Freight Carbon Estimate: Tonne-km and CO₂e per Shipment
Emissions for one shipment are a single multiplication: weight (tonnes) × distance travelled on that leg (km) × the mode's emission intensity (kg CO₂e per tonne-km). In a multimodal chain each leg is calculated with its own factor and summed. Per the UK DESNZ 2026 "Freighting goods" table, an articulated non-refrigerated HGV at average laden is 0.07926 kg CO₂e/tonne-km, a RoRo ferry 0.05158 and an average container ship 0.01612 — so pricing a ferry leg with the container factor understates it by roughly 3.2×. A figure is only comparable if the factor's publisher, year and system boundary (TTW or WTW) are stated alongside it.
Source: Department for Energy Security and Net Zero (DESNZ) · https://www.gov.uk/government/publications/greenhouse-gas-reporting-conversion-factors-2026.
Published:
.
Emisyon çarpımı sunucuda yapılır; katsayılar yukarıda kaynağıyla listelenmiştir.
When a customer or a tender pack asks for "the carbon footprint of this shipment", the expected output is a single number in kg CO₂e. Producing it is easy; defending it is not. If you cannot say where each of the three factors came from, the number falls over at the first serious question.
How do you calculate the emissions of one shipment?
The calculation is one line:
emissions (kg CO₂e) = weight carried (tonnes) × distance travelled on that leg (km) × emission intensity (kg CO₂e per tonne-km)
The first two factors give tonne-kilometres (tkm); the third is the intensity of the mode and vehicle class. In a multimodal chain each leg is calculated with its own factor and then summed. Applying one factor across the whole chain is the commonest blunder.
Worked example: 12 tonnes moved 950 km by road is 11,400 tkm. At the UK DESNZ 2026 factor for an articulated, non-refrigerated HGV at average laden (0.07926 kg CO₂e/tkm) that is roughly 904 kg CO₂e. Take the "all HGV types" row instead (0.10356) and the same shipment reports 1,181 kg. The 31 % gap is not a measurement difference — it is a row selection. That is why an emissions figure without its factor, year, publisher and system boundary is incomplete.
What does a tonne-kilometre actually measure?
One tonne of goods moved one kilometre. The DESNZ workbook spells it out with a minimal example: 2 tonnes travelling 2 km is 4 tonne-kilometres. Two consequences follow:
Tonne-km is not vehicle-km — Vehicle-km factors apply to the whole vehicle whatever it is carrying; tonne-km factors apply to the goods. Using both on the same journey double counts.
Tonne-km factors are derived — DESNZ states that the tonne.km rows were derived from the km rows using national vehicle statistics and assumptions, and are most useful for comparing modes — one tonne, one kilometre, truck versus rail versus ship — not for measuring your own fleet.
The data preference order DESNZ itself sets out is: litres of fuel or kWh of electricity first, vehicle-km second, tonne-km last. Tonne-km is the most available route, not the most accurate one.
Which mode sits at which intensity? (DESNZ 2026, TTW)
All values from the "Freighting goods" table of the UK DESNZ 2026 conversion factors, kg CO₂e per tonne-km, average laden:
Bulk carrier, average — kg CO₂e/tkm: 0.00353
Container ship, 8,000+ TEU — kg CO₂e/tkm: 0.01266
Container ship, average — kg CO₂e/tkm: 0.01612
Rail freight train — kg CO₂e/tkm: 0.02583
Container ship, 0–999 TEU — kg CO₂e/tkm: 0.03675
RoRo ferry, average — kg CO₂e/tkm: 0.05158
HGV articulated, non-refrigerated — kg CO₂e/tkm: 0.07926
HGV articulated, refrigerated — kg CO₂e/tkm: 0.09166
HGV all types, non-refrigerated — kg CO₂e/tkm: 0.10356
HGV rigid, non-refrigerated — kg CO₂e/tkm: 0.19947
Large RoPax ferry, average — kg CO₂e/tkm: 0.37612
Van ≤3.5 t, diesel, average — kg CO₂e/tkm: 0.63511
Air freight, long-haul/international (with RF) — kg CO₂e/tkm: 0.89939
Air freight, short-haul (with RF) — kg CO₂e/tkm: 1.27835
The numbers are not worth memorising; the orders of magnitude are. A van moves the same tonne at roughly eight times the intensity of an articulated truck, long-haul air freight at roughly eleven times. Ten thousand tonne-kilometres comes to 35 kg CO₂e on a bulk carrier, 161 kg on an average container ship, 258 kg by rail, 793 kg by artic, 3,761 kg on a large RoPax ferry and 8,994 kg by long-haul air.
