The accelerating climatic changes and new infrastructure development across the Arctic require more robust risk and environmental assessment, but thus far there is no consistent record of human impact. We provide a first panarctic satellite-based record of expanding infrastructure and anthropogenic impacts along all permafrost affected coasts (100 km buffer, ≈6.2 Mio km²), named the Sentinel-1/2 derived Arctic Coastal Human Impact (SACHI) dataset. The completeness and thematic content goes beyond traditional satellite based approaches as well as other publicly accessible data sources. Three classes are considered: linear transport infrastructure (roads and railways), buildings, and other impacted area. C-band synthetic aperture radar and multi-spectral information (2016–2020) is exploited within a machine learning framework (gradient boosting machines and deep learning) and combined for retrieval with 10 m nominal resolution. In total, an area of 1243 km² constitutes human-built infrastructure as of 2016–2020. Depending on region, SACHI contains 8%–48% more information (human presence) than in OpenStreetMap. 221 (78%) more settlements are identified than in a recently published dataset for this region. 47% is not covered in a global night-time light dataset from 2016. At least 15% (180 km²) correspond to new or increased detectable human impact since 2000 according to a Landsat-based normalized difference vegetation index trend comparison within the analysis extent. Most of the expanded presence occurred in Russia, but also some in Canada and US. 31% and 5% of impacted area associated predominantly with oil/gas and mining industry respectively has appeared after 2000. 55% of the identified human impacted area will be shifting to above 0 ^∘C ground temperature at two meter depth by 2050 if current permafrost warming trends continue at the pace of the last two decades, highlighting the critical importance to better understand how much and where Arctic infrastructure may become threatened by permafrost thaw.

Infrastructure expands rapidly in the Arctic due to industrial development. At the same time, climate change impacts are pronounced in the Arctic. Ground temperatures are, for example, increasing as well as coastal erosion. A consistent account of the current human footprint is needed in order to evaluate the impact on the environments as well as risk for infrastructure. Identification of roads and settlements with satellite data is challenging due to the size of single features and low density of clusters. Spatial resolution and spectral characteristics of satellite data are the main issues regarding their separation. The Copernicus Sentinel-1 and -2 missions recently provided good spatial coverage and at the same time comparably high pixel spacing starting with 10 m for modes available across the entire Arctic. The purpose of this study was to assess the capabilities of both, Sentinel-1 C-band Synthetic Aperture Radar (SAR) and the Sentinel-2 multispectral information for Arctic focused mapping. Settings differ across the Arctic (historic settlements versus industrial, locations on bedrock versus tundra landscapes) and reference data are scarce and inconsistent. The type of features and data scarcity demand specific classification approaches. The machine learning approaches Gradient Boosting Machines (GBM) and deep learning (DL)-based semantic segmentation have been tested. Records for the Alaskan North Slope, Western Greenland, and Svalbard in addition to high-resolution satellite data have been used for validation and calibration. Deep learning is superior to GBM with respect to users accuracy. GBM therefore requires comprehensive postprocessing. SAR provides added value in case of GBM. VV is of benefit for road identification and HH for detection of buildings. Unfortunately, the Sentinel-1 acquisition strategy is varying across the Arctic. The majority is covered in VV+VH only. DL is of benefit for road and building detection but misses large proportions of other human-impacted areas, such as gravel pads which are typical for gas and oil fields. A combination of results from both GBM (Sentinel-1 and -2 combined) and DL (Sentinel-2; Sentinel-1 optional) is therefore suggested for circumpolar mapping.

