Vessel Anchoring Detection in Marine Protected Areas
SAR and very-high-resolution optical imagery can identify vessels anchoring illegally inside MPAs and detect the linear scars anchor chains leave on shallow reef surfaces, giving enforcement agencies evidence that persists long after the vessel has left.
Sensors
- Sentinel-1 SAR (C-band, ESA): 10 m ground range resolution in Interferometric Wide Swath mode, 250 km swath, 6-day exact repeat at the equator (1-3 days with both satellites). Detects stationary vessels as bright point targets regardless of cloud or darkness. Minimum detectable vessel length is approximately 30 m under calm sea conditions; smaller vessels become ambiguous as sea clutter rises.
- Maxar WorldView-3 (optical/SWIR): 31 cm panchromatic, 1.24 m multispectral resolution. In clear, shallow water (typically less than 15 m depth) the panchromatic band resolves anchor chain scars as distinct pale linear features against darker reef substrate. Tasked on demand; revisit over a specific point is 1-4.5 days depending on latitude and cloud probability.
- Planet SuperDove (optical): 3 m multispectral resolution, daily revisit globally. Sufficient to confirm vessel presence and approximate size, and to monitor reef zones for turbidity or bleaching associated with physical disturbance, but not fine enough to resolve individual chain scars. Useful for rapid daily screening before committing to a commercial task.
- Spire spaceborne AIS: Spire operates more than 100 satellites carrying AIS receivers, achieving sub-30-minute global revisit for AIS message collection in most ocean regions. Cross-referencing AIS absence against a SAR detection inside an MPA boundary is the primary method for identifying non-cooperative or deliberately silent vessels.
What the physics gives you for free
A vessel at anchor is, by definition, stationary relative to the seabed. In SAR imagery that stationarity is a gift: the vessel appears as a coherent bright point target with a characteristic radar cross-section, while the surrounding water surface scatters energy away from the sensor. Ship detection in Sentinel-1 IW mode using constant false alarm rate (CFAR) algorithms is well established in the published literature and operationally used by agencies including the European Maritime Safety Agency. The detection floor sits around 30 m vessel length under moderate sea states; a Force 4 sea can push that threshold higher as wave clutter competes with smaller targets.
The anchor chain itself is not visible in SAR. That is where WorldView-3 takes over. In water depths shallower than roughly 10 to 15 m, and with sufficient water clarity, the 31 cm panchromatic band resolves the pale linear scour paths that dragging chain leaves on reef substrate. These scars are spectrally distinct from healthy coral and from sand: coral skeleton exposed by chain abrasion reflects more strongly in the blue-green bands, and the linear geometry is not replicated by natural reef features. The scar persists after the vessel departs, which means imagery collected days or weeks later can still constitute evidence.
Combining fleet screening with boundary polygons
The workflow has two stages. First, Sentinel-1 detections are spatially intersected with MPA boundary polygons published by authorities such as UNEP-WCMC or national marine park agencies. Any detection inside a prohibited anchoring zone that lacks a corresponding AIS transmission is flagged as a priority. Spire AIS data closes the identification loop: if a vessel is broadcasting, its MMSI and flag state are recorded; if it is not, the SAR detection alone is passed to enforcement as a probable dark-vessel event.
Second, flagged locations are queued for tasking of WorldView-3 or, where budget allows, Planet SuperDove for same-day confirmation of vessel presence. If the vessel has already departed, the WorldView-3 image is still requested to document reef-surface condition. The time gap between the SAR detection and the optical collect is the main operational constraint: a fast vessel can exit an MPA in under two hours, so the SAR-to-task latency matters. Sentinel-1 Near Real-Time products are available within one to three hours of acquisition through the Copernicus Data Space, which is a workable window for alert generation.
Reading anchor chain scars: what they show and what they do not
A chain scar in WorldView-3 imagery typically appears as a pale, sinuous or arc-shaped line, 1 to 4 m wide, radiating from a central disturbed patch where the anchor itself contacted the reef. The arc shape reflects the vessel swinging on its rode as wind and current shifted. Multiple overlapping arcs suggest repeated anchoring at the same location, which is useful for establishing a pattern rather than an isolated incident.
The honest limits here matter. Water clarity is the dominant constraint: turbidity from storm runoff, algal bloom or suspended sediment can obscure the bottom entirely. Depths beyond 15 m are generally opaque to optical sensors even in the clearest tropical water. Scars on sand or rubble substrates are harder to distinguish from natural bedform features than scars on live coral. And a scar cannot, by itself, identify which vessel caused it; that inference requires corroborating SAR or AIS data from the same time window. Prosecutors and enforcement officers should treat the imagery as strong supporting evidence, not a standalone conviction.
