CO2 emission from a subarctic mire with different permafrost status

The main goal of this work is to quantify the differences in net ecosystem exchange (NEE) of CO₂, and its component processes, Gross Primary Production (GPP) and ecosystem respiration (Reco) across surfaces with different permafrost status at a sub-arctic mire complex. The study site, Abisko-Stordalen mire in Swedish Lapland, is situated in the mire (68°20' N, 19°30' E). We used data from the Integrated Carbon observation system (ICOS) Ecosystem station (SE-Sto). Palsa plateu and thawing sector data were used from the years 2014-2021, and for fen for 2014.  

We analyzed CO₂ diel cycle for the summer months (June, July, August), to compare it for the three ecosystems. We found out that diel cycle is similar in terms of time but order of magnitude of CO₂ fluxes is different for the fen than for the palsa and the thawing sector. CO₂ responds for the air temperature and incoming radiation were compared among three ecosystems.    

We used three different gapfilling methods: REddyProc (Jena), artificial neural network (ANN), and generalized linear models (GLM) for the annual balance calculation. Gap-filling was made on a half-hour basis, and this allow us to divide fluxes to Reco and GPP.  Comparison of those two NEE components has been done for the palsa plateu and the thawing sector for the whole period.

Of the three surfaces that were examined in this study, the tall sedge fen had the highest CO₂ uptake during the daytime in the summer and the highest net respiration at night as the growing season. However, there were no significant differences in CO₂ fluxes between the palsa plateau and the thawing sector, despite their different permafrost conditions and vegetation characteristics. In contrast, there were significant differences in methane emissions between these systems. Multiple gap filling methods were found to be important for accurate CO₂ emission estimation.