Information on the spatial relationships between disease vectors and environmental factors is fundamental to vector-borne disease control. Although it is well known that mosquito abundance is associated with the amount of rainfall and thus the number of larval breeding sites, the spatial relationship between larval habitat availability and adult mosquito abundance is not clear. We investigated the impact of environmental heterogeneity and larval habitats on the spatial distribution of Anopheles gambiae s. s. and An. funestus adult mosquitoes, the most important malaria vectors in the highlands of western Kenya. Mosquito sampling was conducted in May, August, and November 2002, and February 2003. Geographic information system layers of larval habitats, land use type, human population distribution, house structure, and hydrologic schemes were overlaid with adult mosquito abundance. Correlography was used to determine the spatial autocorrelation in adult mosquito abundance among houses and the cross-correlation between adult mosquito abundance and environmental factors. Getis' G(i)(*)(d) index was used to define focal adult mosquito abundance clusters. We found a significant autocorrelation in the vector population and a significant cross-correlation between the vector population and larval habitat availability. The threshold distances of both autocorrelation and cross-correlation were significantly varied among seasons. Focal clustering analysis revealed that the adult vector population was concentrated along the Yala River Valley where most larval habitats were found. Regression analysis found that distance of a house to the Yala River, age of the house, elevation, house structure, and tree canopy coverage significantly affected adult mosquito abundance. Our results suggest that vector control targeted at malaria transmission hotspots and supplemented by larval control may be an effective approach for epidemic malaria control in the western Kenya highlands.
展开▼
