Road Ecology Center

Large-scale conservation efforts seek to maintain habitat connections so that native wildlife (and plant) species may move across the landscape as necessary to meet their needs to survive and reproduce. Barriers caused by roads and railways pose a significant impediment to wildlife movement at all scales throughout the U.S. Northern Rockies area, and a risk of injury or death to animals whose needs require crossing when traffic is present. In turn, animals on highways pose a risk of injury or death to motorists and property damage to vehicles. As traffic volumes increase, these risks also increase. Bozeman Pass is just beginning to experience significant conflicts with wildlife. In addition to a four-lane freeway (Interstate 90) there are parallel frontage roads and a railway. As traffic volumes continue to increase the problems will only get worse. To plan for inevitable growth in human populations and traffic volumes, and to fulfill the mandates of the Transportation Equity Act for the 21st Century (TEA-21) regarding wildlife needs and public safety, it is imperative that options for wildlife conflict mitigation be started as soon as possible on Bozeman Pass. This study attempts to identify the problem areas for wildlife and human safety at Bozeman Pass and make recommendations about how and where to mitigate wildlife mortality and human safety issues in the connectivity zone. Several moose, mountain lions, black bear, deer, elk, small mammals, and one wolf have been killed by traffic within the past two years. GIS models and maps have been developed for this project to summarize location data for wildlife-vehicle collisions, wildlife movement corridors, wildlife habitat, and potential sites for wildlife crossing structures. GIS models using least-cost-path analysis were compared with the known locations of road-kills and model predictions were close to actual crossing points. Differences between the field data and the model data suggest that the models can be improved by incorporating additional data layers and perhaps by adjusting the weights of model variables.