Pipeline integrity assessment via ground-movement InSAR

Interferometric SAR detects millimetre-scale surface deformation along pipeline corridors weeks or months before a structural failure becomes visible. This page explains which sensors, methods, and detection limits apply.

Sensors

Sentinel-1 (C-band SAR, ESA/Copernicus): 5.6 cm wavelength; Interferometric Wide Swath mode at 5 x 20 m resolution, 250 km swath. Repeat pass every 6 days over most landmasses with both satellites operational. Free archive from 2014. Coherence degrades over vegetation and loose sand; best performance on bare or sparsely vegetated soil.
ICEYE X-band SAR constellation: 9.65 GHz, spotlight mode to approximately 0.25 m resolution; strip mode at 3 m. Revisit configurable to sub-daily over a point target when multiple satellites are tasked. X-band is more sensitive to small deformation signals but loses coherence faster than C-band over time intervals exceeding a few weeks in wet climates.
Capella Space X-band SAR: Spotlight mode at approximately 0.35 m resolution; sliding spotlight at 0.5 m. On-demand tasking with typical revisit of 1 to 3 days over priority crossings. Useful for high-resolution deformation mapping at river crossings, fault zones, or active construction corridors.
COSMO-SkyMed (X-band SAR, ASI): Spotlight mode at approximately 1 m resolution, stripmap at 3 m. Two-satellite constellation with a follow-on second generation pair; repeat pass nominally 16 days but shorter with constellation coordination. Long archive from 2007 enables historical deformation trend analysis.

What InSAR actually measures, and what it does not

Interferometric synthetic aperture radar compares the phase of radar returns from two passes over the same ground. Where the surface has moved between passes, the phase shifts. Processing that shift yields a line-of-sight displacement map with precision in the millimetre range under good conditions. The technique does not image the pipe itself; it images the soil column above and around it. A pipe under stress deforms the ground before it fails, which is precisely why InSAR is useful.

The honest limits matter here. Coherence, the statistical similarity between two SAR images, collapses over dense vegetation, flowing water, and freshly tilled fields. In those conditions InSAR produces noise rather than signal. Atmospheric water vapour introduces apparent deformation artefacts of several centimetres if uncorrected; corrections using ERA5 reanalysis data or GACOS tropospheric models reduce this but do not eliminate it. Line-of-sight geometry means purely horizontal east-west movement is poorly resolved by ascending or descending passes alone; combining both geometries partially addresses this. Minimum detectable displacement in a single interferogram is roughly 5 to 10 mm; time-series methods such as Persistent Scatterer InSAR (PS-InSAR) or Small Baseline Subset (SBAS) push this toward 1 to 2 mm per year over stable, coherent ground.

The failure modes InSAR catches early

Subsidence is the most straightforward signal. Ground settling over a poorly compacted backfill trench, a dissolving evaporite formation, or a dewatered aquifer produces a bowl-shaped deformation pattern centred on the pipeline corridor. Rates above roughly 10 mm per year over a buried steel pipe warrant engineering review; rates above 50 mm per year are considered critical in most pipeline integrity management standards.

Permafrost thaw is a growing concern for Arctic and sub-Arctic pipelines. As the active layer deepens, differential settlement between thawed and frozen ground imposes bending loads on the pipe. Sentinel-1 time-series studies over Siberian and Alaskan terrain have documented seasonal heave and subsidence cycles of 5 to 15 cm annually, with a net downward trend in degrading permafrost zones. Detecting that trend early allows operators to schedule targeted geotechnical surveys rather than blanket inspections.

Lateral displacement at river crossings and slope instability zones is the third category. Slow-moving landslides, which may creep at only a few centimetres per year, are detectable in SBAS time series long before they accelerate to failure. X-band commercial SAR is particularly useful here because shorter wavelengths are more sensitive to small displacements and higher spatial resolution separates the active slide mass from stable ground immediately adjacent.

Building a time series: archive depth and cadence choices

A single interferogram answers the question of whether the ground moved between two specific dates. A time series answers whether movement is accelerating, seasonal, or correlated with external drivers such as rainfall or extraction activity nearby. The Sentinel-1 archive running from 2014 provides a decade of baseline data over most pipeline-bearing regions. That depth is genuinely valuable: it lets analysts distinguish a new anomaly from a pre-existing, stable deformation pattern that predates the pipeline.

For routine corridor monitoring, Sentinel-1 at 6-day revisit is usually sufficient. For critical crossings, fault intersections, or sites already showing anomalous movement, commercial X-band tasking at 1 to 3-day revisit provides earlier warning of acceleration. The practical workflow combines both: Sentinel-1 for the full corridor at low cost, commercial SAR for a small number of flagged locations at higher cadence and resolution. Latency from image acquisition to processed displacement map is typically 24 to 72 hours for near-real-time services, or 1 to 2 weeks for a full PS-InSAR stack reprocessing.

Third-party encroachment: a signal most operators overlook

Ground movement InSAR is primarily discussed in the context of natural hazards, but it is equally sensitive to human activity. Excavation adjacent to a pipeline right-of-way causes localised subsidence and lateral displacement that appears in interferograms within one or two repeat cycles. Dewatering for nearby construction, tunnelling beneath a corridor, and heavy surface loading from stockpiles all produce detectable signatures.

