Offshore supply vessel activity as oil platform operations proxy
Platform supply vessels run on schedules tightly coupled to production and maintenance tempo. Their AIS tracks, corroborated by SAR detections near known platform coordinates, give analysts a reliable operational proxy for installations that publish no direct output data.
Sensors
- Spire Maritime spaceborne AIS: Multi-satellite VHF AIS reception giving global revisit of roughly 20-90 minutes depending on latitude and constellation geometry; captures MMSI, position, speed, heading and voyage data for vessels transmitting Class A transponders, which all PSVs above 300 GT are required to carry under SOLAS
- ICEYE SAR constellation: X-band synthetic aperture radar; strip-mode resolution approximately 3 m, spotlight mode approximately 1 m; revisit to a specific offshore location achievable within 24 hours under tasking; provides vessel detections independent of AIS, including vessels with transponders switched off or malfunctioning
- Sentinel-1 SAR: C-band SAR operated by ESA; Interferometric Wide Swath mode covers 250 km at 5x20 m resolution; revisit 6-12 days at a fixed point in mid-latitudes under the two-satellite configuration, though maritime zones with overlapping passes can see shorter effective revisit; freely available via Copernicus Data Space
- exactEarth spaceborne AIS: Alternative or supplementary commercial AIS constellation; useful for cross-validating Spire track data in high-traffic areas where message collision rates degrade single-source reception, and for historical archive queries
Why a supply boat's schedule is a production signal
Offshore oil and gas platforms are not required to publish real-time production figures. Many do not. But they cannot operate without a steady rhythm of platform supply vessels carrying drilling mud, cement, fuel, water, provisions and equipment from shore bases. That rhythm is tightly coupled to what is actually happening on the platform: a well being drilled demands more frequent calls than a platform in steady production, and a platform undergoing a turnaround or shutdown sees a distinctive spike in maintenance-related cargo runs followed by a quiet period.
PSVs are not casual visitors. They operate on contracted schedules, and their port-call frequency, cargo manifest categories (where visible in port authority data) and standby patterns near a platform are a direct function of operational tempo. That makes them a proxy signal. Not a perfect one, but a consistent and checkable one.
What the AIS record actually contains
Class A AIS transponders, mandatory on commercial vessels above 300 gross tonnes, broadcast position, speed, heading, MMSI and voyage data at intervals of 2-10 seconds when under way. Spire Maritime's constellation of over 100 satellites in low Earth orbit collects these broadcasts globally, with typical revisit gaps of 20-90 minutes depending on latitude and satellite geometry. That is sufficient to reconstruct port departure times, transit durations, time-on-location near a platform and return voyages with reasonable fidelity.
The analytic workflow starts with vessel classification. PSVs have distinctive MMSI type codes (ship type 35 in the ITU classification), typical LOA of 60-100 m, and characteristic low-speed loitering signatures near platform coordinates. Filtering the raw AIS feed to this vessel class against a database of known platform positions reduces noise substantially. From the filtered tracks, three metrics are extracted: port-call frequency per platform per week, mean time-on-location (a proxy for cargo complexity), and standby event frequency, where a vessel holds position within roughly 500 m of the platform without mooring.
Where SAR earns its place in the method
AIS is self-reported. Transponders fail, are misconfigured, or are occasionally switched off. SAR does not care. Both ICEYE X-band and Sentinel-1 C-band detect vessels as radar-bright objects against the sea surface through their normalised radar cross-section. A PSV of 70-90 m length presents a cross-section well above the detection floor of either system in sea states up to around Beaufort 5-6, above which wave clutter begins to mask smaller targets.
ICEYE spotlight mode at approximately 1 m resolution can resolve vessel shape well enough to support type classification by length and beam ratio, providing independent confirmation that the detected object is a PSV rather than a standby safety vessel or a crew boat. Sentinel-1, at 5x20 m in IW mode, detects the vessel but rarely resolves its type; it is more useful for confirming presence or absence at a platform location on a given pass than for classification. The honest limit here is Sentinel-1's 6-12 day revisit at a fixed point, which means it catches a snapshot, not a pattern. ICEYE tasking fills that gap but adds cost.
The corroboration logic is straightforward: when a SAR detection at platform coordinates coincides with an AIS track showing a PSV on location, confidence in both signals rises. When SAR shows a vessel and AIS shows nothing, that is a flag worth investigating. The inverse, AIS track with no SAR detection, is usually a timing mismatch rather than a falsification, given the revisit gaps.