Revisit, cloud and the tropical problem
Many of the world's most sensitive coral reef MPAs sit in the tropics, where cloud cover exceeds 70 percent of days in wet season. SAR is unaffected by cloud, which makes Sentinel-1 the primary screening layer. Optical confirmation is the bottleneck. Planet's daily revisit improves the odds of a cloud-free collect, but 3 m resolution cannot resolve chain scars. WorldView-3 tasking at 31 cm is cloud-dependent, and in persistently overcast conditions a usable collect may take a week or more.
Sentinel-1's 6-day exact repeat (3-day with both satellites active) means a vessel that anchors briefly between passes may not be captured at all. Supplementing with commercial SAR constellations such as Capella Space or ICEYE can reduce revisit to hours, though those data carry separate licensing costs. For high-priority reef zones, a standing tasking order across multiple SAR operators is the only way to approach near-continuous coverage.
Turning detections into enforcement records
An enforcement record needs more than a single image. The minimum useful package is: a georeferenced SAR detection with timestamp and estimated vessel dimensions; any associated AIS record or confirmed absence of AIS; a WorldView-3 or equivalent optical image showing the vessel or the post-departure scar; and a spatial intersection report confirming the coordinates fall inside the protected zone boundary with a documented source for that boundary.
Satellize structures this output as a GIS-ready evidence pack, with all layers in a projection suitable for the relevant jurisdiction's court or administrative process. The Tonga crop-estimation programme gave us direct experience of working within Pacific island government data standards, and the same discipline around provenance and coordinate reference systems applies here. Analysts annotate each detection with the uncertainty radius of the SAR centroid (typically 10 to 20 m for Sentinel-1 IW) so that any positional claim is bounded, not absolute.
What this cannot do on its own
Satellite detection identifies that a vessel was stationary inside a boundary. It does not prove intent, and it cannot confirm whether the vessel's anchor actually contacted the reef rather than a sandy patch nearby. Enforcement agencies should pair satellite evidence with patrol vessel or drone inspection where possible. The satellite layer is most powerful as a prioritisation tool: it tells you where to send the patrol boat, and it documents what the scene looked like before the patrol arrived.
Coverage of very small vessels, kayaks, dive boats and dinghies is beyond the capability of any current spaceborne SAR. These craft cause real reef damage but remain invisible to this method. Drone-based photogrammetry or in-water survey remains the only reliable tool at that scale.
Typical figures
| SAR spatial resolution (Sentinel-1 IW) | 10 m range × 10 m azimuth (after multi-look processing) |
| Optical resolution for scar detection (WorldView-3) | 31 cm panchromatic; 1.24 m multispectral |
| Fleet screening revisit (Sentinel-1, dual satellite) | 1–3 days at mid-latitudes; 6 days exact repeat at equator |
| Optical revisit (WorldView-3, tasked) | 1–4.5 days depending on latitude and off-nadir capacity |
| SAR alert latency (Copernicus NRT) | 1–3 hours post-acquisition |
| Minimum detectable vessel length (SAR, calm seas) | ~30 m; smaller vessels ambiguous above sea state 4 |
| Maximum scar-detection water depth (optical) | ~10–15 m in clear tropical water; shallower in turbid conditions |
| AIS revisit (Spire spaceborne) | Sub-30-minute global average message collection |
| Sentinel-1 archive depth | From April 2014 (Sentinel-1A launch) |
| Delivery formats | GeoTIFF, GeoJSON, KML, Shapefile; PDF evidence summary on request |
Analytics Satellize can run
| MPA anchoring alert | CFAR vessel detection on Sentinel-1 IW, spatial intersection with MPA boundary polygons, AIS cross-reference via Spire | Near-real-time alert (GeoJSON point feature with vessel dimensions, timestamp, AIS status and MPA zone identifier) |
| Dark-vessel flag | SAR detection inside MPA boundary with no corresponding AIS transmission within ±2-hour window | Priority alert appended to anchoring alert feed, flagged for optical tasking |
| Anchor chain scar mapping | WorldView-3 panchromatic and multispectral analysis; spectral unmixing to separate exposed coral skeleton from sand and healthy reef substrate; linear feature extraction | Georeferenced scar polygon layer (Shapefile/GeoJSON) with scar length, width, estimated area disturbed and confidence rating |
| Enforcement evidence pack | Multi-source fusion: SAR detection, AIS record, optical imagery, MPA boundary intersection, SAR centroid uncertainty radius | PDF report with annotated imagery, coordinate tables, provenance chain and uncertainty statements; GIS layers in jurisdiction-specified projection |
| Repeat-offender location analysis | Historical SAR detection archive queried for recurrent stationary targets inside MPA zones; scar overlay compared across multiple WorldView-3 collects | Time-series GIS layer showing detection frequency per location; tabular summary of vessel dimensions and AIS identifiers where available |
| Reef condition change detection | Multi-date Planet SuperDove multispectral comparison in reef zones; anomaly detection on blue-green reflectance indicative of substrate disturbance or bleaching | Monthly change-detection raster with disturbance probability score per 3 m pixel, delivered as GeoTIFF |
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.