Offshore export cable route seabed characterisation from satellite-derived bathymetry
Multispectral satellite imagery can resolve shallow-water bathymetry and broad seabed sediment classes down to roughly 25 m in clear water, giving cable-route planners a low-cost corridor filter before committing to geophysical survey.
Sensors
- Sentinel-2 MSI: 10 m resolution in four visible and near-infrared bands; 5-day revisit at mid-latitudes with two satellites. Blue and green bands (Band 2 at 490 nm, Band 3 at 560 nm) penetrate optically shallow water to roughly 15–20 m in moderately clear seas. Free and openly archived from 2015.
- WorldView-3: Eight multispectral bands at 1.24 m resolution, including a coastal-blue band at 427 nm that improves depth retrieval in cleaner water. Tasked commercially; revisit depends on latitude and tasking priority. The finer spatial resolution resolves individual coral heads and small sediment-type boundaries that Sentinel-2 blurs.
- Planet SuperDove: Eight bands at 3–4 m resolution with near-daily revisit. Useful for tracking turbidity variability across multiple tidal states, which is essential for separating water-column signal from bottom reflectance in radiative transfer inversion.
- ICESat-2 ATL24: NASA's photon-counting lidar measures water-surface and seabed returns along ground tracks spaced roughly 3 km apart at the equator. Depth accuracy is around 0.1–0.3 m in clear water to a maximum of about 40 m. Track spacing is too coarse for engineering survey but provides excellent calibration tie-points for passive-image inversion models.
Why light is a depth gauge
Sunlight entering water is selectively absorbed. Red wavelengths are gone within a few metres. Blue light travels furthest, scattering back upward from the seabed in optically shallow water before the sensor sees it. The ratio of blue to green reflectance carries a bathymetric signal that can be extracted by inverting a radiative transfer model or by fitting an empirical log-ratio against known depth points.
The physics was formalised by Lyzenga in the 1970s and refined by Lee et al. in the Hydrolight-based analytical model now widely cited as the basis for operational satellite-derived bathymetry (SDB). The core equation relates water-leaving radiance to bottom depth, bottom albedo and diffuse attenuation coefficient. Solving it requires assumptions about water clarity, which is where the honest limits begin.
The 25-metre ceiling and what turbidity does to it
In the clearest tropical waters, such as those around Pacific island groups, blue-band penetration reaches approximately 25 m. That ceiling drops sharply with suspended sediment or phytoplankton. In the southern North Sea, where suspended particulate matter routinely exceeds 10 mg per litre, useful depth retrieval may be limited to 8–12 m. In highly turbid estuaries it can fall to 3–5 m or vanish entirely.
This is not a deficiency that better processing can overcome. It is physics. A cable-route planner working in the German Bight or the Firth of Forth must treat SDB as a tool for the shallowest nearshore segment and for identifying where turbidity itself varies seasonally, not as a substitute for multibeam echosounder across the full route corridor.
Tidal state matters too. Images acquired at low water expose more of the intertidal zone and reduce the water-column path length in shallow areas, improving depth retrieval. Multi-date compositing across different tidal phases, combined with a tidal model, is now standard practice for SDB in macrotidal environments.
Sediment proxies: what colour tells you about what is down there
Bottom reflectance, once the water column is removed from the signal, varies with substrate. Carbonate sand reflects strongly across visible wavelengths. Seagrass and macroalgae absorb red and blue light differently from bare sediment, producing distinguishable spectral shapes. Dark patches in shallow water often indicate rocky reef, dense biogenic material or cohesive mud, all of which matter for cable burial feasibility.
WorldView-3's eight-band multispectral suite allows finer substrate discrimination than Sentinel-2's four visible bands. Published studies using WorldView imagery have separated five to seven substrate classes in clear-water environments. Sentinel-2 typically resolves three to four broad classes: carbonate sand, darker sediment or algae, rock, and seagrass. Neither sensor provides grain-size data. Distinguishing fine sand from coarse gravel from bedrock requires ground-truth from even a sparse set of grab samples or video drop-frames.
The practical output for a cable-route team is a corridor-scale substrate classification map with confidence zones, not a burial-assessment report. That distinction matters when presenting the product to a consenting authority.
ICESat-2 as the calibration anchor
Passive multispectral SDB needs depth control points to constrain the inversion. Chart data from national hydrographic offices is the traditional source, but coverage is often patchy or dated in the shallow nearshore zones where cables cross from offshore to landfall. ICESat-2's ATL24 product, released by NASA in 2023, provides photon-counting lidar depth measurements at roughly 0.7 m along-track spacing in water clear enough for the 532 nm green laser to reach the bottom.
The track spacing of approximately 3 km means ICESat-2 cannot substitute for a full multibeam survey. But where tracks cross a proposed corridor, they offer calibration tie-points with sub-metre vertical accuracy that tighten the passive inversion considerably. Combining ICESat-2 control with Sentinel-2 spatial coverage is now an established workflow in the published literature on global shallow-water bathymetry.
