Amphibious assault staging and landing-craft concentration monitoring
Flat-bottomed landing craft and vehicle staging aprons produce distinctive radar and optical signatures that betray amphibious force preparation. This page explains the sensor physics, honest revisit limits, and analytic methods used to monitor coastal marshalling areas.
Sensors
- Maxar WorldView-2: 46 cm panchromatic, 1.85 m multispectral. Eight spectral bands allow vessel-type discrimination by hull colour and deck configuration. Tasked on demand but revisit over a single point is typically 1 to 3.5 days depending on latitude and cloud; not suitable as a sole persistent watch.
- Planet Dove constellation: 3 m multispectral, daily global revisit under clear skies. Sufficient to count large landing craft and detect new vehicle concentrations on staging aprons. Cannot reliably resolve individual wheeled or tracked vehicles; cloud cover breaks the daily cadence unpredictably.
- ICEYE SAR (X-band): Spotlight mode delivers approximately 0.5 m resolution; strip mode approximately 3 m. All-weather, day-night. The high dielectric return from metal hulls and corner-reflector geometry of landing-craft ramps makes them bright SAR targets. Tasking is commercial and latency from collection to delivery is typically under 30 minutes for urgent orders.
- Sentinel-1 C-band SAR: IW mode at 10 m ground resolution, 250 km swath, 6-day repeat at the equator and roughly 3-day repeat at mid-latitudes with both satellites. Free and open. Adequate for detecting clusters of vessels and large staging areas; individual craft below about 20 m length are near the detection floor. Archive extends to 2014.
What a flat-bottomed hull gives away
A landing craft utility (LCU) or LCAC hovercraft presents a geometry that SAR finds difficult to ignore. The flat metal bow ramp, when lowered or even slightly ajar, acts as a dihedral corner reflector, bouncing C-band or X-band energy directly back to the sensor with a radar cross-section disproportionate to the vessel's modest size. A typical LCU of roughly 40 metres length can produce a SAR return comparable to a much larger commercial vessel with a curved hull. That physical fact is not a vulnerability that doctrine can easily close: the ramp is the point of the vessel.
In optical imagery, landing craft are identifiable by hull form rather than superstructure. They sit low in the water, have a distinctive squared-off bow, and when beached or moored bow-on to a ramp, they leave a predictable geometric footprint. At 46 cm resolution (WorldView-2 panchromatic), individual craft are unambiguous. At Planet Dove's 3 m, a cluster of four or more craft moored together becomes a detectable mass even if individual hulls blur together.
The staging apron as the real signal
The vessels themselves are only part of the picture. Amphibious operations require vehicle staging: tanks, infantry fighting vehicles, artillery and logistics trucks must be marshalled, loaded in sequence, and held in a defined order that reflects the landing plan. That process leaves a visible mark on any large paved or compacted-earth apron near the embarkation berths.
Vehicle concentrations on staging areas show up in optical imagery as changes in surface texture and tone. A previously empty concrete apron develops a regular grid of dark rectangles (tracked vehicles) or a denser, lighter pattern (wheeled logistics). Shadow analysis at sub-metre resolution allows approximate height and therefore vehicle-class estimation. Sentinel-1 coherence differencing, comparing phase coherence between passes separated by 6 or 12 days, detects the ground disturbance and metallic clutter introduced by a sudden vehicle concentration even when optical cloud cover is total. This method was demonstrated in published literature for detecting military vehicle movements in Eastern Europe and is not novel.
The combination matters. Vessels alone might indicate an exercise or a routine maintenance period. Vessels plus a staged vehicle concentration plus increased activity at fuel and ammunition handling points constitutes a convergent signature that is harder to explain away.
Revisit gaps are the honest problem
No current open or commercial constellation achieves sub-12-hour revisit over an arbitrary coastal point with guaranteed cloud-free optical coverage. Sentinel-1 provides 3-to-6-day SAR revisit at mid-latitudes. ICEYE can be tasked to sub-daily revisit, but only if the client holds sufficient tasking capacity and the target location is in the satellite's accessible arc at the required time. Planet Dove's daily optical pass is defeated by persistent cloud, which is common over many littoral regions of operational interest.
The practical consequence is that a short-notice sortie window of 6 to 10 hours, a force loading and departing between two SAR passes, will not be captured by any single sensor. The mitigation is sensor fusion: scheduling ICEYE spotlight collections at the highest affordable cadence, using Sentinel-1 as the baseline change detector, and treating Planet Dove as the daily optical confirmation layer. Even then, a determined adversary who loads at night under cloud during a Sentinel-1 gap has a real chance of departing unobserved. That limit should be stated plainly to any client building a monitoring requirement around this use case.
