived environmental predictors. Malaria case data from primary health care facilities across 16 LGAs for the period 2020 to 2023 are obtained from the Kwara State Primary Health Care Development Agency and spatially linked to ward-level administrative units. Environmental predictors including land surface temperature, NDVI, Enhanced Vegetation Index, distance to water bodies, and rainfall anomaly are derived from Landsat 8, MODIS, and CHIRPS datasets. A geographically weighted regression model and a Bayesian spatial model are developed and compared for predictive performance. The Bayesian model outperforms the geographically weighted regression in both predictive accuracy and spatial uncertainty quantification. Results identify high-risk clusters in the Edu, Moro, and Patigi LGAs, characterised by high temperature, proximity to irrigation schemes, and low vegetation maintenance. Keywords: malaria risk, spatial modelling, remote sensing, Kwara State, environmental predictors