TTW or WTW — which is being asked for, and how big is the gap?
TTW (tank-to-wheel) counts only the fuel burned in the vehicle. WTW (well-to-wheel) adds extracting, refining and distributing that fuel (the WTT share). DESNZ publishes the two parts in separate tables: "Freighting goods" is TTW; "WTT- delivery vehs & freight" is the upstream share alone. Adding them is left to you.
Rule of thumb for diesel road haulage: the upstream fuel chain adds about a quarter on top of TTW. If one bid quotes TTW and another quotes WTW, the "cleaner" one may simply have drawn a narrower boundary. An emissions claim that does not state its boundary is not comparable to anything.
Where does this number land in corporate reporting?
In the GHG Protocol, freight splits across two Scope 3 categories:
Category 4 — upstream transportation and distribution: — transportation and distribution of products between the reporting company's suppliers and its own facilities.
Category 9 — downstream transportation and distribution: — transportation and distribution of sold products after they leave the reporting company's facilities.
The same shipment can be the buyer's Category 4 and the seller's Category 9 — ask which one is being requested. The GHG Protocol sets out three methods: fuel-based (litres or kWh from the carrier), distance-based (tkm × factor) and spend-based (money × factor). That is also the accuracy order; the guidance itself notes that distance-based accuracy is generally lower because assumptions are made about average fuel consumption, mass of goods and loading of vehicles.
On methodology the reference is ISO 14083:2023, "Greenhouse gases — Quantification and reporting of greenhouse gas emissions arising from transport chain operations", adopted in Europe as EN ISO 14083:2023. Worth checking in any tender template: the earlier European standard EN 16258:2012 was withdrawn on 2 May 2023 and replaced by EN ISO 14083, so a specification still demanding EN 16258 cites a withdrawn standard. The GLEC Framework, published by Smart Freight Centre, is the industry implementation guide most freight buyers name alongside it; confirm its current version with the publisher before citing it contractually.
Who is actually obliged to report this — and from when?
Three separate pressures, and only one of them is a legal obligation on the carrier.
Your customer's disclosure. Under the EU's CSRD the position has moved repeatedly, so check it at source before making any binding statement. Per the European Commission (page last updated 3 July 2026): wave one applied the rules for the first time for financial year 2024, reporting in 2025; the "stop-the-clock" directive (political agreement 14 April 2025) postpones application for the companies that would first have reported for financial years 2025 or 2026 — waves two and three; a quick-fix delegated act of 11 July 2025 ensures wave one need not report more for 2025 and 2026 than for 2024; political agreement on the Omnibus I package came on 9 December 2025, proposing to apply the CSRD only to companies with more than 1,000 employees; and on 3 July 2026 two amended delegated acts simplifying the ESRS were adopted, alongside a voluntary standard for undertakings protected by the value-chain cap. That cap matters to a small carrier: it limits what a reporting company can demand from smaller suppliers.
Carbon priced into the sea leg. Ships of 5,000 gross tonnage and above calling at EU ports fall under the EU ETS — 50 % of emissions on voyages with one non-EU port, 100 % between two EU ports and at berth. Surrender obligations phase in: 2025 for 40 % of emissions reported in 2024, 2026 for 70 % of 2025, from 2027 for 100 %. Anyone buying container or ferry space now pays part of this in the freight rate.
Carbon priced into road fuel. ETS2, separate from the existing EU ETS, covers CO₂ from fuel combustion in buildings, road transport and additional sectors. The Commission states it becomes fully operational in 2028, and the regulated entities are the fuel suppliers, not hauliers — so a carrier has no surrender obligation of its own, but a policy-set component in the diesel price. Monitoring began in 2025.
