Scope 3 Emissions in Freight: How Shippers Measure Them
Scope 3 emissions in freight look like an accounting problem and behave like a data problem. The method you can use is decided by what your carriers will give you.
Scope 3 emissions in freight fall into two of the fifteen categories, and the dividing line is who pays. If your company purchases the transport service, it is category 4, upstream transportation and distribution. If the freight moves to the end customer and you do not pay for it, it is category 9.
That catches people out, because it means outbound shipping you pay for is upstream. A purchased service is upstream regardless of which direction the goods travel.
Quick answer
Freight sits in Scope 3 category 4 when you pay for the transport and category 9 when you do not, which means most outbound shipping is category 4. Three methods exist. Spend-based is for screening, distance-based is where most enterprises live, and fuel-based is the only one that moves when a carrier buys better trucks.
Which category does freight belong to?
The Greenhouse Gas Protocol published its Corporate Value Chain Standard in 2011, defining 15 categories: 8 upstream and 7 downstream.
Category 4 covers transport between your tier 1 suppliers and your own operations in vehicles you do not own, plus every third-party transport and distribution service you purchase. The technical guidance spells this out, naming inbound logistics, outbound logistics for sold products, and movement between your own facilities.
Category 9 is narrower than most people assume. It covers products sold and moved between your operations and the end consumer only when the reporting company does not pay for it, including retail and storage.
The minimum boundary for both is the same: the scope 1 and scope 2 emissions of your transport providers, allocated to you. Manufacturing the vehicles and building the infrastructure is optional.
The three calculation methods
| Method | What it uses | Where it fits | What it hides |
|---|---|---|---|
| Fuel-based | Litres of fuel, kWh, refrigerant leakage | Dedicated and truckload lanes | Nothing, which is the point |
| Distance-based | Tonnes, kilometres, modal factor | Less-than-truckload, parcel, most networks | Load factor, empty running, vehicle age |
| Spend-based | Freight spend, currency-based factor | Screening and an immaterial tail | Almost everything physical |
Fuel-based multiplies fuel consumed by a fuel emission factor, then adds electricity and refrigerant leakage. The guidance is direct about why it wins: fuel consumption is directly related to emissions, so it is more accurate than distance for carbon dioxide. It works best where a vehicle carries only your goods.
Distance-based multiplies the mass of goods by the distance travelled by an emission factor for that mode, leg by leg. Units are grams or kilograms of carbon dioxide equivalent per tonne-kilometre, where a tonne-kilometre is one tonne moved one kilometre. It is the method for shipments smaller than a full vehicle, which is most of them, and the guidance names its cost plainly: accuracy is lower because assumptions are made about fuel consumption, mass and vehicle loading.
Spend-based multiplies freight spend by a factor expressed per currency unit. It belongs at the bottom of the decision tree.
Allocation defaults are worth knowing. Road, air and rail allocate by mass. Marine allocates by volume, because volume is the binding constraint on a ship. Unladen backhaul is optional.
Which method should you use?
Start with what you already hold.
If your transport management system carries mass, origin, destination and mode per shipment, you can run distance-based today without asking anyone for anything. Most enterprises can.
Fuel-based needs a different organisation to produce data it does not currently produce, which is why it is a project rather than a calculation. The practical route is to move the whole network to distance-based in year one, then pursue fuel data only on your top 10 to 20 lanes by tonne-kilometre.
The accuracy curve is not smooth. It is a step function at the carrier boundary. Everything before that step is cheap and everything after it is a negotiation.
Three structural obstacles sit at that boundary:
- Allocation. In groupage, parcel and less-than-truckload the carrier genuinely cannot attribute fuel to one shipper without an assumption, which reintroduces the averaging you were trying to escape.
- Commercial sensitivity. It is a short step from fuel burn to a carrier's cost base, and carriers know it.
- Regulation. In Europe the rules changed in 2026 in a way that removes your ability to demand the data at all, which is covered below.
Where the numbers mislead
Spend-based estimates fail in the direction that flatters you.
Because the factor is per currency unit, emissions fall when rates fall. Renegotiate a 15 percent rate reduction and you book a 15 percent emissions reduction while moving identical tonnage on identical trucks. Pay more for a cleaner carrier and your reported footprint rises.
That is an inverted incentive, not a rounding error. Confine the method to screening and to a small residual tail.
Modal averages have a subtler problem. A default figure per tonne-kilometre bakes in an assumed load factor, an assumed empty-running rate, an assumed vehicle age and an assumed fuel. Those four things are exactly what a logistics team can change.
