Réponse des peuplements et des arbres à la coupe partielle en forêt boréale mixte de l’est du Canada – approches expérimentale et de modélisation.

Forest management has shifted from a relatively narrow focus on commercial wood supply to greater consideration of the natural dynamics and multiple ecological services provided by forest ecosystems. This recognition has generated interest in ecosystem management approaches based on diversifying and adapting silvicultural practices such as partial harvesting. The SAFE (sylviculture et aménagement forestier écosystémiques) project is a series of stand-level experiments undertaken in the Lake Duparquet Research and Teaching Forest (LDRTF) in the south-eastern Canadian boreal forest. The project was initiated in 1998 and tests the potential of partial harvesting as a tool for ecosystem-based silviculture in trembling aspen (Populus tremuloides Michx.) dominated stands. Previous studies conducted across the Canadian boreal mixedwood forest have indicated that, over the short term, partial harvesting of aspen or mixed-aspen stands can initiate a second cohort of aspen, increase the growth rate of advanced conifer regeneration and maintain most of the structural attributes of mature stands. Over the longer term (12 years), I expected that, following partial harvesting of 50% or more of basal area or through the use of gap harvesting, residual stems would be more susceptible to mortality due to windthrow compared to lighter harvesting prescriptions. This in turn would favor a progressive opening of the canopy, increase canopy gap size and change the dynamics between commercial and competitive species, thus potentially generating a shrub-dominated community. I also expected that more intense partial harvesting prescriptions would accelerate the growth of residual trees by decreasing competition and increasing resource availability as well as accelerating the development of old growth stand attributes by creating growing space for new tree cohorts. 
\nThis thesis is structured around four individual studies conducted in the SAFE project. The first and second studies were conducted in pure aspen stands (93% aspen basal area) while the third and fourth studies also used data from mixed aspen stands (81% aspen basal area). In the first study, we evaluated the effects of partial harvesting on sapling recruitment and residual tree mortality over a twelve year period. Stem analysis and neighborhood competition indices were used in the second study to assess tree-level growth responses over the same period. For the third study, we first identified and characterised, based on the literature, the structural attributes of old-growth trembling aspen (Populus tremuloides Michx.) - dominated stands for boreal mixedwoods. Using inventories conducted in pure aspen and mixed aspen stands, we then assessed the potential of partial harvesting in even-aged aspen-dominated stands to accelerate stand development towards these old-growth attributes. Finally in the fourth study, SORTIE-ND - a spatially-explicit stand dynamics model - was adapted and validated for the region and stand development under a range of partial harvesting scenarios was simulated over a 100-year period. 
\nThe effects of partial harvesting on stand dynamics over a twelve-year period were compared among four treatments: clearcuts (100% basal area (BA) removal); 1/3 partial cut (1/3 PC, 33% BA removal using low thinning); 2/3 partial cut (2/3 PC, 61% BA removal using high thinning) and controls (0% removal). Aspen sapling recruitment increases continuously following clearcut and partial cut treatments and no significant mortality occurred in the sapling layer over the 12-year period. Aspen sapling recruitment was disproportionally greater in the 2/3 partial cuts (56% of aspen sapling density in clearcuts) compared to the 1/3 partial cuts (5% of clearcut densities). Recruitment of conifer saplings increased with time and was significantly higher in the two partial cut treatments than in the clearcut treatment. Mortality of residual merchantable aspen was strongly associated with small stems (10-19.9 cm DBH), regardless of treatment but was initially (1-3 years after treatment) higher in the 2/3 partial cut. Both partial harvesting treatments had the effect of maintaining mountain maple (Acer spicatum Lamb.), a shade-tolerant, high woody shrub, at densities similar to those in control stands, whereas recruitment of mountain maple saplings was negligible in clearcuts. Annual volume increment (AVI) of individual aspen stems was analyzed as a function of treatment, tree social status, pre-treatment growth, time since treatment application (1–12 years) and neighborhood competition. There was no evidence of initial growth lag after partial harvesting. Only the most severe treatment of partial harvesting (2/3 PC) resulted in an increase in volume increment relative to trees in control stands. Annual increase in volume in the 2/3 partial cut was 25.6% higher than controls over 12 years. Annual volume increment of dominant trees was higher by 16.2 dm3yr-1 than that of co-dominants and was proportional to pre-treatment volume growth. 
