Ice-sheet grounding-line migration from tidal InSAR differential phase
Double-differential InSAR detects the tidal flexure zone where grounded ice becomes floating ice, resolving grounding-line position to within a few hundred metres. Retreat of this boundary is among the most direct early indicators of marine ice-sheet instability.
Sensors
- Sentinel-1 IW mode (C-band SAR): 5.6 cm wavelength, 5 x 20 m ground resolution in Interferometric Wide Swath mode, six-day exact repeat at polar latitudes (often shorter due to orbit overlap). The primary workhorse for operational grounding-line monitoring; free and open archive from 2014.
- ERS-1 / ERS-2 tandem archive (C-band SAR): One-day repeat during the 1995-1996 tandem phase produced near-zero temporal decorrelation pairs ideal for tidal differential InSAR. Foundational dataset; Rignot's seminal grounding-line mapping of Antarctica used these pairs. Archive held by ESA.
- TanDEM-X (X-band SAR): 9.6 mm wavelength, bistatic single-pass interferometry at roughly 3 m resolution. Shorter wavelength increases phase sensitivity to small vertical displacements, but X-band is more susceptible to volume scattering in wet snow. Useful for high-resolution flexure-zone characterisation.
- COSMO-SkyMed (X-band SAR): Spotlight mode reaches 1 m resolution; repeat-pass intervals configurable from one to several days. Higher resolution helps resolve narrow flexure zones on smaller outlet glaciers, though X-band coherence over rough or wet ice surfaces can degrade quickly.
Why the grounding line is the number that matters
The grounding line is the boundary beneath an ice sheet where the base lifts off the bedrock and the ice begins to float. Its position is not a static feature. On marine ice sheets resting on beds that slope inland, any retreat exposes the ice to deeper water, which increases buoyancy forces and accelerates flow. This is the marine ice-sheet instability hypothesis, and the grounding line is its observable front line.
Retreat of even a few kilometres per year compounds. The West Antarctic Ice Sheet sits on a retrograde bed in several sectors; Pine Island and Thwaites glaciers have both shown documented grounding-line retreat over the past two decades using exactly the methods described here. Grounding-line position is therefore not a glaciological curiosity. It is a direct input to sea-level rise projections, and monitoring it is a practical requirement for any government or institution with coastal infrastructure to protect.
What a floating roof gives away: the physics of tidal flexure
Grounded ice does not move vertically with the tide. Floating ice does. At the grounding line, there is a transition zone, typically a few kilometres wide, where the ice sheet bends elastically as the ocean tide rises and falls beneath it. This flexure zone is invisible to the eye and to optical sensors, but it produces a measurable, repeatable phase signal in a synthetic aperture radar interferogram.
Double-differential InSAR isolates this signal by forming two interferograms from SAR acquisitions at different tidal states and then differencing those interferograms. Atmospheric and orbital phase contributions largely cancel. What remains is the differential tidal displacement field: a fringe pattern whose inner boundary marks the limit of tidal influence and whose outer boundary approximates the hydrostatic line. The grounding line itself sits between these two, typically identified as the landward limit of detectable flexure. Rignot et al. established this methodology using ERS tandem pairs in the 1990s and it has been applied continuously since.
The method requires coherent SAR pairs, which means temporal baselines short enough that the ice surface has not changed significantly between acquisitions. C-band coherence over slow-moving grounded ice can survive six-day Sentinel-1 repeats in winter; over fast-flowing or wet-surface areas it degrades rapidly. X-band coherence is even more sensitive to surface change, which is why tandem or near-simultaneous acquisitions are preferred for those systems.
From raw phase to a retreating boundary: the processing chain
Processing begins with coregistered SLC (Single Look Complex) pairs. Precise orbit vectors, available for Sentinel-1 within a few days of acquisition, are applied to remove the flat-Earth phase. A digital elevation model, typically TanDEM-X global DEM at 12 m or 30 m posting, removes topographic phase. Two interferograms formed from acquisitions spanning different tidal epochs are then differenced to suppress non-tidal signals.
Phase unwrapping across the flexure zone is the most demanding step. The fringe density near the hinge line can be high, and phase discontinuities from crevassing or surface melt can cause unwrapping errors that propagate into the position estimate. Manual quality control remains standard practice. Once unwrapped, the flexure profile is fitted to an elastic beam model to extract hinge-line position. Uncertainty in the final grounding-line position is typically quoted at 100 to 500 m depending on ice thickness, tidal amplitude and coherence quality. That is honest: the method does not deliver metre-scale precision.
Migration rates require at least two well-separated epochs. Comparing grounding-line positions derived from ERS tandem pairs in the mid-1990s with Sentinel-1 results from the 2010s and 2020s gives multi-decadal retreat rates. For sectors with good coherence and adequate tidal contrast, annual or even seasonal monitoring is achievable with the Sentinel-1 archive.
What the satellites cannot see
Cloud is irrelevant: SAR penetrates it completely. But several other limits are real and worth stating plainly.
