Risky business: dealing with risk in a predator - prey community

Abstract

The Liuwa ecosystem has several ecological properties that affect interactions among large predators, with lions and hyaenas as dominant species and African wild dog and cheetah as subordinate species, and between predator and prey. First, the vegetation structure is highly uniform and typified by open grasslands with good visibility over long distances. Secondly the prey community is heavily dominated by wildebeest, with much lower numbers of zebra, oribi and other species. These characteristics combined with GPS data on a fine spatial scale, and a large observational dataset on both predators and prey enabled us to focus on several little-studied questions about the effects of predation risk in the wild. Interspecific competition between predators can be a strongly limiting force for subordinate predators like cheetahs and African wild dogs. Both species use niche partitioning to reduce the risk of dangerous interactions in different ways that appear to have ramifications for coexistence. Wild dogs showed more dietary and temporal overlap with dominant competitors while cheetahs combine divergence in diet, temporal avoidance and reactive local spatial avoidance to coexist with lions and hyenas in areas of high prey density, even in open habitats. These results provide new insight into the conditions under which partitioning may not allow for coexistence of African wild dog, while it does for cheetah, with dominant predators making wild dogs more prone to competitive exclusion (local extirpation), particularly in open, uniform ecosystems with simple prey communities. Focusing on predator-prey relationships the overall the conclusion is that the assessment of risk by animals is a very fine-tuned process. Our results confirm that both the risky places hypothesis (LT risk) and the risky times hypothesis (ST risk) are important, leading to both reactive and proactive responses. Critically, these effects do not act independently in their effects on the strength of antipredator responses. This interaction presents challenges for the design of research on risk effects. An effect of ST risk could be masked by unmeasured variation in LT risk (or vice versa), and an effect of ST risk might be caused by unmeasured variation in LT risk (or vice versa).