Should we predict wildlife–vehicle collision hotspots from road and landscape characteristics?

Wildlife-vehicle collisions (WVCs) represent a global challenge for human safety and biodiversity conservation. This study investigates the environmental and road characteristics that influence WVC risk, using the province of Parma (Northern Italy) as a representative case study. We analysed 693 collision events recorded between 2013 and 2018, most of which involving roe deer (Capreolus capreolus, 86%) and wild boar (Sus scrofa, 10%). Spatial and temporal analysis revealed how collisions are not randomly distributed, peaking on heavily trafficked major roads where high average speeds coincide with fragmented landscapes. Proximity to croplands and riparian corridors significantly increased collision probability, particularly at the peri-urban interface. Temporal patterns confirm that most WVCs occur during crepuscular and nocturnal hours, aligning with peak ungulate activity, rather than highest traffic volumes. Our research identifies critical road segments from environmental predictors, providing a robust baseline for the implementation of targeted mitigation measures, such as wildlife underpasses and road signage, to enhance both habitats permeability and drivers' safety.