COS 17-5
Diel activity pattern of wolves shapes elk response to spatial predation risk in northern Yellowstone

Monday, August 10, 2015: 2:50 PM
344, Baltimore Convention Center
Michel T. Kohl, Department of Wildland Resources, Utah State University, Logan, UT
Douglas W. Smith, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, WY
Daniel R. Stahler, Yellowstone Center for Resources, National Park Service, Yellowstone National Park, WY
Matthew C. Metz, Wolf Project, Yellowstone Center for Resources, Yellowstone National Park, WY
James D. Forester, Fisheries, Wildlife, and Conservation Biology, University of Minnesota, Saint Paul, MN
Matthew J. Kauffman, Wyoming Cooperative Fish and Wildlife Research Unit, University of Wyoming, Laramie, WY
Daniel R. MacNulty, Department of Wildland Resources, Utah State University, Logan, UT

Most animals exhibit a daily cycle of activity. Despite the ubiquity of this biological process, studies of predator-prey spatial interactions typically assume that the daily activity pattern of predators is constant. We used a novel analysis of hourly wolf (Canis lupus) activity and four indices of spatial predation risk (kill site probability, kill density, openness, and wolf density) to evaluate the influence of wolves on elk (Cervus elaphus) habitat selection in northern Yellowstone National Park (YNP). We quantified habitat selection of adult female elk (n = 27) using a matched case-control logistic regression movement analysis, and calculated hourly wolf activity using GPS step-length data; wolf GPS activity was strongly correlated with direct observations of wolf-elk encounters. We used linear piecewise splines and an information theoretic approach to evaluate whether anti-predator response was related to the magnitude of spatial risk (i.e., risk sensitivity). 


Elk avoided areas where kills were most likely and that were most open only when wolves were highly active (06:00-11:00 and 16:00-20:00 hours) and demonstrated weak, positive selection for these apparently risky places when wolves were less active (11:00-16:00 and 20:00-06:00 hours). Notably, this pattern was evident only at high levels of spatial risk as indexed by kill site probability, kill density, and openness. In contrast, elk selected for places with high wolf densities when wolves were active, a pattern we attribute to inter-pack conflict and a subsequent disconnect between wolf density and kill sites. Wolf-induced shifts in elk habitat selection are a common explanation for the recovery of some woody plant species in YNP. However, our results suggest that elk spatial response only occurs during periods of very high wolf activity (i.e., dawn), a period that accounts for less than 21% of the day. This demonstrates the importance of accounting for the diel activity patterns of predators when assessing prey response to spatial risk, particularly in wild vertebrate systems. Studies that ignore or inadequately represent predator diel activity patterns may fail to accurately identify behavioral mechanisms (e.g., altered habitat selection or vigilance behaviors) and their subsequent ecological effects.