The ongoing rise in temperature caused by climate change has already increased the vegetation coverage and altered the vegetation composition at higher latitudes. To date, it is unclear how these changes influence the large carbon stocks in permafrost soils. This PhD thesis focuses on whether a vegetation shift among spruce trees, tussock-forming graminoids, birch, and alder shrubs could affect soil carbon cycling in permafrost soils through plant litter and root exudation. This was achieved by analyzing 1) properties of plant litter, bulk soils and soil organic matter (SOM) fractions, 2) root exudate composition and release rates, 3) concentration-dependent effects of exuded organic acids on soil carbon and nutrient cycling, 4) turnover of photosynthates in different soil horizons.

The results suggest that particularly shifts between alder shrubs and graminoids can affect SOM properties and stability by differences in litter composition, and that biomass production and soil physical properties may also be important contributing factors. Under alder shrubs most carbon was stored as easily degradable particulate organic matter and thus SOM under alder shrubs may be particularly vulnerable to microbial decomposition. The analysis of root exudates showed that Arctic shrubs and graminoids have a distinctly different root exudate metabolome, despite having similar exudation rates of primary metabolites. Therefore, identification of secondary metabolites and their impact on SOM decomposition is required for a better understanding of how plant shifts affect soil carbon cycling in permafrost soils through root exudation. In addition, the comparison of measured root exudation with previous laboratory soil incubations, where root exudates were simulated by e.g. glucose additions, uncovered that in most studies simulated root exudation corresponded to root exudation by living plants of several growing seasons. Comparing the effects of organic acids in soils at such high concentrations with lower and realistic additions revealed that the use of too high concentrations overemphasized soil carbon losses and artificially increased microbial nutrient demand. Furthermore, carbon allocation was plant- and depth-specific with alder shrubs allocating less carbon into O horizons than birch shrubs. Considering temporal and spatial variation in root exudation could therefore improve model predictions on plant-mediated carbon losses.

All in all, this thesis demonstrated that a shift in plant types has the potential to alter soil carbon cycling though plant litter and root exudates but that not all effects will result in soil carbon losses.