Summer melt on the ice surface destroys C-band coherence within days. Monitoring is therefore largely confined to austral or boreal winter for C-band systems. X-band is worse in this respect. Grounding lines on fast-flowing outlet glaciers, where surface velocities exceed a few hundred metres per year, decorrelate even six-day pairs. For those glaciers, amplitude offset-tracking can estimate ice velocity but cannot resolve the fine-scale flexure signal needed for grounding-line mapping.
Tidal amplitude matters. In regions where the tidal range is less than roughly 0.5 m, the differential phase signal may fall below the noise floor of the interferogram. Parts of East Antarctica have very small tidal ranges and grounding-line detection there is correspondingly harder. Conversely, large tidal ranges in the Amundsen Sea sector help produce clear flexure signals. The method also cannot directly observe the bed geometry beneath the ice, so retreat rates must be interpreted alongside airborne or ship-based bathymetric surveys to assess instability risk.
Operational monitoring and what a programme looks like
A practical grounding-line monitoring programme selects priority sectors based on known retreat risk, available coherent archive, and tidal range. For West Antarctica, the Amundsen Sea Embayment glaciers are the obvious priority. For Greenland, Jakobshavn, Helheim and Kangerdlugssuaq have sufficient archive depth and tidal range to support time-series analysis.
Sentinel-1's six-day repeat and open data policy mean that a systematic Antarctic monitoring run can be maintained at relatively low cost once the processing chain is established. The main operational expense is analyst time for phase-unwrapping quality control and epoch-to-epoch comparison. Outputs are typically delivered as georeferenced polylines (the grounding-line position at each epoch), uncertainty envelopes, and retreat-rate maps in standard GIS formats.
Satellize runs InSAR-based deformation analytics on open constellations and can apply the same differential-phase pipeline to grounding-line monitoring on client licence. The processing architecture is the same one used for surface deformation work in other contexts; the glaciological interpretation layer is what changes. Clients interested in a specific glacier sector or a multi-decade retreat time series using the ERS and Sentinel-1 archive can scope a programme around those boundaries.
Typical figures
| Spatial resolution (Sentinel-1 IW) | 5 x 20 m (single look); multi-looked to ~40-80 m for interferometric products |
| Grounding-line position uncertainty | Typically 100-500 m depending on tidal amplitude, coherence and ice thickness |
| Revisit interval (Sentinel-1 at polar latitudes) | 6 days exact repeat; multiple tracks can reduce effective interval |
| Radar frequency / wavelength | C-band (Sentinel-1, ERS): 5.6 cm; X-band (TanDEM-X, COSMO-SkyMed): 3.1 cm |
| Minimum detectable tidal displacement | Approximately 1-2 cm vertical (C-band phase sensitivity ~2.8 cm per full fringe) |
| Temporal baseline for coherent pairs | 1 day (ERS tandem, TanDEM-X bistatic); 6 days (Sentinel-1); coherence degrades rapidly in summer melt season |
| Archive depth | ERS tandem pairs from 1995-1996; Sentinel-1 from 2014 to present; COSMO-SkyMed from ~2007 |
| Coverage | Both polar ice sheets; Sentinel-1 provides systematic Antarctic coverage; Arctic outlet glaciers covered on repeat |
| Delivery formats | Georeferenced polylines (GeoJSON, Shapefile, GeoPackage), uncertainty envelopes, retreat-rate rasters (GeoTIFF) |
| Latency (Sentinel-1 NRT) | Level-1 SLC data available within ~1 hour of acquisition; processed interferometric products within 24-48 hours of tasking |
Analytics Satellize can run
| Grounding-line position map (single epoch) | Double-differential InSAR tidal flexure detection; elastic beam model fitting to unwrapped phase profile | Georeferenced polyline with positional uncertainty envelope (GeoJSON or Shapefile) |
| Multi-epoch retreat time series | Sequential grounding-line mapping across ERS archive, Sentinel-1 archive and current acquisitions; epoch-to-epoch displacement calculation | Time-series table and animated GIS layer showing retreat progression with annual or seasonal cadence |
| Flexure-zone width and hinge-line characterisation | Elastic beam inversion of differential phase profile; width correlated with ice thickness estimates from published bed topography (BedMachine) | Annotated cross-section report with flexure-zone geometry per glacier sector |
| Coherence quality assessment | Interferometric coherence magnitude mapping across candidate image pairs; seasonal coherence window identification | Acquisition planning calendar identifying optimal winter windows for each target glacier |
| Retreat-rate anomaly alert | Statistical comparison of current-epoch grounding-line position against historical baseline; threshold exceedance flagging | Alert report (PDF and API flag) triggered when retreat exceeds a client-defined threshold in metres per year |
| Integrated retreat and velocity change assessment | Co-registration of grounding-line retreat polylines with surface velocity fields derived from Sentinel-1 offset tracking or InSAR | Combined GIS layer and summary report relating grounding-line migration to upstream velocity acceleration |
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.