| Summary: | Climate warming across the circumpolar north has driven rapid shifts in vegetation productivity and structure, altering the community composition and function of tundra ecosystems. In my MSc thesis, I examined the biophysical factors mediating the effects of climate on vegetation dynamics, and assessed the impact of data type on models of vegetation change. In my first data chapter, I combined field sampling of soils and vegetation and random forests modelling to identify the determinants of spatial heterogeneity in Enhanced Vegetation Index trends derived from the Landsat archive (1984-2016). This analysis showed that over 70% of the Beaufort Delta region has exhibited significant increases in vegetation productivity (greening) from 1984 to 2016. Greening was more common and rapid in lower elevation areas with existing shrub-dominated land cover on till blanket and glaciofluvial deposits. The influence of surficial geology and topography on productivity trends suggests that soil moisture and nutrient availability are mediating the impact of climate warming in the low Arctic tundra. In my second data chapter, I investigated the response of three tundra shrub species (green alder, dwarf birch, and lingonberry) to climate warming using species distribution modelling. In this study, I also explored how data type affects model performance and output. This analysis shows that the use of pseudo-absence data (a common practice in species distribution modelling) results in differences in projected habitat suitability when compared to models parameterized using true absence data. Projections of habitat suitability under a climate warming scenario suggest that shrubs will respond individualistically, likely in response to physiological and ecological differences among species. Overall, my thesis emphasizes the importance of vegetation change at a landscape scale and how larger climate modelling efforts must account for landscape-scale variation in biophysical variables, individualistic responses at the species-level, and ...
|
|---|