LNG terminal throughput inference from vessel activity
Combining AIS tracking with Sentinel-1 SAR and high-resolution optical imagery lets analysts estimate LNG terminal throughput without relying on reported trade data. Draught changes, berth occupancy, and vessel class together reveal whether a carrier loaded or discharged.
Sensors
- Spire Global AIS (space-based): Satellite AIS reception covering open ocean and remote terminals where terrestrial receivers are absent. Spire operates more than 100 LEO cubesats, giving sub-hourly vessel position updates for most major shipping lanes and terminal approaches. Position accuracy is typically within 10 m. Gaps and spoofed positions are the key limitation this workflow is designed to address.
- Sentinel-1 SAR (C-band, ESA): Provides 10 m ground-range-detected imagery in IW mode with a 250 km swath, and 6-day repeat at mid-latitudes (12-day for a single satellite). SAR penetrates cloud and works at night, making it the primary tool for waterline geometry and berth-occupancy confirmation when optical is unavailable. Interferometric coherence can also detect whether a vessel has moved between passes.
- Planet Dove optical: Planet's Dove constellation delivers 3–5 m resolution imagery with near-daily revisit globally. Useful for confirming vessel presence, identifying vessel class by deck geometry, and reading visible loading-arm engagement. Cloud cover is the hard limit; equatorial and monsoon-affected terminals may lose several consecutive days of coverage.
- Maxar WorldView-3: 30 cm panchromatic, 1.24 m multispectral, tasked on demand. At this resolution, LNG carrier class is unambiguous, loading arm position is readable, and hull waterline can be estimated optically to within roughly 0.5 m under calm sea conditions. Tasking latency is typically 1–3 days; cost limits routine use to high-priority events.
Why reported trade data is not enough
National customs agencies, port authorities, and industry data vendors publish LNG cargo figures, but with lags that can reach four to six weeks and with coverage gaps at terminals in jurisdictions with limited reporting obligations. Spot-market traders, sovereign energy ministries, and infrastructure investors often need a faster, independent signal.
Space-based observation does not replace customs data. It produces a parallel, physically grounded estimate that can flag discrepancies, fill gaps, and provide earlier confirmation of cargo movements. The combination is more reliable than either source alone.
What a waterline gives away
A fully laden LNG carrier sits noticeably lower in the water than an empty one. For the largest Q-Flex and Q-Max vessels, the difference between ballast and laden draught can exceed 5 m. For standard 138,000–174,000 m³ vessels, the range is typically 4–6 m. That difference is detectable in SAR imagery through the geometry of the radar return from the hull: a laden vessel presents a taller, more reflective freeboard target, and the waterline position shifts measurably relative to known vessel dimensions.
This is not a precise volumetric measurement. SAR-derived draught estimation carries uncertainty of roughly 0.5–1.5 m depending on sea state, vessel heading relative to the radar look angle, and the availability of reference dimensions from the vessel registry. What it reliably distinguishes is the binary state: did this vessel arrive light and leave heavy, or the reverse? That binary, combined with vessel capacity from public registry data, produces a cargo-volume estimate with uncertainty of perhaps 10–20% for a single voyage.
Reconstructing a port call when AIS goes quiet
AIS transponder manipulation is well documented in LNG shipping. Vessels may disable or spoof their transponders near sanctioned terminals, or simply experience satellite coverage gaps in high-traffic areas where signal collision degrades reception. Spire's constellation reduces but does not eliminate these gaps.
When AIS position reports are absent or inconsistent, SAR and optical imagery provide independent confirmation of vessel presence at berth. A Sentinel-1 pass over a terminal every 6–12 days catches most port calls for large carriers, which typically spend 12–36 hours at berth for loading and 10–24 hours for discharge. Planet Dove's near-daily optical coverage fills the gaps between SAR passes where cloud permits. The workflow cross-references AIS arrival and departure timestamps against imagery timestamps to bound the port-call duration and identify cases where transponder records are implausible.
Building a throughput time series
Individual port calls are aggregated into a monthly throughput estimate per terminal. Each confirmed cargo event is assigned a volume range derived from vessel class and laden/ballast state. The sum across all events gives an estimated export or import volume, expressed in million metric tonnes per annum (MMTPA) equivalent or in LNG cargo equivalents.
Archive depth matters here. Sentinel-1 data is available from 2014, and Planet Dove from roughly 2016 at useful resolution. Spire commercial AIS extends back several years. This allows analysts to construct multi-year baselines, identify seasonal patterns, and detect step changes in terminal utilisation that precede official announcements by weeks.
