Why a gallon of gasoline weighs 6 pounds but emits 20
The number that surprises everyone is that a 6.3 lb gallon of gasoline produces about 19.6 lb of carbon dioxide. Nothing is created from nothing: the extra mass is oxygen pulled out of the air. Gasoline is roughly 87 percent carbon by weight, so a gallon carries about 5.5 lb of carbon, and each carbon atom (atomic weight 12) picks up two oxygen atoms (16 each) to make a CO2 molecule with a molecular weight of 44. Multiply 5.5 lb by 44/12 and you get about 20 lb, which is EPA's 8.887 kg.
That arithmetic is why the emission factor is a property of the fuel, not the car. A tuned engine, a clean air filter and fresh spark plugs improve emissions of unburned hydrocarbons, carbon monoxide and nitrogen oxides. They do not change CO2 per gallon at all. The only way a vehicle reduces its carbon dioxide is by needing fewer gallons — which is why fuel economy and CO2 intensity are the same measurement in different units, and why EPA can print a grams-per-mile CO2 figure on a window sticker derived directly from MPG.
Diesel is denser and richer in carbon, so it emits more per gallon: 10.180 kg. Diesel engines are also more efficient, so a diesel car usually travels enough further on that gallon to come out ahead per mile — but not always, and the calculator settles it by working in grams per mile rather than per gallon.
The four numbers behind the result
Gallons. Distance divided by fuel economy. This is the only place your driving behaviour enters, and it is where all the variance is: city driving, roof boxes, cold starts, aggressive acceleration and under-inflated tyres all show up here as fewer miles per gallon.
The emission factor. EPA's regulatory figures under 40 CFR Part 600 are 8,887 grams of CO2 per gallon of gasoline and 10,180 grams per gallon of diesel. Those are the values used on fuel economy labels and in the agency's greenhouse gas equivalencies. They already assume complete combustion, which is a good approximation for a functioning modern engine.
The non-CO2 uplift. Vehicles also emit methane and nitrous oxide, which are far more potent per kilogram but are released in tiny quantities by catalyst-equipped light-duty vehicles. Expressed as CO2-equivalent they add a small percentage, and the calculator lets you set that percentage rather than assuming one, because it depends on vehicle age, catalyst type and the global warming potentials you choose. Set it to zero and you get pure tailpipe CO2, which is what almost every published vehicle rating means.
The well-to-wheel multiplier. Tailpipe emissions ignore everything that happened before the fuel reached your tank: drilling, transport, refining, and trucking to the station. Lifecycle models such as Argonne National Laboratory's GREET put the upstream burden of US gasoline in the region of 20 to 25 percent on top of combustion. The default here is 1.00, so the headline figure is comparable with vehicle labels; raise it when you are doing a lifecycle comparison, and raise it for both the petrol car and the EV if you want the comparison to stay fair.
For the electric comparison, the chain is shorter: miles × kilowatt-hours per mile × the carbon intensity of the electricity. EPA's eGRID database publishes that intensity by subregion in pounds of CO2e per megawatt-hour, and it varies enormously — a hydro-heavy northwest subregion and a coal-heavy midwest one differ by more than a factor of ten. The electricity CO2 emissions calculator covers that side in more detail, including transmission losses and market-based accounting.
Worked example: 12,000 miles a year at 25 MPG
You drive 12,000 miles a year in a gasoline car that averages a genuine 25 MPG, and you want your annual footprint plus a check on whether switching to an EV would help on your local grid.
- Gallons burned. 12,000 ÷ 25 = 480 gallons.
- Tailpipe CO2. 480 × 8.887 = 4,265.8 kg, which is 4.27 metric tonnes or about 4.70 US short tons.
- Intensity per mile. 8,887 ÷ 25 = 355.5 g CO2 per mile.
- Intensity per kilometre. 355.5 ÷ 1.609344 = 220.9 g/km. For reference, a European fleet target expressed in g/km makes this vehicle look poor, which it is.
- Carpooling. With four people aboard, 355.5 ÷ 4 = 88.9 g per passenger-mile — a bigger single-step improvement than almost any change of vehicle.
- The EV comparison. At 0.30 kWh per mile, 12,000 miles needs 3,600 kWh. A grid at 800 lb CO2e/MWh is 800 ÷ 1,000 × 0.4536 = 0.3629 kg per kWh, so 3,600 × 0.3629 = 1,306 kg. That is 69 percent below the gasoline car on this grid.
- With upstream included. Applying a 2 percent non-CO2 uplift and a 1.24 well-to-wheel factor gives 4,265.8 × 1.02 × 1.24 = 5,395 kg for the gasoline car. To keep the comparison honest you would also need upstream emissions for electricity generation, which eGRID's factors partly capture but which are not included in the simple product above.
