How to use: drag any slider on the left and the bars, totals and history chart update at once. Start from the 2019 values, or press Pre-industrial or 2 × CO₂. Turn on Show equations (top right) to see the formulae behind every bar. Copy link saves your setup in the address.
Key terms on this tab (29)
Atmospheric state
Defaults are the AR6 values for 2019. Drag a slider; every chart updates immediately.
ℹ About this slider
CO₂ bands are strongly saturated near the line centres; extra CO₂ mostly adds absorption in the band
wings, so forcing grows roughly with the logarithm of concentration:
each doubling adds a similar amount (about 3.9 W m⁻² effective).
ℹ About this slider
CH₄ is far less saturated, so forcing grows roughly as the square root of
concentration. It overlaps with N₂O near 7.7–7.8 µm, which the expressions account for.
ℹ About this slider
N₂O forcing also grows roughly as the square root of concentration. Its strong bands lie at the
edges of the 8–12 µm window, and it persists for about 110 years.
–W m⁻² · CO₂+CH₄+N₂O+halogens
–W m⁻² · total anthropogenic
–K · equilibrium warming if this forcing were sustained
Forcing by agent
Bars: best estimate. Whiskers: AR6 5–95 % range, scaled in proportion to each bar when you move a slider.
Non-gas agents use AR6-based emulators calibrated to reproduce AR6 exactly at default settings (equations: tick “Show equations”).
Total anthropogenic forcing, 1850–2019
The black diamond is your current setup placed at 2019 for comparison with the historical AR6 record.
The equations in use
Non-gas agents (emulators)
Equilibrium warming
How the forcing built up, year by year
Net anthropogenic forcing, :
W m⁻²
How to use: press Play or drag the year slider to watch each agent's forcing grow since 1750. Tick Compare with year to hold a second year beside the first.
Key terms on this tab (13)
Coloured bars follow the slider; grey bars stay on the comparison year.
Axes are fixed so bar lengths can be compared across years. Source: IPCC AR6 WG1 Ch. 7 best estimates, relative to 1750;
natural forcings (solar, volcanic) are not shown.
About each agent
Click an agent for a short explanation.
How to use:Molecule spectra shows what each gas absorbs in a laboratory column. Earth's outgoing spectrum shows what escapes to space; change one gas and read the forcing. Method, limits and ideas to try are below the charts.
Key terms on this tab (40)
Click molecules to switch their absorption on and off. Dips are where a gas blocks infrared; the clear region between about 8 and 12.5 µm is the atmospheric window. Most halocarbons absorb right inside it.
Radiation leaving the top of the atmosphere, computed line by line (cross-sections for the halocarbons). The dashed curves are blackbody emission at the surface (288 K) and the cold stratosphere (about 220 K). Change one gas at a time and watch which part of the spectrum responds.
Sky
–W m⁻² · outgoing longwave radiation
–W m⁻² · instantaneous top-of-atmosphere forcing vs the 1750 atmosphere
Where the extra gas removes radiation
Forcing against amount
Why 2 W m⁻² and not 3.9? A worked example for doubled CO₂
What this shows: the forcing on this tab is an instantaneous value at the top of the atmosphere. IPCC AR6 quotes an effective forcing. The steps between them are computed here with the same radiative-transfer model.
The same steps for methane and nitrous oxide (all-sky, 1750 to 2019)
Change
Instantaneous (W m⁻²)
Stratosphere-adjusted, this model
AR6 stratosphere-adjusted
Difference
Method. Fixed dynamical heating: the surface and troposphere are held fixed and the stratospheric temperatures are changed until each layer's longwave cooling returns to its value in the 1750 atmosphere, so solar and dynamical heating balance it again. Only the longwave is recomputed, absorption cross-sections are kept at their reference-temperature values, and the model has 12 stratospheric layers above 12 km. What to read into the numbers: carbon dioxide and nitrous oxide agree with AR6 to about 2 %; the model's methane forcing is about 14 % lower than AR6's. The cause was not isolated (the single idealised profile, the idealised methane vertical profile and band overlap with water vapour are candidates), so treat the methane number as the least certain. The tropospheric adjustment factors (CO₂ +5 %, CH₄ −14 %, N₂O +7 %) are AR6's, not computed here.
