Over the past several decades, a significant body of theory has been developed to help managers and researchers understand how different landscape structures impact population processes. Yet most of this work has focused on exploring the landscape-scale effects of patch size and isolation, but much less has been done using edge effects. This despite decades of research showing edges impact habitat quality for numerous species at a local scale. Indeed, a recent review showed that most area effects are likely scaled-up edge effects. Despite their obvious utility in understanding large-scale distribution patterns, the extrapolation of edge responses to entire landscapes has been rare and is complicated by a theoretical and empirical body of work that is focused on highly simplified edges. The vast majority of edge studies are designed to understand responses along “ideal” edges, which are straight and only have two distinct habitat types on either side of the edge. In reality, edges have complex geometries and individual patches are often surrounded by multiple habitat types. Adding to this complexity, recent research has shown that species should show different edge responses at different edge types and a large review of the empirical literature confirmed that species' edge responses do vary depending on the type of edge studied. We used a modeling approach that includes parameters describing how organisms respond to each unique edge type, and we extrapolated those responses over complex landscape mosaics. Specifically, we modeled the impact of roads on bird distributions in a southeastern managed forest ecosystem. Roads are a valuable modeling tool because they add edge while having a negligible effect on the overall amount of habitat within the study landscape. 

Results/Conclusions

We show that when edge effects are considered in real landscape mosaics, properties long associated with fragmentation emerge. These include threshold effects and highly variable responses that are caused by landscape complexity as opposed to stochasticity. Our results suggest that edge responses could be driving many landscape-level patterns described throughout the fragmentation literature.