Predicting deer–vehicle collisions in an urban area

Collisions with deer and other large animals are increasing, and the resulting economic costs and risks to public safety have made mitigation measures a priority for both city and wildlife managers. We created landscape models to describe and predict deer–vehicle collision (DVCs) within the City of Edmonton, Alberta. Models based on roadside characteristics revealed that DVCs occurred frequently where roadside vegetation was both denser and more diverse, and that DVCs were more likely to occur when the groomed width of roadside right-of-ways was smaller. No DVCs occurred where the width of the vegetation-free or manicured roadside buffer was greater than 40m. Landscape-based models showed that DVCs were more likely in more heterogeneous landscapes where road densities were lower and speed limits were higher, and where non-forested vegetation such as farmland was in closer proximity to larger tracts of forest. These models can help wildlife and transportation managers to identify locations of high collision frequency for mitigation. Modifying certain landscape and roadside habitats can be an effective way to reduce deer–vehicle collisions. ► Deer–vehicle collisions occur where roadside vegetation is more diverse and dense. ► Deer–vehicle collisions occur where road densities are low and speed limits are high. ► Models of deer–vehicle collisions can be helpful in mitigation and prediction. Deer –vehicle collisions can be mitigated by modifying landscape and roadside habitats.