Refrigerant GWP Explained: A Ratio, Not a Product Rating
The gas inside a new aircon has been swapped more than once, and no household was consulted. One number drove each move. Reading it correctly turns a confusing run of changes into a single trend you can plan against.
By Team Snowflake | Updated 16 Sept 2026
What does refrigerant GWP actually measure?
Global warming potential compares one gas against carbon dioxide. It states how much warming a kilogram of that gas causes over a stated period, beside a kilogram of carbon dioxide over the same period. Carbon dioxide is the divisor, so it sits at 1 in every published table.
A ratio behaves nothing like a rating. A rating implies a designed scale with a floor, a ceiling and a good end to aim at. This has none of those; it is arithmetic on a physical property, climbing as far as the chemistry carries it.
The figure is silent on most of what a buyer cares about. Flammability and toxicity are graded separately, in their own safety class, and nothing in the number describes how well the gas cools, what pressure it works at, or how much a system holds. Ozone depletion is a third index, measuring a different harm entirely.
GWP is a property of the substance, not of the machine on your ledge. A charge sealed inside a working circuit emits nothing, whatever its number; what converts the ratio into an emission is the mass that gets out.
Why the gas is weighed rather than counted
GWP is stated per unit of weight, and the detail carries further than it looks. Refrigerant is bought, charged and lost by weight. The plate on an outdoor unit states the factory charge in kilograms, alongside the refrigerant designation.
So the exposure tied to any system is that plate figure multiplied by the share that escapes, and two identical machines can sit far apart on that second term. One holds its charge for a full service life and hands it to a recovery cylinder at the end; the other loses gas through a poor flare joint and gets topped up repeatedly.
Why a GWP figure without a time horizon is incomplete
Every GWP is quoted over a stated number of years, and changing them changes the answer. Take R32, listed as HFC-32 and fitted to most new household splits here. The IPCC Sixth Assessment Report gives it 2,690 over 20 years, 771 over 100 years, and 220 over 500 years: one gas, one report, three answers.
That spread comes out of how long each gas survives up there; the report puts R32's atmospheric lifetime at 5.4 years. Carbon dioxide is the yardstick, and much of any release is still working centuries later. Over a short window the fluorinated gas is mostly still present and the ratio runs high; widened, it has long since broken down while the denominator carries on, and the ratio collapses.
Picking the window is a choice about what to weigh, not a fact anybody found. Singapore settles that choice in law: its Environmental Protection and Management Act 1999 defines the term on a 100-year basis, setting one kilogram of a gas against one kilogram of carbon dioxide, using values NEA publishes.
Assessment reports move the numbers as well. The Fifth Assessment Report puts R32 at 677 over 100 years; the Sixth puts the same molecule at 771 over the same span. Nothing about the gas changed between those publications, only the measurement of it.
That gap is not academic here. 771 sits above the ceiling written for household aircon; 677 sits below it. NEA states that the values it works from align with the Fifth Assessment Report, and the statute points at the list the agency publishes. Two accurate sources can disagree by roughly a hundred points, and only one decides what may be supplied here.