Building a change-detection baseline
Effective monitoring starts with a baseline, not an alert. Sentinel-1's archive from 2014 onward covers most of the world's significant naval and amphibious ports. Establishing a normal distribution of vessel count, vessel type, and staging-apron occupancy for a given port across seasons and years allows anomaly detection to be statistically grounded rather than subjective.
The analytic workflow runs in three layers. First, a vessel-detection algorithm (typically a constant false alarm rate, or CFAR, detector applied to SAR amplitude imagery) counts and locates all metal objects in the water and at berths. Second, optical change detection on the staging apron compares current Planet or WorldView imagery against the baseline. Third, a fusion step flags dates on which both vessel count and apron occupancy exceed their historical 90th-percentile values simultaneously. That conjunction is the alert condition. Each layer has published precedents in the remote-sensing literature; none requires proprietary physics.
Archive depth is a practical asset. If a port has been used for amphibious exercises in prior years, the seasonal pattern of those exercises is recoverable from the Sentinel-1 record and provides a reference against which genuine preparation can be distinguished from routine training.
What the imagery cannot settle
Satellite imagery confirms presence and concentration. It does not confirm intent, readiness, or timeline. A port full of landing craft and a staged vehicle concentration is consistent with an imminent operation, a large-scale exercise, or a deliberate deception. Imagery analysis must be paired with other intelligence streams to carry operational weight.
Resolution limits also constrain vehicle-type identification. At Sentinel-1's 10 m, a tank and a truck are both rectangular bright objects. WorldView-2 at 46 cm allows hull and turret discrimination for main battle tanks, but distinguishing a T-72 variant from a T-80 requires analyst expertise and good sun angle. ICEYE spotlight at 0.5 m begins to allow vehicle-class attribution in SAR, though speckle and foreshortening on sloped surfaces complicate the picture. Clients should not expect vehicle-type census accuracy from SAR alone.
Satellize's analytics work runs on open constellations as the persistent baseline, with commercial tasking added on client licence for surge periods. The Tonga crop-estimation programme demonstrated that consistent revisit scheduling and baseline-relative change detection, rather than raw resolution, is what produces reliable outputs. The same principle applies here, though the stakes and the adversarial context are obviously different.
Typical figures
| Best optical resolution (tasked) | 46 cm panchromatic (Maxar WorldView-2) |
| Best SAR resolution (tasked) | ~0.5 m spotlight (ICEYE X-band) |
| Open SAR resolution | 10 m IW mode (Sentinel-1 C-band, 5.405 GHz) |
| Optical revisit (open) | Daily under clear sky (Planet Dove, 3 m) |
| SAR revisit (open) | 3 to 6 days per point (Sentinel-1, dual-satellite) |
| SAR revisit (commercial surge) | Sub-daily possible with ICEYE tasking allocation |
| Minimum detectable vessel (Sentinel-1) | Approximately 20 m length; smaller craft near or below detection floor |
| Archive depth | Sentinel-1 from 2014; Landsat optical from 1972; Planet from ~2016 |
| Latency (ICEYE urgent tasking) | Under 30 minutes collection-to-delivery for priority orders |
| Delivery formats | GeoTIFF, GeoJSON vessel-count layers, PDF alert report, GIS-ready change polygons |
Analytics Satellize can run
| Vessel count and berth-occupancy time series | CFAR detector applied to SAR amplitude; optical vessel segmentation on Planet and WorldView imagery | Daily GeoJSON layer with vessel count, estimated length class, and berth assignment; weekly trend chart |
| Staging-apron occupancy change alert | Optical change detection (normalised difference of surface reflectance against rolling 90-day baseline); shadow-length vehicle-height estimation | Alert report with annotated before/after image chips and occupancy percentage estimate |
| SAR coherence-loss map for vehicle concentration | Sentinel-1 interferometric coherence differencing between consecutive 6- or 12-day pairs | GeoTIFF coherence-loss layer; flagged polygons where coherence drops below threshold consistent with metallic clutter introduction |
| Convergent-signature fusion alert | Statistical conjunction of vessel-count anomaly and apron-occupancy anomaly against multi-year baseline; threshold set at simultaneous exceedance of 90th percentile on both indicators | Scored alert (low/medium/high) with supporting evidence imagery, delivered within 4 hours of triggering collection |
| Historical exercise-pattern baseline report | Retrospective Sentinel-1 and Planet archive analysis over named port; seasonal decomposition of vessel-count and apron-occupancy time series | PDF baseline report with annotated time series, seasonal norms, and prior exercise signatures for analyst reference |
| Landing-craft type-attribution assessment | Sub-metre SAR and optical hull-geometry analysis; comparison against published vessel dimension databases | Annotated image report with vessel-class estimates and confidence ratings per craft |
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.