Modeling ice storm damage to a mature, mixed-species hardwood forest in eastern Ontario

In January 1998. the worst ice storm of the last century hit regions of southeastern Ontario, Quebec, New Brunswick, and the northeastern United States. Using standard multiple regression and classification tree models, we examined the ice damage suffered by trees in a mature, deciduous forest in eastern Ontario at two scales: plot (5 m radius) and individual tree. Canopy trees were damaged significantly more than mid-story trees and there were significant differences among tree species in damage susceptibility. At the plot scale, the best predictors of damage were average tree size and plot species evenness. Plots dominated by large trees were damaged more than those dominated by small trees and plots with higher species evenness suffered higher levels of damage than,did less even plots, Models incorporating damage to neighbouring plots explained more variance than did models without the neighbour information. At the individual tree scale, damage suffered by the dominant canopy tree species, sugar maple. was best predicted by tree size. Damage suffered by the dominant mid-story tree species, ironwood, was best predicted by neighbour information and, tree size. Disturbances that differentially affect canopy and mid-story layers have been shown to accelerate forest succession by creating light gaps. However, given the species composition and structure of our study forest, we feel that this disturbance will not overly influence forest succession in mature, deciduous forests in eastern Ontario.