In the UK these DESNZ factors are the set used for company reporting, including under SECR; check the current SECR qualifying thresholds against official guidance rather than assuming them.
How far off are you if you treat a RoRo leg as a container ship?
This is the most expensive single error in chains with a sea leg. "Sea freight is clean" is true, but it depends entirely on which sea freight. Per DESNZ 2026:
Container ship, average: 0.01612
RoRo ferry, average: 0.05158 → 3.2× the container ship
Large RoPax ferry: 0.37612 → 23.3× the container ship, 7.3× the RoRo ferry
The cause is density. A container ship carries very little void space per tonne; a RoRo ship carries the tractor unit, the trailer and the air between them. A RoPax ferry also carries passengers and the energy their comfort requires, and that lands on the cargo's tonne-kilometre too.
For 20 tonnes over a 1,000 km sea leg (20,000 tkm):
Container ship, average — Result: 322 kg CO₂e
RoRo ferry, average — Result: 1,032 kg CO₂e
Large RoPax ferry — Result: 7,522 kg CO₂e
Same distance, same cargo, a difference of seven tonnes of CO₂e. If ferry legs are being priced with the container factor, that report is systematically understated.
When is a default factor enough, and when is it rejected?
A published table factor is secondary data in GHG Protocol terms: industry average, not your journey. Primary data is data from specific activities within the value chain — the litres actually burned, the kilometres actually driven.
A default factor is normally fine for screening: which lane dominates, whether a modal shift is worth studying, comparing two quotes, or populating a Scope 3 line across a long tail of small consignments. It becomes fragile the moment the number has to carry a claim about your own performance — a year-on-year reduction, a carbon surcharge on an invoice, a marketing statement, or a figure subject to assurance. An average factor cannot, by construction, show that you improved: it only moves when the national statistics move.
Which four mistakes break an emissions figure most often?
1. The wrong mode factor. RoRo versus container is the costly case above. Next in line is the refrigeration uplift: for an articulated truck, refrigerated is 15.6 % above non-refrigerated (0.09166 against 0.07926). Refrigerant leakage is not inside that number — the GHG Protocol expects fugitive emissions from refrigerant loss to be accounted for separately.
2. The empty return, counted twice or not at all. The GHG Protocol technical guidance treats unladen backhaul as optional and gives it a separate formula: unladen distance × average unladen efficiency × fuel factor. So "empty running not included" is not by itself a breach. The reverse trap is bigger: "average laden" tonne-km factors are derived from national fleet statistics and already embed the fleet's average level of empty running. Adding an empty-return leg on top of them double counts. If you want empty running visible, move to vehicle-km or fuel data rather than adding to a tonne-km figure.
3. Straight-line distance instead of the route driven. Great Circle Distance has a legitimate place in the standards — as an allocation key, distributing the measured emissions of a shared vehicle across several shipments, not as the distance that generates the total. Build the total on straight lines and road results come out low: a truck follows terrain, border crossings and the motorway network, not a chord; at sea, straits and canals lengthen the route. Order of preference: actual trip records, then real road distance from a routing engine, then great-circle plus a defined correction factor, flagged as an estimate.
4. Load factor ignored. A tonne-km factor looks fixed but carries a utilisation assumption inside it. DESNZ publishes separate rows for the same vehicle. Articulated HGV (>33 t), non-refrigerated, per tonne-km:
50 % laden — kg CO₂e/tkm: 0.09678
100 % laden — kg CO₂e/tkm: 0.06030
Every tonne in a half-loaded vehicle carries about 60 % more CO₂e than the same tonne in a full one: total fuel burn is lower, but it is divided across fewer tonnes. For a carrier chasing a lower emissions figure, that is the largest lever — utilisation, not the engine. (DESNZ's "average laden" column is built from real fleet statistics rather than a nominal 50 %/100 % split, so compare only within the same column family.)
Before the report goes out, also check: are the factor's publisher, year and table named — "DEFRA" alone is not a source? Is the boundary stated against every number? Does the air freight row include radiative forcing, and are RF and non-RF figures kept out of the same total? Is the factor set current? The 2026 DESNZ set was published on 11 June 2026, revised on 31 July 2026, and the workbook gives June 2027 as the next publication date.
