Bottom trawling disturbance detection from SAR and optical signatures

Satellite optical and SAR imagery can reveal sediment resuspension plumes and seabed track marks left by demersal trawl gear, providing an independent spatial record of fishing effort that VMS logs alone cannot supply.

Sensors

Sentinel-2 MSI: 10 m resolution in visible and near-infrared bands; 5-day revisit at the equator with both satellites. Blue (Band 2, 490 nm) and green (Band 3, 560 nm) bands are most sensitive to suspended sediment in the upper water column. Freely available via Copernicus Dataspace.
Planet SuperDove: 3 m resolution, daily revisit over most coastal areas. Eight spectral bands including coastal blue (431 nm) improve discrimination of turbidity from phytoplankton. Commercial tasking allows same-day collection when a trawl event is suspected; requires client licence.
Sentinel-1 SAR (C-band, 5.405 GHz): IW mode delivers 10 m resolution, 250 km swath, approximately 6-day revisit per satellite. In water shallower than roughly 20-30 m, bottom trawl gear disturbs the seabed in ways that can alter surface capillary wave structure, producing subtle backscatter anomalies. Cloud-independent. Interpretation requires careful separation from wind and current artefacts.
Landsat 8/9 OLI: 30 m resolution, 16-day revisit per satellite (8-day combined). Coastal aerosol band (Band 1, 443 nm) is specifically designed for water-column work. Longer archive (Landsat 8 from 2013, Landsat 9 from 2021) supports multi-year trend analysis of chronically disturbed grounds.

What a trawl leaves behind, and how long it lasts

When demersal trawl gear contacts the seabed it does two distinct things: it mechanically disturbs benthic habitat, and it injects fine sediment into the water column. The sediment signal is what remote sensing can see. In shallow, optically shallow water, typically less than 20-30 m, resuspended particles scatter incoming sunlight back upward through the water column and appear as elevated radiance in the blue and green bands of multispectral imagery. The resulting plume is elongated in the direction of trawl travel and can persist for several hours under calm conditions, long enough for a satellite overpass to capture it.

The track marks themselves, parallel furrows in the seabed, are occasionally visible in very shallow, clear water through direct bottom reflectance. Published work using Sentinel-2 imagery over the North Sea and Mediterranean has demonstrated detection of plumes in water depths of 5-25 m. Beyond 30 m the water column attenuates the signal to the point where it is indistinguishable from background variability. That is a firm physical ceiling, not a processing limitation.

Reading turbidity anomalies without crying wolf

The core optical method is a turbidity or remote-sensing reflectance anomaly: compare the observed blue-green reflectance in a candidate pixel against a baseline derived from recent cloud-free composites of the same location under similar tidal and wind conditions. A statistically significant positive anomaly, elongated and co-located with a vessel track from AIS or VMS, constitutes a detection event. Without the co-location check, river plumes, natural resuspension from storm swell, and algal blooms all produce false positives.

Tidal phase matters enormously. Spring tides in naturally turbid estuaries can raise background suspended sediment concentration by an order of magnitude. Any credible detection workflow must condition anomaly thresholds on tidal state, ideally using a hydrodynamic model or a tidal-phase-stratified baseline. Ignoring this produces alert rates that are operationally useless.

Atmospheric correction is the other common failure point. Over water, small errors in aerosol optical depth retrieval translate directly into reflectance errors in the blue band. The ESA ACOLITE processor and NASA SeaDAS are the standard open tools for coastal water atmospheric correction; both are well-documented and freely available.

What SAR adds, and where it falls short

C-band SAR backscatter from the sea surface is governed primarily by centimetre-scale capillary and short gravity waves driven by wind. In very shallow water, bottom roughness and sediment mobility can modulate surface roughness indirectly, and trawl-induced bottom disturbance has been proposed as a detectable signal in Sentinel-1 imagery. The physics is plausible: disturbed sediment changes bed roughness and can alter near-bottom turbulence, which in turn affects surface wave damping.

In practice, the SAR signal is weak and easily confused with wind shadow, current shear, and surfactant films from vessel traffic. SAR is most useful here not as a standalone detector but as a corroborating layer: if optical imagery shows a turbidity plume and SAR shows a co-located backscatter anomaly, confidence rises. Treating SAR as primary evidence in the absence of optical confirmation is not currently supported by peer-reviewed detection performance figures at operational scale.

Honest limits of the method

Depth is the dominant constraint. Below roughly 30 m, optical detection of sediment resuspension is not reliable with current freely available sensors. This excludes the majority of continental shelf trawling grounds, where depths of 50-200 m are common. The method is genuinely useful in shallow coastal zones, estuaries, and shelf-sea areas with naturally low background turbidity.

