Sensors
- Sentinel-1 (C-band SAR, ESA): 5 m x 20 m ground resolution in Interferometric Wide Swath mode; 6-day revisit at mid-latitudes, 3-day with both satellites. C-band (5.405 GHz) backscatter is sensitive to centimetre-scale sea-surface roughness caused by near-surface current anomalies over scoured seabed. All-weather, day-night. Free and open.
- TerraSAR-X / TanDEM-X (X-band SAR, Airbus): Stripmap mode delivers 3 m resolution; Spotlight up to 1 m. X-band is more sensitive to fine surface capillary waves than C-band, improving detection of subtle current-roughness signatures in calm sea states. Tasked commercially; 11-day repeat orbit but rapid-revisit programming available.
- Sentinel-2 (multispectral, ESA): 10 m resolution in visible bands (Blue, Green, Red, NIR). Bathymetric inversion using the Stumpf log-ratio method is viable in waters shallower than roughly 15 m and low turbidity. 5-day revisit with both satellites. Cloud cover is a hard limit; North Sea conditions mean usable acquisitions are infrequent in winter.
- Planet SuperDove: 3 m resolution, 8 spectral bands including coastal blue, which penetrates shallow water better than standard blue. Near-daily revisit. Useful for bathymetric inversion in very shallow nearshore arrays and for cross-validating Sentinel-2 depth estimates. Commercial licence required.
Why scour is a silent structural threat
Scour is the erosion of seabed sediment around a foundation caused by flow acceleration and vortex shedding. Around a monopile, the obstruction forces tidal and wave-driven currents to accelerate and spiral downward in a horseshoe vortex pattern. That vortex lifts and carries away sand and gravel. In soft sandy seabeds typical of the southern North Sea and the Irish Sea, scour holes can reach depths of one to two pile diameters and develop within a single storm event.
The structural consequence is a reduction in lateral soil resistance. Design codes such as DNV-ST-0126 require scour protection and periodic monitoring precisely because undetected scour changes the foundation's effective embedment depth and shifts its natural frequency. A turbine whose natural frequency drifts toward the rotor's excitation frequency faces accelerated fatigue. Physical sonar surveys are the definitive measurement tool, but they are expensive, weather-dependent and typically run annually at best. The gap between surveys is where satellite monitoring earns its place.
What a scoured seabed does to the water surface
The physical link between seabed scour and SAR imagery is indirect but well-established in the oceanographic literature. A scour hole changes the local bathymetry, which modifies the vertical structure of tidal currents above it. Shallower effective water depth over the scour rim accelerates flow; the deeper pit at the centre can create a local deceleration zone. These current gradients produce convergence and divergence of surface water, which in turn modulates the amplitude of short-period surface gravity and capillary waves.
SAR measures normalised radar cross-section (NRCS), which is dominated by Bragg-resonant centimetre-scale waves at C-band and X-band. Areas of current convergence suppress these waves, producing a dark patch in the image. Divergence zones roughen the surface and appear bright. The resulting anomaly pattern around a monopile is typically a dipole or horseshoe shape that correlates with the tidal flow direction at the time of acquisition. This is the same physical mechanism used to detect internal waves, submarine sand dunes and shallow bathymetric features in open-ocean SAR imagery.
The critical caveat: the anomaly is a proxy, not a direct depth measurement. Wind speed matters enormously. Below about 2 m/s, the sea surface is too smooth for reliable Bragg scattering and anomalies disappear. Above roughly 12 to 14 m/s, wave breaking dominates and the scour signal is buried in noise. The detection window is real but narrow, and any operational programme must select acquisitions by sea-state as well as by calendar.
Multispectral bathymetric inversion: shallower waters only
Where wind array foundations sit in water depths below about 15 m and turbidity is low, satellite-derived bathymetry (SDB) offers a more direct measurement. The Stumpf log-ratio method uses the differential attenuation of blue and green light with depth to estimate water depth from Sentinel-2 or SuperDove imagery. Published validation studies report accuracies of 0.5 to 1.5 m root-mean-square error in suitable conditions.
Applied to repeat imagery around a foundation, SDB can in principle detect a deepening scour hole as a change in the retrieved depth field. In practice, turbidity variation between acquisitions introduces errors that are comparable to the scour signal itself, especially after storms when suspended sediment is elevated. SDB for scour monitoring therefore works best as a seasonal or annual cross-check against sonar data rather than as a storm-event early-warning tool. That role belongs to SAR.
Revisit cadence and the storm-event problem
Scour can develop to critical depth in 12 to 48 hours during a major storm. Sentinel-1's 6-day single-satellite repeat (3-day with both) is adequate for trend monitoring but will miss a rapid storm-event scour unless an acquisition happens to fall in the right sea-state window shortly after the event. Adding commercial TerraSAR-X tasking to the programme allows targeted acquisitions in the 24 to 72 hour window after a named storm, when conditions have moderated enough to fall within the Bragg-scattering wind window.
