Carbon Accounting: Boundaries, Factors and What A1–A3 Leaves Out

Quantity times factor, and the factor is the whole problem

Every carbon calculation here has the same form: a quantity of material or energy, multiplied by a factor expressing emissions per unit, summed. It is arithmetic a spreadsheet does, and it is not where the difficulty lies.

The difficulty is that an emission factor is a summary of an entire supply chain, produced under a set of methodological choices — what was included, how shared processes were divided between products, which year and which region the data came from. Two factors for nominally the same material can differ substantially without either being wrong, because they answer slightly different questions.

So a carbon figure is only meaningful accompanied by its boundary and its source. A number on its own — even a precise one — cannot be compared with another number, and the most common misuse of these calculators would be to compare a result from one against a manufacturer's claim computed on a different basis.

A1–A3 is the factory gate, and that is a long way from a building

Life-cycle standards divide a building's emissions into labelled modules, and the most frequently quoted figure covers only the first three: raw material supply, transport to the manufacturing plant, and manufacturing. That is CRADLE TO GATE, and it stops at the factory door.

Everything after it has its own module. Transport to site is separate. Installation, including the waste generated while installing, is separate. Use, maintenance, repair and replacement over the building's life are separate, and for a material replaced three times in sixty years that module can exceed the original. Demolition, transport and disposal or recovery are separate again.

This is why a low cradle-to-gate figure does not settle an argument. A lightweight imported product can beat a heavy local one at the gate and lose once transport is added; a material with a low initial figure and a fifteen-year life loses to a heavier one lasting sixty. A comparison confined to A1–A3 is legitimate only when the things being compared have similar transport, similar installation waste and similar lives — and saying so is part of reporting the result.

The modules also explain a common confusion in reported totals. An "upfront" carbon figure covers the modules up to practical completion; a "whole life" figure includes use and end of life; and the two are routinely quoted with the same label. Where these calculators report a boundary, they name the modules.

Concrete: clinker is nearly all of it

In a typical concrete mix the cement dominates the cradle-to-gate total, and within cement it is the CLINKER that dominates — the aggregates and water are minor by comparison despite being most of the mass. That single fact directs all the useful effort: reducing cement content, or replacing part of it, moves the answer, and changing the aggregate barely does.

It also explains why specifying a higher strength class than the structure needs is expensive in carbon as well as money. Strength is bought largely with cement, so an over-specified mix carries a proportionate penalty in every cubic metre.

Two mechanisms reduce clinker without reducing performance. Later-age strength specification — assessing at fifty-six or ninety days rather than twenty-eight — allows slower-reacting mixes that use less clinker, and suits elements that are not loaded early. And supplementary cementitious materials replace a fraction of the clinker directly, which is the subject of the next section.

The by-product question: why a low-cement mix is partly an accounting choice

Ground granulated blast-furnace slag and fly ash are not manufactured for concrete. They are by-products of steelmaking and coal-fired power generation, and the convention in most published data is to allocate little or none of the upstream process's emissions to them, on the basis that they are wastes the primary process would have produced anyway.

That convention is what makes a high-replacement mix look dramatically better, and it is a METHODOLOGICAL choice rather than a physical fact. Under a different allocation — by economic value, say, which is the approach life-cycle standards prefer when a stream has a market — these materials would carry a share of their parent process and the advantage would shrink.

The practical consequence is arriving anyway. As steel production routes change and coal generation closes, the supply of both materials falls while demand for them rises, and a material with a price, a shortage and a competitive market is harder to describe as a waste. Anyone relying on high replacement levels for a carbon target should know that the number rests on an allocation rule as much as on chemistry.

None of which makes replacement a bad idea — it reduces clinker demand, which is a real reduction in a real process. It means the size of the benefit is contested, and a calculator that reports it should say so rather than presenting the accounting convention as a measurement. The pages here do.

Operational carbon: the factor moves under you

Multiplying a building's annual electricity use by today's grid emission factor and then by thirty years is wrong in a predictable direction. Electricity grids in most regions are decarbonising, sometimes rapidly, so the factor in year thirty will be a fraction of today's and the sum overstates the total, often substantially.

This reverses comparisons. A heat pump assessed against today's grid can look similar to a gas boiler and will diverge sharply over the life of the equipment, because one fuel's factor is falling and the other's is roughly fixed. A comparison made on a single year's factors is not a comparison of the two systems; it is a snapshot of the grid.

There is a second distinction that reporting and decision-making handle differently. An AVERAGE grid factor describes the mix as a whole and is what emissions reporting uses. A MARGINAL factor describes what actually changes when one more unit is consumed — which is the relevant quantity for deciding whether to add or remove load, and can be much higher than the average in a grid where the marginal generator is fossil-fired.

Timing matters for the same reason. A kilowatt-hour at midday in a solar-heavy grid and one at the evening peak carry very different marginal emissions, and annual-average accounting cannot see the difference. Where a decision is about shifting load rather than reducing it, an annual factor answers nothing.

Diversion and recycled content are weighed on different scales

A construction waste diversion rate is normally computed by MASS, and construction waste is dominated by dense inert material — concrete, brick, soil, aggregate. Crushing and reusing that stream can carry a project past a high headline diversion rate while almost everything light is still being landfilled: plasterboard, insulation, packaging, plastics, timber offcuts.

The rate is therefore an honest measure of what it measures and a poor proxy for environmental outcome, because mass and impact are not proportional. Reporting diversion by mass alongside the streams diverted is the useful form; a single percentage is the one that can be gamed without breaking any rule.

Recycled content carries an equal-and-opposite trap. In the credit systems where it is most often calculated, the percentage is weighted by COST rather than by mass — recycled content value, not recycled content mass — so the two quantities share a name and answer different questions. A material that is heavy, cheap and recycled contributes far less to a cost-weighted percentage than its tonnage suggests.

Neither figure should be read as a carbon result. Diverting material avoids disposal and may avoid primary production, but the saving depends entirely on what the diverted material displaces, and a diversion percentage contains no information about that.

What the arithmetic cannot settle

Biogenic carbon is the largest open question in the field. Timber stores carbon that the tree removed from the atmosphere, and how that storage is counted — as a negative at harvest, as a credit only while the product remains in service, with or without the timing of regrowth — changes a timber building's reported total dramatically. Different standards make different choices, and a comparison across them is not valid.

Concrete carbonates over its life, reabsorbing a fraction of the carbon dioxide released when its cement was made, and whether and how that is credited is a further methodological choice. So is the treatment of material recovered at end of life, which can be credited to the building being demolished or to the one that uses it next, but not honestly to both.

And any reduction claim needs a BASELINE. A mix, a specification or a design is lower carbon than something, and unless that something is stated the claim is unfalsifiable. The calculators here compute totals and differences from the inputs given; they are estimating and comparison tools for a design stage, and a reportable assessment is produced to a named standard with product-specific declarations, which is a different exercise from the arithmetic on this page.