Skip to content
PunarGen

Why two published carbon factors for the same material disagree

Seven published libraries price steel between 1.56 and 2.86 kg CO₂e per kg — an 84% spread. None of them is a mistake. Here is what actually moves a factor, and how to tell which one answers your question.

6 min read · last checked

Look up the carbon factor for steel and you will find a number. Look it up again somewhere else and you will find a different one. This is the point at which most people conclude that one of the sources is wrong, pick the one they like, and move on.

Neither is wrong. Here are seven, all published, all competently produced, side by side — the full table is on the steel comparison page.

kg CO₂e/kgLibraryWhat the source calls it
1.56USLCI (United States)Steel, hot rolled coil
2.07ELCD 3.2 (Europe)Steel hot rolled coil, blast furnace route, 2–7 mm
2.14BAFU:2026 (Switzerland)Steel, low-alloyed, at plant
2.21Base Carbone (France)New steel or sheet metal · manufacture
2.29Ökobaudat (Germany)Steel and iron
2.78EF node (Europe)Steel sheet, primary and secondary mix
2.86DESNZ 2026 (UK)steel cans — primary

Highest is 84% above lowest. Read the third column before the first.

The third column is doing most of the work. A hot rolled coil, a low-alloyed billet and a steel can are not the same product, and a figure is an answer to whatever question its row label asks.

1. System boundary — usually the largest term

A cradle-to-gate figure ends at the factory gate, but which gate? PET makes this visible because two of the four published figures say so in their own names:

kg CO₂e/kgLibraryWhat the source calls it
2.80BAFUPolyethylene terephthalate, granulate, amorphous
3.27Base CarbonePolyethylene terephthalate · PET · new
3.44ELCD 3.2PET granulate, production mix, at plant
3.86DESNZ 2026PET (incl. forming) — primary

Three are granulate. One says "incl. forming" — it has already paid for turning the pellets into a bottle.

The 38% spread on PET is not four laboratories disagreeing about polyester. Three of them stopped at the pellet and one carried on to the moulded article. If your bill of materials is in granulate, the DESNZ figure counts a step you have not reached; if your BOM is finished bottles and you use a granulate figure, you are missing one. Neither is an error until you pick the wrong one.

2. Geography — the grid the material was made on

Steel is largely an energy product, and energy is local. A tonne made on a coal-heavy grid and a tonne made on a hydro-heavy one differ before anything about the steelmaking changes. The same is true of every electricity-intensive material — aluminium most of all.

This is why the assessment engine behind this site consults exactly one library per question rather than searching all eight and returning whatever hits. A figure borrowed from the wrong geography is not an approximation of the right one; it is a description of a different supply chain.

3. Characterisation method

The seven steel figures are characterised under six different methods — TRACI 2.2, ReCiPe 2016, IPCC 2021, EF 3.1 and others. For global warming the methods agree closely, so this moves a carbon number least. It matters enormously the moment you look at any other indicator: ReCiPe reports fossil resource use in kg oil eq and EN 15804 reports it in MJ, about 42 MJ to the kilogram, and the two have been confused by people who should know better.

Two figures characterised under different methods should never be summed, averaged or compared to the second decimal place. On the comparison pages they sit in separate rows for that reason.

4. And the one that is not a disagreement at all

Sometimes a spread is enormous and the explanation is that the sources are talking about different things. Cement:

kg CO₂e/kgLibraryWhat the source calls it
0.0506Base CarboneGravel/cement mix · manufacture
0.73BAFUCEM I cement, at plant
0.93AusLCIGeneral purpose cement

A factor of 18. A gravel/cement mix is mostly gravel — it is not cement.

Two of these are cement and one is a mix that contains some. The same happens with leather, where one library prices a raw beef hide at 0.35 and another a polyurethane leather lacquer at 6.74. Neither pair is a disagreement; the material name is simply doing a job it cannot do.

An apparent order-of-magnitude disagreement between reputable factor sets is almost always two different products wearing one name. That is also exactly how a wrong number ends up in somebody’s report — the gap looks like uncertainty, so it gets averaged.

Of the 41 materials priced by more than one library here, 7 are wide enough to fall into this category and are flagged as such rather than presented as disagreements. The remaining 34 are genuine comparisons.

So which figure should you use?

The one whose boundary and geography match your question — not the lowest, and never an average of them. Averaging two figures built on different boundaries produces a number that is true of nothing and cannot be defended to anyone who asks how it was derived.

In practice that means reading the row label first and the number second, which is the opposite of what a spreadsheet encourages. Every figure on this site carries its source row, its geography, its method and its licence for that reason.

Check it yourself

Every figure above is on this site with its source, geography, method and licence.

Use this

These are the mistakes the assessment engine is built to avoid making. To watch it not make them on a product of your own, run a bill of materials through the free carbon calculator — no sign-up, and it runs in your browser. For the whole picture, including the process and its scrap, model the whole route in the prototype, or see what TrueGauge does for a manufacturer.

Early access

Request early access.

Both products are pre-launch. We're working with an early group of manufacturers, architects, builders and sustainability teams to shape what gets built first — tell us which platform is relevant to you.

  • Shape the roadmap

    Tell us which decisions you need supported, and what the data has to do to be usable.

  • First access

    Be first in line as each platform opens to early users.

  • A direct line to the team

    Early conversations happen with the people building it, not a sales funnel.

Register interest

Which platform? *

This registers interest in early access — both platforms are pre-launch, so this isn't a product signup or a purchase.