Thawing permafrost exerts a major forcing on carbon cycling in peatlands along the margins of the Arctic. Our aim in this work is to quantify the differences in net ecosystem exchange (NEE) of CO₂ and its component processes &mdash; Gross Primary Production (GPP) and ecosystem respiration (Reco) &mdash; across surfaces with different permafrost status in a sub-arctic mire complex. The study site, the Abisko-Stordalen mire in Arctic Sweden, is located at 68&deg;20' N, 19&deg;30' E. We used data from two eddy covariance towers, with footprints covering areas with different permafrost status, i.e. mostly permafrost covered palsa plateau (2014-2021), partly thawed bog-palsa mosaic (2014-2021), and non-permafrost rich fen (2014). The bi-modal wind direction pattern at the measurement site helped to derive surface type specific data data set from each tower.</p> <p>The diel cycle of carbon dioxide (CO₂) fluxes during the summer months (June, July, and August) were similar in their pattern, but the magnitudes of CO₂ fluxes at the fen exhibited higher fluxes than the palsa and thawing sectors. The CO₂ fluxes between the palsa plateau and partly thawed area were like each other. CO₂ responses to air temperature and incoming solar radiation indicated highest Reco and GPP and any given temperature and light level at the non-permafrost fen.</p> <p>Analysis of the gapfilled NEE at the palsa plateau and partially thawed sector, for which multi-year time series is available, showed that 2016 had significantly higher fluxes than the multi-year average. The likely reason for this is the longer growing season in 2016. This is also supported by the locally measured NDVI values, suggesting greater vegetation development during that year.</p> <p>In summary, the tall sedge fen had the highest daytime CO₂ uptake in summer and the highest net respiration at night during the growing season. The palsa plateau and partially thawed sector exhibited significantly lower CO₂ fluxes. There were no significant differences in NEE, GPP, or Reco between these two sectors, despite their differing permafrost conditions and vegetation characteristics. This is in contrast with significant differences in methane emissions between these three systems reported in previous studies.