Post-cyclone wave damage assessment for offshore aquaculture infrastructure
Tropical cyclones can destroy marine cage arrays and longline systems before any vessel reaches the site. SAR coherence change detection and very-high-resolution optical imagery together provide the earliest credible evidence of structural loss, displaced equipment and debris fields.
Sensors
- Sentinel-1 SAR (IW mode, C-band): 5 x 20 m ground resolution in Interferometric Wide Swath mode; 6-day repeat at mid-latitudes, 12-day at the equator with a single satellite. Penetrates cloud and rain bands immediately after landfall. Coherence change detection between pre- and post-event acquisitions flags surfaces that have physically moved or changed dielectric properties.
- Maxar WorldView-3: 0.31 m panchromatic, 1.24 m multispectral. Can resolve individual cage collar sections, mooring lines and debris items when cloud clears. Tasking latency after a request is typically 1-3 days depending on orbital geometry and residual cloud.
- Planet SkySat: 0.5 m resolution, up to 12 tasked passes per day over a target. Useful for rapid revisit once the cloud deck breaks, enabling near-daily damage progression monitoring in the days after a storm.
- Capella Space X-band SAR: Spotlight mode down to approximately 0.35 m resolution. X-band backscatter is more sensitive to small metallic and plastic structures than C-band, improving detection of individual cage components and mooring hardware. Commercial tasking with sub-24-hour latency in many tropical regions.
What coherence actually measures, and why it matters here
SAR coherence quantifies how similar the phase relationship between two radar acquisitions is over the same patch of ground. When a surface is undisturbed, coherence is high. When structures move, collapse or are replaced by open water, coherence drops sharply. For offshore aquaculture, this is useful because cage collars, net pen frames and longline floats are relatively stable radar scatterers in calm conditions, so a pre-storm baseline acquisition gives a clean coherence reference.
After a cyclone, areas where cages have been displaced or submerged show low coherence, while intact infrastructure retains it. The technique does not require sunlight or clear skies, which matters enormously in the 24-72 hours after landfall when cloud cover is near-total. Sentinel-1's C-band (5.405 GHz) penetrates typical post-storm cloud without attenuation.
The residual swell problem: where interpretation gets difficult
The honest complication is this: a rough sea surface also decorrelates SAR phase. Residual swell and whitecapping after a cyclone can produce low-coherence patches that look, in the data, like displaced infrastructure. Separating genuine structural loss from surface-roughness decorrelation requires several checks.
First, the spatial pattern matters. Swell-driven decorrelation is spatially continuous and aligned with wave propagation direction. Cage damage produces compact, discrete low-coherence patches at the known coordinates of farm structures. Second, backscatter intensity change is an independent signal: a submerged or absent cage produces a different backscatter signature than a rough water surface. Third, where archive imagery exists, the pre-storm backscatter footprint of the installation can be used as a mask to constrain where damage detections are physically plausible.
In practice, the first SAR pass after a storm is best treated as a triage layer, identifying candidate damage sites for follow-up with very-high-resolution optical once cloud breaks. Treating it as definitive without optical confirmation overstates certainty.
Very-high-resolution optical: what 0.3 metres actually resolves
At WorldView-3's 0.31 m panchromatic resolution, individual cage collar sections (typically 10-50 m diameter for offshore salmon or tuna operations) are clearly resolved. Analysts can distinguish a cage that has capsized but remains moored from one that has broken free and drifted. Mooring buoys, feed pipes and net pen walkways are visible as distinct objects. Debris fields, including fragments of HDPE pipe, net material and loose floats, appear as irregular bright or dark objects against the sea surface.
SkySat's value is in revisit frequency rather than resolution. At 0.5 m, it cannot match WorldView-3 for fine structural detail, but the ability to task multiple passes per day means analysts can track debris drift, confirm that a cage previously identified as displaced has not been recovered, and monitor whether secondary hazards such as net entanglement zones are expanding. For insurance or government loss-assessment purposes, a time series of SkySat imagery in the 72 hours after cloud clearance is often more useful than a single WorldView-3 scene.
Building the pre-event baseline before cyclone season opens
The quality of post-event analysis depends almost entirely on what was collected before the storm. A farm with no satellite baseline is a farm where analysts are working from first principles, estimating what was there from licensing records or operator-supplied diagrams. That introduces ambiguity that is hard to resolve in a loss-adjustment context.
A pre-season baseline should include at minimum one cloud-free very-high-resolution optical acquisition showing cage positions and mooring layout, and two or three Sentinel-1 acquisitions under calm conditions to establish a coherence reference and a typical backscatter fingerprint for the installation. Sentinel-1 data is freely available through the Copernicus Data Space, so the cost of building this archive is primarily analytical rather than data-acquisition. Capella Space X-band imagery, if included in the baseline, adds sensitivity to smaller structural elements that C-band may not resolve clearly.
