The objective of this thesis was to characterize the microbiology of subsurface fracture water at a location within the Canadian Shield considered as a prospective host site for a geologic repository. Repository performance and long-term safety predictions for geologic confinement of radioactive waste rely on understanding the natural microbial processes that occur within a host formation; however, the inaccessibility of the crystalline terrestrial subsurface means that this habitat is difficult to explore and as such is largely unknown. The area characterized is located within the boundary of the Chalk River Laboratories site, situated within the Central Gneissic Belt of the Grenville province (formed 1.5 and 1.0 ba years before present); the site is also situated within the Ottawa-Bonnechere graben (formed 0.5 ba before present). Fracture water was accessed from one 34 m deep open drill hole and six cased and sealed drill holes; the Westbay Multilevel Groundwater Monitoring system preserved the natural fracture flow enabling discrete fracture water sampling at multiple depths via each of the sealed drill holes. This thesis combined multiple datasets in an exploratory analysis for drivers of subsurface microbial assembly within an ecological framework of selection, dispersal, drift and diversification across two spatial extents: 25 km3 and 1 km3. The outcomes of the multivariate analyses, null models and generalized linear models identified prospective source waters and distributional relationships of total and viable cell counts with fracture water sulfate and manganese. Random processes (dispersal, drift and diversification) explain close to 50% of the variance of the phylogenetic beta diversity among the component taxa. Selection associated with differential abundance of the 16S rRNA gene V4 region and spatial and environmental factors identified sulfate and organic carbon plus manganese and a spatial coefficient. Selection by differential abundance was not a major driver of community assembly; accounting for ~4% each of the total abundances linked to sulfate and manganese. At a spatial scale less than 1 km3, it may be possible to identify either greater significance for sulfate and manganese or to identify additional environmental factors linked to selection. Demonstrated metabolism included nitrate reduction (common across all sampling locations and at each sampling campaign) and sulfate reduction (observed at one sampling campaign). The distributions of total and viable cell counts correspond with the distributions of sulfate and manganese, respectively. The fracture water was a source of sulfate and manganese, but not a source of nitrate. A limited analysis of rock porewater identified sulfate and nitrogen compounds (ammonia, nitrite and nitrate) in parts per million concentrations, suggesting that the rock is a source of these compounds, and this finding warrants an assessment of rock weathering as a driver of selection at these sampling locations. Phylogenetic relationships across sampling locations showed that the component taxa were more closely related than expected by chance; this pattern suggests that, at the spatial scale of the analysis, competitive exclusion was not a driver of subsurface community assembly. Co-existence of close relatives may be biologically relevant—and thus be a sign of diversification—or this pattern may reflect the 16S rRNA gene copy number combined with intra-genomic heterogeneity greater than the 97% sequence similarity threshold for binning sequencing reads into organizational taxonomical units (OTUs). The distribution of genes for energy metabolism was uniform across all sampling locations. A metatranscriptome assessment would help differentiate between the genes present from the genes expressed. Testing for a wider range of demonstrated metabolic capabilities would support RNA-level gene expression analyses. Overall, applying the ecological framework of four main drivers of community assembly show that, at the spatial scale of the sampling, up to 50% of the variance among community dynamics reflect randomness. Approximately ~1% of the total abundance was linked to measured metabolism; at smaller spatial scales, ~8% of the total 16S rRNA gene abundance was linked to differential abundances across—potentially connected--sampling locations. By sampling at smaller spatial scales, therefore, it may be possible to discern additional metabolic and selective processes. These data will inform models for the performance and long-term safety of geological repositories.

Non-native species and climate change pose serious threats to global biodiversity. However, the roles of climate, dispersal, and competition are difficult to disentangle in heterogeneous landscapes. We combine empirical data and theory to examine how these forces influence the spread of non-native species in Lake Baikal. We analyze the potential for Daphnia longispina to establish in Lake Baikal, potentially threatening an endemic, cryophillic copepod Epischurella baikalensis. We collected field samples to establish current community composition and compared them to model predictions informed by flow rates, present-day temperatures, and temperature projections. Our data and model agree that expansion is currently limited by dispersal. However, projected increases in temperature reverse this effect, allowing D. longispina to establish in Lake Baikal’s main basin. A strong negative impact emerges from the interaction between climate change and dispersal, outweighing their independent effects. Climate, dispersal, and competition have complex, interactive effects on expansion with important implications for global biodiversity.