Capacity looks like the simplest number on an asset page. It causes more wrong analysis than anything else on the site.
The reason is that “capacity” is not one quantity. It is a family of related quantities that get quoted with the same word, and four specific confusions between them account for nearly every disagreement you will find between two published figures.
One: the unit
LNG capacity appears in at least four incompatible units, and they measure genuinely different physical things.
Mtpa is mass of LNG per year. bcm/y is volume of gas per year. Cubic metres is volume of liquid, used for ships and tanks rather than plants. MMBtu is energy, which is what the cargo is actually priced in.
Converting between them requires an assumption about gas composition, because a rich cargo carries more energy per tonne than a lean one. This site uses 1 Mtpa equals 1.36 bcm/y throughout, which is the convention Global Energy Monitor uses. Another source using 1.38 will produce figures around 1.5 per cent different for no reason other than the assumption, and neither is wrong.
The unit converter sets out the factors used here and what each one assumes.
Two: the status
This is the error that produces the worst numbers, because the arithmetic is easy and the result is confidently wrong.
An LNG terminal in a tracker may have trains that are operating, under construction, and merely proposed, all at once. Reporting a single figure for the site means adding those together, and the sum describes nothing: not the capacity today, not the capacity if everything is built, and certainly not a forecast, because a large share of proposed capacity is never built at all.
So every capacity figure here is broken out by status, the totals are never combined, and the status explainer sets out what each of the nine values means. When you see a headline figure for global LNG capacity somewhere else, the first question to ask is which statuses it includes.
Three: nameplate against output
Nameplate is a design specification. It says what the plant was built to produce under stated reference conditions.
Real plants routinely beat it. Refrigeration works better in cold weather, so a plant in a cold climate can run several per cent above nameplate through winter, and operators progressively find capacity through debottlenecking that was never in the original design. Equally, real plants routinely miss it: heat, maintenance, feed gas constraints and unplanned outages all take output down.
Neither direction is an error in the data. It means the number is a specification, and using it as a measurement of what was produced last year will be wrong in an unpredictable direction.
Four: repeated figures summed
This one is specific to how trackers record projects made of parts, and it is easy to miss.
Global Energy Monitor records pipelines segment by segment, and a project-level figure — the merged length, or the design capacity of the whole system — is frequently repeated on every segment row. Sum those rows naively and a two-segment project doubles, a five-segment project quintuples.
The ingest here handles it explicitly: where every segment reports the identical value, it is counted once and labelled repeated; where the values differ, they are genuinely per-segment and are summed. Any pipeline page showing a length or capacity says which basis was used. It is worth checking the same question of any other aggregate you meet.
What to check before using a number
Four questions, in order.
Which unit, and converted with what assumption? If the figure crossed between mass, volume and energy, an assumption was made somewhere.
Which statuses does it include? If the answer is “all of them”, the figure is not usable for anything.
Nameplate or actual? If it is nameplate, it does not tell you what was produced.
Is it an aggregate, and how was it aggregated? Sums across parts of a project are where double counting lives.
What this site does
Capacity is reported by status and never combined across statuses. Mtpa and bcm/y appear together on terminal pages so no conversion is hidden. Pipeline lengths and capacities state whether they were summed or counted once. A field the tracker leaves empty renders as “Not reported”, which means the source does not carry it — never that the value is zero.
None of that makes the figures more precise than their source. It makes the imprecision visible, which is the most a directory can honestly offer.