| Summary: | Snow depth is an essential climate variable critical to global energy balance, basin scale hydrology, vegetation change and human livelihoods. It holds special significance in Indigenous northern communities such as those found in Nunatsiavut, NunatuKavut and Nitassinan (Labrador) where changes to snow depth can be hazardous to infrastructure, travel and cultural activities. The energy exchange between snow and ground thermal regime is of particular concern as permafrost distribution has been found to be highly influenced by snow onset/melt and redistribution in Labrador. Our understanding of snow and the impact it has on vegetation and permafrost is hindered by large spatial and temporal gaps in snow measurement and biases toward urban centres that may not be representative of environmental conditions. This thesis expands our understanding of snow characteristics through development of a new, low-cost snow observation technique and through the application of a numerical model that links snow variability to ecosystem and ground thermal processes. This thesis introduces the snow characterization with light and temperature method (SCLT) for determining snow depth using vertically arranged light and temperature loggers. SCLT data was collected for one year at six remote field sites located in forest and shrub-tundra environments in eastern Labrador. Three different approaches to analyze SCLT data are presented and results are compared to a temperature-only approach applied by prior studies. The results show that SCLT can definitively be used to estimate snow depth accurately. A sensitivity analysis is then performed using the Northern Ecosystem Soil Temperature (NEST) model to consider snow-vegetation-permafrost interactions in Labrador. Preliminary simulations at two sites in coastal Labrador show no significant ground temperature warming over 1979-2018 at the top of permafrost/ base of the freeze-thaw layer for most snow/ vegetation conditions. Findings support previous research that wind scouring controls ...
|
|---|