Polar Research and Resupply Vessel Expedition Tracking

AIS drops out above 70°N. Cloud and darkness eliminate optical cover for months at a time. SAR and satellite-relayed AIS together maintain positional continuity for research icebreakers and resupply ships, but separating vessel returns from pressure-ridge clutter demands more than a threshold.

Sensors

ICEYE SAR constellation: X-band (9.65 GHz) synthetic-aperture radar. Spotlight mode delivers approximately 0.5 m ground resolution; strip mode approximately 3 m. Revisit at high latitudes can reach sub-daily due to orbital geometry favouring polar passes. All-weather, all-hours imaging regardless of cloud or polar night.
Sentinel-1 SAR (ESA): C-band (5.405 GHz) SAR, 10 m resolution in Interferometric Wide Swath mode with a 250 km swath. Repeat cycle of 6 days per satellite (12 days for a single satellite); polar regions receive more frequent coverage due to orbit convergence. Free and open archive from 2014.
Spire Global AIS (satellite-relayed): Spire operates over 100 LEO satellites carrying AIS receivers. At high latitudes, orbital geometry increases pass frequency, improving AIS message capture rates compared with mid-latitudes. Positional latency is typically under 90 minutes for a given vessel, though message collision in busy straits can degrade decode rates.
Iridium-relayed AIS: Iridium's 66-satellite LEO constellation provides near-global AIS relay including polar regions above 70°N and 70°S. Aireon, hosted on Iridium NEXT, achieves global AIS coverage with a median latency of around 90 seconds for position reports, filling the gap where terrestrial and lower-inclination satellite AIS struggles.
Spire GNSS Radio Occultation: GNSS-RO profiles atmospheric refractivity, providing temperature and moisture soundings that feed into sea-ice and weather forecasting products used to contextualise vessel routing. Not a vessel-detection sensor, but operationally relevant for route-risk assessment in polar passages.

Why AIS fails precisely where you need it most

AIS was designed for coastal and open-ocean collision avoidance. Its VHF signal (161.975 MHz and 162.025 MHz) propagates line-of-sight, so terrestrial base stations lose contact with vessels beyond roughly 40 nautical miles offshore. Satellite AIS extends that range, but the polar regions introduce a compounding problem: the same high latitudes that attract research and resupply missions also produce the densest satellite pass frequency, which sounds like an advantage until you account for message collision. When multiple vessels transmit simultaneously on the same channel, decode rates fall. The Arctic shipping lanes around the Northern Sea Route and the waters approaching Antarctica are not yet congested enough to saturate AIS systematically, but individual vessels operating far from any lane, as polar research ships routinely do, can go unheard for hours.

The practical consequence is a position record that looks like a dotted line rather than a track. A vessel that reported at 0300 UTC and next appears at 0900 UTC has, in AIS terms, been invisible for six hours. For a resupply mission threading through moving pack ice, that gap is operationally significant. Iridium-hosted AIS (via Aireon) addresses the relay problem with genuine near-global coverage, but it cannot decode a transmission that was never made: vessels with malfunctioning or deliberately disabled transponders remain dark regardless of relay infrastructure.

What a floating pressure ridge gives away, and why that matters for detection

SAR fills the optical gap in polar night and cloud, but it introduces a different problem. Sea ice is not a quiet background. First-year ice, multi-year ice, pressure ridges, brash ice and icebergs all produce radar backscatter, and some of it overlaps with the backscatter signature of a steel hull. A pressure ridge can return a signal comparable to a small vessel in C-band imagery. An iceberg with a steep, faceted face will produce a very bright point return in X-band spotlight mode that, stripped of context, resembles a ship.

The standard approach to vessel detection in open water, constant false-alarm rate (CFAR) thresholding against a local background estimate, degrades significantly in heterogeneous ice scenes. The background is no longer statistically uniform. Analysts must apply ice-type classification as a prior step, segmenting the scene into open water, nilas, first-year ice, and deformed ice zones before applying detection logic. Sentinel-1's dual-polarisation (HH+HV) data is particularly useful here: sea ice and vessel hulls have different polarimetric ratios, and the combination of backscatter intensity with cross-polarisation ratio reduces ambiguity. ICEYE's higher-resolution spotlight imagery allows analysts to look for vessel-specific features such as superstructure geometry, shadow length consistent with hull freeboard, and the characteristic bright-dark-bright pattern of a ship's bow. None of these cues are infallible, but together they push detection confidence well above what intensity thresholding alone achieves.

Fusing SAR detections with satellite AIS: where the arithmetic gets interesting

A SAR detection without an AIS correlation is a candidate vessel. An AIS position without a SAR confirmation is an asserted position. Fusing the two requires accounting for the time offset between the SAR acquisition and the nearest AIS timestamp, then projecting the vessel's expected position forward or backward using reported speed and heading. For a vessel travelling at 8 knots, a two-hour offset places the AIS-projected position roughly 16 nautical miles from the acquisition time position. In open water that is a manageable search radius. In ice, where a vessel may be beset, stopped, or moving at 1 to 2 knots through leads, the geometry tightens considerably.

When a SAR detection falls within the projected position envelope and the radar cross-section is consistent with the vessel class, the fusion is straightforward. When it does not match, the analyst faces three possibilities: the AIS position was fabricated or stale, the SAR detection is a false positive from ice clutter, or the vessel has deviated significantly from its reported track. In polar research contexts, the third explanation is common and benign. Ships follow leads, not straight lines. The fusion algorithm must tolerate large heading deviations without flagging them as anomalies.

