Structure of forests used by small mammals in the industrially damaged landscape of Sudbury, Ontario, Canada

We studied forest cover and use by small mammals in the fall of 2001 and 2002 in patches of the industrially damaged landscape of Sudbury, Ontario, Canada (46°30′N, 81°00′W). Three common forest patch types in the region were surveyed: semi-barren terrain, a protected arboretum, and a 30-year-old conifer plantation. Each year, four sites per treatment were selected in a 10 km2 area in the vicinity of the Laurentian University campus. At each site, plant cover (percent low-shrub-herbaceous layer (0–1 m); percent shrub layer (1–3 m); and stand height was surveyed in four 5-m radius plots while number of trees were tallied using a Basal Area Factor 2 prism. Small mammal plots (10 m × 10 m) were set on site with mouse- and squirrel-size live traps for 3 days during 4 weeks in early fall. A total of 398 mammals of 9 species (deer mouse (Peromyscus maniculatus); short-tailed shrew (Blarina brevicauda); masked shrew (Sorex cinereus), eastern chipmunk (Tamias striatus), red squirrel (Tamiasciurus hudsonicus), northern flying squirrel (Glaucomys sabrinus); Gapper's red-backed vole (Clethrionomys gapperi); meadow vole (Microtus pennsylvanicus), house mouse (Mus musculus)) were captured in 1265 trap-nights and 938 trap-days (for diurnal species). There were significant differences in the structural habitat variables among treatments, though corresponding trends in abundance, activity and diversity of mammals were not significant. At a smaller scale we found: (a) a significant positive relationship between site-specific stand height, tree basal area, mammal richness, diversity, short-tailed shrew abundance and flying squirrel activity; (b) a positive effect of percent shrub cover on red squirrel activity; and (c) a negative effect of shrub cover on deer mouse abundance. Our data provides evidence that habitat selection processes occurred at a small scale within the landscape; mammal diversity increased with forest cover recovery, though both positive and negative specific responses to patch structure were detected. We suggest that these small-scale effects may be meaningful for the recovery of fragmented landscapes where suitable (i.e. structured) forest patches are small, within a low quality habitat matrix.