Conference Agenda

The Arctic is undergoing rapid warming, with Svalbard experiencing temperature increases well above the global average. Permafrost degradation, active-layer thickening, and slope instability increasingly affect ground stability in both natural and built environments. These processes generate measurable surface deformation signals that require systematic spatial observation. Reliable, spatially extensive monitoring of ground displacement is critical for geohazard assessment, infrastructure planning, and understanding climate-driven landscape processes. While operational InSAR Ground Motion Services (GMS) exist for mainland Norway and continental Europe, Svalbard remains outside existing operational frameworks due to environmental and processing constraints.

The InSAR Svalbard Ground Motion Service was developed to provide systematic deformation monitoring for High Arctic conditions. The service is based on Sentinel-1 C-band data and applies a Small Baseline Subset (SBAS) multi-temporal processing strategy tailored limited snow-free seasons and strong seasonal deformation signals. Processing strategies were further refined to account for non-linear thaw-freeze dynamics and spatially heterogeneous ground behavior characteristic of continuous permafrost terrain.

To quantify ground displacement at multiple temporal scales in permafrost environments, two complementary processing schemes were implemented. Seasonal time series quantify cumulative line-of-sight (LOS) displacement within individual snow-free periods (2016–2024), associated with thaw subsidence, frost heave, and seasonal slope creep. From these series, metrics such as maximum seasonal displacement and the timing of peak displacement (Day of Year) are derived to characterize intra-seasonal behavior. In parallel, interannual time series quantify long-term LOS velocity trends between consecutive snow-free seasons (2018–2024), enabling detection of gradual subsidence, persistent slope movement, and long-term permafrost change.

The current release covers five pilot areas in Western and Central Spitsbergen, including Longyearbyen, Ny-Ålesund, Svea, Hornsund, and Kapp Linné. These areas include main settlements, research stations, geomorphologically active slopes, and cultural heritage sites. Products are provided as open-access CSV datasets and are accessible through an interactive WebGIS platform developed in dialogue with end-users, including local authorities and research institutions in Svarlbard.

Results demonstrate that the service can resolve both rapid seasonal deformation and subtle interannual trends across diverse permafrost terrains. Observed displacement patterns illustrate how systematic InSAR monitoring can contribute to the assessment of sediment characteristics and ground-ice conditions in permafrost terrain. The InSAR Svalbard Ground Motion Service represents the first operational InSAR-based deformation monitoring framework established in a continuous permafrost setting in the High Arctic. Ongoing development will expand spatial coverage, incorporate new satellite acquisitions, and further refine time-series analysis to support long-term hazard assessment and climate change monitoring in Svalbard.

As satellite Earth observation systems mature into long-term operational infrastructures, understanding their environmental limitations becomes as important as demonstrating their capabilities. In polar and subpolar regions, rapid climate-driven transformations and expanding human activity demand reliable, repeatable monitoring tools capable of functioning under highly variable surface conditions. Yet, extreme seasonal forcing, particularly in freeze-thaw dominated terrains, challenges the temporal stability assumptions that underlie interferometric analysis. Establishing quantitative performance bounds for
radar coherence in such environments is therefore critical for defining when measurements can be trusted
and when caution is required.

Satellite radar interferometry (InSAR) is widely used for operational deformation monitoring, however, its reliability in high-latitude freeze-thaw environments remains strongly season-dependent due to rapid changes in surface dielectric properties, snow accumulation, meltwater infiltration, and permafrost dynamics. These processes introduce temporal decorrelation that may obscure true surface displacement signals. A structured, quantitative assessment of coherence stability under such conditions is therefore required to define the operational limits of C-band InSAR in Arctic environments. This study evaluates season-dependent Sentinel-1 coherence reliability over Longyearbyen, Svalbard, a compact Arctic testbed characterized by exposed bedrock, glaciers, thaw-sensitive permafrost, and critical infrastructure. A total of 334 Sentinel-1 SLC bursts acquired between 2019 and 2024 were processed using a single relative orbit to ensure geometric consistency. All acquisitions were IW-mode, in HH-HV polarization, reflecting Sentinel-1’s high latitude acquisition configuration over Svalbard, Longyearbyen during the study period. Short baseline (12-day) interferograms were generated to isolate seasonal decorrelation effects across multiple freeze-thaw cycles.