Rain-on-snow (ROS) events occur across many regions of the terrestrial Arctic in mid-winter. Snowpack properties are changing, and in extreme cases ice layers form which affect wildlife, vegetation and soils beyond the duration of the event. Specifically, satellite microwave observations have been shown to provide insight into known events. Only Ku-band radar (scatterometer) has been applied so far across the entire Arctic. Data availability at this frequency is limited, however. The utility of other frequencies from passive and active systems needs to be explored to develop a concept for long-term monitoring. The latter are of specific interest as they can be potentially provided at higher spatial resolution. Radar records have been shown to capture the associated snow structure change based on time-series analyses. This approach is also applicable when data gaps exist and has capabilities to evaluate the impact severity of events. Active as well as passive microwave sensors can also detect wet snow at the timing of an ROS event if an acquisition is available. The wet snow retrieval methodology is, however, rather mature compared to the identification of snow structure change since ambiguous scattering behaviour needs consideration. C-band radar is of special interest due to good data availability including a range of nominal spatial resolutions (10 m–12.5 km). Scatterometer and SAR (synthetic aperture radar) data have therefore been investigated. The temperature dependence of C-band backscatter at VV (V – vertical) polarization observable down to −40 ^∘C is identified as a major issue for ROS retrieval but can be addressed by a combination with a passive microwave wet snow indicator (demonstrated for Metop ASCAT – Advanced Scatterometer – and SMOS – Soil Moisture and Ocean Salinity). Results were compared to in situ observations (snowpit records, caribou migration data) and Ku-band products. Ice crusts were found in the snowpack after detected events (overall accuracy 82 %). The more crusts (events) there are, the higher the winter season backscatter increase at C-band will be. ROS events captured on the Yamal and Seward peninsulas have had severe impacts on reindeer and caribou, respectively, due to ice crust formation. SAR specifically from Sentinel-1 is promising regarding ice layer identification at better spatial details for all available polarizations. The fusion of multiple types of microwave satellite observations is suggested for the creation of a climate data record, but the consideration of performance differences due to spatial and temporal cover, as well as microwave frequency, is crucial. Retrieval is most robust in the tundra biome, where results are comparable between sensors. Records can be used to identify extremes and to apply the results for impact studies at regional scale.

Shallow lakes are common across the entire Arctic. They play an important role as methane sources and wildlife habitats, and they are also associated with thermokarst processes which are characteristic of permafrost environments. Many lakes freeze to the ground along their rims and often over the entire extent during winter time. Knowledge on the spatial patterns of ground-fast and floating ice is important as it relates to methane release, talik formation and hydrological processes, but no circumpolar account of this phenomenon is currently available. Previous studies have shown that ground-fast ice can easily be detected using C-band Synthetic Aperture Radar (SAR) backscatter intensity data acquired from satellites. A major challenge is that backscatter intensity varies across the satellite scenes due to incidence angle effects. Circumpolar application therefore requires the inclusion of incidence angle dependencies into the detection algorithm. An approach using ENVISAT ASAR Wide Swath data (approximately 120 m spatial resolution) has therefore been developed supported by bathymetric measurements for lakes in Siberia. This approach was then further applied across the entire Arctic for late winter 2008. Ground-fast ice fraction has been derived for (1) two million lake objects larger than 0.025 km² (post-processed GlobeLand30), (2) a 50 x 50 km grid and (3) within certain zones relevant for climate studies (permafrost type, last glacial maximum, Yedoma). Especially lakes smaller than approximately 0.1 km² may freeze completely to the ground. The proportion of ground-fast ice increases with increasing soil organic carbon content in the proximity of the lakes. This underlines the importance of such lakes for emission studies and the need to map the occurrence of ground-fast lake ice. Clusters of variable fractions of ground-fast ice occur especially in Yedoma regions of Eastern Siberia and Alaska. This reflects the nature of thaw lake dynamics. Analyses of lake depth measurements from several sites suggest that the used method yields the potential to utilize ground-fast lake ice information over larger areas with respect to landscape development, but results need to be treated with care, specifically for larger lakes and along river courses.

Satellite‐derived surface soil moisture data are available for the Arctic, but detailed validation is still lacking. Previous studies have shown low correlations between in situ and modeled data. It is hypothesized that soil temperature variations after soil thaw impact MetOp ASCAT satellite‐derived surface soil moisture (SSM) measurements in wet tundra environments, as C band backscatter is sensitive to changes in dielectric properties. We compare in situ measurements of water content within the active layer at four sites across the Arctic in Alaska (Barrow, Sagwon, Toolik) and Siberia (Tiksi), taken in the spring after thawing and in autumn prior to freezing. In addition to the long‐term measurement fields, where sensors are installed deeper in the ground, we designed a monitoring setup for measuring moisture very close to the surface in the Lena River Delta, Siberia. The volumetric water content (VWC) and soil temperature sensors were placed in the moss organic layer in order to account for the limited penetration depth of the radar signal. ASCAT SSM variations are generally very small, in line with the low variability of in situ VWC. Short‐term changes after complete thawing of the upper organic layer, however, seem to be mostly influenced by soil temperature. Correlations between SSM and in situ VWC are generally very low, or even negative. Mean standard deviation matching results in a comparably high root‐mean‐square error (on average 11%) for predictions of VWC. Further investigations and measurement networks are needed to clarify factors causing temporal variation of C band backscatter in tundra regions.