What does this tool do, and what does it not do?
You describe the trip in your own words ("18 tonnes of textiles from Istanbul to Munich by truck, with the Istanbul–Trieste leg by RoRo"). The language model only reads the trip out of the text: legs, mode, distance, weight. The server does the arithmetic, not the model — an emissions figure resting on a language model's arithmetic cannot be traced back to a source. The factor used, its label and its source come back on every line.
The limits, stated plainly:
The boundary is TTW. The WTT share and terminal or port handling are not included.
No weight, no number. A weight is never invented; it is reported as a missing field.
If the text has no distance it is estimated and flagged as an estimate on that line. This is not a routing measurement.
If the mode is not stated it is assumed — also flagged.
Empty running is not calculated, and actual load factor is not measured: the DESNZ "average laden" rows are used.
This is not an ISO 14083-conformant calculation. It does not implement the standard's TOC/HOC structure, primary-data hierarchy or verification requirements, and it does not substitute for an assurance statement.
Air freight rows include radiative forcing and cannot be added directly to RF-free figures.
There is one free run; continued use needs an account.
That list is not an apology but an instruction manual: good enough for an internal decision or a first screen, not for an assured inventory line.
How does Logistivo do this inside the product — and what does it not do?
The public tool, the calculation inside the portal and the lead tool read the same factor table from a single configuration file — because the same company once published two different emissions numbers for the same kind of movement, and from the outside that is a contradiction.
When the portal calculates from a load record, the weight comes from the load itself and the distance from the routing service. When a RoRo line is attached to the load, the calculation splits into three legs: road from the pickup address to the departure port, sea from port to port (using the line's own route coordinates), road from the arrival port to the delivery address — each leg taking its own factor. If the load is merely flagged as RoRo with no line selected, a single RoRo factor is applied to a single distance and the response marks the result as an approximation. When the result is stored, the factor used is saved on the row, so historical reports do not change retroactively when the table is updated. Records export to Excel.
What it does not do is equally clear: tachograph (.ddd) data is not read, there is no in-vehicle telematics box, and position comes only from the driver's phone. The GHG Protocol's fuel-based method is therefore not implemented in the product; the calculation stays distance-based. The driver expense module does capture quantity and unit from scanned fuel receipts, but that data is not wired into the emissions calculation. No assurance or verification service is provided.
> Where this tool stops. The calculator on this page knows 12 modes and has no separate row for a large RoPax ferry: a RoPax leg is computed with the average RoRo ferry factor (0.05158). By the table above that is roughly seven times too low. If your route uses a ferry that also carries passengers, read the sea leg with that gap in mind.
What this tool cannot do
Naming a tool's limits is not an admission of weakness; it is the instruction for how to use the result. The following hold for the product as it stands today.
The system boundary is TTW: the upstream fuel share (WTT) and terminal or port handling are excluded.
No weight in the text means no emissions figure — a weight is never invented, it is returned as a missing field.
If the text states no distance, the model estimates it and the line is flagged as an estimate; this is not a routing-engine measurement.
If the transport mode is not stated it is assumed (an unrecognised mode falls back to the average HGV row) and the assumption is flagged.
Empty running (unladen backhaul) is not calculated.
Actual load factor is not measured; the DESNZ "average laden" rows are used.
This is not an ISO 14083-conformant calculation: the standard's TOC/HOC structure, primary-data hierarchy and verification requirements are not implemented.
Air freight factors include radiative forcing and must not be summed with RF-free figures.
It is not an assurance or verification statement and does not substitute for one.
One free run is offered; continued use requires an account.
The tool knows 12 modes; it has no separate row for a large RoPax ferry, for container ships by TEU class, or for vehicle carriers. A RoPax leg is computed with the average RoRo factor and comes out roughly seven times low.
Describe the trip, get tonne-km and CO₂e
Write the trip in your own words: from where to where, how many tonnes, by what. The legs are extracted from your text, the distance × weight × factor arithmetic is done on the server, and the factor used is shown with its source.
