Abstract:
Of the many negative effects roads can have on wildlife, wildlife-vehicle collisions are the most devastating. Efforts to predict where wildlife cross roads are vital for mitigation and prevention efforts. In this study, a Geographic Information Systems (GIS)-based approach was used to evaluate the relationship between wildlife-vehicle collisions, road traffic volume, and wildlife habitat suitability near roads. Road characteristics that potentially affect driver visibility and travel speeds, including the slope and curviness of the roads, were also considered. The robustness of the results was evaluated by varying the maximum length of the road segments in the spatial data and the distance from the roads used to estimate a wildlife crossing index based on habitat suitability. The case study evaluated moose (Alces alces) and black bear (Ursus americanus) collisions in VT from 1990 to 2006 for all roads in the state, three major roadways, and four functional classes of roadways. Habitat suitability had the most consistent results across models, as road segments with better suitability had a higher collision density. The robustness analysis showed that as the buffer distance used to estimate the wildlife crossing index increased, the explained variation of wildlife-vehicle collision density increased as well. Road traffic volume demonstrated mixed results across models, as higher volume was associated with more collisions in the models with all roads, but was associated with fewer collisions in the roadway-specific and functional classification models. The length of the road segments in the spatial data layer affected the predictive power of the models, suggesting that scale may be an important factor in characterizing these relationships. The results offer an improved understanding of wildlife-vehicle collisions, which can potentially be used to develop mitigation and prevention efforts aimed at reducing the negative effects of roads on wildlife.