| A figure you might be quoted | Where it comes from | What it can settle |
|---|---|---|
| 677 for R32 | IPCC Fifth Assessment Report, 100-year basis, which NEA aligns to | Whether that gas clears the household ceiling here |
| 771 for the same molecule | IPCC Sixth Assessment Report, 100-year basis | The current scientific estimate, which is not the local test |
| 2,690 for the same molecule | IPCC Sixth Assessment Report, 20-year basis | Near-term warming weight, used in climate work |
| 220 for the same molecule | IPCC Sixth Assessment Report, 500-year basis | How little of it remains after five centuries |
| A bare number with no source attached | Nothing anybody can go and check | Nothing at all. Ask for the report and the horizon first |
- A figure you might be quoted
- 677 for R32
- Where it comes from
- IPCC Fifth Assessment Report, 100-year basis, which NEA aligns to
- What it can settle
- Whether that gas clears the household ceiling here
- A figure you might be quoted
- 771 for the same molecule
- Where it comes from
- IPCC Sixth Assessment Report, 100-year basis
- What it can settle
- The current scientific estimate, which is not the local test
- A figure you might be quoted
- 2,690 for the same molecule
- Where it comes from
- IPCC Sixth Assessment Report, 20-year basis
- What it can settle
- Near-term warming weight, used in climate work
- A figure you might be quoted
- 220 for the same molecule
- Where it comes from
- IPCC Sixth Assessment Report, 500-year basis
- What it can settle
- How little of it remains after five centuries
- A figure you might be quoted
- A bare number with no source attached
- Where it comes from
- Nothing anybody can go and check
- What it can settle
- Nothing at all. Ask for the report and the horizon first
A mixture carries a figure of its own
Some refrigerants are single compounds and others are mixtures of two or three. The statutory definition covers both, naming a blend or mixture of greenhouse gases alongside a single gas. A mixture therefore has one GWP attached to it, belonging to the mixture rather than anything inside it.
The slip worth avoiding is reading a component's figure off a table and treating it as the mixture's; the two are rarely the same, and the gap can move a product across a limit.
The 750 ceiling, and why the showroom changed without you
From 1 October 2022, a household aircon may not be supplied in Singapore if its refrigerant carries a GWP above 750. That restriction lives in the same Act. It binds importers, manufacturers and sellers rather than owners. Supplying a regulated good that fails it carries a fine of up to $10,000 on conviction.
The ceiling reaches a defined class of machine, not a brand or a price band. NEA writes a household air-conditioner here as a single-phase, non-ducted room unit rated at 17.6 kW or lower, with variable refrigerant flow equipment left out. Ordinary household splits sit well inside that.
750 is not a national figure either. It is written per category, and the categories diverge sharply: a household refrigerator sits at 15 under the same scheme, as does an electrically driven water-cooled chiller of 1,055 kW or more. The number follows what the equipment is and how much gas it carries.
Nothing about the rule looks backwards: NEA is explicit that the restriction reaches new sale or supply from that date onward, and equipment already installed beforehand is untouched. Suppliers were given until 30 September 2023 to clear non-compliant stock, so no owner picked up a duty then.
A ceiling written as a number erases everything above it from new supply. Manufacturers do not argue with a threshold; they withdraw the ranges that fail and engineer around whatever clears. A shopper met all of that as a changed model line and an unfamiliar sticker, with no visible link to a government table.
Three rules, three different measures
The GWP ceiling gets confused with the energy schemes constantly, and they are separate instruments doing separate jobs. This one governs the gas, measured as a warming ratio. The minimum efficiency standard governs how much electricity a model may draw for the cooling it produces, a ratio of an entirely different kind. The labelling scheme governs what must be printed on the unit and registered.
A model satisfies all three before it can be put in front of you. Clearing any one says nothing about the other two. Each of the three has its own guide on this site. Treating them as one single approval is how a shopper ends up believing a compliant refrigerant implies an efficient machine.
What GWP leaves out of the environmental account
The ratio accounts for what escapes, never for what the system consumes. A split system in a Singapore flat runs across much of the year, and the electricity behind that is a continuing cost for its service life. Refrigerant is a stock that counts only once it leaves the circuit, and which of the two ends up larger depends on how much gas gets out and how much power the machine draws.
The escaped weight is the term almost nobody quotes: GWP is stated per kilogram, so the real emission is leaked weight times the ratio. A circuit that holds its charge and is emptied into a recovery cylinder has released close to nothing; the same system losing gas yearly and vented on removal has released its charge several times over. Where an old charge ends up is covered separately.
Push back when a low GWP figure is offered as the environmental case for a purchase. It is one input, already constrained by law, and every model lawfully in front of you cleared the same bar. What separates two machines is the efficiency figure, capacity matched honestly to the room, and an installation tight enough not to leak.
