Every rule in LNG safety — exclusion zones, gas detection, ESD systems, the shape of a terminal itself — follows from a short sequence of physics. Understanding that sequence once makes the rest of the rules obvious rather than arbitrary.
The sequence
Spill. LNG escapes containment, onto ground, onto water, or into the air.
Rapid boiling. The liquid is roughly 160 degrees colder than its surroundings, so it boils violently on contact — with concrete, with soil, with water. On water this boiling can be fast enough to produce a rapid phase transition, a physical explosion with no fire involved.
A cold, heavy cloud. The vapour leaving the pool is far colder than ambient air, and cold gas is denser than warm gas. So immediately after a release, the cloud is heavier than air and hugs the ground, spreading rather than rising.
Warming and dispersion. As the cloud entrains ambient air, it warms. Methane itself is lighter than air at normal temperature, so once the cloud has warmed enough — somewhere around minus 110 degrees Celsius — it becomes buoyant and lifts away, diluting as it goes.
A flammable window. Only where the mixture sits between roughly 5 and 15 per cent methane in air will it ignite. Closer to the source the mixture is too rich; far enough away it is too lean. The flammable region is therefore a band, not the whole cloud, and it moves and shrinks as the cloud disperses.
Why this is different from other hydrocarbon spills
LPG and petrol vapours are heavier than air because of their molecular weight, and they stay heavier as they warm. A propane cloud pools in low ground and lingers.
LNG’s heaviness is a temperature effect, not a molecular one, and it fades as the cloud warms. That is why LNG hazard modelling treats dispersion as a race between cooling-driven sinking and warming-driven lifting, rather than assuming persistent ground-hugging behaviour throughout.
It is also why LNG incidents differ in character from other fuel spills: the danger zone is transient and shifts with wind and weather far more than a design based on a static heavy-gas assumption would suggest.
What actually causes harm
Three distinct mechanisms, each requiring a different part of the sequence above.
Thermal radiation from a fire — either a pool fire at the spill site or a flash fire if the drifting vapour cloud finds an ignition source before it disperses. This is the dominant hazard in siting analysis, because heat travels well beyond the fire itself.
Asphyxiation, in confined or low-lying spaces where vapour has displaced oxygen. This does not require ignition at all; it is a hazard from the gas itself.
Cryogenic contact, causing burns to skin and embrittling materials it touches. Localised, but immediate for anyone or anything nearby during a spill.
Notably absent from that list: LNG does not burn as a liquid under any circumstances relevant to an incident, and there is no scenario in ordinary handling where the liquid itself ignites.
Why the industry’s response looks the way it does
Gas detectors are placed to catch a cold, low-lying cloud before it reaches an ignition source, which is why they sit low and near likely release points rather than high in a structure.
Exclusion zones are sized from dispersion modelling of exactly this sequence: how far the cloud could plausibly travel while still within the flammable band.
Emergency shutdown systems exist to stop a release fast, because every stage above scales with how much escapes and for how long.
None of this is exotic engineering. It is a direct, close reading of what a spill actually does, applied consistently across ships, terminals and pipelines.
Where this sits in the broader picture
This module sets out the physics; the following ones in this track cover the specific hazards (rapid phase transition and pool fires), the codes that govern design (the IGC Code and SIGTTO), how exclusion zones are actually set, and the separate question of emissions, which is an environmental rather than a safety concern and should not be conflated with it.