Knowledge base · 02-scopes

02 - Scope 1, 2, 3

Status. Draft v0.1 · First draft: 17-03-2026 · Pre-discussion.

Why this matters. Every regulation in carbon accounting - the GHG Protocol, ISO 14064, CDP, CSRD, SBTi, CBAM - uses the same three-bucket structure for organisational emissions: Scope 1, Scope 2, Scope 3. The buckets are not symmetric. Scope 1 is the smallest in surface area but the most directly measured. Scope 2 is small in count but has a peculiar dual-reporting requirement that swings answers by an order of magnitude. Scope 3 is the largest, the messiest, and the one most companies still cannot count honestly. For an integrated steel mill, the answer to which scope dominates depends on the production route. For a buyer of steel, the Scope 3 of one company is the Scope 1+2 of another - so the accounting only adds up across the value chain if everyone uses the same definitions and shares the same evidence. That last sentence is the whole reason the GHG Protocol exists. It is also the reason a trusted-data platform like SIDK has a market.

1. Why scopes exist at all

A company emits greenhouse gases in a lot of places. The chimneys of its own boilers. The power plant burning gas that supplies its purchased electricity. The mine that supplied the iron ore. The truck that brought the ore to port. The customer who melted the steel to make a car. The driver of the car. The scrapyard that eventually crushed it.

If every company tried to count “all emissions it is connected to”, the same molecules of CO₂ would be counted by dozens of companies and the global total would be nonsense. The opposite mistake - every company counting only its own chimney - would leave the embedded carbon in supply chains uncounted and let the worst polluters off the hook.

The GHG Protocol Corporate Accounting and Reporting Standard ¹, first published in 2001 and the de-facto global framework ever since, solves this with a three-bucket partition. Each bucket has a clear owner. Across the global economy, every emission falls into Scope 1 for exactly one company (the one whose facility emitted it), into Scope 2 for at most one company (the one whose purchased electricity caused it), and may legitimately appear in many companies’ Scope 3 (everyone connected to that emission via their value chain). The arithmetic works because Scope 1 + Scope 2 globally equals total energy-related emissions; Scope 3 is overlap by design, used for decisions, not for adding up a planetary total.

2. Scope 1 - direct emissions

Definition. Greenhouse gases released directly from sources owned or controlled by the company. Fuel burnt in the company’s boilers. Process emissions from its furnaces and kilns. Fugitive leaks from its refrigerants and gas seals. Emissions from vehicles it owns. If a meter, a vent, or a chimney is on a piece of equipment the company runs, the gases coming out of it are Scope 1.

How it’s measured. Three tiers, roughly:

Direct measurement. Continuous Emissions Monitoring Systems (CEMS) on the chimney. A stack analyser samples flue-gas composition; a flow meter measures volume; you integrate over time and get kgCO₂ directly. The gold standard, used for major stacks under regulated regimes (EU ETS Tier 4, EPA Part 75). Expensive, calibration-heavy, but unambiguous.
Fuel-combustion calculation. Meter the fuel going in (m³ of gas, tonnes of coal); multiply by an emission factor; integrate over the period. The everyday workhorse for boilers, burners, reheaters. Cheaper and more practical than CEMS for thousands of small sources.
Mass-balance. For processes like blast furnaces where carbon enters in multiple streams and leaves in multiple streams, you account for carbon in vs carbon out, and the unbalanced carbon is what was emitted as CO₂. This is the only correct method for integrated metallurgical sites.

These are not three different answers to the same question - they are three different problems. The Carbon Engine picks among them per source, with audited fallback. We cover this in 06 - The Carbon Engine.

For steel. Scope 1 is dominant for the integrated BF-BOF route. If coke and sinter are made on the same site, Scope 1 represents roughly 70% of total hot-metal emissions ². For the scrap-EAF route, Scope 1 is small (only the auxiliary fuels and small process emissions), and the centre of gravity shifts to Scope 2.

3. Scope 2 - purchased energy, and the dual-reporting problem

Definition. Emissions from purchased electricity, steam, heat, and cooling - energy generated off-site by someone else, but consumed by the company. The emissions physically happen at the generator. The company is held responsible because its purchasing decisions caused them.

So far, so simple. Where it gets interesting is which emission factor you apply to the kWh you bought.