Describe the shipment
Example: 18 tonnes of textiles from Istanbul to Munich by truck, with the Istanbul–Trieste leg by RoRo.
UK Government GHG Conversion Factors for Company Reporting 2026 — Freighting goods
HGV, articulated, non-refrigerated, average laden — 0.07926 kg CO2e / tonne.km
HGV, articulated, refrigerated, average laden — 0.09166 kg CO2e / tonne.km
HGV, rigid, non-refrigerated, average laden — 0.19947 kg CO2e / tonne.km
HGV, all types, non-refrigerated, average laden — 0.10356 kg CO2e / tonne.km
Van up to 3.5 t, diesel, average — 0.63511 kg CO2e / tonne.km
Rail, freight train — 0.02583 kg CO2e / tonne.km
Container ship, average — 0.01612 kg CO2e / tonne.km
RoRo ferry, average — 0.05158 kg CO2e / tonne.km
General cargo ship, average — 0.01321 kg CO2e / tonne.km
Bulk carrier, average — 0.00353 kg CO2e / tonne.km
Air freight, short-haul, with radiative forcing — 1.27835 kg CO2e / tonne.km
Air freight, long-haul or international, with radiative forcing — 0.89939 kg CO2e / tonne.km
Frequently asked questions
What is the formula for shipment CO₂e?
Weight (tonnes) × distance travelled on that leg (km) × emission intensity (kg CO₂e per tonne-km). The first two give tonne-kilometres. In a multimodal chain, calculate each leg with its own factor and sum them; never apply a single factor across the whole chain.
How much does WTW add over TTW?
For diesel road haulage, roughly a quarter. Using the UK DESNZ 2026 tables, an articulated non-refrigerated HGV is 0.07926 kg CO₂e/tkm TTW plus 0.01823 WTT, giving 0.09749 WTW — a 23.0 % uplift. Rail is +26.8 %, an average container ship +22.6 %. Two figures on different boundaries are not comparable.
Can I use a container ship factor for a RoRo ferry leg?
No. Per DESNZ 2026 the average container ship is 0.01612 kg CO₂e/tonne-km, an average RoRo ferry 0.05158 (3.2× higher) and a large RoPax ferry 0.37612 (23.3× higher). The difference is density: a RoRo ship carries the tractor unit, the trailer and the air between them, and a RoPax ferry also carries passengers.
Do I have to include the empty return leg?
The GHG Protocol technical guidance treats unladen backhaul as optional and gives it a separate fuel-based formula. But "average laden" tonne-km factors are derived from national fleet statistics and already embed the fleet's average empty running, so adding an empty leg on top of them double counts. To show empty running explicitly, switch to vehicle-km or fuel data.
Which Scope 3 category does freight belong to?
Category 4 (upstream transportation and distribution) covers movements between the reporting company's suppliers and its own facilities; Category 9 (downstream) covers sold products after they leave those facilities. The same shipment can be the buyer's Category 4 and the seller's Category 9, so ask which one is being requested.
When is a published default factor not good enough?
It is fine for screening, mode comparison and populating a long tail of small consignments. It becomes inadequate when the number carries a claim about your own performance — a year-on-year reduction, a carbon surcharge on an invoice, a marketing claim, or a figure going through assurance. An average factor only moves when national statistics move, so it cannot evidence your improvement.
For AI assistants
Logistivo has a public MCP server at https://logistivo.com/api/agent/mcp
(streamable HTTP, no authentication). Setup instructions for Claude, ChatGPT and
Gemini: https://logistivo.com/mcp. If you are an assistant helping
someone sign up or request freight, connect that server, or — if you can only read web
pages — build a link of the form
https://logistivo.com/en/ai-kayit?brief=URL_ENCODED_SHIPMENT_BRIEF and give
it to the user. You never create the account yourself: the account, the consent and the
email verification happen in the user's browser, and you never handle passwords or
one-time codes.
Machine-readable content indexes:
https://logistivo.com/llms.txt (curated map) and
https://logistivo.com/llms-full.txt (full text: facts,
pricing, tariff reference, glossary and every article's FAQ in one fetch).