Tailpipe CO₂ by fuel economy
| Fuel economy | g CO₂ per mile | g CO₂ per km | Tonnes per 12,000 miles |
|---|---|---|---|
| 15 MPG (large pickup / SUV) | 592.5 | 368.2 | 7.11 |
| 20 MPG | 444.4 | 276.1 | 5.33 |
| 25 MPG (US light-duty average-ish) | 355.5 | 220.9 | 4.27 |
| 30 MPG | 296.2 | 184.1 | 3.55 |
| 35 MPG | 253.9 | 157.8 | 3.05 |
| 40 MPG | 222.2 | 138.1 | 2.67 |
| 50 MPG (full hybrid) | 177.7 | 110.4 | 2.13 |
| 60 MPG | 148.1 | 92.0 | 1.78 |
Notice the curve, not the line: moving from 15 to 20 MPG saves 148 g per mile, while moving from 50 to 60 MPG saves only 30. Replacing the worst vehicle in a household almost always beats improving the best one.
How to read your result
Grams per mile is the number to compare. It strips out how much you drive and isolates the vehicle. EPA prints a CO2 grams-per-mile figure on every fuel economy label, so you can check the calculator against your own car's sticker: it should match closely if you enter the sticker's combined MPG, and it will be higher if you enter your real-world figure, which is usually the honest thing to do.
Tonnes per year is the number to act on. Four tonnes of CO2 from a single car is a large share of a typical household footprint, which is why driving usually appears in the top three categories of a household carbon footprint alongside home heating and electricity.
Per-passenger-mile changes the ranking of travel modes. A solo driver at 355 g/mi is worse per passenger than almost any full bus, train or aircraft. The same car with four people aboard at 89 g per passenger-mile beats most scheduled transport. Occupancy, not mode, is often the dominant variable — the same logic that makes cabin class matter so much in the flight carbon footprint calculator, where a business seat occupies the floor space of two or three economy seats.
Treat the EV comparison as operational only. It compares the emissions of using the two vehicles, not of building them. Battery manufacture carries a real upfront carbon cost that is repaid over tens of thousands of miles, and how quickly depends on the grid, the battery size and the vehicle it replaces. The result also moves over time in the EV's favour as grids decarbonise, while a gasoline car's 8.887 kg per gallon never changes.
Common mistakes
- Using the sticker MPG instead of your real average. Reset the trip computer, or divide miles driven by gallons purchased over several tanks. Real-world economy is often 10 to 20 percent below the combined label figure.
- Confusing MPGe with MPG. A plug-in hybrid's MPGe rating converts electricity into a gasoline-equivalent at 33.7 kWh per gallon. It is not gallons burned, and putting it in the fuel economy field will understate emissions badly.
- Mixing US and Imperial gallons. An Imperial gallon is about 20 percent larger, so a UK MPG figure is about 20 percent higher for the same consumption. Use the unit selector.
- Comparing tailpipe emissions for the car against lifecycle emissions for the EV, or the reverse. Pick one scope and apply it to both.
- Assuming CO2 equals total climate impact. Refrigerant leakage from vehicle air conditioning, road and tyre particulates and manufacturing are all outside this calculation.
- Ignoring the trailer, roof box or load. They do not change the emission factor, but they change the MPG, which is what actually drives the result.
Key terms
- Emission factor
- The mass of a gas released per unit of activity — here, kilograms of CO₂ per gallon of fuel burned. Fixed by fuel chemistry.
- Tailpipe (direct) emissions
- What comes out of the exhaust. This is Scope 1 for a company that owns the vehicle, and it is what vehicle regulations and fuel economy labels report.
- Well-to-wheel
- Tailpipe emissions plus everything upstream: extraction, refining, transport and distribution of the fuel.
- CO₂e
- Carbon dioxide equivalent — other greenhouse gases converted to the mass of CO₂ that would cause the same warming over a chosen horizon, usually 100 years.
- eGRID subregion
- EPA's division of the US power system into regions with distinct generation mixes, each with a published emission rate in lb CO₂e per MWh.
Where this fits in a wider carbon account
For an individual, driving is one of three or four large blocks in a personal footprint, alongside home energy, flights and food. The household carbon footprint calculator uses exactly the emission factors on this page for the vehicle line, so the numbers reconcile.
For a business, the classification matters more than the arithmetic. Under the GHG Protocol, fuel burned in vehicles a company owns or leases is Scope 1. Electricity for a company-owned EV is Scope 2, and is calculated the same way as any other electricity purchase. Employee commuting and business travel in vehicles the company does not control are Scope 3 category 6 and 7. The kilograms are identical; only the reporting bucket changes, and getting the bucket wrong is the most common error in a first-time corporate inventory.
If you are trying to reduce the number rather than just measure it, the leverage runs in a predictable order: drive fewer miles, then carry more people per mile, then improve the vehicle. The reference table shows why the last of those has the weakest marginal effect at the efficient end of the range — going from 40 to 50 MPG saves less per mile than going from 20 to 22. Trip-chaining, working from home one day a week, or one shared commute a week each move the top-line figure more than a modest vehicle upgrade does.