Take a look at the carbon footprint on almost any product label and you'll see a number followed by "CO2e." A flight, a steak, a t-shirt, a kilowatt-hour of electricity. Same unit every time. It's convenient. It's also where almost everyone gets tripped up, because two activities reporting the same tonnage of CO2e can have wildly different consequences for the climate—and nobody bothers to tell you that.
I ran into this the first time I tried to make sense of my own company's footprint. We had a nice round figure, a few dozen tonnes of CO2 equivalent, and I felt pretty good about it. Then a colleague who actually understands atmospheric chemistry asked me one question: "How much of that is methane?" I didn't know. That's the moment I realized that understanding CO2 equivalent and greenhouse gas metrics isn't about the number at all—it's about what the number is hiding.
Key takeaways
- CO2e is a conversion unit, not a measurement. It folds different gases into a single figure using a multiplier called the global warming potential (GWP).
- The GWP multiplier depends entirely on the time horizon you pick. Methane is roughly 28 times CO2 over 100 years, but around 80 times over 20 years.
- Two companies can report identical CO2e totals and have completely different climate impacts if one is methane-heavy and the other is CO2-heavy.
- Choosing GWP100 versus GWP20 is a policy decision, not a physics fact. It changes who looks clean and who looks dirty.
- If you only track the aggregate number, you'll optimize the wrong things.
Why CO2 equivalent is a translation, not a measurement
Here's the thing nobody puts on the label: you cannot weigh a gas and get a climate impact. You weigh a gas, and you get a mass. Turning that mass into something comparable across methane, nitrous oxide, refrigerants and carbon dioxide requires a series of judgment calls.
Each gas traps heat differently. A molecule of CO2 hangs around for centuries and absorbs a modest amount of infrared radiation. A molecule of methane is a far more aggressive heat-trapper while it exists—but it breaks down in the atmosphere within a couple of decades. Nitrous oxide sits somewhere in between, and it lingers. Sulfur hexafluoride, used in electrical switchgear, is almost absurdly potent.
The multiplier you never see
To compare them, climate scientists define a reference: one unit of CO2, over a set period. Then they ask, "over that same period, how much more heat does one unit of this other gas trap?" The answer is the global warming potential, or GWP. Multiply the mass of any gas by its GWP and you get CO2e.
The formula, in practice:
CO2e = activity data × emission factor × GWP
Activity data is what you actually did—litres of diesel, cubic metres of gas, kilograms of refrigerant topped up. The emission factor converts that into a mass of the specific gas. The GWP converts that gas into CO2e. Three separate assumptions stacked on top of each other, and the last one is the one most people never examine.
I'll be honest: for the first couple of years of doing this, I treated the GWP as a fixed constant, like a conversion rate for currency. It isn't. It's a choice, and the choice has consequences.
GWP100 versus GWP20: the horizon changes everything
When you see a GWP figure quoted—for methane, say—there's a hidden variable attached. Over what period? Because methane's impact is front-loaded. It hits hard and fades fast. CO2 hits gently and stays for millennia.
So the same kilogram of methane looks like two very different problems depending on your window:
| Gas | Approximate GWP over 20 years | Approximate GWP over 100 years |
|---|---|---|
| Carbon dioxide (CO2) | 1 | 1 |
| Methane (CH4) | ~80 | ~28 |
| Nitrous oxide (N2O) | ~270 | ~265 |
Those numbers are non-attributed orders of magnitude, and they shift slightly with each assessment cycle—but the pattern is stable and it's the whole point. Methane's multiplier nearly triples when you shrink the horizon.
Why the standard uses 100 years
Most reporting frameworks default to GWP100. It's the convention, and there's a decent reason: a century is roughly the span over which the climate system responds to a pulse of emissions, and it's long enough to smooth out short-term noise. It's also, frankly, convenient for anyone whose emissions are methane-heavy, because it makes those emissions look less than a third as bad as a 20-year view would.
