Ferry route frequency and passenger demand proxying from AIS
Spaceborne AIS data from Spire Global and exactEarth lets analysts reconstruct actual ferry departure frequencies on island and coastal routes, compare them against published timetables, and infer demand suppression or cancellations at scale.
Sensors
- Spire Global spaceborne AIS: Constellation of over 110 low-Earth-orbit satellites carrying AIS receivers. Provides global coverage including oceanic routes beyond VHF shore-station range. Typical message latency under 90 minutes for any given location; revisit sufficient to capture most port-departure events on routes with hourly or better frequency.
- exactEarth spaceborne AIS: Dedicated maritime AIS constellation offering near-real-time vessel position feeds. Particularly strong on high-traffic coastal corridors and island chains in Europe and Southeast Asia. Historical archive extends back to approximately 2011, enabling multi-year trend analysis.
- Orbcomm spaceborne AIS: Supplementary AIS data stream from Orbcomm's satellite network. Useful for cross-validation of departure events and for filling gaps in single-provider coverage on remote Pacific or Caribbean routes where message collision rates are lower.
- Terrestrial AIS (VDES shore stations): Shore-based VHF Data Exchange System receivers cover ports and coastal waters typically within 40 to 74 kilometres of the coast. Essential for high-resolution sub-minute position logging during berthing and departure manoeuvres, but blind to open-ocean ferry routes.
What a departure record actually contains
Every AIS-equipped vessel broadcasts its MMSI, position, speed over ground, course, navigational status and timestamp. Class A transponders, mandatory on passenger vessels above 500 gross tonnes under SOLAS, transmit at intervals of two to ten seconds when under way and every three minutes at anchor. That cadence is far denser than spaceborne receivers can capture on a single pass, but the aggregate of multiple satellite passes per day is enough to reconstruct port-departure events with reasonable confidence.
The analytic logic is straightforward. A vessel's navigational status field switches from 'moored' (status 5) to 'under way using engine' (status 0) at departure. Speed over ground rises above a threshold, typically one to two knots, and the vessel track diverges from the quay. Combining these signals across all passes for a known MMSI produces a timestamped departure log. Repeat this for every vessel assigned to a route and you have an observed timetable.
Turning raw departures into a frequency signal
The raw departure log becomes analytically useful when compared against a published schedule. Most ferry operators in regulated markets, including EU cabotage routes and Pacific island public-service-obligation routes, file timetables with port authorities or transport ministries. Digitising those schedules and matching them against observed departures produces a simple binary per sailing: operated or missed.
Aggregated over a week or month, the ratio of observed to scheduled sailings is a cancellation rate. Sustained low rates on a route can indicate vessel unavailability, weather-driven suppression, or a commercial decision to reduce frequency in response to weak demand. The direction of inference matters: a ferry operator cutting sailings is a leading indicator of financial stress or route withdrawal, not a lagging one. Investors, port authorities and island governments have used exactly this kind of frequency monitoring to anticipate service changes before formal announcements.
One honest limit: AIS does not carry passenger counts. Frequency is a proxy for demand, not a direct measure of it. A vessel may sail full or empty; the satellite cannot tell. The inference is probabilistic. Routes with high frequency relative to vessel capacity are more likely to be commercially viable than routes with chronic cancellations, but the signal needs to be read alongside ticket-price data, seasonal calendars and population statistics to be actionable.
Where shore-based AIS fails and satellites take over
The practical advantage of spaceborne AIS is coverage. Terrestrial VHF receivers saturate in high-traffic areas and go silent beyond roughly 74 kilometres offshore. For a ferry crossing between two islands 200 kilometres apart, a shore station at each port captures the departure and arrival but misses the transit entirely. Spaceborne receivers have no such horizon.
This matters for island nations in the Pacific, Caribbean and Indian Ocean, where inter-island ferries are often the only affordable transport link and where shore-station infrastructure is sparse or absent. Spire Global's constellation, for example, provides multiple passes per day over every ocean basin, giving analysts enough position fixes to confirm whether a vessel completed its crossing or turned back. A vessel that departs but does not arrive within the expected transit window is a safety and reliability signal, not just a commercial one.
Message collision is the main technical constraint. In busy straits, many vessels transmit simultaneously on the two AIS channels (161.975 MHz and 162.025 MHz), and satellite receivers decode only a fraction of messages. Collision rates in the English Channel or Singapore Strait can cause significant position gaps even with a full spaceborne constellation. Ferry routes in remote island chains rarely face this problem; routes through congested coastal waters may need terrestrial AIS blended in to fill gaps.
