Matrix quality and disturbance frequency drive evolution of species behavior at habitat boundaries

Previous theoretical studies suggest that a species' landscape should influence the evolution of its dispersal characteristics, because landscape structure affects the costs and benefits of dispersal. However, these studies have not considered the evolution of boundary crossing, that is, the tendency of animals to cross from habitat to nonhabitat ("matrix"). It is important to understand this dispersal behavior, because of its effects on the probability of population persistence. Boundary-crossing behavior drives the rate of interaction with matrix, and thus, it influences the rate of movement among populations and the risk of dispersal mortality. We used an individual-based, spatially explicit model to simulate the evolution of boundary crossing in response to landscape structure. Our simulations predict higher evolved probabilities of boundary crossing in landscapes with more habitat, less fragmented habitat, higher-quality matrix, and more frequent disturbances (i.e., fewer generations between local population extinction events). Unexpectedly, our simulations also suggest that matrix quality and disturbance frequency have much stronger effects on the evolution of boundary crossing than either habitat amount or habitat fragmentation. Our results suggest that boundary-crossing responses are most affected by the costs of dispersal through matrix and the benefits of escaping local extinction events. Evolution of optimal behavior at habitat boundaries in response to the landscape may have implications for species in human-altered landscapes, because this behavior may become suboptimal if the landscape changes faster than the species' evolutionary response to that change. Understanding how matrix quality and habitat disturbance drive evolution of behavior at boundaries, and how this in turn influences the extinction risk of species in human-altered landscapes should help us identify species of conservation concern and target them for management.