Most prey living in natural systems cope with multiple predators. These predators often hunt in different places or at different times, limiting prey escape options, and leading to questions about how prey manage conflicting predation pressures. Here we tested the hypothesis that spatiotemporal differences in predator hunting behavior, defined as differences in the spatial distribution and timing of predator activity, provide vacant predator niches that serve as refugia for prey in a multi-predator environment. For example, cursorial predators may hunt in open areas during day whereas ambush predators hunt in forested areas at night. Thus, prey could select for risky places during the safe times provided by lulls in predator activity. We tested for this potential predator avoidance mechanism by quantifying adult female elk (n = 27) habitat selection in northern Yellowstone National Park within the first decade following wolf reintroduction (2000-2004). We quantified spatial predation risk from known cougar- (n = 257) and wolf-killed (n = 400) elk locations. We quantified diel activity from wolf (n = 21) and cougar (n = 6) GPS-collared individuals. This allowed us to evaluate whether elk habitat selection was a function of the spatio-temporal predation risk imposed by wolves, cougars, or both.
Wolves and cougars partitioned themselves across spatial (openness, topographic roughness) and temporal (diel activity) axes. Wolves were crepuscular hunters that killed elk in flat, open areas, whereas cougars were nocturnal hunters that killed elk in rough, forested areas. These differences defined two vacant predator niches (open-night and forested-day) that elk regularly utilized. Elk selected for the open-night niche when wolf activity was low, and for the forested-day niche when cougar activity was low. Although prey commonly use space and time to manage predation risk, this is the first study to identify how prey integrate these axes to simultaneously avoid multiple predators. Given that resource partitioning is common in predator communities, this avoidance mechanism may occur in a variety of ecological contexts. Moreover, ecological theory predicts that increasing predator diversity will improve ecological function via prey suppression. However, the ability of behaviorally-sophisticated prey to navigate multi-predator landscapes may minimize predation risk and its cascading ecological effects. Thus, our results suggest that the ability of prey to exploit vacant predator niches may be an important process to consider when predicting the effects of predator diversity on ecosystems.