| Summary: | Degree: Master of Science Abstract: Mortality caused by collisions with trains affects numerous species globally and has increased recently to threaten the grizzly bear (Ursus arctos) population protected in Banff and Yoho National Parks, Canada. Although train collisions are the ultimate cause of mortality, the more proximate processes that cause grizzly bears to use the railway are unclear. No previous study has assessed the relative strength of grizzly bear attraction to the railway compared to other linear features in the study area, including power lines, a secondary road and the right-of-way along the TransCanada Highway fence. According to local knowledge, male bears make greater use of the rail than females, but there has been no quantitative assessment of this putative tendency. I hypothesized that selection for the railway is strongest in spring and fall, when forage resources are limited. I also hypothesized that male bears would select the railway more strongly than females to increase one or more of forage selection, travel efficiency or access to social dominance and mating opportunities. Using GPS collar data collected by Parks Canada, I modeled individual and population-level resource selection functions to compare selection by feature type and season. I also conducted repeated vegetation sampling in each linear feature type and season to quantify the availability of bear forage. Percent cover of vegetation available to grizzly bears was higher in linear features than adjacent forests, lower on the rail than other feature types and lower in the fall than the spring. At the population level, grizzly bears increased selection for the railway in spring and fall and the power line in summer. I did not find evidence of consistent sexual segregation within or among linear features and selection for linear features was highly variable even within classes of individuals. My results suggest that individual experiences and attributes, as well as temporal and spatial features of local landscapes, have ...
|
|---|