Methodology & Sources

Introduction#

Different flight carbon calculators can give wildly different answers for the same trip — sometimes off by a factor of three — because they make different, scientifically defensible choices about what to include. Showing the work is the only honest move.

This calculator is built on Google's Travel Impact ModelAn open, peer-reviewed methodology that estimates per-passenger CO₂ emissions using aircraft type, route, and cabin class. (TIM) — the same methodology Google Flights uses. We chose it because it is transparent, open-source, and maintained in collaboration with academic and industry partners rather than a private black box.[Google TIM]

Every number we use is sourced below. The default RFI multiplierRadiative Forcing Index: a multiplier that converts CO₂ to CO₂-equivalent by accounting for non-CO₂ warming effects of aviation. , the cabin-class weights, the emissions factors for ground transport — every figure is citable, and where we let you adjust a number, we explain the range.

This page is long but searchable. If you want to verify a specific number quickly, use the quick reference table below.

Quick Reference Table#

Every constant the calculator uses, in one place. Click column headers to sort. Use the search box to filter.

Flight Emissions — the Travel Impact Model#

The Travel Impact Model (TIM) takes origin airport, destination airport, aircraft type, and cabin class, and returns kilograms of CO₂ emitted per passenger for that specific flight. It combines published fuel-burn curves for each aircraft type with typical load factors (how full planes usually are on that route) and cabin-class floor-space weighting.[Google TIM]

What TIM includes#

TIM measures well-to-wake CO₂Emissions covering the full fuel lifecycle: extracting, refining, and transporting jet fuel to the airport, plus burning it in the engine. — not just the CO₂ burned in the engine, but also the upstream emissions from producing and delivering jet fuel to the aircraft. Roughly 80% of the total comes from in-flight combustion; ~20% from fuel production and distribution.

TIM is also aircraft-specific. A Boeing 787 on a route gets a different number than a 767 on the same route, because fuel-burn curves differ by airframe. Google maintains the model as airline fleets evolve, which means the numbers update when real-world efficiency changes. Importantly, since TIM is the same methodology used by Google Flights, our number for a given flight matches what you'd see there.

Cabin class weighting#

A business-class seat occupies roughly the floor space of three economy seats; a first-class suite takes roughly four. The aircraft burns the same fuel regardless of how its interior is configured, so per-passenger emissions scale with how much of the plane's real estate that passenger occupies. TIM's weights are approximately:

• 1.0× — Economy
• ~1.5× — Premium economy
• ~2.9× — Business
• ~4.0× — First

These are fleet averages. Specific cabin configurations on specific aircraft vary. A particularly dense business class (e.g., herringbone-style seating) will have a lower floor-space ratio than a full flat-bed suite.

Load factors#

"Per passenger" emissions depend on how full the plane is. If a flight is unusually empty, each passenger's share of the fuel burn is higher; on a packed plane, lower. TIM uses route-typical load factors because you cannot know in advance exactly how full your specific flight will be. This is a reasonable and standard approximation.

Fallback estimation#

When TIM cannot resolve a specific route — for example, an unusual airport pair with no historical data — we fall back to distance-based estimates. Short-haul flights are more carbon-intensive per kilometer because takeoff and climb consume a disproportionate share of total fuel. Our three-tier factors:

• Short-haul (<1,500 km): ~158 g CO₂/pax-km
• Medium-haul (1,500–4,000 km): ~130 g CO₂/pax-km
• Long-haul (>4,000 km): ~115 g CO₂/pax-km

These derive from TIM's published averages and broadly match ICCT's independent estimates.[ICCT 2020]

Non-CO₂ Effects and the CO₂e Multiplier#

This is the most technically dense and most disputed part of aviation's climate impact. We don't shy away from explaining it.

Why aviation is climate-special#

Burning jet fuel at altitude does things ground-level combustion doesn't:

Contrails. Water vapor from engine exhaust condenses around soot particles at altitude, forming thin cirrus-like clouds. These clouds trap outgoing heat at night more than they reflect incoming sunlight during the day; the net effect is warming. Contrail-induced cirrusIce clouds formed by aircraft exhaust. Their warming effect depends on altitude, humidity, time of day, and atmospheric conditions — making them the most uncertain non-CO₂ effect.