Run your whole network on defaults and consolidation, backhaul optimisation, fleet renewal and renewable diesel adoption all become invisible. You will have decarbonised and reported nothing.
There is a tiering subtlety here too. A figure supplied by a carrier that was produced by applying a default intensity to a tonne-kilometre count is still modelled data. The data quality tier follows the underlying method, not the sender.
What the standards actually require
ISO 14083:2023, published in March 2023, is the transport-chain standard. It measures freight activity in tonne-kilometres on a well-to-wheel basis, covering the fuel's production as well as its combustion. Its structure breaks a transport chain into elements, groups similar operations into categories, derives an emission intensity for each, and multiplies by activity. Logistics hubs are in scope alongside road, rail, air, sea, inland waterway, cable and pipeline. The free practitioner guide published by CLECAT is a usable substitute if you do not want to buy the standard.
The GLEC Framework, published by Smart Freight Centre, is the operational layer on top: default intensity factors, allocation rules and data quality tiers for multimodal logistics. Version 3.2 arrived in October 2025 and is aligned to ISO 14083, the Greenhouse Gas Protocol and the major disclosure platforms.
The Environmental Protection Agency's SmartWay partnership is the most underused free resource in this space. Its programme page, updated on 4 September 2026, reports more than 4,000 partners and 55.4 billion dollars in partner fuel savings. For a shipper, its value is carrier-specific performance data, which is one rung above generic modal defaults.
What changed recently
Four things, and the direction is not uniform.
Europe standardised the method, voluntarily. Regulation (EU) 2026/1030, known as CountEmissionsEU, was adopted on 29 April 2026 and took effect in June 2026, confirming EN ISO 14083:2023 as the reference methodology. It applies from December 2030. It standardises how you calculate if you disclose. It does not require you to disclose.
European reporting narrowed sharply. The Omnibus directive amending CSRD came into force on 18 March 2026, cutting scope to entities above both 1,000 employees and 450 million euro turnover, with first application to financial years beginning 1 January 2027. Critically for freight, in-scope companies are now prohibited from requiring sustainability information from value-chain undertakings below 1,000 employees. That describes most European road hauliers. Carrier data is a commercial negotiation now, not a compliance demand, so budget for contract clauses at renewal rather than for data requests.
The US federal rule is effectively dead. The Securities and Exchange Commission proposed a rescission of its climate disclosure rules, published in the Federal Register on 3 June 2026 with comments closing on 3 August 2026. Do not build a freight programme around it. State-level requirements are a separate matter and are still moving, so confirm current dates with the regulator directly rather than from secondary coverage.
The standard itself is being rewritten. The Greenhouse Gas Protocol published its Phase 1 progress update on 31 March 2026. Proposals include requiring companies to account for at least 95 percent of total required Scope 3 emissions, capping exclusions at 5 percent. Public comment is expected in the second quarter of 2027 with final text targeted for the end of 2028, so nothing changes for this reporting year.
The honest read across all four: the compliance pressure that was supposed to force this work has softened, while the methodological consensus has hardened. If you are still measuring freight, it is now because you want the number to be useful, not because someone is making you.
That is a better reason anyway.
For another example of how we write, see our other writing samples.
Frequently asked
- Is outbound freight I pay for category 4 or category 9?
- Category 4. The test is not direction, it is who purchases the service. The technical guidance puts outbound logistics services purchased by the reporting company in the upstream category because they are a purchased service. Category 9 applies when the goods move to the end customer and you are not paying the freight.
- How do I calculate freight emissions when carriers will not give me fuel data?
- Use the distance-based method. Multiply mass in tonnes by distance in kilometres by a modal emission factor, leg by leg, then aggregate. The technical guidance explicitly directs you there when carrier fuel data cannot easily be obtained, along with the caveat that accuracy is lower because of the assumptions baked into the factor.
- What is the difference between the GLEC Framework and ISO 14083?
- ISO 14083 is the standard that defines the method. The GLEC Framework, published by Smart Freight Centre, is the implementation guidance and the source of default factors and data quality tiers. Version 3.2 is conformant with the standard, so you are not choosing between them.
- What is a tonne-kilometre and why do consultants use it?
- One tonne of goods moved one kilometre. It is the allocation unit that makes a 2 tonne shipment on a 20 tonne truck attributable to you, and it is the activity measure ISO 14083 requires for freight.
- Do I still have to report Scope 3 under CSRD?
- Only if you clear both thresholds set by the 2026 Omnibus directive: more than 1,000 employees and more than 450 million euro in turnover. First reporting covers financial years starting on or after 1 January 2027. Note also that in-scope companies can no longer require data from value-chain partners below 1,000 employees.