From data to a loss estimate: what satellite analysis can and cannot deliver
Satellite analysis can reliably establish which structures were present before the storm, which are absent, displaced or visibly damaged afterwards, and the approximate extent of debris fields. It can do this within 24-48 hours of a Sentinel-1 acquisition, and within hours of cloud clearance for optical sensors.
What it cannot do is establish the condition of submerged nets, the viability of surviving stock, or the mechanical integrity of mooring systems that appear visually intact. A cage that looks undamaged from 500 km altitude may have a compromised anchor chain. Satellite assessment is therefore a first-order triage and documentation tool, not a replacement for in-water survey. The most defensible loss-assessment workflow combines satellite-derived damage mapping with targeted vessel survey, using the satellite layer to prioritise where vessels go first and to provide independent documentary evidence for insurers or government compensation schemes.
Satellize has applied similar SAR and optical change-detection methods in Pacific island contexts, including work adjacent to the Kingdom of Tonga crop-estimation programme, where post-cyclone agricultural loss mapping raised analogous questions about separating storm damage from background variability in remotely sensed signals.
Activation under the Copernicus Emergency Management Service
For government clients, it is worth knowing that the Copernicus Emergency Management Service (CEMS) can be activated by eligible national authorities after declared disasters, providing rapid mapping products at no direct cost. CEMS grading maps have been used for coastal infrastructure damage after cyclone events in the Pacific and Indian Ocean regions. The limitation is that CEMS products are general-purpose and may not be georeferenced to aquaculture licence boundaries or individual farm structures. A purpose-built analysis, using the farm operator's own site coordinates and a pre-event baseline, will almost always produce more actionable outputs for a specific aquaculture operator or insurer.
The International Charter on Space and Major Disasters offers a parallel activation route for member agencies, again at no direct cost to the activating authority. Both mechanisms are worth understanding before cyclone season, not after.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 5 x 20 m (range x azimuth) |
| SAR spatial resolution (Capella Space X-band spotlight) | Approximately 0.35 m (published specification) |
| Optical resolution (WorldView-3) | 0.31 m panchromatic, 1.24 m multispectral |
| Optical resolution (Planet SkySat) | 0.5 m |
| Sentinel-1 revisit (equatorial) | 12 days single satellite; 6 days with two-satellite constellation |
| First SAR acquisition after landfall | Typically within 12-24 hours of storm passage, cloud-independent |
| Minimum detectable displaced structure (SAR coherence) | Cage arrays of approximately 20 m diameter or larger; individual mooring buoys not reliably resolved at C-band |
| Minimum detectable debris item (VHR optical) | Objects approximately 0.5-1 m across in calm post-storm sea conditions at WorldView-3 resolution |
| Sentinel-1 archive depth | From October 2014 (Sentinel-1A launch); freely accessible via Copernicus Data Space |
| Delivery formats | GeoTIFF damage maps, GeoJSON structure-status layers, PDF loss-assessment report, KMZ for field navigation |
Analytics Satellize can run
| SAR coherence change map | Interferometric coherence differencing between pre- and post-event Sentinel-1 IW acquisitions; spatial masking to licensed farm footprints | GeoTIFF and GeoJSON layer showing coherence loss by structure, delivered within 24-48 hours of post-event SAR acquisition |
| Backscatter anomaly map | Sigma-nought intensity comparison between pre-storm baseline stack and post-event acquisition; z-score thresholding against calm-water reference | GeoTIFF overlay distinguishing absent structures from rough-surface decorrelation artefacts |
| Structure-status classification | Object-based image analysis on VHR optical (WorldView-3 or SkySat) fused with SAR coherence layer; per-cage status assigned as intact, displaced, partially submerged or absent | GeoJSON feature layer with per-structure status attributes, suitable for import into farm management GIS |
| Debris field extent and drift tracking | Multi-date SkySat time series; change detection on floating object signatures combined with modelled surface current vectors from Copernicus Marine Service | Animated GeoTIFF time series and debris drift forecast polygon, updated daily while cloud permits |
| Pre-season baseline package | Calm-condition Sentinel-1 coherence stack (minimum three acquisitions) plus one cloud-free VHR optical scene; georeferenced to operator licence boundary | Archived GeoTIFF baseline set with metadata, held for rapid differencing against any future post-event acquisition |
| Insurance-grade loss documentation report | Structured comparison of pre- and post-event imagery with annotated change evidence; narrative interpretation of ambiguous detections with stated confidence levels | PDF report with georeferenced evidence plates, per-structure loss table and explicit statement of analytical limitations for use in insurance or government compensation processes |
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.