Species and Stand Dynamics in the Mixed Woods of Quebec's Southern Boreal Forest

In order to test whether changes in forest composition observed while sampling portions of the landscape originating from different fires may be explained by stand-level processes, I reconstructed species and stand dynamics for mesic sites in the mixed-boreal forests of Quebec. The stands studied are located in the forest surrounding Lake Duparquet in northwestern Quebec and originated from eight successive fires representing a chronosequence of 230 yr. Tree composition and ecological characteristics were assessed in 313 quadrats distributed systematically within the burned areas; one representative stand per fire, presenting the average species composition for clay soils with a moderate moisture regime, was selected for detailed dendroecological analysis. All living and dead stems (> 1 cm dbh) were mapped and cut down in a 20 × 20 m quadrat. Cross sections collected at the root collar and at every meter were analyzed using standard dendrochronological techniques. The age of the stand before the last fire was estimated using snags and logs located in, or in the vicinity of, each quadrat. A PCA (principal-components analysis) ordination of all quadrats shows a gradual change with time since fire from stands dominated by hardwoods (aspen and paper birch), to mixed stands with an important white spruce component, to coniferous stands dominated by balsam fir and white cedar. Despite variations in the age and composition of the stands before the last fire, they generally follow a common pattern characterized by post-fire hardwood dominance. Age structures show successive waves of aspen, birch, and fir recruitment, corresponding respectively to the post-fire cohort, the gradual dismissal of the first aspen cohort, and spruce budworm outbreaks. Each of these waves corresponds to a decrease in hardwoods and an increase in the conifer component of the stand. Suppressed white spruce individuals, recruited primarily as part of the post-fire cohort, experienced growth releases following the dismissal of the first aspen cohorts. White cedar increased in abundance late in succession and tended to outcompete balsam fir in old coniferous stands severely affected by spruce budworm outbreaks. The observed multi-cohort process contrasts with the simple replacement of hardwoods by conifers as suggested by observations of differential growth rates. The patchy distribution of conifer regeneration together with the poorer ability of conifers to rapidly fill gaps may explain why hardwoods are successfully recruited. This gradual rate of change, occurring over several generations, results in the maintenance of an important hardwood component in most stands even in the absence of fire. After 230 yr, stands are still mainly composed of trees originating from punctuated events such as fire, the gradual dismissal of the first aspen cohort, or spruce budworm outbreaks. This contrasts with old-growth boreal forests dominated by spruces in which a quasi-equilibrium is maintained by small canopy disturbances. The study confirms that changes in forest composition observed while sampling portions of the landscape originating from different fires may be explained by simple processes occurring at the stand level.