DataSheet_1_Bison and elk spatiotemporal interactions in Elk Island National Park.docx

Elk Island National Park (EINP) is a fenced park in the Beaver Hills UNESCO Biosphere Reserve of central Alberta where aspen parkland is being conserved. This area maintains high densities of native ungulates including elk (Cervus elaphus), bison (Bison bison bison, B. bison athabascae), moose (Alces alces), white-tailed deer (Odocoileus virginianus), and mule deer (O. hemionus). Our objective was to evaluate spatiotemporal overlap and interspecific interactions between bison and elk in EINP using camera traps. Bison and elk have overlapping habitat use and diet and are the primary focus of ungulate management in EINP. We examined distributions of both species in relation to season, landscape characteristics, anthropogenic features, and heterospecifics using generalized linear models (GLMs). We then examined seasonal daily activity patterns of bison and elk and calculated the degree of overlap. The spatial analysis revealed that bison counts were positively associated with higher proportions of open habitats across seasons and in areas farther from water in summer and fall but had no associations with distance to water during winter. Bison removal year (years during which bison removals were conducted by Parks Canada) was a significant predictor variable for bison counts in winter when the bison roundup takes place. Elk avoided areas with high linear feature density across seasons. During fall and winter, we observed higher elk counts associated with bison presence. Temporal activity patterns revealed that elk were crepuscular in all three seasons, but bison activity patterns varied with diurnal activity being more common in the summer, crepuscular activity in winter, and intermediate activity patterns during fall. Coefficients of overlap between elk and bison were high in all three seasons with the greatest difference in daily activity patterns in summer and the highest overlap in winter when both species showed strong crepuscular activity. Despite the fenced perimeter in EINP resulting in high ungulate densities, limited dispersal, and low predation, our data show patterns of habitat use and interactions between bison and elk that were similar to those in other systems. Spatiotemporal partitioning does not appear necessary for coexistence of bison and elk in the aspen parkland.