Research on green roofs has shown their potential to moderate stormwater runoff and excessive heat in large cities. This has supported the implementation of a bylaw in Toronto, Canada which mandates large scale green roof construction. However, designers lack knowledge on appropriate plant selection for local green roofs. European research has demonstrated that healthy plant cover can significantly improve stormwater capture and cooling on green roofs. This thesis employed two methodologies to select native plants suitable for green roofs. Plants species were first identified from Ontario ecosystems with conditions similar to extensive green roofs including alvars, rock barrens, talus and cliffs. Secondly, plant surveys of existing extensive green roofs in Toronto and other southern Ontario locations to discover plant species already in use. Formal testing of the plant species generated from both methodologies can eventually provide designers with knowledge of dependable plant assemblages for extensive green roofs in Toronto.

Ontario’s rock barrens landscape consists of exposed bedrock ridges which host a mosaic of thin lichen- and moss- covered soil patches, forested valleys, beaver ponds, and depressional wetlands. Peat-filled ephemeral wetlands within bedrock depressions act as gatekeepers to hydrological connectivity between their small headwater catchments and the rest of the landscape downstream through strong fill-and-spill dynamics. We developed a water balance model, RHO, with inputs of precipitation and potential evapotranspiration (PET) to better understand the factors impacting water table (WT) and storage dynamics and in turn the hydrological connectivity of ephemeral wetlands. Field surveys were conducted at six wetlands to obtain and determine the variability in measurable site characteristics, in particular the wetland depression morphometry, to parameterize RHO. Three sites were used in a calibration and validation procedure where modelled WTs were compared to measured WT data from the snow-free seasons for each site to determine the best parameter values. We show that RHO is capable of predicting WT dynamics with inputs of precipitation and PET, when parameterized for specific sites. Wildfire disturbance is known to increase the run-off from hillslopes and remove surface organic soils through combustion. To predict the impacts of wildfire disturbance on ephemeral wetland hydrological connectivity, a generic model wetland depression was parameterized in RHO and used to predict the changes in hydrological connectivity under various wildfire scenarios and test the sensitivity of modelled connectedness to impacted parameters. Modelled results show that connectivity increases under all scenarios tested, and that changes to connectivity are primarily due to increases in run-in. Water balance models, like RHO, can be used to better understand the hydrological connectivity of wetlands in a rock barrens landscape. These models are useful in predicting impacts on the hydrological connectivity, and hydrological ecosystem services, from disturbances such as wildfire and can inform future field research experimental designs. / Thesis / Master of Science (MSc) / Rock barrens landscapes provide several important ecosystem services, which are influenced by hydrological flow paths and water storage on the landscape. Central to these hydrological dynamics is the storage and discharge of water in small wetlands which form in bedrock depressions. Here we develop a simple hydrological model to simulate the water storage and discharge of rock barrens wetlands. We then use this model to explore how wildfire disturbance is likely to change the supply of water to the rest of the landscape by simulating several different scenarios and testing which changes in the model have the largest impact on the water supply. We show that wetlands discharge more water after wildfire disturbance, mainly because of increases in run-off from areas upstream of the impacted wetlands. This modelling approach helps us better understand how wildfire is likely to impact the ecosystem services of a rock barrens landscapes.

Lichen and moss are the dominant ground cover on the Canadian Shield rock barrens of eastern Georgian Bay, and they provide many ecosystem services. Lichen and moss mats are essential for developing and accumulating soil on the bedrock landscape, and as the mats establish they moderate soil temperature and reduce soil water losses, thereby improving the microclimate for more complex vegetation. In addition to pioneering ecosystem succession, the lichen and moss mats provide essential nesting habitat for turtle species-at-risk. These lichen and moss mats are not well understood on rock barrens landscapes, and as such this thesis aims to increase knowledge of the growth, persistence and restoration approaches for these valuable ecosystem resources. We quantified the ecohydrological controls on the growth of lichen and moss mats by measuring the CO2 exchange of lichen and moss under varying environmental conditions. From these results we determined that key growth periods for lichen and moss were during the wet portions of the growing season (spring and fall), and that growth was limited or non-existent during the dry period (summer). Further, we determined that soil moisture was the most important control on lichen and moss CO2 exchange, and that this relationship differed among cover type (lichen, moss, mix of lichen and moss). Moss was able to continue CO2 uptake at a lower water content than lichen, suggesting that lichen would have a greater decline in productivity under drier conditions. A decline in lichen and moss productivity would also likely lead to a decline in soil development through chemical weathering which, in turn, could affect the availability of turtle nesting habitat. We also used CO2 exchange measurements to compare lichen and moss productivity between natural and transplanted mats. Transplanting in-tact patches of lichen and moss has not been widely studied, and as such we tested this approach on a rock barrens landscape. We determined that natural and transplant productivity did not differ for lichen, and that there were some differences between treatments for mixed and moss plots. We also used the tea bag index method to compare relative decomposition rates between treatments (natural, transplant), where we found that decomposition rates did not differ. Our results indicate that it is feasible to remove lichen mats from the footprint of a planned disturbance such as construction and transplant them successfully to nearby undisturbed areas. This approach would restore the lichen cover and the ecosystem services that lichens provide immediately rather than waiting decades for natural regrowth or fragment establishment. As a whole, this thesis will increase knowledge of both the growth and persistence, as well as the restoration of lichen and moss on rock barrens landscapes. Given that lichens and mosses of these genera grow globally, our findings can be applied widely to enhance and protect lichen and moss mats, and the ecosystem services they provide / Thesis / Master of Science (MSc)