Honest limits of the proxy
PSV frequency is a lagging and indirect indicator. A spike in calls could mean a new drilling campaign, an unplanned equipment failure, or a scheduled five-year inspection. Distinguishing these requires context: rig-move AIS tracks from anchor-handling vessels, drilling permit filings where public, or seismic survey vessel presence in the area. The method is most reliable when used to detect relative change at a platform over time, or to compare operational tempo across a peer group of platforms in the same basin, rather than to infer absolute production volumes.
AIS coverage degrades in high-traffic areas due to VHF message collision, and in polar regions due to satellite geometry. Spire's constellation mitigates but does not eliminate this. Weather has no effect on AIS or SAR, which is one genuine advantage over optical methods for offshore work. Cloud is irrelevant here.
Building an operational tempo index
The practical output is a per-platform weekly index: normalised PSV call frequency, weighted by time-on-location, cross-checked against SAR detections where tasking budget allows. Run across a basin, this produces a ranked list of platforms by apparent activity level. Changes of more than one standard deviation from a platform's own 90-day baseline are flagged as events worth investigating.
Satellize applies this type of AIS-plus-SAR fusion methodology as part of its broader satellite-data analytics work; the same analytic pipeline that supports the Tonga crop-estimation programme's change-detection logic underpins the vessel classification and temporal indexing here, adapted for maritime rather than agricultural assets. Clients typically receive the index as a weekly GIS layer with an accompanying structured data feed, queryable by platform identifier or basin.
What this cannot replace
No satellite method replaces a direct production meter, a regulatory filing or an operator disclosure. What it provides is an independent, continuous, non-cooperative signal that is available whether or not the operator chooses to communicate. For energy analysts, commodity traders, infrastructure investors and government regulators monitoring foreign basin activity, that independence is the point. The method is transparent about what it measures: vessel behaviour, not barrels. The inference from one to the other is probabilistic and should be presented as such.
Typical figures
| AIS position accuracy | GPS-derived; typically better than 10 m for Class A transponders under open-sky conditions |
| Spire Maritime AIS revisit (mid-latitude) | Approximately 20-90 minutes per location depending on satellite geometry and constellation load |
| ICEYE SAR resolution (spotlight mode) | Approximately 1 m; strip mode approximately 3 m |
| Sentinel-1 SAR resolution (IW mode) | 5 x 20 m range x azimuth; 250 km swath |
| Sentinel-1 revisit at fixed point | 6-12 days under two-satellite configuration; shorter in overlap zones |
| ICEYE tasked revisit | Sub-24-hour to a specific offshore location under commercial tasking |
| Minimum detectable vessel (SAR) | Approximately 20-30 m length in Beaufort 3-4; detection degrades above Beaufort 5-6 for smaller targets |
| AIS archive depth | Spire Maritime historical archive extends several years; exactEarth offers comparable depth |
| Delivery formats | GeoJSON or GeoPackage vessel track layers, CSV event tables, weekly structured data feed, platform-level tempo index |
| Latency (AIS-derived index) | Near-real-time to 24-hour lag depending on data pipeline configuration |
Analytics Satellize can run
| Per-platform PSV call frequency index | AIS track segmentation by platform proximity zone (typically 500 m radius), vessel-type filtering by MMSI ship-type code and LOA distribution | Weekly GIS layer and CSV table; one row per platform per week with call count, mean time-on-location and deviation from 90-day baseline |
| Basin-wide operational tempo ranking | Cross-platform normalisation of PSV frequency index; z-score ranking against basin peer group | Ranked platform list with tier classification (elevated, normal, suppressed activity); delivered as structured report or API feed |
| SAR vessel detection at platform coordinates | Constant False Alarm Rate (CFAR) detection on ICEYE or Sentinel-1 SAR imagery; detections cross-referenced to platform location database | Per-pass detection report with vessel centroid coordinates, estimated length from radar shadow or bright-target extent, and AIS match status |
| AIS-SAR fusion corroboration score | Temporal and spatial join of AIS track positions to SAR detections within configurable time and distance windows; match rate computed per platform per period | Corroboration confidence layer appended to tempo index; flags unmatched SAR detections for further investigation |
| Standby event detection | Identification of loitering signatures in AIS tracks: speed below 1 knot, position variance within 500 m of platform coordinates, duration above configurable threshold | Standby event log with start time, duration and vessel MMSI; interpretable as maintenance or weather-hold indicator |
| Platform activity change alert | Statistical process control on rolling 90-day PSV frequency baseline; alert triggered at greater than one standard deviation shift sustained over five or more consecutive days | Email or webhook alert with platform identifier, basin, magnitude of shift and supporting AIS track visualisation |
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.