Neighboring edges: Interacting edge effects from linear disturbances in treed fens.

Questions: Edge influence on forest biodiversity is an important environmental effect associated with habitat fragmentation, but extrapolating the influence of edges across the broader landscape has been difficult, especially for situations where multiple edges exist in close proximity. We asked whether there were differences in edge effects between two types (3 m vs 8 m width) of low‐severity linear disturbance (seismic lines) and whether there were interactions of edge effects when seismic lines occur in dense networks; that is, do multiple narrow seismic lines have a stronger or weaker edge influence than a single narrow seismic line. Location: Treed peatlands in northeastern Alberta, Canada. Methods: Seismic lines are created during oil and gas exploration and are responsible for dissection of boreal forests in western Canada. We sampled vascular plants along transects perpendicular to seismic lines in moderate‐rich and poor treed fens. We used the "Randomization Test of Edge Influence" (RTEI) to calculate the magnitude and distance of edge effects and then compared these between narrower (3 m) versus wider (8 m) lines and between single narrow lines versus multiple narrow lines (parallel and ~50 m apart). Results: In moderate‐rich fens, we found a positive edge influence on understorey diversity from both wide and narrow seismic lines. We also found a weakening edge interaction on diversity, that is, single narrow seismic lines had a stronger edge influence on diversity than did multiple narrow seismic lines. In treed poor fens, multiple narrow seismic lines had a negative edge effect on tree density, understorey abundance, richness, and composition. In addition, we found strengthening edge interactions in treed poor fens on tree density, graminoid cover, and understorey composition. Conclusions: Even narrow linear disturbances, such as seismic lines, can have significant edge effects and these are exacerbated when lines occur in dense networks. [ABSTRACT FROM AUTHOR]