Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4496298 | Journal of Theoretical Biology | 2013 | 11 Pages |
highlights•Predation rates can depend upon vegetation biomass.•Vegetation-dependent predation can cause alternative stable states.•Vegetation-dependent predation can cause unstable dynamics.
Predator–prey interactions play out in a physical matrix defined at least to some extent by the quantity and architecture of plants. Ambush predators for instance might more effectively encounter and capture prey in thick vegetation than in thin, open vegetation with good visibility. Yet these vegetation attributes are themselves potentially influenced by the intensity of herbivory, which in turn reflects the magnitude of predation. This intertwining of trophic interactions and basal biomass leads to a largely unexplored feedback in food chain dynamics. In this paper, we examine a number of extensions of standard predator–prey and food chain models that incorporate this kind of “bottom–up” influence on trophic interactions. Even simple models reveal a range of interesting behaviors. For instance, vegetation-dependent attack rates can generate alternative stable states, in one of which the predator effectively limits herbivores to low numbers, freeing plants of substantial regulation by herbivory, and in the other of which the predator is either absent, or present as an ineffective top trophic level. This can occur both in models in which the top predator is a specialist, completely dependent upon the herbivore, and in models in which the top predator is a generalist, inflicting mortality upon the focal prey species, but itself sustained in numbers by alternative prey. Another possibility is that the feedback between basal biomass and attack rates destabilizes the trophic interactions, leading to sustained oscillations in food chains. Comparable effects can arise if handling times depend upon vegetation biomass. These models illustrate the rich complexity of effects that can arise even in relatively simple community models, once one takes into account the range of potential feedback effects that are potentially present.