The honest caveat is that throughput estimates derived this way are probabilistic ranges, not audited figures. Uncertainty accumulates from vessel-class misidentification (rare with WorldView-3 imagery, more common with 3 m optical), from draught estimation error, and from port calls that fall entirely between satellite passes. For major export hubs with high vessel traffic and good AIS coverage, the aggregate monthly estimate is typically within 10–15% of subsequently reported figures. For smaller or more opaque terminals, uncertainty is wider.
Where the method reaches its limits
Persistent cloud cover is the most common operational frustration. Several major LNG import terminals in South and Southeast Asia experience weeks of cloud during monsoon season, forcing reliance on SAR alone. SAR handles berth occupancy well but gives less confident vessel-class identification than optical at equivalent resolution.
Terminals with multiple berths and high simultaneous occupancy are harder to parse. When three or four vessels are present at once, attributing cargo events to specific berths requires careful geometric registration of imagery to terminal infrastructure maps. Errors in that registration propagate into the throughput estimate.
Sanctioned terminals present a specific challenge: vessels calling at them are more likely to have manipulated AIS, to use ship-to-ship transfers offshore rather than berth calls, and to use older, less well-documented vessels whose registry dimensions are uncertain. The method degrades in exactly the environments where independent intelligence is most valuable.
Putting it into practice
A practical monitoring programme for a single major terminal typically requires weekly SAR tasking or Sentinel-1 scheduling, daily Planet Dove tasking with cloud-contingency protocols, and continuous AIS ingestion from a space-based provider such as Spire. Periodic WorldView-3 tasking, triggered by ambiguous optical or SAR results, resolves edge cases.
Satellize structures analytics of this type as a recurring feed: each confirmed port call generates a structured event record with vessel identifier, estimated laden state, volume range, and confidence flag. Monthly summary reports aggregate these into throughput estimates with explicit uncertainty bounds. The Overhead column has covered the public methodology behind similar approaches for readers who want the technical background before commissioning an engagement. A first step for prospective clients is a terminal-specific feasibility review, which assesses historical imagery availability, AIS coverage quality, and expected uncertainty for the specific geography.
Typical figures
| SAR spatial resolution (Sentinel-1 IW mode) | 10 m ground range, 250 km swath |
| SAR revisit (Sentinel-1, single satellite) | 6 days at mid-latitudes; 12 days near equator with single satellite |
| Optical resolution (Planet Dove) | 3–5 m, near-daily global revisit |
| Optical resolution (Maxar WorldView-3) | 30 cm panchromatic, 1.24 m multispectral, tasked on demand |
| AIS position accuracy (Spire satellite AIS) | Typically within 10 m; update frequency sub-hourly for major shipping lanes |
| Draught estimation uncertainty (SAR-derived) | 0.5–1.5 m depending on sea state and vessel heading |
| Single-cargo volume uncertainty | Approximately 10–20% for standard LNG carriers with known registry dimensions |
| Monthly throughput estimate accuracy (major hubs) | Typically within 10–15% of subsequently reported figures |
| Archive depth | Sentinel-1 from 2014; Planet Dove from ~2016; Spire AIS several years |
| Delivery formats | GeoJSON event feed, GeoTIFF imagery chips, CSV throughput time series, PDF monthly report |
Analytics Satellize can run
| Port-call event log | AIS track fusion with SAR and optical berth-occupancy detection; vessel identification from registry cross-reference | Structured GeoJSON feed: vessel MMSI, arrival/departure bounds, berth ID, confidence flag, per confirmed call |
| Laden/ballast state classification | SAR waterline geometry analysis against published vessel class dimensions; binary classification with uncertainty score | Per-event laden-state flag (loaded/discharged/uncertain) appended to port-call log |
| Single-voyage cargo volume estimate | Vessel class capacity from public registry combined with laden-state classification; probabilistic volume range | Volume estimate in cubic metres LNG with explicit uncertainty bounds, per port call |
| Monthly terminal throughput time series | Aggregation of per-voyage estimates; gap-filling protocol for missed passes; comparison against prior-period baseline | CSV time series and PDF summary report with MMTPA equivalent and confidence interval |
| AIS anomaly detection | Cross-validation of AIS position records against SAR/optical imagery; flagging of transponder-off periods and implausible track segments | Alert feed of AIS anomalies at monitored terminals, with supporting imagery chips |
| Multi-terminal flow balance | Pairing of export and import events by voyage reconstruction across terminal pairs; net flow estimation | Trade-flow matrix (GIS layer and spreadsheet) for selected terminal pairs, updated monthly |
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.