Permafrost is warming globally, which leads to widespread permafrost-thaw, inducing highly dynamic local thaw disturbances impacting their direct surrounding and ultimately resulting in widespread permafrost degradation and loss on local to regional scales. Permafrost-thaw disturbances range from local, rapid single events to regional-scale gradual processes, pulse and press disturbances, respectively. Both disturbance types directly affect the surrounding landscape, ecosystems and infrastructure and further impact hydrological systems, soil carbon accumulation and decomposition and greenhouse gas emissions. Retrogressive thaw slumps (RTS) are highly dynamic pulse permafrost disturbance features. Although mainly locally occurring, they significantly impact the surrounding biogeochemistry, hydrology, and geomorphology. Moreover, RTS influence soil organic carbon stocks significantly as they erode ice-rich permafrost soils deeply and mobilise large volumes of soil and carbon on an annual scale. Previous studies focused on mapping RTS at local sites and deriving dynamic information from individual time snapshots but a continuous and high temporal assessment of the disturbance dynamics as well as the number and distribution of RTS on a pan-arctic scale are still missing. Hence, mapping and monitoring of permafrost disturbance dynamics is essential for assessing and quantifying their impacts and further incorporating the dynamics in large-scale climate models. Remote sensing offers a wide range of assessments possibilities and especially the Landsat archive has been used extensively for high spatial, large-scale land cover changes, as it contains the longest, continuous image archive with data for more than 48 years. However, optical remote sensing is restricted in northern high latitudes due to persistent cloud cover and specific environmental factors such as frequent snow and ice cover, short growing season length, and challenging solar geometry. Despite the continuous acquisition scheme of the Landsat missions, we still have considerable temporal data gaps in northern high latitudes and especially at coastal sites. We therefore propose to combine imagery from the Landsat archive with the newly available Sentinel-2 images. Sentinel-2 is part of ESA’s recently launched Copernicus Earth observation programme. The Sentinel-2 mission consists of two satellites, Sentinel-2A and Sentinel-2B, which both contain a multispectral imager (MSI) sensor system on board and acquire images since 2015 and 2017, respectively. Both satellites together have a revisit time of 5 days at the equator, which is a huge temporal improvement to Landsat’s 16-day revisit cycle. When relying on both, Landsat 8 and Sentinel-2, the combined sensor-independent revisit time shortens to ~2.9 days and to even less at the high latitudes, increasing the likelihood of acquiring cloud-free images from particular observation periods such as the summer season drastically. We want to leverage this advantage of an enhanced optical image collection archive and propose a combined Landsat and Sentinel-2 image assessment approach in time series analysis for northern high latitudes. Our objective is to assess the possibilities of combining Landsat and Sentinel-2 data to ensure a gap-free time series analysis to perform high spatial and temporal permafrost disturbance assessments in northern high latitudes. To address this, we focused on three sub-objectives: 1. Assessing the comparability and compatibility of Landsat 8 and Sentinel-2 data in northern high latitudes, 2. Combining Landsat and Sentinel-2 data together in annual mosaics for time series analysis, and 3. Conducting a time series analysis with Landsat and Sentinel-2 data to assess permafrost disturbance dynamics. First, Sentinel-2 and Landsat 8, the current Landsat missions’ satellite since 2013, are orbiting and acquiring images simultaneously. For an in-depth comparison, to assess their spectral comparability and compatibility with one another, we identified overlapping same-day acquisitions at three sites in Eastern Siberia in 2016. For each same-day image pair we conducted a spectral comparison for all corresponding spectral bands (visible to shortwave infrared) and performed a pixel-based ordinary least square regression analysis to assess the spectral agreement of Landsat 8 and Sentinel-2 bands. The results showed an overall good fit between Landsat 8 and Sentinel-2 bands. The regression coefficients underlined this as well and provided the coefficients for a bandpass transformation. We adjusted the Sentinel-2 images with the coefficients to resemble the Landsat 8 images, which increased their spectral agreement and showed and almost perfect alignment between the Landsat 8 and adjusted Sentinel-2 images. Besides the individual site assessments, we combined data from all three sites, performed the same analysis and derived band transformation coefficients representative for Eastern Siberia, enabling a combined use of Landsat 8 and Sentinel-2. Besides the spectral adjustment of Sentinel-2 to Landsat, the general limitations of optical remote sensing in northern high latitudes due to persistent cloud cover still restricts the application of temporally-dense time series analysis. We therefore assessed the possibility to integrate Landsat and Sentinel-2 data in a combined annual mosaic workflow, enhancing data coverage to achieve better and more reliable time series analysis. Annual mosaics combine several images to achieve a full coverage across large spatial scales, avoiding clouds, cloud shadows and data gaps. Time series disturbance algorithms, such as LandTrendr, rely on annual gap-free mosaics, which cannot be achieved by Landsat-only data in northern high latitudes. We therefore, adapted the Landsat mosaicking workflow to incorporate Sentinel-2 images, included the spectral bandpass transformation and created combined Landsat+Sentinel-2 mosaics. We evaluated the improved data availability for annual mosaics when using both Landsat and Sentinel-2 and also assessed the improved mosaic coverage and maintained good mosaic quality. For twelve representative sites across Siberia, the image availability increased drastically with the addition of Sentinel-2 and likewise did the cloud-free pixel number, demonstrating that a combined Landsat and Sentinel-2 image collection enhances the database for the annual mosaics greatly. Furthermore, the Landsat+Sentinel-2 mosaics showed a more reliable spatial coverage of the study sites compared to Landsat-only mosaics. While Landsat+Sentinel-2 mosaics always fully cover a study site (100%) the Landsat-only mosaics reached only 27.2% and 58.1% coverage for Sobo-Sise and East Taymyr in 2017. A spectral comparison between the Landsat+Sentinel-2 mosaics and individual cloud-free Landsat 8 and Sentinel-2 images for all twelve sites showed high spectral agreement between the input images and the mosaic output, illustrating a good quality of the combined mosaics. Our analysis showed that especially the most northern, coastal sites benefit from a combined Landsat and Sentinel-2 mosaic workflow. The combined workflow allows to create high spatial resolution input mosaics for high temporal continuous time series analysis. For the first time, permafrost disturbance dynamics can be assessed on an annual temporal scale across large regions, with algorithms such as LandTrendr. Thirdly, we adapted and applied LandTrendr to use the Landsat+Sentinel-2 mosaics as input and detected the thaw dynamics of RTS. LandTrendr applies a temporal segmentation of spectral data on a pixel-by-pixel basis, that catches both long-term trends and abrupt changes. Abrupt disturbances such as RTS have specific temporal disturbance trajectories which are used to spatially map the disturbances. Our study presents a continuous time series analysis from 1999 to 2019, determining the RTS dynamics based on disturbance year, magnitude of disturbance, recovery duration and rate of change at high spatial resolution across Siberia on an annual scale. This is the first comprehensive analysis of RTS dynamics providing the possibility to identify years with higher thaw dynamics, which can be further related to auxiliary data, such as air temperature and precipitation, to identify main thaw drivers. Characterisation of permafrost thaw disturbances and recovery are key processes describing the thaw dynamics affecting the permafrost carbon cycle. Thus, spatially and temporally explicit knowledge of these processes and their drivers are critical for understanding large scale carbon cycles.