An empirical evaluation of the determinants of moose-vehicle collisions on the island of Newfoundland, Canada

Abstract

Moose-vehicle collisions (MVCs) are a problem throughout the circumpolar range of moose, but are especially prevalent on the island of Newfoundland, Canada. I designed a field study which determined that a common MVC mitigation strategy, roadside vegetation cutting, does not attract moose into roadside areas to browse. I also conducted a spatial analysis and identified small scale MVC hotspots scattered throughout the island, and medium and large scale MVC hotspots on primary roads and on the Avalon Peninsula. Finally, I used model selection to identify the best spatial predictors of the probability of occurrence of MVCs in Newfoundland. Specifically, primary roads, straight roads, decreased distance to large cities, and decreased distance to mining areas are associated with areas of high MVCs rates. This research provides managers with a basis for i) continuing roadside vegetation cutting and ii) implementing MVC mitigation strategies in strategic areas to reduce the number of MVCs.