There are three major explanatory factors for increased abundance at an edge: spillover/mass effects, the edge as an enhanced habitat and complementary resource distribution. Spillover is proximity based, where individuals will disperse into a non-habitat area across an edge. Abundance is increased at this edge because the edge habitat itself is more similar to the patch interiors. Edges can be enhanced habitats and display increased abundance when resources are unique to an edge or absent/rare in the adjacent patches. This particular factor can also start a chain of increased abundance, drawing predators to prey (i.e. particular seed <-- seed predator <-- bird of prey, though “resources” are not limited to food). A complementary resource distribution describes a situation where two patches have different, desirable resources, which makes living on the edge of the two patches advantageous: equal access to both.
Complementary resource distribution is well illustrated by the brown-headed cowbird. The cowbird is measured at greater abundances at particular edges for ease of access to two vital resources. It forages in pastures and meadows and parasitizes songbirds in forests, famously laying its eggs in other birds’ nests and sloughing the work of incubating and raising the cowbird’s chick on the unsuspecting songbird mother, whose chicks are often pushed out of the nest completely by the invader. It’s important to note, however, that some studies of these birds have not found strong evidence for the preference of pasture-forest edges, which complicates the evaluation of edge responses. (I’ll get back to this idea later in the post.)
Edge avoidance is a bit more straightforward. When a decreased abundance is observed at an edge, it’s typically because the edge has become hostile to the organism because of a reduction of a particular resource. But, as the authors explain, it’s important to realize that there will probably still be a dispersion of individuals across the edge. They don’t just stop at the theoretical boundary. In this case, there’s a gradient of organism-specific habitat quality that can be graphed across the edge, from high quality and high abundance to low quality and relatively low abundance.
Building on this framework, Ries and Sisk have published a paper in Oikos recently suggesting a more dynamic interpretation of how species respond to edges. Instead of using the terminology “edge-loving” or “edge-avoiding” to describe an organism’s response to an edge, they suggest using a kind of sensitivity spectrum for evaluation, trying to place response on a scale of insensitive to sensitive. Generally, sensitivity is determined by the consistency of response at an edge expected to generate a response. An edge sensitive species is one that displays a response at one or more edges. An edge insensitive species rarely or never responds.
A species will not always respond in the same way to all habitat edges, and can sometimes respond differently to the same kind of edge. Edge effects are “idiosyncratic” in that sense, but according to the authors, not to a point where they are unpredictable. Returning to what I said earlier about cowbirds, the edge preference is not always demonstrable because edges are not always the same, they differ in composition and orientation. Over 60 years of literature, species are described as being unresponsive to edges approximately 70 percent of the time. It should be of interest to ecologists exactly why species do not respond at certain edges, why they are insensitive. (The authors also suspect that the phenomenon might also be explained by low statistical power.)
While many species have been identified as regularly responding to edges, no species has yet been identified as being edge insensitive. This is due to a general bias against negative results, and because most edge studies do not mea sure responses at enough edge types to support a determina tion of insensitivity. We suggest that one future focus for edge research and literature reviews should be comparing observed responses with predicted responses, thereby allow ing the determination of the edge sensitivity for a suite of species.
The authors cite a study done on butterflies to test the extent of variability considering what they call "extrinsic" factors like edge composition and contrast with adjacent patches. Twelve edge types were studied around the San Pedro River, 15 butterfly species at each edge type. Twelve out of the fifteen species showed positive, negative and neutral responses depending on the edge type, reaffirming the idea that species should not be expected to respond the same way to every type.
Ries, L., & Sisk, T. (2010). What is an edge species? The implications of sensitivity to habitat edges Oikos DOI: 10.1111/j.1600-0706.2010.18414.x
Ries, L., & Sisk, T. (2004). A PREDICTIVE MODEL OF EDGE EFFECTS Ecology, 85 (11), 2917-2926 DOI: 10.1890/03-8021