What Features of the Landscape and Highway Influence Ungulate Vehicle Collisions in the Watersheds of the Central Canadian Rocky Mountains: A Fine-Scale Perspective?

Wildlife-vehicle collisions represent an additive source of mortality to wildlife populations, in addition to other mortality, such as predation and disease. The trends of increasing traffic volumes and road densities will only magnify the mortality impacts of roads on large mammals and other vertebrates. In this study, we examined the descriptive and spatial aspects of ungulate-vehicle collisions (UVCs) in the Central Canadian Rocky Mountains (CCRMs). We then specifically addressed the landscape and highway characteristics associated with the UVCs in four major watersheds: the Bow Valley, Kananaskis Valley, Kicking Horse Valley, and Kootenay Valley, each with differing road-types, topography, and habitat. We grouped the factors associated with vehicle collisions into three groups: combined, landscape-animal, and highway-vehicular-animal. The combined model included all variables, the landscape-animal model included factors that influence whether an animal makes it to the roadway, and the road-vehicular model included factors that influence the probability of an interaction between the animal and the vehicle. Between 1999 and 2003 all kill sites were initially measured with a Global Positioning System (GPS) (accuracy m) and later revisited to measure all field measurements. Many other studies have looked at the factors associated with wildlife vehicle collisions; however, our study is unique in that we were able to revisit exact collision sites (accuracy m). There were a total of 546 ungulate mortalities on all highways in the watershed with the majority occurring in the Bow Valley followed by the Kicking Horse Valley, and Kananaskis Valley, and the least occurring in Kootenay Valley. The distribution of kills was correlated with the traffic volumes on each road-type. Further, UVC distributions differed significantly from random distributions along all road types in each watershed. Type of habitat was the most important variable in explaining UVCs in the combined, landscape and Bow watershed models. UVCs were less likely to occur in open water, rock, and closed coniferous forest relative to open habitat. The proportion of open vegetation in the Bow Valley positively influenced wildlife mortality, while in the Kicking Horse watershed it negatively influenced mortality. Width and traffic volume were significantly positively correlated with the occurrence of UVCs in the combined model and Bow model, respectively. Elevation was a significant factor in the combined, landscape, Bow, and Kootenay watersheds, having a negative correlation on ungulate mortality. The proportion of open habitat positively contributed to kills in the Bow; whereas, it negatively influenced kills in the Kicking Horse. The three grouped models were ranked differently in their ability to predict the observed likelihood for UVCs. The combined model was the most important model in predicting the occurrence of UVCs, followed by the landscape model, and lastly the road-vehicular-animal model. Our findings show that kills do not occur randomly in the landscape. Different scales of analysis, i.e., ecoregion or watershed perspective, can influence which variables are important in contributing to the spatial distribution of UVCs. Further, different groups of variables, i.e., roads and motorist related factors, or landscape and animal behavior factors, may contribute differently to the spatial occurrence of UVCs. The factors contributing to UVCs along each landscape and highway are critical for developing knowledge-based mitigation for reducing effects of vehicle collisions on large animal populations and increasing public safety on highways.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.