Spatial complexity enhances predictability in food webs.

The prediction of an ecosystem's response to an environmental disturbance or the artificial control of ecosystems is a challenging task in ecology. Ecological theory predicts that disturbances frequently result in unexpected responses between interacting species due to the many indirect interactions within a complex community. However, such indeterminacy appears to be unusual in nature. Here using a meta-community food web, I show that spatiality is key to resolving this disparity. A moderate level of spatial coupling strength between habitats due to species migration increases the possibility of expected responses to press perturbation or predictability. Moreover, predictability increases with increasing spatial complexity, as measured by the number of local food webs and their connectivity. A meta-community network can attenuate the propagation of disturbances through indirect pathways due to species emigration to other habitats, thereby preserving the expected effect on the interacting species. These results suggest that the isolation of communities due to habitat destruction decreases the predictability of communities, thereby complicating the control of ecosystems.