Interferometric processing was conducted in ESA SNAP using the Sentinel-1 TOPS workflow, including
precise orbit correction, burst splitting, back-geocoding with enhanced spectral diversity, interferogram
formation, topographic phase removal using the Copernicus 30 m DEM, multilooking, and terrain correction. Coherence statistics were evaluated seasonally to quantify temporal stability across surface
Types.

Based on seasonal coherence distributions, a three-tier operational usability classification is proposed:
Reliable (coherence > 0.5), Usable with Caution (0.3-0.5), and Not Reliable (< 0.3). This framework
translates coherence variability into practical decision-support guidance for Arctic monitoring applications.
Results indicate pronounced seasonal variability. Mean summer coherence during early years (2019-
2021) decreased to approximately 0.31 over snow and melt-affected surfaces, while winter coherence
in later years increased to values approaching 0.56 over exposed and infrastructure-dominated areas.
These findings confirm that freeze-thaw transitions represent a dominant limitation on phase stability,
with winter acquisitions providing substantially more reliable interferometric performance.

By quantitatively linking environmental forcing to C-band coherence stability, this study defines operational reliability thresholds for Sentinel-1 IW, HH-HV acquisitions in freeze-thaw dominated terrain and provides a transferable evaluation methodology for other high-latitude monitoring scenarios.

Keywords: Sentinel-1, InSAR, coherence, Arctic monitoring, freeze-thaw, Svalbard, TOPS, SNAP

Accurate and spatially comprehensive monitoring of seasonal snow depth (SD) is crucial for understanding and predicting hydrological processes, energy balance, and ecological dynamics in mountainous regions. In-situ measurements, while precise, are spatially limited and logistically challenging. Alternatively, LiDAR sensors can provide highly accurate measurements but are constrained by limited spatial and temporal coverage due to sparse sampling and low revisit frequencies. In this context, interferometric synthetic aperture radar (InSAR) systems offer a valuable alternative for assessing SD by providing high-resolution data over extended areas, independent of weather conditions or illumination. In this research, we address key associated challenges for deriving accurate SD from Digital Surface Models (DSMs) generated from Synthetic Aperture Radar Interferometry (InSAR), specifically focusing on mitigating errors introduced by radar wave penetration into the snowpack and ensuring precise DSM calibration. Current approaches rely on simplified assumptions regarding the compensation for the penetration depth, often averaging DSMs acquired under varying conditions, which introduces substantial uncertainty, particularly in complex alpine environments. Similarly, accurate DSM require a precise calibration due to the presence of residual offsets and tilts, which is expensive and impractical for large-area monitoring.

This paper outlines a novel, end-to-end framework for robust SD estimation from InSAR data. The core strategy involves differencing precisely corrected DSMs, one representing snow-free conditions and the other the snow surface. A critical innovation lies in the correction through a three-stage methodology designed to tackle the limitations of existing techniques. First, we develop an automated and reference-free approach for precise mutual calibration of InSAR-derived DSMs utilizing Persistent Scatterer Candidates (PSC) from Sentinel-1 data to establish natural calibration tie-points. This method eliminates the dependency on costly and potentially unavailable GPS or LiDAR ground truth data, while also compensating for residual offsets and tilts. Exploiting Sentinel-1’s global repeat-pass acquisition capability in conjunction with high-resolution TanDEM-X data ensures scalability and quality. Second, we implement a data-driven, deep learning (DL)-based approach to accurately estimate and correct for the radar penetration bias in snow-covered areas. Preliminary results on the Greenland Ice Sheet achieve state-of-the-art performance (RMSE = 0.65m, R² = 0.90) for X-band data, demonstrating the potential of this approach. Finally, after compensating the DSMs for both the calibration offset and tilts and the radar penetration, these corrected DSMs are used to perform DSM differencing to retrieve SD.

The estimation of the radar wave penetration into the snowpack can be adapted to multiple frequencies (from X- to C- and L-band) and potentially used in future InSAR missions such as Harmony and ROSE-L. For transferring the framework to C-band, we are considering the usage of historical STRM and Sentinel-1 data. For L-band, available airborne F-SAR data can be used, in view of extending the DL model with upcoming NISAR/ALOS-4 data.

The proposed framework enables reliable, large- scale SD estimation from InSAR DSM differencing, improving the understanding of cryospheric processes and leveraging the capabilities of current and future InSAR missions.