SAR coherence change, a related but distinct technique, flags surface disturbance directly: freshly disturbed ground loses coherence relative to the previous image, producing a bright anomaly in a coherence difference map. Combining coherence change with deformation time series gives operators both an early warning of activity and a quantitative measure of whether that activity is moving the ground toward the pipe. This is covered separately in the pipeline construction progress monitoring page in this library.

Honest assessment of where the method struggles

Tropical and equatorial pipeline routes present the hardest conditions. Dense rainforest canopy destroys coherence almost completely at C-band over intervals longer than a few days. X-band is worse in this respect, not better. L-band SAR, from ALOS-2 (JAXA) or the forthcoming NISAR mission, penetrates vegetation more effectively and maintains coherence over longer intervals, making it the preferred choice for forested corridors. NISAR is scheduled for launch in 2024 and will provide free L-band and S-band data globally.

Subsea pipeline sections are outside InSAR's reach entirely. Offshore riser bases and nearshore shallow-water sections can sometimes be monitored where the seabed is exposed at low tide, but this is an edge case. Operators should treat InSAR as a tool for the onshore and near-shore terrestrial sections of their networks, and use acoustic or ROV-based methods for the rest.

Satellize runs PS-InSAR and SBAS workflows on Sentinel-1 open data and coordinates commercial tasking for flagged sites. Its Overhead column has published worked examples of corridor deformation analysis using publicly available data.

Turning displacement maps into actionable alerts

Raw deformation maps are not operationally useful without a threshold framework tied to the specific pipe material, diameter, burial depth, and soil type. A 20 mm displacement along a 48-inch, high-pressure gas transmission line in stiff clay carries different risk implications than the same displacement along a 6-inch, low-pressure distribution main in soft alluvium. The analytics layer should translate displacement rates into strain estimates using published beam-on-elastic-foundation models, then compare those estimates against the operator's integrity management thresholds.

Deliverables that work in practice are GIS layers showing displacement velocity in mm per year, colour-coded against operator-defined risk bands, with point alerts triggered when any segment exceeds a configurable threshold. Integrating those alerts with the operator's existing SCADA or GIS platform, rather than requiring a separate interface, is what determines whether the data actually drives inspection decisions. The next step for any operator evaluating this approach is a corridor baseline assessment: a PS-InSAR stack over the full route using the Sentinel-1 archive, delivered as a GIS layer with annotated anomaly locations for engineering review.

Typical figures

Spatial resolution (Sentinel-1 IW mode)	5 x 20 m (range x azimuth); multi-looked to ~14 x 14 m in standard processing
Spatial resolution (ICEYE / Capella spotlight)	0.25 to 0.5 m; practical InSAR processing typically aggregates to 1 to 3 m
Repeat pass / revisit (Sentinel-1)	6 days (both satellites); 12 days (single satellite)
Repeat pass / revisit (commercial X-band)	1 to 3 days for tasked priority sites; sub-daily possible with multiple satellites
Minimum detectable displacement (single interferogram)	~5 to 10 mm line-of-sight under good coherence conditions
Minimum detectable displacement rate (PS-InSAR time series)	~1 to 2 mm/year over coherent, stable ground with sufficient stack depth
Radar frequency / wavelength	C-band (5.6 cm, Sentinel-1); X-band (3.1 cm, ICEYE, Capella, COSMO-SkyMed); L-band (23.6 cm, ALOS-2)
Archive depth (Sentinel-1)	From April 2014 (Sentinel-1A launch); continuous systematic acquisition over most land
Processing latency	24 to 72 hours for near-real-time single interferogram; 1 to 2 weeks for full PS-InSAR stack
Typical deliverable formats	GeoTIFF displacement velocity maps, GeoPackage / Shapefile anomaly layers, time-series CSV per point, alert feeds via API or email

Analytics Satellize can run

Corridor baseline deformation map	SBAS InSAR time series on Sentinel-1 archive; atmospheric correction via ERA5 or GACOS	GeoTIFF velocity map (mm/year) and GIS layer with annotated anomaly polygons, delivered as a one-off report
Persistent Scatterer displacement time series	PS-InSAR on Sentinel-1 or commercial X-band stack; scatterer density dependent on surface type	CSV time series per persistent scatterer point, with displacement plotted against rainfall or extraction event data where relevant
Critical-crossing high-cadence monitoring	Repeat commercial X-band tasking (ICEYE or Capella) at 1 to 3-day intervals; single-pair interferograms processed on acquisition	Near-real-time displacement alert (email or API) when threshold exceeded at a named crossing
Permafrost thaw trend assessment	Multi-year SBAS stack separating seasonal heave/subsidence cycle from net downward trend; comparison against ERA5 soil temperature data	Annual trend report with mapped thaw-settlement zones colour-coded by severity band
Third-party encroachment flag	SAR coherence change detection combined with deformation anomaly screening; two-pass coherence differencing on Sentinel-1 6-day pairs	Weekly coherence-change GIS layer with flagged disturbance polygons within a configurable buffer around the pipeline centreline
Strain proxy estimation	Displacement rate converted to longitudinal strain using published beam-on-elastic-foundation approximations; output compared against operator-supplied integrity thresholds	Risk-banded segment table (green / amber / red) integrated into operator GIS or delivered as a structured report for engineering review

Sources

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.