Description: Permafrost temperatures increase globally and lead to extensive permafrost thaw. With increasing air temperatures, changing precipitation regimes, increased intensity and frequency of extreme events, and disturbances such as wildfires, permafrost is increasingly vulnerable to thawing. Permafrost thaw either occurs gradually over decades or is initiated through rapid and abrupt permafrost disturbance processes, which can develop within a few days to years. The impact of such rapid disturbances on Arctic-Boreal ecosystems can be drastic on local to regional-scale and a global impact has been suggested through soil carbon mobilisation. Retrogressive thaw slumps (RTS) are highly dynamic and abrupt permafrost disturbance features that result from slope failure from thawing of ice-rich permafrost. Although RTS are small-scale features, they often occur in clusters, significantly impacting the surrounding landscapes and ecosystems at high temporal scales. The occurrence and thawing activity of RTS signifies an increased permafrost vulnerability. So far, previous assessments focused on mapping RTS with remote sensing imagery at local-scale and deriving temporal information from few snapshots in time. A continuous and high temporal assessment of the disturbance dynamics at a representative larger-scale is still missing. Thus, our main objective was to map and monitor RTS on a continental-scale and assess their annual temporal thaw dynamics in Northeast Siberia. We adapted and parametrised LandTrendr, an automated temporal segmentation algorithm, to identify the abrupt annual RTS disturbances in Landsat and Sentinel-2 time series data. Additionally, we applied spectral and spatial masks and a machine learning classification for improved RTS mapping. Our results show the first continental-scale mapped RTS distribution in NE Siberia and their annual thaw dynamics from. Overall, the RTS thaw dynamics steadily increased in NE Siberia during the assessment period. At the same time, local assessments revealed distinct periods of increased and decreased thawing dynamics. This indicates spatiotemporal variability in thaw dynamics and a strong connection to local drivers. Overall, our results highlight increased permafrost thaw, its large-scale impact and heightened permafrost vulnerability to thaw.

Description: Soils in the permafrost region have acted as car- bon sinks for thousands of years. As a result of global warming, permafrost soils are thawing and will potentially release greenhouse gases (GHGs) such as methane (CH4) and carbon dioxide (CO2). However, small-scale spatial heterogeneities of GHG production have been neglected in previous incubation studies. Here, we used an anaerobic incubation experiment to simulate permafrost thaw along a transect from upland Yedoma to the floodplain on Kurungnakh Island. Potential CO2 and CH4 production was measured during incubation of the active layer and permafrost soils at 4 and 20 ◦C, first for 60 d (approximate length of the growing season) and then continuing for 1 year. An assessment of methanogen abundance was performed in parallel for the first 60 d. Yedoma samples from upland and slope cores remained in a lag phase during the growing season simulation, while those located in the floodplain showed high production of CH4 (6.5 × 103 μg CH4-C g−1 C) and CO2 (6.9 × 103 μg CO2-C g−1 C) at 20 ◦C. The Yedoma samples from the permafrost layer started producing CH4 after 6 months of incubation. We conclude that landscape position is a key factor triggering CH4 production during the growing season time on Kurungnakh Island.