The dual-reporting requirement. Since the GHG Protocol’s 2015 Scope 2 Guidance ³, any company in a market with retail-level energy choice must report Scope 2 both ways:

Location-based. The kWh is multiplied by the average emission intensity of the local grid. If you are on the Indian grid, you use India’s grid factor. If your factory is in France, you use France’s much lower one. This is the “where you are” answer.
Market-based. The kWh is multiplied by the emission factor of whatever contract you signed for that electricity. If you bought a Power Purchase Agreement (PPA) from a wind farm, that contract has its own factor - possibly zero. If you bought an unbundled Renewable Energy Certificate (REC), the certificate retires a unit of green generation and your kWh is treated as carrying its attributes. This is the “what you bought” answer.

The two numbers can diverge dramatically. The Carbon Engine document gives a worked example for an H2-DRI steel plant: 66 kgCO₂e/t under market-based accounting (with green PPAs in place), 2,510 kgCO₂e/t under location-based on the same grid. The same physical kWh of electricity. Both numbers are GHG-Protocol-correct. Both must be reported.

The Quality Criteria. A market-based number is only legitimate if the contractual instrument behind it meets the eight Scope 2 Quality Criteria ⁴. The criteria require that the instrument conveys attributes only once (no double-counting across markets or buyers), that geography and vintage match the consumption period, that the attributes are exclusive to one buyer, and that origin is traceable. A REC that fails any criterion cannot be used. This is the discipline that prevents market-based reporting from sliding into greenwash.

The Carbon Engine treats Quality-Criteria validation as a Factor Registry concern, not an engine concern: the engine reads a qualityCriteriaStatus field on the resolved factor and flags downstream consumers if the instrument is FAILED or NOT_VALIDATED. That’s the seam between “factor governance” and “calculation”; we’ll come back to it in 06.

An ongoing revision. GHG Protocol opened a public consultation on Scope 2 in late 2024, with revised guidance expected to take effect in 2027 ⁵. The headline proposed change is hourly matching - requiring market-based instruments to match consumption hour-by-hour rather than annually. If finalised, this materially tightens what counts as a “green” kWh and will reshape how PPAs are designed and how platforms must track them. We are tracking the revision; the platform’s architecture (per-resolution emission factors, hourly telemetry, evidence provenance) is already shaped to support it.

For steel. Scope 2 is dominant for scrap-EAF and electrified routes (EAF, induction). It is small for traditional BF-BOF, where most energy is internal to the process. The H2-DRI route is uniquely sensitive: the whole green credential of the route hangs on the market-based Scope 2 number, and on whether the Quality Criteria are met for the renewable-power contracts.

4. Scope 3 - the value chain

Definition. Indirect emissions in the company’s value chain that are not Scope 2 - everything that is not the company’s own facility and not its purchased energy. The Corporate Value Chain (Scope 3) Standard ⁶, published 2011, partitions Scope 3 into 15 categories, eight upstream and seven downstream.

The 15 categories (numbered as in the standard):

Upstream (1–8): purchased goods and services, capital goods, fuel- and energy-related activities not in Scope 1/2, upstream transportation and distribution, waste generated in operations, business travel, employee commuting, upstream leased assets.

Downstream (9–15): downstream transportation and distribution, processing of sold products, use of sold products, end-of-life treatment of sold products, downstream leased assets, franchises, investments.

Why it’s the hardest. Three reasons.

First, the data is not yours. Your Category 1 (purchased goods and services) is your suppliers’ Scope 1 + 2. To count it accurately, you need primary data from your suppliers. They have no obligation to share it, and even if they do, you have no way to verify it. In practice most companies fall back on spend-based estimates (currency × category-average factor) or activity-based estimates with industry-default factors - both of which are crude.

Second, the boundaries are slippery. Category 11 (use of sold products) for a steel company is ambiguous: do you count the operational emissions of every car ever made from your steel? The standard offers methodologies but does not enforce one answer; different companies make different reasonable choices, and the resulting numbers are not directly comparable.

Third, the totals are huge. For most manufacturers, Scope 3 is 70–90% of the total carbon footprint. Reducing your own Scope 1+2 by half changes very little if Scope 3 is untouched. This is what makes Scope 3 the regulatory frontier (CSRD now requires it; SBTi requires targets for it; CBAM will eventually broaden into it).

