Natural and anthropogenic causes have produced changes in tree dominance from coniferous to broadleaf deciduous forests, generating shifts in litter inputs and plant understory composition. The impact of changes in canopy-associated factors on belowground microbial communities remain poorly understood. The objective of this study was to better understand how abiotic and biotic factors in black spruce and trembling aspen forests shape soil microbial community structure.

Boreal forests are increasingly changing in tree dominance from coniferous to broadleaf deciduous forests, due to natural fires, human land use and climate change. In the boreal forest of northwestern Quebec, coniferous forests dominated by black spruce trees (Picea mariana (Mill.) Britton, Sterns & Poggenb.) experience an expansion of broadleaf forests, producing areas dominated by trembling aspen trees (Populus tremuloides Michx.). These two forest types can be considered as two alternative stable states, with similar permanent abiotic conditions originating from the same disturbance trigger, and only differing in tree dominance. Alternative stable states defined by tree dominance shape local conditions for plant and microbial communities in the understory, via their effects on environmental factors, such as litter deposition, light conditions, and soil nutrient availability. Consequently, increasing changes in tree dominance in boreal forests may differentially affect understory vegetation and associated microbial communities. The general objective of this thesis was to understand how tree dominance of coniferous and broadleaf deciduous forests influenced understory vegetation and associated microbial communities in the boreal forest. This was analyzed in three chapters. In Chapter 1, we analyzed the main factors related with tree dominance driving changes in understory vegetation between forest types. We performed a 5-year in situ experiment using alternative stable states as a guiding theoretical framework including two approaches: 1) the ecosystem approach, manipulating environmental conditions of light, litter and nutrients; and 2) the community approach, exchanging understory communities between forest types. In Chapter 2, we evaluated the relative importance of factors related to tree dominance shaping microbial community structure. First, we analyzed differences in bacterial communities between microhabitats (soil microbiome vs. tree phyllosphere) and forest types (black spruce vs. trembling aspen forests). Secondly, we analyzed if bacterial and fungal communities were affected by changes in factors associated with tree dominance, including shifts in litter deposition and in understory transplantations between forest types, and correlated them with abiotic and biotic factors. Finally, in Chapter 3, we used genetic sequencing to identify phyllosphere bacteria associated with the two most abundant feather mosses in both forest types, in order to determine if moss-host species or tree dominance drive bacterial diversity in the moss phyllosphere. Results of Chapter 1 indicate that the plant understory composition in trembling aspen stands was both resistant to shifts in local conditions and resilient in the alternative state dominated by black spruce. In contrast, the abundance of mosses and ericaceous plants that typically compose black spruce stands was negatively affected by a physical effect of broadleaf addition and they were less resilient in trembling aspen stands as they were invaded by local plants over time. Despite the fact that shifts in litter deposition and in understory vegetation exchanged between forest types were important field manipulations that altered understory vegetation after 5 years, in the results of Chapter 2 we did not observe a significant effect of treatments on the soil microbial α-diversity, relative abundance, and community composition. These bacterial and fungal communities in the mineral soil were resilient to the aboveground changes made by the treatments and were only shaped by the underlying effects of forest type. While microbial communities were influenced by forest type, they were also highly specific to their microhabitat, i.e. in the tree phyllosphere or soil microbiome of each forest type. Finally, in Chapter 3, the overall composition of moss phyllosphere bacterial communities was defined by the interaction of both forest type and host species, although the abundance of most bacterial phyla was most strongly influenced by forest type. The moss phyllosphere in trembling aspen forests harboured a higher relative abundance of diazotrophic bacteria, including ecologically important cyanobacteria, compared to black spruce forests. Our project advanced the understanding of two representative forest types in the boreal area, by highlighting the effect of tree dominance of black spruce and trembling aspen forests on understory vegetation (Chapter 1), tree phyllosphere and soil microbial communities (Chapter 2), and moss bacterial phyllosphere (Chapter 3) in the boreal forest of eastern Canada. Forests dominated by trembling aspen trees harbour an abundant and more resilient plant understory and a more diverse soil microbial community than black spruce forests, which is useful in forest management to better cope with increasing changes in the boreal biome.

The composition of ecologically important moss-associated bacterial communities seems to be mainly driven by host species but may also be shaped by environmental conditions related with tree dominance. The moss phyllosphere has been studied in coniferous forests while broadleaf forests remain understudied. To determine if host species or environmental conditions defined by tree dominance drives the bacterial diversity in the moss phyllosphere, we used 16S rRNA gene amplicon sequencing to quantify changes in bacterial communities as a function of host species (Pleurozium schreberi and Ptilium crista-castrensis) and forest type (coniferous black spruce versus deciduous broadleaf trembling aspen) in eastern Canada. The overall composition of moss phyllosphere was defined by the interaction of both factors, though most of the bacterial phyla were determined by a strong effect of forest type. Bacterial α-diversity was highest in spruce forests, while there was greater turnover (β-diversity) and higher γ-diversity in aspen forests. Unexpectedly, Cyanobacteria were much more relatively abundant in aspen than in spruce forests, with the cyanobacteria family Nostocaceae differing the most between forest types. Our results advance the understanding of moss-associated microbial communities among coniferous and broadleaf deciduous forests, which are important with the increasing changes in tree dominance in the boreal system.