Statistical analysis of habitat distribution of host-seeking vector mosquitoes of human diseases based on land cover composition: Case study of vector inhabitation along the eastern coast of Lake Biwa, Japan

This study examined the determinants of spatial variation of host-seeking vector mosquitoes of Japanese encephalitis, Culex tritaeniorhynchus, and malaria vector mosquitoes, the Anopheles sinensis group, to estimate the distribution of mosquito habitats along the eastern coast of Lake Biwa, Japan. Data for the statistical analysis were obtained from a mosquito survey conducted using 20 traps during May to October 2009. For explanatory variables, we constructed a dataset of land cover in the study area by classifying satellite images of ALOS. We employed partial least squares (PLS) regression to analyze the relationship between the number of mosquitoes collected and land cover composition around trap sites. The results indicate that the number of host-seeking Cx. tritaeniorhynchus was larger when the areal components of the rural-agricultural landscape (e.g., paddy fields) were larger around the trap sites. For the An. sinensis group, the number of host-seeking mosquitoes was larger when the areal components of the wetland landscape (e.g., water bodies and reed beds) were larger around the trap sites. The potential habitat maps derived by fitting the PLS regression model into a GIS environment indicated that suitable habitats for Cx. tritaeniorhynchus are mainly distributed across plains and that for the An. sinensis group are concentrated around water bodies. We interpreted these results to mean that the distribution of the estimated potential habitats of host-seeking mosquitoes reflects the behavioral response of mosquitoes to spatially heterogeneous environments. Thus, the mapping of potential habitats estimated using datasets of vector mosquitoes and land cover may provide fundamental information for planning effective control strategies for vector mosquitoes.