A practical monitoring design uses Sentinel-1 as the continuous baseline, flagging anomalies that exceed a threshold calibrated against the site's sonar survey archive. TerraSAR-X is triggered on a storm-event basis. This keeps commercial tasking costs proportionate while ensuring that the highest-risk periods receive higher-resolution coverage. No satellite programme eliminates the need for sonar surveys; it extends the interval between them and prioritises which foundations to survey first.
Calibration, honest limits and what the data cannot tell you
SAR-based scour detection is an anomaly-detection method, not a bathymetric measurement. Without calibration against at least one sonar survey at the site, it is impossible to convert a backscatter anomaly into a scour depth estimate. The calibration relationship is also site-specific: tidal range, sediment grain size, foundation diameter and local current regime all affect how a given scour geometry projects onto the surface roughness field.
Cloud cover does not affect SAR, which is a genuine advantage over optical methods in the North Sea or the Baltic. But rain cells at C-band introduce false roughness signatures that can mimic scour anomalies. Automated processing pipelines need rain-flag masking. Vessel wakes around active construction or maintenance operations create bright linear features that can obscure foundation-scale anomalies if the geometry is unfortunate.
Satellize runs SAR anomaly detection on Sentinel-1 open-archive data and can add commercial tasking for storm-event response. The Tonga crop-estimation programme demonstrated the team's approach to calibrating satellite-derived proxies against ground-truth data, which is the same discipline that makes scour monitoring credible rather than decorative. Operators considering a new monitoring programme should plan for a sonar calibration campaign in the first season; the satellite data then earns its keep in the years that follow.
From pixels to an inspection priority list
The analytic output that matters to an asset manager is not a map of backscatter anomalies. It is a ranked list of foundations showing which sites have exhibited the largest or most persistent anomalies since the last sonar survey, with an estimate of confidence based on the number of valid acquisitions and the sea-state conditions at each pass.
Feeding that ranked list into an inspection scheduling tool lets operators direct their survey vessels to the highest-risk foundations first, rather than working through an array in geographic sequence. In a large array of 100 or more turbines, even a modest improvement in inspection targeting has a material effect on the cost and timeliness of maintenance. The satellite data does not replace engineering judgement; it informs where that judgement should be applied next.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 5 m range x 20 m azimuth (ground range) |
| SAR spatial resolution (TerraSAR-X Spotlight) | ~1 m, commercially tasked |
| SAR revisit (Sentinel-1, both satellites) | 3 days at mid-latitudes; 6 days single satellite |
| Optical resolution (Sentinel-2 visible bands) | 10 m; Planet SuperDove 3 m |
| Optical revisit | 5 days (Sentinel-2); near-daily (Planet SuperDove, commercial) |
| SAR operating frequency | C-band 5.405 GHz (Sentinel-1); X-band 9.65 GHz (TerraSAR-X) |
| Effective wind-speed detection window for SAR scour anomaly | Approximately 2 to 12 m/s; outside this range signal reliability degrades |
| Bathymetric inversion depth limit (SDB) | Approximately 15 m in low-turbidity conditions; accuracy 0.5 to 1.5 m RMSE |
| Sentinel-1 open archive depth | From 2014; free via Copernicus Data Space |
| Latency (Sentinel-1 NRT) | Typically 1 to 3 hours after acquisition via Copernicus NRT service |
Analytics Satellize can run
| Per-foundation SAR backscatter anomaly score | Normalised radar cross-section time-series differencing against site-specific baseline; Bragg-scattering roughness anomaly detection | Monthly GIS layer (GeoTIFF or GeoJSON) with anomaly magnitude and acquisition metadata per foundation |
| Storm-event scour alert | Triggered TerraSAR-X acquisition within 72 hours of storm threshold exceedance; rapid anomaly scoring against pre-storm baseline | Alert report within 24 hours of image delivery, naming foundations with anomaly score above calibrated threshold |
| Satellite-derived bathymetry change map | Stumpf log-ratio inversion on Sentinel-2 or SuperDove blue/green bands; multi-date depth differencing | Seasonal depth-change raster clipped to array footprint, with turbidity-flag mask and confidence layer |
| Inspection priority ranking | Composite scoring of anomaly persistence, magnitude and recency across all valid SAR acquisitions since last sonar survey | Ranked foundation list in CSV or PDF, suitable for direct input to survey vessel scheduling |
| Calibration support dataset | Co-registration of sonar survey depth measurements with contemporaneous SAR anomaly scores to derive site-specific transfer function | Calibration report with transfer function parameters and uncertainty bounds; updated after each sonar campaign |
| Long-term scour trend report | Annual time-series analysis of Sentinel-1 archive anomaly scores per foundation; trend fitting and outlier identification | Annual PDF report with per-foundation trend charts and array-wide heatmap |
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.