Method, limits and ideas
Data and method. H₂O, CO₂, CH₄, N₂O and O₃: HITRAN2020 absorption lines (Gordon et al. 2022) with Voigt line shapes. Water vapour also includes the MT_CKD 4.3 continuum (Mlawer et al. 2012; AER). Halocarbons: HITRAN infrared cross-sections for 17 molecules. Layer-by-layer radiative transfer through the 1976 US Standard Atmosphere (idealised water-vapour and ozone profiles; halocarbons uniformly mixed). Only derived spectra are shown.
All-sky estimate. Clouds are represented by one grey deck (cover 0.67) whose opacity is tuned so the model's longwave cloud effect equals the satellite value of 27.9 W m⁻² (CERES EBAF; Loeb et al. 2018). It is a global-mean estimate, not a cloud model, and clear-sky values are shown for comparison.
Limitations. One atmospheric profile, no line mixing, one gas varied at a time, instantaneous forcing at the top of the atmosphere without stratospheric adjustment (so it is smaller than AR6's effective forcing for CO₂; the worked example above adds the missing steps). Halocarbon spectra exist only over the wavenumber ranges measured in the laboratory; outside them the lines are left blank, not drawn as transparent.
Ideas to try
Choose CFC-12, HFC-134a or SF₆ and slide the column from 10¹⁶ to 10¹⁹ molecules cm⁻²: the bands sit inside the window, which is why a few parts per trillion matter.
Raise CO₂ over many orders of magnitude and watch the 15 µm band flatten while its wings keep growing.
In the outgoing spectrum compare the forcing of 1000× today's SF₆ with the straight-line (AR6) prediction: strong bands begin to saturate.
Switch between clear sky and the all-sky estimate: window absorbers lose about a third of their forcing to clouds.
From forcing to warming
How to use: pick a scenario, then change the climate sensitivity and the ocean mixing. The model turns the forcing of that scenario into temperature in a fraction of a second, and compares it with thermometer records and with what the IPCC assessed.
Key terms on this tab (23)
Scenario
Warming after CO₂ doubles and the climate has fully settled. AR6: best estimate 3 K, likely 2.5–4 K.
How fast heat moves into the deeper ocean. Faster mixing delays surface warming.
–K
–K
–K · equilibrium climate sensitivity (ECS) of your model
–K · transient climate response (TCR); AR6 best estimate 1.8 K, likely 1.4–2.2 K
Global warming
Thermometer record: HadCRUT5 (Met Office Hadley Centre and University of East Anglia), re-based to 1850–1900. The projections start from the same forcing as the FaIR model; only the way the climate responds changes with your sliders.
Forcing behind the scenario
Effective radiative forcing by agent, relative to 1750. Aerosol cooling shrinks as air pollution is cleaned up, which is why warming can speed up even when CO₂ emissions are flat. Natural forcing (solar, volcanic) is the same in every scenario after 2019 in this tool.
Carbon dioxide
Observed: NOAA Global Monitoring Laboratory, Mauna Loa annual mean.
How this works, and what it leaves out
The model. A three-layer energy balance model: a thin surface layer exchanging heat with a mid and a deep ocean. It is the temperature part of FaIR v2 (Leach et al. 2021), with the heat capacities and exchange rates set to the median of the AR6-constrained calibration (Smith et al. 2024). ECS is changed by altering the climate-feedback parameter only.
The forcing. From one FaIR run per scenario using RCMIP emissions, with the same median climate parameters. It is a single deterministic run, not the full uncertainty ensemble, so the grey and coloured lines have no uncertainty band. The AR6 assessed range is shown for 2081–2100 instead.
Check against AR6. The reference setting gives 1.36, 1.73, 2.59, 3.44 and 4.27 K for 2081–2100 in the five scenarios, inside the AR6 very likely ranges and 0.1 K below its best estimates. For 2011–2020 it gives 1.00 K, inside the AR6 assessed 0.95–1.20 K but about 0.1 K below the thermometer record (HadCRUT5 1.12 K). ECS is 3.0 K and TCR 1.7 K.
Limits. The sliders change how the climate responds, not the emissions: concentrations and forcing stay as in the FaIR run. There is no regional climate, no weather and no tipping points. Past 2025 these are scenarios, not predictions.
