Spatial role of land cover on West Nile virus disease in Europe

We analyzed West Nile virus (WNV) disease incidence across European provinces from 2005 to 2019. Using spatial regression models, we quantified how land-cover gradients, climatic conditions, and socio-demographic variables jointly shape spatial heterogeneity in WNV disease incidence in humans. Shrubland cover showed the strongest and most spatially consistent positive association with human WNV disease incidence, whereas forest cover generally exhibited a negative relationship; urban and cropland areas had weaker, regionally variable effects. Climatic factors-particularly warm summer temperatures and seasonal moisture balance-emerged as dominant predictors, while socio-economic variables contributed little at this scale. The spatially adaptive spatially lagged geographically weighted regression (GWR-SL) model revealed pronounced regional variation in these associations, highlighting that WNV disease drivers are highly context dependent. These findings underscore the value of integrating land-cover and climatic information into targeted surveillance and vector-control strategies, while acknowledging limitations related to land-cover aggregation, climatic averaging, and underreporting.