For steel. A typical integrated mill’s Scope 3 is dominated by Category 1 (coking coal, iron ore, alloying elements) and Category 11 (the emissions associated with whatever the steel ends up being used for - though the standard allows a “sold final product” exclusion in some cases). A scrap-EAF mill’s Scope 3 is dominated by the scrap supply chain (collection, transport) and electricity-related Category 3.

The platform’s stance on Scope 3 is the same as for the other scopes: the evidence tier is recorded alongside the number. A Scope 3 emission computed from a verified supplier Product Carbon Footprint (PCF) carries that fact in its provenance; a Scope 3 emission computed from a category-default factor carries that fact. Two steel mills with the same headline Scope 3 number but different evidence quality are not the same; a sophisticated buyer or regulator can see the difference; the platform exposes it.

5. Scope versus boundary - the confusion that bites everyone

A trap that catches new entrants, and one worth flagging early: scope is not the same as boundary.

Scope is a categorisation of an emission by who controls the source: direct (1), purchased energy (2), or value chain (3).
Boundary is the question the calculation is answering: gate-to-gate, cradle-to-gate, the whole site for a year, a specific Lot, a CBAM production process, an EPD module A1-A3.

A single emission has one scope but can appear inside many different boundaries - or be excluded from them. The Scope 1 emission from a blast furnace is part of the BF’s gate-to-gate boundary, part of the mill’s site-operational boundary, part of a downstream coil’s cradle-to-gate boundary (via lineage), and part of a buyer’s Category 1 (upstream Scope 3).

The Carbon Engine treats boundary as a first-class input to every calculation: same subject (a Lot), different boundaries, different CalculationResults, all retained, all queryable. Scope is then a property of each individual emission component inside that result. The two concepts are orthogonal; the engine carries both, and the verifier auditing the claim can navigate both. See 02a - Boundaries for the full vocabulary (cradle-to-gate, cradle-to-grave, EN 15804 modules, CBAM production process), and 06 - The Carbon Engine for how the engine implements it.

6. Why this is a platform problem

Three reasons the scope structure becomes a platform problem, not a spreadsheet problem:

Provenance per component. Every emission inside a scope has a tier - directly measured, fuel-meter-derived, mass-balance, supplier-declared, supplier-PCF, category-default. The platform records the tier and shows it. Two suppliers reporting the same number with different evidence quality are not interchangeable.

Dual Scope 2 by construction. The Carbon Engine always computes both location-based and market-based values for any boundary that includes Scope 2, regardless of which one the caller asked for. Both are stored on the CalculationResult so a future view under a different regime doesn’t require recomputation. This is a deliberate small-cost-now / big-value-later choice - Quality-Criteria-driven regime mismatch is one of the more painful operational surprises for steel mills, and the platform pre-empts it.

Cross-company Scope 3 reconciliation. Your Scope 3 is someone else’s Scope 1+2. If both of you are on the platform and the supplier has issued a verified Product Carbon Footprint (PCF) for the lot you bought, the buyer’s Scope 3 number for that lot can be the supplier’s PCF, with full lineage walkable end-to-end. This is what the passport and lineage machinery in SIDK is for, and it is what makes the platform’s Scope 3 numbers different in kind from a spreadsheet’s. We cover this in 05 - SIDK and 06 - The Carbon Engine.

The next doc, 02a - Boundaries, pairs with this one - scope and boundary are the two orthogonal axes of any carbon claim, and stakeholders comparing third-party numbers need fluency in both. After that, 03 - Regulations overview is where the scope-and-boundary structure meets the regulatory regimes that turn “we should count this” into “we must count this, attested by an accredited verifier, by a deadline”.

02 - Scope 1, 2, 3

1. Why scopes exist at all

2. Scope 1 - direct emissions

3. Scope 2 - purchased energy, and the dual-reporting problem

4. Scope 3 - the value chain

5. Scope versus boundary - the confusion that bites everyone

6. Why this is a platform problem

References & further reading

Primary sources (cited inline)

Further reading (not cited above, but worth your time)

02 - Scope 1, 2, 3

1. Why scopes exist at all

2. Scope 1 - direct emissions

3. Scope 2 - purchased energy, and the dual-reporting problem

4. Scope 3 - the value chain

5. Scope versus boundary - the confusion that bites everyone

6. Why this is a platform problem

References & further reading

Primary sources (cited inline)

Further reading (not cited above, but worth your time)

Footnotes