Assessment of how environmental variables and vegetation structure influence peatland carbon fluxes

Sammanfattning: Boreal peatlands are important ecosystems from a carbon balance perspective because they are natural carbon sinks as a result of millennia of peat formation. However, due to climate change, these ecosystems are susceptible to changes in hydrology and temperature. To predict future changes in boreal peatland carbon balance, I need to understand their future role as carbon sinks and their potential to mitigate climate change effects. One way to model and predict these future changes is using plant community traits and environmental variables. In this study, I assessed which environmental variables and plant community characteristics are the best predictors of CO2 and CH4 fluxes in a boreal bog. The importance of variables was studied within a peatland and across time (mid-August to the beginning of November) to capture variation in plant communities and microclimate, and abiotic conditions that change over the seasons. In addition, I investigated if CH4 flux measured under dark conditions is underestimated compared to light conditions due to possible transportation through open stomata. To carry out these research questions, I used 16 permanent collars that were distributed over different microsites in Harbo peatland (ca. 40 km northwest of Uppsala). CO2 and CH4 fluxes were measured using the closed chamber technique. Plant community characteristics included shrub cover, sedge shoot density, moss capitula density (i.e. shoots per unit area), moss biomass. Environmental variables measured were soil and air temperature, Photosynthetic Active Radiation (PAR), and water table depth. For CO2 fluxes, I found that PAR, peat moss biomass, and peat moss capitula density are important predictor variables, and the bog was a carbon source at most measuring points. For CH4 fluxes, the seasonality (difference in the gas emission from the summer to autumn periods) effect was rather small. I found that all of the studied predictor variables had a significant, but generally small effect on CH4 fluxes, except for sedge shoot density, which was the most important variable affecting CH4 emission. Finally, I concluded that CH4 emissions were not generally underestimated when measured in dark compared to light conditions.  ­