\nBased on a literature review, it was determined that compared to mature, even-aged stands, old-growth aspen stands have lower merchantable stem densities and basal area, more large aspen stems and higher stem size variability, more than one cohort of trees, greater percentage area occupied by gaps and more and larger snags and downed wood. Inventories conducted over a 12-year post-treatment period indicate that while the partial harvesting treatments applied in this study successfully retained most of the structural attributes of mature aspen stands (untreated controls), they did not generally “accelerate succession” toward old-growth in the 12-year time interval. Nonetheless, overall results do suggest that by promoting irregularities in both horizontal and vertical structures, high-intensity partial harvesting will accelerate stand development towards what could be characterised as old-growth aspen-dominated mixedwoods. 
\nThe results of simulations with SORTIE-ND indicate that following the mortality of the first cohort of aspen, white spruce (Picea glauca [Moench] Voss) maintained dominance in un-harvested controls of pure aspen stands whereas balsam fir (Abies balsamea (L.) Mill.) dominated in mixed aspen stands. All gap cuts and 80% dispersed cuts favoured recruitment of aspen over conifers. At year 100 of simulation runs, the 1,600 m2 gap cut resulted in highest stand basal areas in both pure aspen and mixed aspen stands with 38.0 and 34.1 m2.ha-1, respectively, of which 18% and 28%, again respectively, was composed of tolerant conifers. 
\nThe overall results of the thesis indicate that partial harvesting is a viable silvicultural option for trembling aspen-dominated boreal mixedwoods of Eastern Canada. This practice can be used to improve the growth of large aspen trees and to promote old-growth attributes. However, residual tree mortality immediately after treatments applied in this study and limited conifer recruitment bring into question the general potential of partial harvesting in these stand types. I argue that adapting partial harvesting treatments (intensity and spatial configuration of tree removal) based on pre-harvest stand conditions (e.g. stand age, stem size distribution, presence of conifer seed trees and advanced regeneration, and presence of woody shrubs) is the key to ensuring success of partial harvesting treatment. There has been some debate surrounding absolute retention levels to be applied in partial harvesting; however, modelling results suggest that both stand structure and timber production rates are strongly influenced not only by retention levels but also by spatial configuration of residual trees. 
\n
\nLa gestion des forêts est passée d’une dynamique productiviste visant principalement l'approvisionnement en bois commercial à une meilleure intégration de la dynamique forestière naturelle et des multiples services écologiques des forêts. Cette évolution s’est traduite par des approches d’aménagement écosystémiques qui préconisent la diversification et l’adaptation des pratiques sylvicoles incluant le recours à des coupes partielles. Le projet SAFE (sylviculture et aménagement forestier écosystémiques) comprend une série d'expériences sylvicoles conduites l’échelle du peuplement en forêt boréale mixte de l’Est canadien. Le projet, initié en 1998 dans la forêt d’enseignement et de recherche du lac Duparquet, vise à valider le potentiel sylvicole des coupes partielles appliquées à peuplements équiennes matures dominés par le Peuplier faux-tremble (Populus tremuloides Michx.). Des études antérieures menées en forêt boréale mixte canadienne indiquent qu’à court terme la coupe partielle appliquée à des peuplements purs ou mélangés dominés par le P. faux-tremble initie une deuxième cohorte de P. faux-tremble, augmente la croissance de la régénération résineuse préétablie et maintient la plupart des attributs structurels de peuplements matures. J’ai émis l’hypothèse que, sur le plus long terme (12 ans), la mortalité les tiges résiduelles due à la chablis serait plus élevée après un prélèvement de 50% ou plus de la surface terrière (ST) ou suite à une coupe par trouées comparativement à un prélèvement moins fort. Cette mortalité favoriserait l’ouverture progressive de la canopée, augmenterait la dimension des trouées et altérerait la dynamique entre les espèces commerciales et concurrentes, favorisant ainsi la strate arbustive. J’ai aussi émis l’hypothèse qu’un prélèvement plus élevé permettrait d'accélérer la croissance des arbres résiduels en diminuant la compétition et en accroissant la disponibilité des ressources et permettrait aussi d'accélérer le développement des attributs structurels caractéristiques des peuplements plus âgés ou anciens en créant l’espace nécessaire à l’établissement de nouvelles cohortes d'arbres. 
\nLa thèse est structurée autour de quatre études individuell

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.