Seasonality, weather and the demand-suppression question
Ferry frequency is not constant. Most operators run enhanced summer schedules and reduced winter ones. Any frequency analysis that does not account for published seasonal timetables will misread a planned reduction as a cancellation. The correct baseline is the schedule in force for the period under analysis, not an annual average.
Weather cancellations are a separate category. A cluster of missed sailings during a recorded storm is operationally normal. A cluster of missed sailings in calm conditions is not. Wind and wave data from ERA5 reanalysis or Copernicus Marine Service can be joined to the AIS departure log to separate weather-forced gaps from unexplained ones. This distinction is important for any client trying to assess operator reliability rather than just frequency.
Demand suppression is the subtler inference. If an operator quietly reduces frequency below the published schedule over several months, and the reduction correlates with a period of economic contraction or population decline on the served island, the AIS record provides evidence that the route is under commercial pressure. This is the kind of signal that a regional development bank or a transport ministry might use to decide whether a public-service-obligation subsidy is warranted.
Building the analytic pipeline
The pipeline has four stages. First, vessel identification: compile a list of MMSIs for every ferry operating on the routes of interest, cross-referenced against Lloyd's or IHS Markit vessel registries to confirm vessel type and gross tonnage. Second, AIS ingestion: pull historical and live position feeds from one or more spaceborne providers, filtered to the geographic bounding boxes around each port pair. Third, event detection: apply the departure and arrival detection logic to produce timestamped port-call records. Fourth, schedule comparison: match observed events against the digitised timetable and compute frequency metrics.
Deliverables can be as simple as a weekly CSV of operated versus scheduled sailings per route, or as complex as a live dashboard showing real-time vessel positions alongside a running cancellation-rate indicator. Archive depth varies by provider: exactEarth holds data from around 2011, giving over a decade of history for trend analysis. Spire's commercial archive is somewhat shorter but updated continuously.
Satellize runs this class of AIS analytics as part of its broader satellite-data services. The same pipeline architecture that underpins the Tonga crop-estimation programme, where small-island data scarcity is the central challenge, translates directly to island ferry monitoring, where the problem is identical in structure if different in sensor.
Typical figures
| AIS message types used | Class A position reports (types 1, 2, 3); static and voyage data (type 5); navigational status field |
| Spaceborne revisit (single provider) | Typically 2 to 6 passes per day over any given ocean location; higher at higher latitudes |
| Position fix latency (spaceborne) | Under 90 minutes for Spire Global; near-real-time streams available from exactEarth at higher data cost |
| Positional accuracy | AIS self-reported GPS position: typically better than 10 metres; spaceborne reception does not degrade this |
| Coverage | Global, including routes beyond 74 km from any VHF shore station; no polar exclusion for LEO AIS |
| Minimum detectable vessel | Any SOLAS-mandated Class A transponder (passenger vessels over 500 GT, all vessels over 300 GT on international voyages) |
| Archive depth | exactEarth: from approximately 2011; Spire Global: from approximately 2016; Orbcomm: from approximately 2008 |
| Key constraint: message collision | Decode rates fall in high-traffic straits; can drop below 50% of transmissions in the busiest corridors |
| Delivery formats | CSV position logs, GeoJSON vessel tracks, port-call event tables, schedule-comparison dashboards |
Analytics Satellize can run
| Observed departure frequency per route per week | AIS navigational-status transition detection and speed-threshold filtering to identify port-departure events | Weekly CSV or dashboard table: scheduled vs observed sailings, cancellation rate per route |
| Multi-year seasonal frequency baseline | Historical AIS archive aggregation segmented by calendar week; comparison against digitised operator timetables | Annual frequency profile per route, exportable as time-series chart or GIS layer |
| Weather-adjusted cancellation index | AIS gap detection joined to ERA5 or Copernicus Marine Service wind and wave reanalysis for the same period | Monthly report distinguishing weather-forced from unexplained cancellations per route |
| Vessel utilisation and fleet assignment tracking | MMSI-level voyage reconstruction to identify which vessel operated which sailing; cross-reference against vessel registry for capacity data | Fleet assignment log; flag events where a smaller substitute vessel replaced a scheduled vessel |
| Route withdrawal early-warning signal | Rolling 30-day cancellation rate threshold alert; sustained rates above a defined level trigger notification | Automated alert feed to client dashboard or API endpoint when a route crosses the threshold |
| Inter-island connectivity index | Aggregate operated sailing counts across all routes serving a named island or port cluster, normalised by population or published service-obligation targets | Monthly connectivity scorecard per island; suitable for transport ministry reporting or subsidy review |
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.