NOx at altitude. Aircraft emit nitrogen oxides at cruising altitude, where they participate in different atmospheric chemistry than at ground level. The net result is increased short-lived ozone (a warming agent) and decreased methane (a cooling agent), with warming dominating.

Water vapor in the stratosphere. A smaller additional warming contribution at very high cruising altitudes.

Together, these non-CO₂ effects mean that aviation's total warming impact is roughly three times its CO₂ alone — the central estimate of Lee et al. 2021 using the GWP* method, with a 5–95% confidence range of 1.9× to 4.7×. Net aviation effective radiative forcing in 2018 was +100.9 mW/m² (5–95% range: 55–145 mW/m²), with contrail cirrus contributing +57.4 mW/m², CO₂ +34.3 mW/m², and NOx +17.5 mW/m². The contrail cirrus ERF/RF ratio of 0.42 indicates contrails are less efficient at surface warming than older RFI-based work suggested, which is part of why the confidence range is wide.[Lee et al. 2021]

What CO₂e means#

CO₂-equivalent (CO₂e)A unit that converts non-CO₂ warming effects into 'how much extra CO₂ would have produced the same warming.' Multiplying CO₂ by an RFI or ERF multiplier gives you CO₂e. is a unit that converts non-CO₂ warming effects into a CO₂ equivalent. Multiplying the raw CO₂ figure by an RFIRadiative Forcing Index. The ratio of total aviation radiative forcing to that from CO₂ alone. Sometimes replaced by ERF (Effective Radiative Forcing) in more recent literature. or ERFEffective Radiative Forcing. A more refined measure of warming impact that accounts for rapid climate adjustments. Lee et al. 2021 uses ERF. multiplier gives you CO₂e.

The four presets#

• 1.9× Conservative (ICCT / IPCC lower bound). Used when you want to be cautious about overstating non-CO₂ effects. Also the low end of the Lee et al. 2021 5–95% likelihood range.[ICCT 2020]
• 2.0× UK DEFRA standard. The multiplier in the UK government's official Greenhouse Gas Conversion Factors used by corporations for reporting. Close to the lower bound; widely accepted for accounting purposes.[UK DEFRA]
• 3.0× Lee et al. 2021 central estimate (our default). The central finding of Lee, D. S. et al., "The contribution of global aviation to anthropogenic climate forcing for 2000 to 2018," Atmospheric Environment, vol. 244, 2021. Using the GWP* warming-equivalent method, the paper concludes that aviation emissions are currently warming the climate at approximately three times the rate of aviation's CO₂ emissions alone. In 2018, non-CO₂ effects (contrail cirrus, NOx, water vapor) accounted for ~66% of aviation's net effective radiative forcing. The paper's 5–95% likelihood range is 1.9× to 4.7×.[Lee et al. 2021]
• 4.7× Upper bound of confidence range. The high end of the Lee et al. 2021 5–95% likelihood interval. Use this when you want the most-aggressive defensible estimate under current peer-reviewed science. Note that the true value for any specific flight is sensitive to altitude, time of day, atmospheric humidity, and engine type, and may sit anywhere within the 1.9× to 4.7× range.[Lee et al. 2021]

Scientific uncertainty#

The science here is still evolving. Contrail-driven warming in particular depends on atmospheric conditions and could in principle be reduced by operational changes — routing around certain humid air masses, for instance. The honest summary: we know non-CO₂ effects matter and roughly triple the warming impact on average (Lee et al. 2021 central estimate), but the 5–95% confidence range of 1.9× to 4.7× is wide. We don't know the exact number for any specific flight.

This is why the multiplier is user-adjustable. There is no single "right" answer, and pretending otherwise would be dishonest.

Ground Transport Comparison Data#

Gas car#

EPA's "Greenhouse Gas Emissions from a Typical Passenger Vehicle" gives 404 g CO₂/mile for an average US passenger car (2023 figure, light-duty fleet, model year 2022 averages). This covers tailpipe CO₂ only. Including upstream fuel production brings the well-to-wheels figure to roughly 485 g/mile.[EPA 2023]

All driving comparisons assume solo occupancy. With multiple passengers the per-person emissions scale down nearly linearly.

