Spatial patterns of animals are driven primarily by the distribution of resources and the presence of predators, however these factors can differ in their magnitude of effect and predictability. Although prey may have a reliable knowledge of the current distribution of resources, their perception of the spatial and temporal distribution of predation risk must be updated more frequently. Prey rarely live in single predator environments and are expected to evaluate relative predation risk from multiple predators simultaneously. Studies that focus on single predator- prey interactions can therefore be inadequate for understanding antipredator responses in multi-predator systems. Here we examined the distribution of seven African ungulate species in the fenced Karongwe Game Reserve, South Africa, as a function of predation risk from all large carnivore species (lion, leopard, cheetah, African wild dog, and spotted hyena). Using observed kill data, we generated ungulate-specific predictions of relative predation risk and of riskiness of habitats. To determine how ungulates minimize predation risk at the landscape level, we explicitly tested five hypotheses consisting of strategies that reduce the probability of encountering predators, and the probability of being killed.
Results/Conclusions
We found that ungulates distinguished among potential predators, and showed a distribution pattern that simultaneously minimized predation risk from at least some of the predators. All ungulate species avoided risky habitats and most selected safer habitats, thereby reducing their probability of being killed. To reduce the probability of encountering predators, most of the smaller species (impala, warthog, waterbuck, kudu) avoided the space-use of all predators, while the larger species (wildebeest, zebra, giraffe) only avoided areas where lion and leopard space-use were high. The strength of avoidance for the space-use of predators did not correspond to the relative predation threat from those predators. Instead, ungulates used a simpler behavioral rule of avoiding the activity areas of sit-and-pursue predators (lion and leopard), but not those of cursorial predators (cheetah and African wild dog). In general, selection and avoidance of habitats was stronger than avoidance of the predator activity areas. We expect similar decision rules to drive the distribution pattern of prey in other multi-predator systems, especially when predators differ in their hunting modes. However, such antipredator decision rules are likely to change at different spatial scales, and we see this as an important avenue for future research, especially when predation risk from multiple predators becomes more variable and when resource requirements become more constrained.