Ice-condition context changes what a track means

A vessel's position is only half the operational picture. Whether it is making progress, beset, or sheltering behind a grounded iceberg depends on the ice conditions surrounding it. Sentinel-1 wide-swath imagery, processed through established ice-classification algorithms, provides a synoptic view of ice concentration and type across the scene at 10 m resolution. The European Centre for Medium-Range Weather Forecasts and national ice services including the Norwegian Ice Service and the Canadian Ice Service publish ice charts derived partly from Sentinel-1 that cover the Arctic and Southern Ocean on daily to weekly cycles.

Overlaying a vessel track on an ice-condition mosaic answers questions that position alone cannot. A vessel stationary for 18 hours in a region of 9/10 ice concentration is almost certainly beset. The same vessel stationary in open water near a research station is resupplying. The distinction matters for mission planning, insurance notification, and search-and-rescue readiness. Ice-condition context also informs forward route assessment: a corridor that was navigable during the last Sentinel-1 pass may have closed due to wind-driven ice compaction in the intervening 12 hours, a risk that GNSS-RO atmospheric data and numerical weather prediction can help quantify.

Honest limits of the method

Even with ICEYE tasking on demand, SAR revisit at a specific location in polar regions is not continuous. A vessel can move significantly between passes. Spotlight mode, which provides the resolution needed to distinguish hull geometry from ice clutter, covers a scene of roughly 5 km by 5 km per acquisition. Tasking a moving vessel requires predicting its position, which depends on the quality of the last known fix. If AIS has been silent for several hours and the vessel is navigating through ice at variable speed, the positional uncertainty can exceed the spotlight footprint.

Cloud is irrelevant to SAR, but orbital geometry is not. Both ICEYE and Sentinel-1 are in sun-synchronous or near-polar orbits, which means pass frequency over the high Arctic and Antarctic is higher than at mid-latitudes, but the passes are not evenly distributed through the day. Gaps of four to six hours between any satellite SAR pass over a given point are common. For a vessel in distress, that is a long time. The system described here is a monitoring and intelligence tool, not a real-time safety system. Search-and-rescue operations require dedicated assets and direct communication links that satellite imagery cannot replace.

Satellize runs polar-region SAR analytics as part of its maritime intelligence service, applying ice-classification preprocessing before vessel detection to reduce the false-positive rate that naive CFAR thresholding produces in heterogeneous ice scenes.

What the delivered product looks like

The operational output for a polar expedition monitoring programme is a fused track layer updated on each new SAR acquisition or AIS capture, whichever arrives first. Each position fix carries a confidence score reflecting whether it came from AIS, SAR, or both, and whether the SAR detection was confirmed against an ice-classified background or flagged as ambiguous. Ice-condition overlays from Sentinel-1 mosaics are attached as a separate layer, timestamped to the acquisition. Route deviation alerts fire when a vessel's SAR-confirmed position diverges from its filed expedition plan by more than a configurable threshold.

For programme managers and national Antarctic or Arctic operators, this is not a surveillance product in the adversarial sense. It is a situational awareness feed for missions that operate in some of the most communication-degraded environments on Earth, where the cost of losing track of a vessel is measured not in commercial loss but in lives and very expensive hardware.

Typical figures

SAR spatial resolution (ICEYE Spotlight)	Approximately 0.5 m
SAR spatial resolution (Sentinel-1 IW mode)	10 m, 250 km swath
SAR revisit at polar latitudes (ICEYE)	Sub-daily at latitudes above 70° due to orbital convergence; exact frequency depends on constellation tasking
SAR revisit (Sentinel-1, single satellite)	12-day repeat at equator; more frequent at high latitudes due to orbit overlap
AIS positional latency (Iridium/Aireon)	Median approximately 90 seconds for position reports where transponder is active
AIS positional latency (Spire LEO)	Typically under 90 minutes per vessel per pass at high latitudes
Minimum detectable vessel (SAR, open water)	Approximately 10 m length in C-band; smaller targets detectable in X-band spotlight but ice clutter raises false-positive rate
Sentinel-1 archive depth	From 2014 (Sentinel-1A launch); accessible via Copernicus Data Space
ICEYE archive depth	From 2019; commercial access via tasking or archive licence
Delivery formats	GeoJSON track layers, GeoTIFF ice-classification mosaics, alert feeds (JSON/API), PDF expedition reports

Analytics Satellize can run

Fused vessel position track	SAR CFAR detection on ice-classified background, fused with satellite AIS using time-offset position projection	GeoJSON track layer with per-fix confidence score, updated on each new SAR acquisition or AIS capture
Ice-condition mosaic	Sentinel-1 dual-polarisation (HH+HV) classification into open water, nilas, first-year and deformed ice using published polarimetric ratios	GeoTIFF mosaic timestamped to acquisition, delivered alongside vessel track layer
Beset-vessel alert	Velocity-from-track calculation; stationary flag triggered when vessel speed falls below 0.5 knots for a configurable duration in high ice-concentration zone	Automated alert (JSON/API) with last confirmed position, ice concentration at location, and hours stationary
Route deviation report	Comparison of SAR-confirmed positions against filed expedition plan waypoints; deviation calculated in nautical miles	Alert when deviation exceeds operator-defined threshold; weekly PDF summary for programme managers
Iceberg and pressure-ridge clutter classification	X-band spotlight texture analysis and shadow-length measurement to distinguish hull geometry from ice features	Annotated SAR scene with candidate vessels flagged, ambiguous returns labelled, and confirmed false positives suppressed
AIS gap analysis	Time-series gap detection in AIS message record; cross-referenced against SAR detections to determine whether silence reflects relay failure or transponder absence	Gap log with probable cause classification (relay gap, transponder off, vessel absent from scene) delivered as tabular report

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.