Sensors
- Sentinel-1 (C-band SAR): Repeat-pass interferometric pairs at 6-day revisit (12-day for a single satellite) over most high-latitude areas. Ground-range resolution of 5 x 20 m in IW mode. Phase coherence across pairs allows detection of line-of-sight displacement to roughly 3-5 mm per interferogram under good coherence conditions. Permafrost monitoring studies routinely use time-series stacking (SBAS or PS-InSAR) to separate seasonal heave from long-term subsidence trends.
- TerraSAR-X / TanDEM-X (X-band SAR): 11-day repeat, spotlight mode down to 1 m resolution, stripmap at 3 m. X-band is more sensitive to surface scattering and loses coherence faster over vegetated tundra than C-band, but provides finer spatial detail over built infrastructure. Useful for site-specific structural assessments where Sentinel-1's 5 m pixel is too coarse.
- MODIS Land Surface Temperature (Terra/Aqua): 1 km spatial resolution, twice-daily overpass per satellite. MOD11A1 and MYD11A1 daily LST products, with MOD11A2 8-day composites reducing cloud contamination. Used to compute freeze-thaw cycle frequency, thawing degree-days, and multi-year warming trends that drive active-layer deepening. Archive extends to 2000.
- VIIRS (Suomi NPP / NOAA-20): 375 m resolution in I-bands, 750 m in M-bands. Daily global coverage. VIIRS LST products complement MODIS with slightly finer spatial detail and extend the thermal record forward as MODIS instruments age. Particularly useful for identifying localised warm anomalies around industrial heat sources that accelerate thaw.
- Landsat 8/9 (TIRS thermal): 100 m thermal infrared resolution (resampled to 30 m in products), 16-day repeat. Provides longer archive context back to 1984 in the broader Landsat family. TIRS band 10 (10.6-11.19 µm) supports surface temperature retrieval for multi-decadal trend analysis at the landscape scale.
What permafrost failure actually looks like from orbit
Permafrost does not fail suddenly. The sequence is: warming summers deepen the active layer, ice-rich ground melts, volume is lost, and the surface subsides. That subsidence can run at 1-3 cm per year in ice-rich terrain and accelerate sharply when drainage is disrupted. Thermokarst ponds form, expand, and coalesce. On a map, this looks like a rash of dark irregular water bodies spreading across what was previously dry ground.
From a Sentinel-1 interferometric pair, the same process appears as a coherent phase gradient: a bowl-shaped deformation pattern centred on the thawing zone. The seasonal signal is also diagnostic. Permafrost ground heaves in winter as the active layer refreezes and subsides each summer thaw. The amplitude of that annual cycle, and whether it is shrinking or growing, tells you whether the permafrost table is stable or retreating. A site showing net residual subsidence after each annual cycle is losing ground volume permanently.
Why InSAR is the right instrument, and where it fails
InSAR works by comparing the phase of radar returns across two passes of the same satellite. If the ground has moved between passes, the phase shifts. At C-band (5.6 cm wavelength), one full phase cycle corresponds to roughly 2.8 cm of line-of-sight displacement, and sub-centimetre precision is achievable through time-series averaging. That is the sensitivity needed for permafrost monitoring, where structural risk accumulates over years of millimetre-scale losses.
The honest limits matter. Coherence collapses over dense vegetation and wet snow, which are common in boreal and Arctic environments precisely when summer thaw is most active. Analysts work around this by selecting pairs from late summer when vegetation is dry and snow-free, and by using persistent-scatterer or SBAS approaches that identify coherent pixels (rock outcrops, gravel pads, infrastructure surfaces) within otherwise incoherent scenes. Over open tundra, coherence is generally adequate. Over boreal forest, it is often not, and optical or thermal methods must carry the analysis.
Atmospheric water vapour also introduces phase delays that can mimic deformation signals of 1-2 cm. Corrections using ERA5 reanalysis data or GACOS (Generic Atmospheric Correction Online Service) are standard practice but add processing complexity and residual uncertainty.
Thermal products: the climate context InSAR cannot provide
InSAR tells you the ground moved. It does not tell you why, or how the risk will evolve. MODIS and VIIRS LST products fill that gap by quantifying the thermal forcing driving active-layer change. Thawing degree-days, computed from daily LST above 0°C accumulated over the summer season, are the primary driver of active-layer thickness in most permafrost models. A site where thawing degree-days have increased by 15-20% over two decades faces a structurally different risk profile than a site where they have been stable.
The combination is powerful for prioritisation. A property portfolio spanning hundreds of sites across northern Canada, Siberia, Alaska, or the Tibetan Plateau can be screened thermally at low cost to identify which sites warrant full InSAR time-series analysis. Sites with both accelerating thermal forcing and detectable InSAR subsidence move to the front of the inspection queue.
What the deformation signal means for a property or infrastructure owner
A seasonal heave-subsidence amplitude of 2-5 cm with no net trend is consistent with a stable active layer. Net residual subsidence of 1-3 cm per year, persisting across multiple annual cycles, indicates active ice loss and is the threshold at which foundation engineers typically recommend site investigation. Rates above 5 cm per year have been documented in ice-rich terrain in Siberia and are associated with rapid structural compromise.
