Multiple edge effects and their implications in fragmented landscapes

1. Habitat edges are thought to explain much of the negative effects arising from habitat fragmentation; however, progress has been limited in extrapolating edge effects to different situations because ecologists still do not understand if and how multiple edges interact within fragments. It also remains controversial whether edge effects govern patch-size effects, such as area sensitivity, observed in many migratory songbirds. 2. I examined how multiple edges within fragments may intensify edge responses by investigating spatial distributions of an area-sensitive songbird that breeds in temperate grasslands of North America, the bobolink (Dolichonyx oryzivorus Linnaeus). I tested whether bobolinks avoid edges and whether avoidance is stronger near two edges (double-edge plots) than near only one edge (single-edge plots). I subsequently linked bobolink distributions to landscape maps that vary in the amount of habitat and degree of fragmentation to explore some potential implications of multiple edges on patch- and landscape-level distributions. 3. Multiple edges appeared to influence the magnitude of observed edge effects, in which the probability of bobolink occurrence was four times lower in double-edge plots and two times lower in single-edge plots than in the interior of grasslands. Within single-edge plots, the probability of occurrence increased with increasing distance from edge. Within double-edge plots, the probability of occurrence increased as a function of the nearest and next-nearest distances from edges. Multiple edges also appeared to increase the extent of edge effects, or distance of edge influence, which was estimated to be approximately 11-33% greater in double-edge plots than in single-edge plots, depending on the next-nearest distance from edge. 4. Extrapolating local bird distributions to landscape models suggests that edge effects can have strong influences on large-scale distributions and that models incorporating multiple edge effects are different to simple nearest-edge models only in highly fragmented landscapes, regardless of landscape composition. Furthermore, edge effects can lead to patch-size effects similar to empirical patterns of area sensitivity observed in this species. I conclude that edge effects can be intensified when multiple edges collide, a feature that permeates many fragmented landscapes.