Edge effects at clearcut, fire and riparian buffers in the boreal forest of Northwestern Ontario
Stewart, Katherine Janet
Master of Science
SubjectBuffer zones (Ecosystem management) Ontario, Northwestern
Clearcutting (Environmental aspects Ontario, Northwestern)
Riparian ecology (Ontario, Northwestern)
MetadataShow full item record
In the boreal forest of northwestern Ontario clearcutting and fire are two common edge-creating disturbances. Fundamental knowledge regarding the ability of fire and clearcut edges to minimize edge effects, preserve interior habitat, and provide sources of vegetative growth is lacking. Development of sustainable forest management strategies that emulate natural disturbance relies upon such information. This thesis examines edge effects in the boreal forest of northwestern Ontario at conifer clearcut, deciduous clearcut, and conifer 6re edges. Riparian buffers with an upland clearcut edge and an upland-riparian ecotone were also studied. Edge effects were explored on a number of scales ranging from landscape-level to stand-level to small-scale bryophyte response. Residual patch, core area and edge were assessed in 1000 ha and 250 ha windows in both clearcut and fire disturbance. At the stand-level canopy and understory conditions were sampled along transects (60 m) placed across edges and buffers, in comparison to transects in the interior forest or at undisturbed stream edges. Edge characteristics and the depth of edge influence (DEI) were determined using the critical values approach, multiple response permutation procedure, analysis of covariance and other nonparametric tests. Conifer and deciduous clearcut edges had many similarities. Species response across conifer fire edges was different 6om clearcut edges due to shading provided by standing dead trees and a pre-existing moisture gradient at burnt edges. Most buffers maintained a similar species composition to undisturbed stream edges, but changes in species abundance were detected at the stream edge. The DEI was greatly decreased for most response variables 10 m past the edge; however, a significant DEI was found for some response variables at 40 m or up to the stream edge, which was the greatest distance measured.