Retaining wall lateral displacement from InSAR
Repeat-pass SAR interferometry detects millimetre-scale displacement in retaining walls and sheet-pile structures, but line-of-sight geometry means lateral motion is only partially captured. Combining ascending and descending passes resolves the ambiguity; inclinometers resolve the rest.
Sensors
- Sentinel-1 (C-band, ESA): 5.6 cm wavelength, 5 x 20 m ground range resolution in Interferometric Wide Swath mode, 6-day repeat at mid-latitudes with both satellites active. Persistent-scatterer processing achieves 1-2 mm line-of-sight precision on coherent targets. Free and open archive from 2014.
- COSMO-SkyMed (X-band, ASI): 3.1 cm wavelength, spotlight modes down to 1 m resolution, revisit as short as 12 hours with the four-satellite constellation. Shorter wavelength improves sensitivity to small displacements but increases atmospheric phase noise in humid conditions.
- TerraSAR-X / TanDEM-X (X-band, DLR): Spotlight mode at 1-2 m resolution, 11-day repeat for a single satellite. High coherence on engineered metal structures such as sheet piles. Staring-spotlight mode achieves sub-millimetre precision on strong point scatterers.
- ICEYE (X-band, commercial): Sub-3 m resolution in strip mode, flexible tasking with revisit configurable to daily or better for a given site. Useful for rapid-onset monitoring where archive depth is less important than acquisition cadence.
Why InSAR sees walls differently from the ground
A SAR satellite measures the round-trip phase delay between itself and a target on the ground. That delay encodes displacement only along the line of sight, which is inclined at roughly 30 to 46 degrees from vertical depending on the sensor and the selected beam. The implication for retaining walls is awkward: the motion engineers worry about most, lateral wall translation toward an excavation, is largely horizontal. A purely horizontal displacement of 10 mm produces a line-of-sight signal of only 5 to 7 mm, depending on incidence angle. The satellite does not ignore lateral motion; it just sees a projection of it.
Vertical settlement, by contrast, is well-captured. A wall that settles 10 mm produces a line-of-sight signal close to 10 mm on a steep-incidence geometry. This asymmetry matters when interpreting results: a quiet InSAR time series does not guarantee the wall is stationary. It may be rotating or translating in a direction the satellite cannot fully resolve.
Ascending and descending passes: the geometry that makes decomposition possible
Sentinel-1 and COSMO-SkyMed both acquire from ascending (south-to-north) and descending (north-to-south) orbits, looking right in each case. Because the two look directions are roughly opposite in the east-west plane, a target that moves horizontally east-west will appear to move toward the satellite on one pass and away on the other. Combining the two line-of-sight measurements allows decomposition into a vertical component and an east-west horizontal component. North-south motion remains almost invisible to right-looking sensors at mid-latitudes, which is a genuine limitation, not a footnote.
For a retaining wall oriented north-south, the lateral displacement of interest is east-west, and the two-pass decomposition works well. For a wall oriented east-west, the wall's lateral motion is north-south, and InSAR cannot recover it reliably. Site geometry must be assessed before committing to a satellite-only monitoring programme. In practice, many urban excavation walls have mixed orientations, and the analyst must identify which wall segments are observable and which are not.
Persistent-scatterer processing and what sheet piles offer
Standard differential InSAR loses coherence over time as the ground surface changes between acquisitions. Retaining walls and sheet piles are ideal persistent scatterers: the corrugated steel faces produce strong, stable radar returns across dozens of repeat passes. Persistent-scatterer InSAR (PS-InSAR) identifies these stable point targets, models and removes atmospheric phase delay using the spatial and temporal structure of the phase field, and estimates displacement time series at each scatterer.
With Sentinel-1 C-band data, published studies of urban infrastructure report line-of-sight precision of 1 to 2 mm per epoch on good persistent scatterers. COSMO-SkyMed and TerraSAR-X in spotlight mode can reach sub-millimetre precision on strong metallic targets. The archive depth matters: Sentinel-1 data runs from 2014, giving a decade of baseline before any new monitoring programme starts. For a wall installed in 2018, that archive may already contain the construction-phase displacement history.
One practical limit is spatial density. PS-InSAR needs many coherent pixels to solve the atmospheric model. A short, isolated wall in a vegetated or low-density area may yield too few scatterers for reliable processing. Urban and industrial settings with dense hard surfaces are where the method performs best.
The rotation-translation ambiguity and what inclinometers resolve
A retaining wall can fail in two geometrically distinct ways. Pure translation means the entire wall moves laterally as a rigid body. Rotation means the wall pivots, typically bulging outward at mid-height while the top and toe move less. Both modes produce line-of-sight displacement signals, but the spatial pattern differs: rotation produces a displacement gradient along the wall height, translation does not.
