A cargo of LNG is worth somewhere in the tens of millions of dollars. What the buyer pays is decided by a measurement taken over a few hours at the ship’s manifold, and nobody measures the thing being sold directly.
Energy is calculated, not measured
LNG is sold by energy, in MMBtu, because energy is what the buyer wants. There is no instrument that reads energy off a tank.
So it is computed. Volume comes from the tank gauges. Density comes from an equation using the measured temperature and the analysed composition. Multiply the two for mass, multiply mass by the gross calorific value derived from the same composition, and you have energy. Then subtract: vapour displaced back to shore during loading, and gas consumed by the ship’s engines during discharge.
Every input to that chain carries uncertainty, and they accumulate.
What is measured
Level. Certified gauges, usually radar or float, read the liquid surface in each tank to a few millimetres. On a tank tens of metres across, a few millimetres is a substantial volume, which is why the gauges are calibrated instruments subject to survey rather than operational indicators.
Temperature. Sensors at several heights in each tank, because the liquid is not uniform. Temperature feeds the density calculation and confirms the cargo is at the expected condition.
Pressure and vapour temperature. Needed to account for the vapour space, which contains a real quantity of methane that has to be included on one side of the transaction or the other.
Composition. A sample is drawn and analysed by gas chromatograph. This is where the calorific value comes from, and it matters more than anything else: two cargoes of identical volume can differ by several per cent in energy depending on how much ethane and propane they carry.
Trim and list. The ship’s attitude changes the relationship between level and volume, so both are recorded and corrected for.
Where the error lives
Composition analysis is the largest single contributor to uncertainty in most transfers, because a small sample has to represent a very large volume, and sampling a boiling cryogenic liquid without changing it is genuinely hard. Sample vaporisers, line purging and repeat analyses all exist to manage that.
Level measurement is next, for the reason above: the tanks are enormous and the lever arm on any error is long.
Overall uncertainty typically lands within a fraction of a per cent. That sounds reassuring until it is multiplied out, at which point a routine measurement uncertainty is worth more than most people’s houses. Hence independent surveyors, witnessed procedures, and contracts that specify in advance how a disagreement is resolved.
Ageing complicates it
The cargo that arrives is not quite the cargo that loaded. Methane boils off preferentially through the voyage, so what remains is progressively heavier and higher in calorific value per unit volume — ageing, and on a long voyage it is measurable.
That is a real effect rather than a measurement artefact, and it is why loading and discharge are measured independently rather than one being inferred from the other. The difference between them is not error; it is boil-off consumed, boil-off vented or reliquefied, and composition change, and the contract has to say who bears each.
Why it appears on this site only indirectly
Custody transfer data is commercial, cargo by cargo, and none of it is public. Nothing here reports it.
Where it does surface is in the unit conventions. The reason this site quotes capacity in Mtpa and bcm/y together, states the conversion factors it assumes, and warns that two sources converting with different assumed compositions will disagree, is the same reason custody transfer needs a chromatograph: a cubic metre of LNG is not a fixed amount of energy, and any figure that implies it is has hidden an assumption.