Cloud cover is a secondary but significant constraint for optical sensors. In persistently overcast regions, such as the North Sea in winter, useful optical imagery may be available only a few days per month. SAR fills part of this gap but with the interpretive caveats noted above. Revisit frequency also limits detection of short-duration events: a vessel that trawls for two hours and departs before the next Sentinel-2 overpass leaves no optical trace, even in ideal conditions.

Finally, the method produces evidence of disturbance, not legal proof of illegal fishing. A plume co-located with a vessel that has a valid licence and is operating within its permitted grounds is not an infringement. The analytical output is an independent spatial record to be compared against VMS, AIS, and licensing data, not a substitute for them.

Turning detections into a monitoring programme

A practical monitoring system combines three layers: a near-real-time optical anomaly alert (triggered within 24-48 hours of a cloud-free overpass), a retrospective monthly disturbance map showing cumulative trawl pressure across a management area, and a cross-reference table matching detected plume events to VMS or AIS records for the same time window. The last layer is where the independent value lies. Events with no corresponding licensed vessel in the VMS log are the cases a fisheries enforcement authority needs to investigate.

Archive depth matters for baseline construction. Landsat data from 2013 onward and Sentinel-2 data from 2015 onward allow analysts to characterise the natural turbidity regime of a site before building anomaly thresholds. For sites with genuinely high and variable background turbidity, it may be necessary to conclude that optical detection is not feasible and redirect effort toward AIS gap analysis or SAR vessel detection instead. Satellize runs this kind of feasibility scoping as a standard first step before committing to a monitoring architecture, drawing on the same open-constellation workflows used in its Tonga crop-estimation programme.

Integrating satellite evidence with existing fisheries oversight

VMS provides position and speed but is controlled by the vessel operator and subject to manipulation or technical failure. Satellite imagery is collected by a third party and cannot be altered after the fact. That independence is the core argument for including it in a fisheries monitoring framework, not as a replacement for VMS but as a check on it.

Several regional fisheries management organisations and national agencies have begun integrating remote-sensing turbidity analysis into their compliance workflows, particularly in the Mediterranean and South-East Asian shelf seas where shallow trawling grounds overlap with protected areas. The Global Fishing Watch platform publishes AIS-derived trawling effort maps that can be used as a comparison layer. Where satellite-detected plumes appear in areas with no AIS-recorded trawling, the discrepancy is itself an investigative lead.

Typical figures

Optical spatial resolution	10 m (Sentinel-2), 30 m (Landsat 8/9), 3 m (Planet SuperDove on licence)
SAR spatial resolution	10 m (Sentinel-1 IW mode)
Revisit (optical, free)	5 days at equator (Sentinel-2 combined); 8 days combined (Landsat 8+9)
Revisit (SAR, free)	Approximately 6 days per Sentinel-1 satellite; ~3 days with both
Key spectral bands	Coastal blue ~443 nm, blue ~490 nm, green ~560 nm for turbidity; C-band 5.405 GHz for SAR
Maximum detection depth	Approximately 20-30 m for optical plume detection; SAR bottom-roughness signal limited to <20 m
Minimum detectable plume width	Approximately 30-50 m at Sentinel-2 resolution under low background turbidity conditions
Alert latency	24-48 hours from satellite overpass to analyst-reviewed alert, subject to cloud cover
Archive depth	Sentinel-2 from 2015; Landsat 8 from 2013; Sentinel-1 from 2014
Delivery formats	GeoTIFF anomaly rasters, GeoJSON event polygons, CSV cross-reference table, PDF monthly report

Analytics Satellize can run

Near-real-time turbidity anomaly alert	Per-pixel remote-sensing reflectance anomaly against tidal-phase-stratified baseline (Sentinel-2 Bands 2/3); ACOLITE atmospheric correction	GeoJSON alert polygon with confidence score, delivered within 48 hours of cloud-free overpass
Monthly cumulative trawl-pressure map	Aggregation of confirmed plume-detection events over rolling 30-day window, normalised by cloud-free observation frequency	GeoTIFF raster of disturbance frequency per grid cell, suitable for GIS import
VMS/AIS cross-reference report	Spatial and temporal join of satellite detection events against client-supplied VMS or public AIS records; discrepancy flagging	CSV table and PDF summary identifying unmatched detection events for enforcement review
Background turbidity regime characterisation	Multi-year Sentinel-2 and Landsat time-series analysis of natural suspended sediment variability, stratified by tidal phase and season	Site feasibility report with detection probability estimates and recommended monitoring configuration
SAR backscatter anomaly layer	Sentinel-1 IW VV/VH ratio change detection against rolling 30-day baseline; co-registration with optical plume detections	GeoTIFF corroborating layer flagged as secondary evidence only; delivered alongside optical alert
Protected area incursion log	Intersection of confirmed plume polygons with client-supplied MPA boundary files; filtering for events with no corresponding licensed VMS track	Monthly incursion log in GeoJSON and PDF, formatted for submission to fisheries authority

Sources

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.