For built infrastructure, the spatial pattern matters as much as the magnitude. Differential subsidence across a building footprint, rather than uniform settlement, generates the shear stresses that crack walls and buckle floors. A 3 cm total displacement is manageable if uniform; the same displacement concentrated across 10 metres of a pipeline or runway is a failure mode. InSAR at Sentinel-1 resolution can resolve differential signals across most large structures, though TerraSAR-X spotlight mode is needed for precise assessment of individual building footprints.
Valuation implications follow directly. Lenders financing infrastructure in permafrost zones increasingly require evidence of ground stability. Insurers pricing long-duration policies on northern assets need quantified subsidence rates, not qualitative hazard maps. Satellite-derived deformation time series, validated against ground truth where available, are becoming a standard input to that underwriting process.
Building an operational monitoring programme
A practical monitoring stack for a northern property or infrastructure portfolio combines three layers. First, a baseline thermal screening using the full MODIS archive (2000 to present) to classify sites by cumulative thermal forcing change. Second, a retrospective InSAR time series using the Sentinel-1 archive (from 2014 in most areas) to establish historical deformation rates at priority sites. Third, a forward-looking alert system that flags new interferograms exceeding a defined displacement threshold, typically 5-10 mm per 12-day pair, at monitored locations.
Satellize runs this kind of stacked analysis on open constellations, adding commercial TerraSAR-X tasking where site-level resolution is required. The Tonga crop-estimation programme demonstrated the same principle of combining open thermal products with higher-resolution tasking at decision-relevant sites. The architecture transfers directly to northern monitoring.
Cloud cover is less of a problem for SAR-based monitoring than for optical programmes, but it remains a constraint for the MODIS and VIIRS thermal inputs. Persistent cloud in maritime Arctic environments can reduce usable LST observations to fewer than 30 clear days per summer season, which limits the precision of thawing degree-day calculations. Analysts should treat LST-derived thermal indices as climatological context rather than precise annual measurements in heavily clouded regions.
Typical figures
| Spatial resolution (InSAR, Sentinel-1 IW) | 5 x 20 m ground range; multi-looked to ~14 m for deformation products |
| Spatial resolution (InSAR, TerraSAR-X spotlight) | 1-3 m; suitable for individual structure assessment |
| Minimum detectable displacement (InSAR time series) | 3-5 mm per interferogram; sub-centimetre with SBAS/PS stacking over multiple pairs |
| Revisit interval (Sentinel-1) | 6 days (two-satellite constellation); 12 days single satellite |
| Thermal product resolution (MODIS LST) | 1 km; 8-day composites reduce cloud gaps |
| Thermal product resolution (VIIRS LST) | 375 m (I-band); daily global coverage |
| Archive depth | Sentinel-1 InSAR from 2014; MODIS LST from 2000; Landsat thermal from 1984 |
| SAR frequency / wavelength | C-band 5.405 GHz / 5.6 cm (Sentinel-1); X-band 9.65 GHz / 3.1 cm (TerraSAR-X) |
| Coverage | Global; Sentinel-1 acquisition mode prioritises Arctic and sub-Arctic regions |
| Delivery formats | GeoTIFF deformation maps, GeoPackage vector alerts, CSV time-series per site, PDF risk summary |
Analytics Satellize can run
| Site-level annual subsidence rate | SBAS InSAR time-series inversion over Sentinel-1 archive, atmospheric correction via ERA5 | GeoTIFF displacement velocity map and per-site CSV with confidence intervals |
| Seasonal heave-subsidence amplitude | Harmonic decomposition of InSAR displacement time series to separate annual cycle from secular trend | GIS layer with amplitude and trend classified by risk tier |
| Thermokarst pond expansion rate | Multi-temporal optical water-body mapping (Sentinel-2 NDWI or Landsat) combined with InSAR subsidence extent | Annual change polygon layer with area and rate statistics |
| Thawing degree-day trend index | MODIS MOD11A1 daily LST accumulation above 0°C, multi-year linear regression per site | Site screening table ranked by thermal forcing change, exportable to Excel or GIS |
| Differential subsidence across structure footprint | TerraSAR-X PS-InSAR over individual building or infrastructure footprint; tilt and gradient computation | Per-structure deformation gradient report for engineering review |
| Portfolio risk tier classification | Combined scoring of InSAR subsidence rate, thermal trend index, and ice-content proxy (soil type overlay) | Ranked asset list with risk tier (stable / watch / elevated / critical) and recommended action |
| Displacement threshold alert | Automated interferogram processing flagging pixels exceeding operator-defined displacement per cycle | Email or API alert with location, magnitude, and link to current interferogram |
Who does the work
We can get this done for you. Satellize runs its own analyst desk and a strong science team. You do not buy a data feed and work out what it means; our people source the imagery, run the analysis described on this page, and hand you the answer with its confidence limits stated. Discuss this requirement.