Replacement is the argument to watch for, and none of this supports it. Building and disposing of a system carries its own cost, and the old charge is released only if somebody chooses to. Nothing in the ceiling asks an owner to do anything, and anybody using it to press for replacing a machine that still cools has travelled well past what the rule says.
What the number tells a buyer, and what it tells an owner
For a purchase today the question has largely been answered before you walk in. Every household split that may lawfully be sold here sits under 750, so the gas is not a lever anybody pulls between two compliant models. Which gas a system can physically take is a separate question, answered elsewhere on this site.
Read GWP as a supply signal rather than a product feature. A gas sitting well above the current ceiling is one whose new production is being squeezed by international agreement, and that surfaces as price and waiting time on a refill. It reaches an owner at one moment: the day a sealed circuit is opened and refilled.
Holding a system on a heavier gas is not a problem on a schedule. A sealed charge emits nothing, and swapping out a working machine to change its gas carries costs of its own. What genuinely shifts is the arithmetic on one particular fault: once the circuit needs refilling, the gas going back in comes from a supply that keeps shrinking.
Ask which refrigerant the model uses, then ask for its number and its source. A seller who names the gas, states the GWP and can say which report and window produced it is describing something checkable. A figure with no report and no horizon behind it has not been sourced.
- Ask which assessment report a quoted figure came from. The Fifth and Sixth disagree about the same molecules, and only one governs supply here.
- Ask over how many years. A 20-year figure and a 100-year figure for one gas can differ more than threefold.
- Treat the rule as a minimum, not a selling point. Every new household split on sale here already sits under the ceiling.
- Ask what the system holds in kilograms, not only what gas. The ratio is per kilogram, so the charge weight is half the answer.
- On an installed system, keep the attention on whether the circuit is tight. A charge that never escapes never becomes an emission.
| Where you are standing | What GWP decides there | What to weigh instead |
|---|---|---|
| Comparing new household splits | Nothing between them. All of them cleared the same ceiling | The efficiency figure, and whether the sizing fits the room |
| Living with an older machine that holds a heavier gas | Nothing at all while the circuit stays shut | Ordinary servicing, and cooling that starts to fade |
| Facing a leak on that older system | The refill comes from a supply being deliberately reduced | Getting the leak found and proven before gas goes back in |
| Having a system taken away | The whole charge is decided in a single afternoon | Whether that charge is recovered or simply let go |
| Handed a GWP number by a seller | Nothing, until the report and the window are named | Ask which assessment report, and over how many years |
- Where you are standing
- Comparing new household splits
- What GWP decides there
- Nothing between them. All of them cleared the same ceiling
- What to weigh instead
- The efficiency figure, and whether the sizing fits the room
- Where you are standing
- Living with an older machine that holds a heavier gas
- What GWP decides there
- Nothing at all while the circuit stays shut
- What to weigh instead
- Ordinary servicing, and cooling that starts to fade
- Where you are standing
- Facing a leak on that older system
- What GWP decides there
- The refill comes from a supply being deliberately reduced
- What to weigh instead
- Getting the leak found and proven before gas goes back in
- Where you are standing
- Having a system taken away
- What GWP decides there
- The whole charge is decided in a single afternoon
- What to weigh instead
- Whether that charge is recovered or simply let go
- Where you are standing
- Handed a GWP number by a seller
- What GWP decides there
- Nothing, until the report and the window are named
- What to weigh instead
- Ask which assessment report, and over how many years
Common questions
What does a GWP limit mean for a new aircon in Singapore?
Why does the same refrigerant carry two different GWP figures?
Does a high GWP affect how well an aircon cools?
Sources
- Registered Goods Product Search (Air-Conditioner)
National Environment Agency · Checked
Registered household models and their declared refrigerants.
Ready to get started?
Tell us what’s going on. Symptoms, setup, photos, anything we should know. We’ll assess and come back with the right next step.