VWTVWT
← All insights

How to Work Out Carbon per Shipment in Thailand

Published November 11, 2026 Β· 9 min read

A customer sends a spreadsheet. One column wants kilograms of CO2 for every delivery you made them last year. Nobody in the transport team has ever produced that number, and the carrier, when asked, offers a percentage it cannot explain.

The good news is that this is arithmetic, not science. Thailand publishes its own official factors, they are free, and a transport manager with fuel records or trip records can produce a defensible figure in an afternoon. The bad news is that the three most common ways of doing it are all wrong in the same direction, and one of them is wrong by about 18 per cent.

Where this fits

The fuel surcharge article takes the Thai pump price apart, including the two diesel blends that matter below. This one is about turning litres or kilometres into a carbon number somebody else will check.

There are only two methods

Fuel-based. Count the litres, multiply by a factor per litre. This is the better method whenever you can get litres, because fuel burnt is what emissions actually follow. Your own fleet has this in the fuel card data.

Distance-based. Count tonne-kilometres, multiply by a factor per tonne-kilometre. This is what you use for a hired truck when nobody will tell you the litres.

The two are not alternatives in quality. The first measures, the second estimates from an average vehicle. Use the first when you can, and keep the second to check it.

The factor almost everybody gets wrong

The textbook rule of thumb for a fleet is that a litre of diesel makes about 3 kilograms of CO2. Convenient, and too high for Thailand.

Two words first, because the spreadsheet will use them. Scope 1 means emissions from fuel your own company burns, so your own trucks sit there. Emissions from a truck you hired are somebody else's Scope 1 and your Scope 3, the part of your footprint that happens in other people's operations. The arithmetic is identical either way. Only the box it goes in changes.

Thailand's greenhouse gas management organisation publishes a factor list for organisational footprints, in a version marked for use from 1 January 2026. For on-road vehicles it gives two ways to treat diesel, and the difference is the thing people miss.

What you use per litre of diesel kgCO2e
Diesel, whole litre counted as fossil 2.7403
7% blend, the fossil part, reported in Scope 1 2.5504
7% blend, the plant-derived part, reported separately 0.1151
20% blend, the fossil part, reported in Scope 1 2.1976
20% blend, the plant-derived part, reported separately 0.3288

Ordinary high-speed diesel at a Thai pump is B7, a 7 per cent biodiesel blend. The carbon in that 7 per cent came out of the air last season when the crop grew, so returning it to the air is not treated as adding anything. The official list therefore puts 2.5504 in your reported Scope 1 and the remaining 0.1151 on a separate line, called biogenic, that is disclosed but sits outside the scopes.

So the three common errors:

  • Using the rule of thumb of 3.0 overstates your reported Scope 1 by about 18 per cent.
  • Using the plain diesel figure of 2.7403, which is in the same official table, overstates it by about 7 per cent.
  • Using 2.5504 but forgetting the biogenic line means your total carbon is understated even though Scope 1 is right.

If your carrier runs B20, the reported figure drops again, to 2.1976. That is worth asking, because it changes the answer by 14 per cent and costs you nothing to find out.

The Thai table for a hired truck

When litres are not available, the product footprint guideline carries a table built from Thailand's own life cycle inventory. It is organised by vehicle, by whether the run is normal or rough, and by how full the truck is. These are the normal-running rows, in kgCO2e:

Vehicle Empty, per km Half load, per tonne-km Three-quarters, per tonne-km Full, per tonne-km
6-wheel, 8.5 t 0.4238 0.1237 0.0868 0.0672
6-wheel, 11 t 0.4882 0.1074 0.0762 0.0609
10-wheel, 16 t 0.5851 0.0964 0.0685 0.0529
Semi-trailer, 32 t 0.8612 0.0797 0.0565 0.0441

Two things fall straight out of that table.

A full 10-wheel moves a tonne of goods a kilometre for 0.0529. The same truck at half load does it for 0.0964, which is 82 per cent more per tonne. Load factor, not vehicle choice, is the big lever.

Smaller is worse per tonne. A full 8.5-tonne 6-wheel is 27 per cent higher per tonne-kilometre than a full 10-wheel, and a full semi-trailer is 17 per cent lower. That is the carbon version of the arithmetic in the two trucks or one question, and it points the same way.

