Greenhouse gas emissions from thawing permafrost in arctic ecosystems may amplify globalwarming, yet estimates of the rate of carbon release, and the proportion of carbon released as methane (CH 4 ) or carbon dioxide (CO 2 ), have a high degree of uncertainty. There are many areas where no measurements exist, and few year-round or long-term records. Existing year-round eddy covariance measurements of arctic CH 4 fluxes suggest that nongrowing season emissions make up a significant proportion of tundra systems emissions on an annual basis. Here we present continuous CH 4 flux measurements made at Eight Mile Lake, an upland tundra ecosystem undergoing permafrost degradation in Interior Alaska. We found net CH 4 emissions throughout the year (1.2 ∓ 0.011 g C-CH 4 m 2 /yr) that made up 61% of total radiative forcing from annual C emissions (CO 2 and CH 4 ; 32.3 g C m 2 /yr) when taking into account the greenhouse warming potential of CH 4 relative to CO 2 . Nongrowing season emissions accounted for 50% of the annual CH 4 budget, characterized by large pulse emissions. These were related to abrupt increases in air and shallow soil temperatures rather than consistent emissions during the zero curtain-a period of the fall/early winter season when subsurface soil temperatures remain near the 0°C freezing point. Weekly growing season CH 4 emissions in 2016 and 2017 were significantly related with thaw depth, and the magnitude of CH 4 emissions between these seasons was proportional to the rate of active layer thaw throughout the season.Arctic tundra systems account for an estimated 19 Tg/yr of global CH 4 emissions at present, but there are large uncertainties (8-29 Tg/yr) owing to scarce spatial coverage and the paucity of year-round or TAYLOR ET AL. 2695 Key Points: • Arctic permafrost tundra ecosystem is an annual net source of CH 4 • Nongrowing season emissions make up 45% of annual CH 4 loss • Thaw depth is a significant driver of growing season CH 4 emissions Data gaps were filled using the marginal distribution sampling (MDS) algorithm (Reichstein et al., 2005) in REddyProc (Wutzler et al., 2018) that excluded data >2 standard deviations from the overall mean of the

Permafrost thaw is typically measured with active layer thickness, or the maximum seasonal thaw measured from the ground surface. However, previous work has shown that this measurement alone fails to account for ground subsidence and therefore underestimates permafrost thaw. To determine the impact of subsidence on observed permafrost thaw and thawed soil carbon stocks, we quantified subsidence using high‐accuracy GPS and identified its environmental drivers in a permafrost warming experiment near the southern limit of permafrost in Alaska. With permafrost temperatures near 0°C, 10.8 cm of subsidence was observed in control plots over 9 years. Experimental air and soil warming increased subsidence by five times and created inundated microsites. Across treatments, ice and soil loss drove 85–91% and 9–15% of subsidence, respectively. Accounting for subsidence, permafrost thawed between 19% (control) and 49% (warming) deeper than active layer thickness indicated, and the amount of newly thawed carbon within the active layer was between 37% (control) and 113% (warming) greater. As additional carbon thaws as the active layer deepens, carbon fluxes to the atmosphere and lateral transport of carbon in groundwater could increase. The magnitude of this impact is uncertain at the landscape scale, though, due to limited subsidence measurements. Therefore, to determine the full extent of permafrost thaw across the circumpolar region and its feedback on the carbon cycle, it is necessary to quantify subsidence more broadly across the circumpolar region.

Warming of the Arctic can stimulate microbial decomposition and release of permafrost soil carbon (C) as greenhouse gases, and thus has the potential to influence climate change. At the same time, plant growth can be stimulated and offset C release. This study presents a 15‐year time series comprising chamber and eddy covariance measurements of net ecosystem C exchange in a tundra ecosystem in Alaska where permafrost has been degrading due to regional warming. The site was a carbon dioxide source to the atmosphere with a cumulative total loss of 781.6 g C m−2 over the study period. Both gross primary productivity (GPP) and ecosystem respiration (Reco) were already likely higher than historical levels such that increases in Reco losses overwhelmed GPP gains in most years. This shift to a net C source to the atmosphere likely started in the early 1990s when permafrost was observed to warm and thaw at the site. Shifts in the plant community occur more slowly and are likely to constrain future GPP increases as compared to more rapid shifts in the microbial community that contribute to increased Reco. Observed rates suggest that cumulative net soil C loss of 4.18–10.00 kg C m−2—8%–20% of the current active layer soil C pool—could occur from 2020 to the end of the century. This amount of permafrost C loss to the atmosphere represents a significant accelerating feedback to climate change if it were to occur at a similar magnitude across the permafrost region.