InSAR measures displacement at the wall face as seen from above at a fixed incidence angle. It cannot directly observe the wall profile at depth. A satellite time series showing 4 mm of line-of-sight change at the top of a sheet-pile wall is consistent with either 6 mm of lateral translation or a rotation of roughly 0.05 degrees, or some combination. Without inclinometer readings at depth, the two modes cannot be separated from satellite data alone. This is not a deficiency of the method; it is a physical limit of surface observation. The honest use of InSAR for retaining walls is as an early-warning and spatial-coverage tool, not as a replacement for embedded instrumentation.
Practical monitoring workflow for an infrastructure project
A sensible programme combines a pre-construction baseline from the Sentinel-1 archive, active monitoring during excavation using a commercial X-band sensor for higher resolution and flexible revisit, and post-construction settlement tracking on open data. The commercial sensor is tasked on a schedule tied to construction milestones: each major excavation stage, prop installation, and basement slab pour. Alerts trigger when cumulative line-of-sight displacement at any persistent scatterer on the wall face exceeds a project-defined threshold, typically 5 to 10 mm.
Satellize can structure this as a continuous analytics feed, ingesting new acquisitions as they arrive, updating the PS-InSAR time series, and issuing georeferenced displacement maps in GeoTIFF or GeoJSON format. For a project requiring formal reporting, outputs can be formatted to align with geotechnical monitoring standards, though the satellite data sits alongside, not above, the inclinometer and settlement-point record.
Where the method reaches its limits
Several conditions degrade or defeat the approach. Rapid displacement, more than roughly a quarter of the radar wavelength between passes, causes phase unwrapping errors. At C-band (5.6 cm wavelength), that ceiling is about 14 mm per 6-day interval. A wall moving faster than that, which can happen during a prop failure or unexpected groundwater ingress, will produce decorrelated fringes rather than a clean displacement signal. X-band sensors with shorter wavelengths have a lower ceiling still, around 8 mm per pass at 3.1 cm.
Construction activity itself is disruptive. Plant and equipment moving across the site between passes decorrelates the surrounding area, reducing the number of usable persistent scatterers near the wall face. Newly installed walls with fresh paint or galvanising may behave differently as scatterers than weathered steel. And if the wall is obscured from the satellite's look direction by an adjacent building or a tall crane, that segment simply cannot be observed. None of these are reasons to avoid InSAR monitoring; they are reasons to design the monitoring programme with the geometry of the specific site in front of you.
Typical figures
| Spatial resolution (Sentinel-1 IW mode) | 5 x 20 m ground range; PS-InSAR point targets at sub-pixel precision |
| Spatial resolution (COSMO-SkyMed / TerraSAR-X spotlight) | 1 to 3 m; staring-spotlight down to ~0.25 m |
| Revisit (Sentinel-1, dual satellite) | 6 days at mid-latitudes |
| Revisit (COSMO-SkyMed constellation) | 12 hours minimum; typically 1 to 4 days for tasked monitoring |
| Line-of-sight displacement precision (PS-InSAR) | 1 to 2 mm per epoch (C-band); sub-millimetre achievable on strong X-band scatterers |
| Rapid-displacement ceiling (phase continuity) | ~14 mm per pass at C-band; ~8 mm per pass at X-band |
| Horizontal motion sensitivity | East-west component recoverable from ascending/descending combination; north-south component not reliably detectable |
| Archive depth (Sentinel-1) | From April 2014 (Sentinel-1A launch) |
| Delivery formats | GeoTIFF displacement maps, GeoJSON point time series, CSV epoch tables, PDF monitoring reports |
| Latency after acquisition | Sentinel-1: typically 1 to 3 hours to Copernicus data hub; processed output within 24 hours for routine monitoring |
Analytics Satellize can run
| PS-InSAR displacement time series | Persistent-scatterer InSAR (Ferretti et al. method class, published 2001); atmospheric correction using spatial-temporal filtering | GeoJSON point layer with per-scatterer displacement history, updated each acquisition cycle |
| Ascending/descending displacement decomposition | Two-geometry vector decomposition to separate vertical and east-west horizontal components from paired line-of-sight measurements | GeoTIFF rasters of vertical and horizontal displacement, epoch by epoch |
| Wall-segment displacement profile | Spatial aggregation of PS points along wall alignment, interpolated to uniform spacing | Per-wall-segment displacement chart in PDF and CSV, suitable for geotechnical reporting |
| Threshold-breach alert | Automated comparison of cumulative line-of-sight displacement against project-defined thresholds at each persistent scatterer | Email or API alert with scatterer ID, coordinates, displacement value and acquisition date |
| Pre-construction baseline assessment | Retrospective PS-InSAR processing of Sentinel-1 archive back to 2014 for the site footprint | Baseline displacement report identifying any pre-existing movement trends before works commence |
| Coherence change map | Interferometric coherence differencing between pre-construction and active-construction epochs to identify areas of surface disturbance affecting scatterer density | GeoTIFF coherence map flagging wall segments with degraded observability |
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.