The rough-running rows, for lanes that are not mostly sealed highway, run between about 10 and 27 per cent above the normal rows depending on the vehicle. Pick the one that matches the lane, not the one that flatters it.

Sanity-check any row before you use it

Divide a full-load factor by the vehicle's rating, then by the blend factor, and you get the litres per kilometre it assumes. The 10-wheel row implies 3.01 kilometres per litre and the 6-wheel 4.46, both of which are believable. Do this once per row you rely on, because the published table has rows that fail it:

  • The 4-wheel pickup rows imply about 2.6 kilometres per litre loaded, while the same vehicle's own empty row implies 8.2. No pickup does both. The loaded rows for that class are left out of the table above.
  • An empty 18-wheel full-trailer is listed at 0.2289 per kilometre, below an empty 6-wheel.
  • One 20-wheel half-load row is labelled per kilometre while carrying a per-tonne-kilometre value.

None of that makes the table unusable. It makes the check worth thirty seconds.

One lane, both methods

A 10-wheel carries 12 tonnes 450 kilometres and comes back empty. Full load for that vehicle class is 16 tonnes, so the truck is at three-quarters.

Fuel-based. Measured Thai averages, from on-board devices on 400 tanker trucks across 8,794 trips, are 3.11 kilometres per litre loaded and 4.31 empty.

  • Loaded: 450 / 3.11 = 144.7 litres
  • Empty: 450 / 4.31 = 104.4 litres
  • Total 249.1 litres, at 2.5504 = 635 kg CO2e in Scope 1, plus 29 kg biogenic reported separately
  • Per tonne of goods: 635 / 12 = 52.9 kg

Distance-based. Using the table above.

  • Loaded leg: 0.0685 Γ— 12 tonnes Γ— 450 km = 370 kg
  • Empty return: 0.5851 Γ— 450 km = 263 kg
  • Total 633 kg, or 52.8 kg per tonne

The two methods land within half a per cent of each other, and that is not luck. The three-quarters row at 12 tonnes works out to 0.822 kg per kilometre, which at the blend factor implies 3.10 kilometres per litre, and the empty row implies 4.36. The official table and the trucks agree. That gives you a check worth keeping: if a carrier claims a fuel economy far from about 3.1 loaded and 4.3 empty for a 10-wheel, one of the numbers in front of you is wrong.

The empty leg is not a detail

On that lane the return trip is 263 of 633 kilograms, or 42 per cent of the total. A carrier that reports only loaded kilometres hands you a number that is roughly 40 per cent too small, and it will look perfectly reasonable.

The guideline does not leave this to judgement. In its own worked rule for waste haulage it counts the loaded leg on a full 16-tonne 10-wheel and then tells you to count the empty return as well.

That makes the carbon case for the thing the cost case already wanted. A return load does not just cut the rate. It removes the leg that was carrying 42 per cent of the footprint, and it puts the truck's whole day against goods instead of against air.

What to ask the carrier for

Ask for inputs, never for a carbon figure. A figure you cannot rebuild is a figure you cannot defend.

  1. Litres, by lane or by month. If they will not give litres, ask for kilometres and their own fleet average fuel economy, and check it against 3.1 and 4.3.
  2. B7 or B20. It moves the reported number by 14 per cent.
  3. Vehicle class and full-load rating. You need the rating to work out the load percentage, and the table is organised by it.
  4. Loaded and empty kilometres separately. Without the split you cannot do the 42 per cent.
  5. Whether the lane runs normal or rough. The rough-running rows are materially higher, and a site on an unsealed access road is not a motorway run.

Put those five in the tender document and you never have to ask again.

Where the evidence stops

The factor list and the vehicle table are official and current, and both were read in Thai. The method is ordinary and the arithmetic above is reproducible from the two of them.

Three limits worth stating plainly. The measured fuel economy comes from tanker trucks, so it is a good Thai benchmark for a rigid 10-wheel on a highway and not a figure to force onto a refrigerated body or an urban multi-drop round. The vehicle table is an average per class from a national inventory, so it cannot tell you about your own trucks, only about the class. And nothing here rests on a reporting obligation: Thai disclosure rules for listed companies are being revised, the detail is not settled, and an article that tied the arithmetic to a rule still in motion would be out of date before the rule arrived. The arithmetic does not depend on it.

What the customer wants is a number with a method behind it. That you can produce today.