Habitat fragmentation caused by Utah road systems contributes to collisions of vehicles with Rocky Mountain mule deer (Odocoileus hemionus hemionus), causing death, injury to humans and animals and costing up to 7.5 million dollars each year in damage. Our research focuses on identification of components of the natural and built environment that contribute to deer-vehicle collisions (DVC).
Our hypotheses were that vegetation, topographic roughness, and status of fences along roadways affect WVC density and distribution.
Our team gathered data on fence status along highways in southwestern Utah, and we used topographic roughness and vegetation cover data from LANDFIRE. We used field and remotely sensed data to predict the density of DVCs along Highways 18, 20, 21, 56, and 130. We managed the data by building a series of multiple regression models in the R environment, and used model selection and model averaging to identify the role of each covariate in predicting DVCs.
Preliminary results based on Highway 56 suggest that of the variables we tested, fencing and topographic roughness were the most important predictors of the presence of DVC hotspots.
This work will allow us to make recommendations to road planners on the locations of potential mitigation structures,, such as appropriately sized culverts or revised fencing, which could significantly reduce risk to drivers and wildlife.