That's not a conspiracy. It's a legitimate methodological choice with a real political edge to it. And I'd argue that if you're serious about near-term warming—the kind that plays out within your own strategic planning horizon—GWP100 flatters you. I've made that mistake myself. Our first inventory used GWP100 exclusively, and the report looked calm. Recalculating on a 20-year basis was uncomfortable.
What about GTP and GWP*?
Two alternatives deserve a mention, because they're where the field is actually moving.
- GTP (global temperature change potential) measures the temperature response at a chosen date rather than cumulative heat trapped. Subtler, harder to communicate, rarely used in voluntary reporting.
- GWP* is designed specifically for short-lived gases like methane. Instead of treating a steady methane source as a permanent warming commitment, it accounts for the fact that constant methane emissions stabilize temperature rather than pushing it up indefinitely.
- Neither has displaced GWP100 in mainstream corporate accounting. Yet.
GWP* is genuinely interesting and genuinely contested. It can make a stable herd of cattle look climate-neutral, which is either an accurate reflection of physics or a very convenient accounting trick, depending on whom you ask. I lean toward "partly both."
A worked example: from activity data to CO2e
A mid-sized warehouse I worked with burned roughly 30,000 litres of diesel in its yard vehicles over a year. Diesel combustion releases about 2.7 kg of CO2 per litre. So the CO2 portion alone was around 81 tonnes.
But that's not the whole story, and this is the part the label hides. Diesel combustion also produces methane and nitrous oxide, in small quantities. Multiply those masses by their GWPs and you add a few more tonnes of CO2e—small in this case, because the masses are tiny.
Now flip it. Same warehouse, but suppose it also leaks refrigerant from a cold-storage unit. Say 4 kg of an HFC escapes over the year. That refrigerant has a GWP in the thousands. Four kilograms becomes tens of tonnes of CO2e—more than a quarter of the diesel figure, from a leak nobody noticed.
Which number got attention in the original report? The diesel. Because it was big and obvious and had a nice line item. The refrigerant leak was a rounding error in the spreadsheet and a serious problem in reality. That's what happens when you let one aggregate figure do all the talking.
Why you should never manage a single CO2e number
Once you understand that CO2e is a weighted sum, the strategic implications become obvious—and slightly uncomfortable. A total tells you nothing about the shape of your emissions. Two organizations, both at 1,000 tonnes CO2e:
- Company A: 980 tonnes from CO2, 20 tonnes from a bit of methane.
- Company B: 400 tonnes from CO2, the rest from methane and refrigerant losses.
Annual reports would present these identically. But Company B has a much sharper near-term warming effect, and its emissions are far more fixable—plug the leaks, switch the fuel, and the number collapses fast. Company A's CO2 is stickier.
So here's my opinion, and I'll defend it: any carbon inventory that reports only a total, without a gas-by-gas breakdown and the GWP basis used, should be treated as decorative. It's not lying. It's just not saying anything useful.
The practical checklist
- Ask which GWP basis was used, and which assessment cycle. If the answer is vague, the number is vague.
- Request the split by gas, at least CO2 / CH4 / N2O / fluorinated gases.
- Check whether refrigerant top-ups are included. In my experience they're the most commonly omitted item and often the most concentrated hit.
- Run the same inventory on a 20-year horizon once, as a stress test. It costs an afternoon and it changes how you prioritize.
None of this requires a climate science degree. It requires asking one extra question at the point where someone hands you a number.
A metric is a lens, not a verdict
The phrase "one tonne of CO2 equivalent" sounds like a fact about the world. It's closer to a fact about a spreadsheet—a useful, standardized, deeply imperfect translation of physical reality into something you can compare and act on.
That's not a reason to distrust it. It's a reason to read it like an engineer reads a spec sheet: with attention to the assumptions underneath. The GWP is the assumption that matters most, and it's the one printed in the smallest font, if it's printed at all.
So the next time a supplier, a label, or your own sustainability report hands you a round number in tonnes of CO2e, ask yourself: a hundred years from when, and weighted toward which gas? Whoever can answer that question clearly is the person you want running your carbon accounting. Everyone else is just reading the label.