Snag dynamics in partially harvested and unmanaged northern hardwood forests

Abstract

We used data from hardwood-dominated permanent sample plots in Ontario to estimate the probability of a tree falling during the 5 year period in which it dies ("tree fall"), and likewise the 5 year probability of snag fall. Tree fall probabilities ranged from 5% to 31% across species, with smaller dead trees more likely to be downed than larger ones. Expected half-lives (median time from death to fall) for 25 cm diameter snags varied from 5 to 13 years among species. Fall rates were higher for 10 cm diameter snags but relatively constant for 20–60 cm diameter snags. Recent harvesting substantially increased the probabilities of both tree fall and snag fall, with the former effect most pronounced for small individuals. We used these estimated fall rates to simulate snag dynamics in uneven-aged sugar maple (Acer saccharum Marsh.) stands. Mean snag densities were 32 and 50 snags/ha in selection-management and old-growth scenarios, respectively. Fifty-four percent of this difference was attributable to the lower density of live trees in the selection-management scenario, while 31% was attributable to losses of snags during harvesting. Silvicultural practices that strategically increase tree mortality rates, together with snag retention, as far as safety permits, during harvesting, could represent an effective approach to snag management under the selection system.