Hybrid car#

EPA fueleconomy.gov averages for hybrid passenger vehicles come in around ~190 g CO₂/mile (model year 2023 average, including plug-in hybrids in electric-dominant mode). Wide range across models — a Prius is much lower; a hybrid SUV is much higher.[EPA fueleconomy.gov]

Electric car#

EV emissions depend entirely on the grid powering them. Using the US national grid average from EPA's eGRID 2022 data (~386 kg CO₂ per MWh) and a typical EV efficiency of ~3.5 miles per kWh, an average US EV emits roughly ~100 g CO₂/mile well-to-wheels. The wheels-to-wheels figure is zero; all upstream emissions come from generating the electricity.[EPA eGRID 2022]

State-level variation is significant. In Vermont or Washington (~90% clean generation), an EV can run below 30 g/mile. In coal-heavy states, above 200 g/mile. This calculator uses the national average for v1.

Intercity bus#

Greyhound and similar carriers, when reasonably full, run around ~30 g CO₂/pax-km.[Our World in Data 2023] Buses are arguably the most underrated low-carbon option in the US, because they fill seats efficiently. A near-empty bus is worse; a full bus is the lowest-carbon motorized option for most trips.

Amtrak (US)#

Amtrak's 2023 Sustainability Report states roughly ~85 g CO₂/pax-km system-wide.[Amtrak 2023] Long-distance diesel routes (Coast Starlight, California Zephyr) are higher. Northeast Corridor electrified service is significantly lower — Acela runs at approximately ~50 g CO₂/pax-km, with year-over-year improvements as the regional grid decarbonizes.

Spain AVE#

Renfe's high-speed AVE network runs on Spain's electrified grid, which has substantial renewable penetration. Per-passenger emissions are approximately ~28 g CO₂/pax-km.[Renfe Sustainability Report][EEA]

Other international high-speed rail#

Grid composition dominates electric rail emissions. The same train technology operates at radically different per-passenger footprints depending on where its electricity comes from:

• France TGV: ~3 g CO₂/pax-km. The French grid is ~70% nuclear. SNCF reports this figure with regular EU verification.[SNCF]
• Japan Shinkansen: ~7 g CO₂/pax-km. Mix of nuclear, renewable, and gas-fired generation; JR Central reports this figure.[JR Central]
• Eurostar (London–Paris/Brussels): ~6 g CO₂/pax-km. Reflects the cross-Channel mix of UK and French grids.[Eurostar / EEA]

Bike and walking#

Direct emissions are effectively zero. There is a small food-based footprint — roughly ~5 g CO₂e/km from the extra calories burned, varying with diet.[Our World in Data 2023] We list bike and walking as "≈99% fewer emissions" rather than zero to be technically honest while keeping the comparison meaningful.

Equivalencies#

The equivalencies section translates abstract kg of CO₂e into familiar reference points. All use the CO₂e figure (i.e., after applying the RFI multiplier), not raw CO₂.

Miles driven in an average gas car#

Total trip CO₂e ÷ 404 g/mile = miles driven. This is the most intuitive equivalency because most users have a feel for what 1,000 miles of driving represents.[EPA 2023]

Gallons of gasoline#

Gasoline emits 8.887 kg CO₂/gallon when burned.[EPA] The average US light-duty vehicle gets 28.2 MPG (EPA, 2023 model year). We show gallons of gas so you can substitute your own car's MPG if you prefer.

Trees absorbing CO₂#

A typical mature tree absorbs roughly 21.77 kg CO₂/year.[US Forest Service / EPA] This number varies hugely by species, age, climate, and growing conditions. It is a rough intuition-building figure, not a precise scientific measurement.

Months of US household electricity#

Average US household electricity consumption is ~890 kWh/month.[EIA] At the national grid average of ~386 g CO₂/kWh,[EPA eGRID 2022] that is roughly ~344 kg CO₂/month per household. So: trip CO₂e ÷ 344 = months of electricity. Note this is electricity only, not total household energy (which would include gas heating).