The recently launched NASA–ISRO Synthetic Aperture Radar (NISAR) mission will provide global L-band and regional S-band time-series measurements of polarimetric radar backscatter and InSAR-derived ground displacements. While the mission focuses on the operational generation of Level-1 and Level-2 data products, these products must also satisfy Level-3 science requirements defined by the Solid Earth, Ecosystems, and Cryospheric Science communities.

As part of the Solid Earth requirements, NISAR must enable semi-monthly measurements of surface deformation in permafrost-affected regions at 100 m spatial resolution during snow-free months. The mission is required to measure these displacements with an accuracy of 4(1 + L^1/2) mm or better, over length scales of 0.1 km < L < 50 km, across 80% of selected regions, and within any 90-day interval.

To validate the NISAR permafrost displacement requirement, we implemented a two-pronged approach. In Approach 1, we compare InSAR-derived surface displacements derived from NISAR Level-2 products with field observations collected at four permafrost validation sites in Alaska. At each site, we acquire dense in-situ measurements of seasonal permafrost displacement using repeated leveling and kinematic phase-based GNSS surveys. Observations are collected at 150 locations per site—50 along each of three transects within a single 100 m NISAR product pixel. This spatial sampling density is necessary to evaluate the NISAR permafrost displacement requirement with the necessary accuracy.

In Approach 2, we analyze NISAR InSAR data over regions with negligible ground displacement. Although stable terrain is uncommon in permafrost regions, we identified exposed bedrock sites on the Canadian Shield and the North Slope of Alaska for this purpose. Tropospheric and ionospheric correction layers are applied to raw phase time series, and spatial structure functions are computed to quantify InSAR noise as a function of length scale. These structure functions are compared to the scale-dependent NISAR permafrost displacement requirement to assess compliance at the 80th percentile.

This paper will present the selected validation sites and summarize the field protocols and analytical methodologies implemented for both approaches. Initial results from the NISAR permafrost requirement assessment will be presented, including comparisons between in-situ measurements and InSAR-derived surface displacements, as well as structure-function analyses of phase noise, both using Sentinel-1 data as a proxy alongside early NISAR observations. The paper will also describe planned field activities for the upcoming thaw season and outline the timeline for completing the overall validation effort.

Spaceborne Synthetic Aperture RADAR data are a key asset in the generation of geoinformation products for ice sheets in Antarctica and Greenland. SAR interferometry in particular proves crucial to generate ice velocity and grounding line information.

In Antarctica, observations started in the early 1990’s and early campaigns were regionally limited. While a backscatter map of the continent was achieved with a single satellite in 1997 (RADARSAT Antarctic Mapping Project - RAMP), full interferometric coverage would take another decade and a coordinated effort of three missions in support of the International Polar Year (IPY 2007-09). The launch of the Sentinel-1 constellation fundamentally changed data availability, because ESA committed to ongoing acquisitions in coastal Antarctica thus generating an archive more than a decade deep and going strong, with satellites 3 and 4 commissioned and a second generation of satellites in development. This coastal coverage continues to be augmented by acquisitions from other missions increasing area coverage, or providing acquisitions better suited for the analysis of fast glaciers. Together, the available missions form a virtual constellation that exceeds the information content of any single mission and provides a long term data record. Acquisition plan coordination, where available, allows targeted acquisitions to each mission strengths and responsible use of space assets. This virtual constellation has been serving the science community for many years now with shifting capabilities and new opportunities as missions undergo generational changes and new missions come online. In recent years, the commercial sector has started to contribute to the constellation and added the capability to collect fast (1-day) repeat InSAR data. Dense time series in fast changing environments opens the door for new research, the data are therefore particularly suited for fast glaciers. The launch of NISAR represents another leap in data availability, as this left looking science mission will cover most of the continent with ascending and descending acquisitions.

Here we provide a summary of our group's long-standing effort to provide Earth System Data Records for Antarctica based on the virtual SAR constellation. We highlight contributions and strength of the various missions and show examples of multi-mission based products like our most recently published 33 year grounding line record. Our results show that 77% of the grounding line has remained stable for the observation period. Grounding line retreat is concentrated in some key regions where retreat ranges from 10 to more than 40km. Based on our findings, we make recommendations for future acquisitions of missions that do not already have ongoing acquisition plans in place. This work is funded by NASA.