Sensors
- Sentinel-1 SAR (C-band, ESA): Interferometric Wide Swath mode gives 10 m ground resolution across a 250 km swath. Repeat pass over European waters is typically 6 days per satellite, 3 days with both Sentinel-1A and -1B active. Detects vessels as bright point targets regardless of AIS status, but turbine structures and their radar shadows create clutter zones that can mask small vessels within roughly 100–300 m of a turbine mast, depending on incidence angle.
- exactEarth satellite AIS: Space-based AIS aggregates Class A and Class B transponder messages globally. Provides vessel identity, position, speed, course and declared voyage data. Coverage latency varies from near-real-time to 30-minute gaps in high-density areas where message collision degrades reception. Fishing vessels under 15 m are not legally required to carry AIS, so absence of a signal is not absence of a vessel.
- Planet SuperDove (PlanetScope): 3 m resolution multispectral imagery at up to daily revisit. Confirms vessel type by hull geometry and deck structure where cloud cover permits. Useful for distinguishing construction support vessels from fishing vessels when SAR returns are ambiguous. Daytime only; cloud cover over the North Sea and similar latitudes limits clear-sky acquisition to a fraction of passes.
- Sentinel-2 MSI (ESA): 10 m resolution in visible and near-infrared bands, 5-day revisit at mid-latitudes with both satellites. Covers large wind farm zones in a single tile. Useful for mapping sediment plumes from construction activity and confirming gross vessel presence, though vessel detection at 10 m is limited to craft above roughly 20–25 m in length.
What a turbine array does to radar
A wind turbine is an excellent radar reflector. Its steel tower, nacelle and rotating blades produce a strong backscatter return in Sentinel-1 C-band imagery, typically saturating pixels at 10 m resolution. The problem is what happens immediately behind it. In SAR geometry, each turbine casts a radar shadow that can extend several hundred metres downrange, the exact length depending on incidence angle and turbine height. A modern offshore turbine with a hub height of 90–120 m can shadow an area large enough to conceal a 20 m fishing vessel entirely.
Large wind farms compound this. Arrays of 50 to 200 turbines arranged in regular grids produce overlapping shadow corridors that form systematic blind spots. Vessels that are either unaware of this geometry or deliberately exploiting it can transit through these corridors with reduced probability of SAR detection. This is not a theoretical concern: the same physics that makes wind farms visible from orbit makes the spaces between turbines harder to monitor than open water.
Choke points, exclusion zones and the vessels that ignore both
Offshore wind farm safety zones are typically defined by national maritime authorities as 500 m exclusion radii around each turbine during construction, with operational zones varying by jurisdiction. In practice, fishing vessels frequently transit within or through these zones, either because the boundaries are poorly marked at sea or because the fishing grounds predate the wind farm. Construction phases introduce additional vessel categories: cable-lay ships, jack-up barges, crew transfer vessels and heavy-lift crane vessels, each with different manoeuvrability and collision risk profiles.
The geometry of a large array forces non-construction traffic into defined corridors between turbine rows. These corridors become de facto choke points where vessel density is higher than the surrounding sea, and where a single slow-moving or anchored vessel can create conflict with faster-moving traffic. Mapping these corridors from first principles, using the published turbine coordinates from consent documentation, allows conflict risk to be assessed before an incident rather than after.
SAR detection inside the array: what works and what does not
Sentinel-1 can detect vessels as small as roughly 10–20 m in length in open water under moderate sea states, based on published detection studies using CFAR (Constant False Alarm Rate) algorithms. Inside a wind farm array, that floor rises. Turbine clutter and shadow contaminate the background statistics that CFAR relies on, so the practical minimum detectable vessel size increases to perhaps 30–50 m in heavily cluttered zones. Smaller fishing vessels in these areas require either a favourable incidence angle, a calm sea state, or corroboration from optical imagery.
AIS correlation is the primary method for distinguishing construction fleet traffic from fishing or transit vessels. A vessel appearing in SAR but absent from AIS within the wind farm boundary is not automatically a violation: small fishing vessels below 15 m are exempt from AIS carriage requirements in many jurisdictions. The analytic task is to classify that residual population by size and behaviour, using SAR cross-section and, where available, optical confirmation.