Distance and Routing#

Great-circle distance for flights#

Flight distances use the great-circle distanceThe shortest path between two points on a sphere. It is the theoretical minimum flight distance, ignoring actual routing, weather, and air traffic control. between origin and destination airports, computed from their latitude/longitude coordinates in the OpenFlights dataset.[OpenFlights] Actual flight paths are often longer due to air traffic routing, prevailing winds, and weather deviations. We do not add a routing uplift factor because TIM's per-passenger emissions figure already accounts for typical routing efficiency.

Ground distance estimates#

Ground transport comparisons also use great-circle distance for parity. Real road distances are typically 10–20% longer than great-circle. In a future iteration we would integrate a routing API to get actual road distance; for v1, the order-of-magnitude comparison is the point, and the directional conclusion (flying is much more carbon-intensive) does not change with a 10–20% distance adjustment.

How We Handle Uncertainty#

Not all of our numbers carry the same confidence. We try to be explicit about which are which.

Essentially certain: CO₂ per gallon of gasoline (a chemistry fact), great-circle distance between two airports.

Well-established with real uncertainty: The RFI/ERF multiplier for non-CO₂ effects, contrail warming. The science here is peer-reviewed and mainstream, but the confidence range is wide. We make the multiplier user-adjustable precisely because there is no single defensible answer.

Useful approximations with wide variation: Tree absorption (varies by species, age, climate), household electricity CO₂ (varies by state and grid mix), hybrid and EV emissions (vary by model and grid).

Where we can let you adjust a number, we do. Where we cannot, we pick the most-cited mainstream figure and disclose the source. We update the underlying data when major sources publish updates.

The tone we aim for: honest about what we know and what we don't, rather than projecting false precision.

Limitations#

Surfacing what this tool cannot do builds trust in what it can. Here is what the calculator does not account for:

• Flight-specific routing (winds, ATC delays) — only typical routing is modeled.
• The specific load factor on your flight — only route averages.
• Lifecycle emissions from building, maintaining, or disposing of aircraft, cars, or trains. These are real but small relative to operating emissions.
• Induced demand effects — e.g., your booking making a route more profitable and encouraging more flights.
• Radiative forcing variations by time of day or specific atmospheric conditions during your flight.
• Corporate Scope 3 reporting. For that, use a certified accounting framework; this tool is for personal intuition-building, not compliance.

Full Reference Bibliography#

1. Amtrak. 2023 Sustainability Report. amtrak.com/sustainability
2. Atmosfair gGmbH. Emissions Calculator Methodology. atmosfair.de
3. EPA. Greenhouse Gas Emissions from a Typical Passenger Vehicle. US Environmental Protection Agency, 2023. epa.gov
4. EPA. eGRID 2022 Summary Tables. US Environmental Protection Agency, 2024. epa.gov/egrid
5. EPA. Greenhouse Gases Equivalencies Calculator — Calculations and References. epa.gov
6. EIA. Frequently Asked Questions — How much electricity does an American home use? US Energy Information Administration. eia.gov
7. European Environment Agency. Rail and waterborne — best for low-carbon motorised transport. EEA, 2021. eea.europa.eu
8. Google. Travel Impact Model. github.com/google/travel-impact-model
9. ICCT. CO2 emissions from commercial aviation: 2013, 2018, and 2019. International Council on Clean Transportation, 2020. theicct.org
10. Lee, D. S., Fahey, D. W., Skowron, A., et al. The contribution of global aviation to anthropogenic climate forcing for 2000 to 2018. Atmospheric Environment, 244, 2021. sciencedirect.com
11. OpenFlights. Airports database. openflights.org/data.html
12. Our World in Data. Which form of transport has the smallest carbon footprint? Hannah Ritchie, 2023. ourworldindata.org
13. Renfe. Sustainability Reports and AVE service data. renfe.com
14. SNCF. Environmental Performance of TGV. sncf.com
15. UK DEFRA/BEIS. Greenhouse Gas Conversion Factors for Company Reporting. Annual publication. gov.uk
16. Union of Concerned Scientists. Driving Cleaner: How Electric Cars and Pickups Beat Gasoline on Lifetime Global Warming Emissions. 2022. ucsusa.org
17. US Forest Service. Tree carbon sequestration estimates.

Version and Changelog#

When a constant changes by more than 5%, we note what changed and why.

• 2026-05 — Initial publication.