What satellite-derived bathymetry cannot do
It cannot see through turbid water. It cannot resolve sub-metre seabed features. It cannot determine sediment bearing strength, shear resistance or burial depth for cable protection. It gives no information on boulders smaller than the pixel footprint. It does not detect unexploded ordnance, pipelines or existing cable crossings below the seabed surface.
For engineering-grade cable burial assessment, a geotechnical programme combining multibeam sonar, sub-bottom profiler and sediment coring remains mandatory. The value of satellite-derived characterisation is in corridor screening: narrowing a 10 km-wide candidate corridor to two or three 1–2 km strips that are worth surveying in detail, and flagging the segments where turbidity or depth will make even that screening unreliable.
Satellize applies this workflow operationally, running Sentinel-2 and WorldView-3 inversion pipelines alongside ICESat-2 calibration for preliminary route assessments. The Tonga crop-estimation programme demonstrated the organisation's approach to combining open and commercial satellite data for sovereign clients in island environments, a context directly analogous to nearshore cable-route work in Pacific and Caribbean jurisdictions.
Fitting SDB into the consenting timeline
Offshore cable consenting in most jurisdictions requires a Marine Environmental Impact Assessment that includes a seabed description. Satellite-derived bathymetry and substrate maps, clearly labelled with their depth and turbidity limits, can populate the preliminary seabed characterisation chapter and demonstrate that the applicant has considered alternative routes before selecting one for detailed survey. Regulators in several North Sea jurisdictions have accepted SDB products as supporting evidence in scoping submissions, provided the uncertainty bounds are explicit.
The economic case is straightforward. A geophysical survey vessel day costs roughly £30,000–£80,000 depending on vessel class and mobilisation distance. Satellite-derived corridor screening that eliminates two of five candidate routes before the vessel is hired reduces survey scope materially. The satellite analysis is not free, but it is fast: a Sentinel-2 archive analysis covering a 50 km corridor can be completed in days rather than the weeks a survey mobilisation requires.
Typical figures
| Spatial resolution (passive multispectral) | 1.24 m (WorldView-3), 3–4 m (Planet SuperDove), 10 m (Sentinel-2 visible bands) |
| Depth retrieval limit (clear tropical water) | Up to ~25 m (passive); up to ~40 m (ICESat-2 lidar) |
| Depth retrieval limit (turbid northern seas) | Typically 8–12 m; may fall to 3–5 m in high-sediment environments |
| Depth accuracy (SDB passive, calibrated) | ±0.5–1.5 m RMSE in clear water; degrades with turbidity |
| ICESat-2 along-track spacing | ~0.7 m; cross-track spacing ~3 km at equator |
| Sentinel-2 revisit | 5 days (two-satellite constellation) at mid-latitudes |
| Spectral bands used | Coastal blue (~427–490 nm), green (~560 nm), red (~665 nm), NIR for water masking |
| Sentinel-2 archive depth | From 2015 (Sentinel-2A launch); Sentinel-2B from 2017 |
| Substrate classes distinguishable | 3–4 broad classes (Sentinel-2); 5–7 in clear water (WorldView-3) |
| Minimum resolvable seabed feature | ~3× pixel size; sub-pixel features not detected |
Analytics Satellize can run
| Satellite-derived bathymetry grid | Radiative transfer inversion (Lee et al. analytical model) or log-ratio empirical method (Lyzenga), calibrated against ICESat-2 ATL24 or chart control points | GeoTIFF depth grid with per-pixel uncertainty layer; GIS-ready, EPSG-specified |
| Seabed substrate classification map | Bottom-reflectance retrieval after water-column correction; supervised classification against sparse ground-truth (grab samples, video, existing survey data) | Vector polygon layer of substrate classes with confidence ratings; PDF summary for consenting submissions |
| Turbidity variability assessment | Multi-date Sentinel-2 or SuperDove time series; suspended particulate matter proxy from red and NIR band ratios; tidal-phase stratification using published tidal model | Seasonal turbidity climatology map showing depth-retrieval reliability zones; tabular summary by corridor segment |
| Corridor comparison scorecard | Overlay of bathymetry, substrate class, turbidity reliability and depth-limit flags across candidate route corridors; ranked by survey-priority score | Single-page corridor comparison table and annotated map for engineering and consenting teams |
| ICESat-2 track extraction and calibration report | NASA ATL24 product query and filtering for cloud-free, low-turbidity passes over the study area; quality-flagged depth points exported as calibration dataset | Calibration point shapefile with depth, uncertainty and acquisition metadata; integration note for passive SDB workflow |
| Archive change detection for seabed stability | Multi-year Sentinel-2 bottom-reflectance time series to identify migrating sandbanks or shifting sediment boundaries in the shallow corridor | Change-detection map with flagged unstable zones; relevant for cable burial depth planning |
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.