Fusing SAR, AIS and optical into a conflict picture
The working method is a three-layer stack. First, SAR detections from each Sentinel-1 pass are extracted using CFAR, with turbine locations masked using the consented array coordinates. Second, exactEarth AIS positions are time-matched to the SAR acquisition window, typically within plus or minus 15 minutes, and each SAR detection is assigned to an AIS track where the positional agreement is within 200–300 m. Third, unmatched SAR detections, those with no corresponding AIS signal, are flagged for further classification.
Planet SuperDove imagery, when cloud-free, provides the classification step for flagged targets. Hull aspect ratio, deck clutter and wake geometry at 3 m resolution are sufficient to separate a 15 m fishing vessel from a 40 m crew transfer vessel in most cases. The output is a conflict density map showing vessel trajectories, exclusion zone dwell events and unidentified vessel positions, updated on each Sentinel-1 pass. Satellize runs this fusion pipeline on open Sentinel data with commercial AIS and optional Planet tasking added on client licence.
Honest limits: cloud cover over the North Sea averages around 70–80% of days, which means optical confirmation is intermittent. SAR revisit at 3 days is adequate for trend analysis but will miss short-duration incursions. The shadow masking problem is real and not fully solvable at current Sentinel-1 resolution; it requires acknowledgement in any risk assessment delivered to a client.
What the output is actually used for
Wind farm operators and their marine warranty surveyors need evidence of exclusion zone compliance for insurance and regulatory purposes. Port and harbour authorities managing the surrounding sea area need to understand where fishing vessel traffic has been displaced to, since a wind farm that displaces fishing effort into a shipping lane creates a different risk from one that displaces it into open water. Consent authorities reviewing new array applications need historical traffic density data for the proposed site.
The conflict map also has a temporal dimension that matters. Construction phases, typically 18–36 months for a large array, generate the highest vessel density and the most heterogeneous traffic mix. Operational phases reduce construction vessel activity but do not eliminate it: operations and maintenance vessels make regular visits, and the fishing displacement effect persists. A monitoring programme that distinguishes these phases in the archive produces a more useful risk picture than a single snapshot.
Typical figures
| SAR spatial resolution (Sentinel-1 IW mode) | 10 m ground range, 250 km swath |
| SAR revisit (European waters, dual satellite) | 3 days typical; 6 days single satellite |
| Optical resolution (Planet SuperDove) | 3 m multispectral; up to daily revisit, cloud-dependent |
| Optical resolution (Sentinel-2 MSI) | 10 m visible/NIR; 5-day revisit at mid-latitudes |
| Minimum detectable vessel (SAR, open water) | ~10–20 m length under moderate sea state |
| Minimum detectable vessel (SAR, within array clutter zone) | ~30–50 m; smaller vessels require optical confirmation |
| AIS message latency (exactEarth space-based) | Near-real-time to ~30 min in high-density areas |
| Sentinel-1 archive depth | From October 2014 (Sentinel-1A launch) |
| Radar frequency / band | C-band, 5.405 GHz |
| Delivery formats | GeoTIFF conflict density rasters, GeoJSON vessel event logs, PDF compliance reports |
Analytics Satellize can run
| Exclusion zone violation log | SAR-AIS temporal matching with consented array boundary polygon overlay; unmatched or boundary-crossing detections flagged | Timestamped GeoJSON event feed of confirmed and probable violations, updated per Sentinel-1 pass |
| Vessel traffic density heatmap | Kernel density estimation applied to SAR detection centroids and AIS track interpolations over a rolling time window | GeoTIFF raster layer showing vessel density per square kilometre, configurable for construction versus operational phase periods |
| Unidentified vessel classification report | CFAR detection of AIS-dark targets; size estimation from SAR cross-section; optical hull geometry classification where Planet imagery is available | PDF report with vessel thumbnails, estimated dimensions and probable type assignments for each unmatched SAR detection |
| Radar shadow corridor map | Geometric shadow modelling from published turbine coordinates and hub heights, computed per Sentinel-1 incidence angle | GIS polygon layer of shadow zones per acquisition, usable as a detection-confidence qualifier in downstream analysis |
| Displaced fishing effort analysis | Comparison of pre- and post-construction vessel density distributions in the surrounding sea area, using Sentinel-1 archive from 2014 onwards | Time-series chart and spatial difference map showing fishing vessel redistribution around the array boundary |
| Construction phase traffic mix report | AIS vessel type classification combined with SAR size estimation to separate construction fleet from incidental traffic; temporal segmentation by project phase | Monthly summary report with vessel category counts, peak density dates and identified conflict corridor locations |
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.