Comparing gases: GWP and GTP
How to use: choose a gas to see its global warming potential (GWP) and temperature change potential (GTP) at several time horizons, how they change with the horizon, and where the gas sits among 249 others. The calculator below turns emissions into CO₂-equivalent.
More terms used on this tab (13)
GWP: global warming potential
Imagine releasing 1 kg of a gas today. It adds extra radiative forcing for as long as it stays in the air. The GWP adds up (integrates) that extra forcing over a chosen number of years, the time horizon, and divides by the same quantity for 1 kg of CO₂. So GWP-100 = 27.9 for methane means 1 kg of methane traps as much heat over 100 years as 27.9 kg of CO₂.
The numbers 20, 100 and 500
They are time horizons in years. GWP-20 stresses the near term, so it is large for gases that disappear quickly (methane: 81). GWP-100 is the most widely used, the one in national inventories and the Kyoto Protocol (methane: 27.9). GWP-500 stresses the long term, so it favours gases that last (methane: 7.95). No horizon is the “right” one: choosing it is a value judgement about how much the near term matters compared with the long term.
GTP: global temperature change potential
Instead of adding up forcing, the GTP asks one question: how much warmer is the planet in a single chosen year after the pulse, compared with the same mass of CO₂? GTP-50 is 50 years after the emission, GTP-100 is 100 years after. Because it ignores what happened before that year, it forgets short-lived gases faster than the GWP does.
Related terms
AGWP and AGTP are the absolute versions before dividing by CO₂ (W m⁻² yr per kg, and K per kg). CO₂-equivalent is the emission multiplied by the metric: 1 Mt of methane is 27.9 Mt CO₂e at GWP-100. Pulse means a one-off emission. The values here include the extra CO₂ released because the warming weakens carbon uptake (the climate–carbon feedback), as in AR6. See the Glossary for lifetime and radiative efficiency.
–GWP-20 · 20-year horizon
–GWP-100 · 100-year horizon
–GWP-500 · 500-year horizon
–GTP-50 · temperature at year 50
–GTP-100 · temperature at year 100
–lifetimeRE – W m⁻² ppb⁻¹
How the metric changes with the time horizon
Compare with
Lines are computed here; diamonds are the AR6 table values. A gas that is gone within years (methane) loses most of its GWP as the horizon lengthens; a gas that lasts millennia (SF₆, CF₄) keeps its potency. GWP adds up forcing over the whole horizon, while GTP asks only about the temperature at one moment, so it forgets short-lived gases faster.
Why gases differ: lifetime against radiative efficiency
Every gas in AR6 Table 7.SM.7 that has a lifetime and a radiative efficiency. Up and to the right means a stronger absorber that also stays longer, so a higher GWP. Per molecule, halocarbons absorb far more than CO₂ (see the Spectra tab for the reason: they sit in the atmospheric window).
CO₂-equivalent calculator
Gas
Emission (Mt)
Factor
Mt CO₂e
Enter emissions in megatonnes (1 Mt = 10⁹ kg). The bar chart is logarithmic because the gases differ by orders of magnitude. CO₂-equivalent depends on the metric and horizon you pick; there is no single "right" one.
Method and limits
Method. For a 1 kg pulse, the absolute GWP is the forcing added up over the horizon and the absolute GTP is the temperature at the horizon, from a two-timescale climate response. Dividing by the same quantity for CO₂ (Joos et al. 2013 carbon cycle) gives the potential. The implementation follows IPCC AR6 WG1 Section 7.6.1 and includes the climate–carbon feedback, the CH₄ and N₂O adjustments and the ozone terms.
Checked against AR6. CO₂'s absolute values match the AR6 table within 0.1 %. For the 136 halogenated gases with a lifetime of at least a year, all five metrics here fall within 1.6 % of AR6 Table 7.SM.7 (135 of them within 1 %). Methane reproduces 27.9 and nitrous oxide 273.
Limits. Methane uses the non-fossil value (27.9); AR6 gives a higher fossil-methane value (29.8) that includes the CO₂ from its oxidation. Gases living under a year, and two with no usable formula, are shown with AR6 values only. Metrics are conventions for comparing emissions, not predictions of warming.
Glossary and naming guide
How to use: decode a gas name in the first box, look up any of the 22 molecules used on the Spectra tab in the table, or search the terms below by word or topic.
Decode a halocarbon name
CFC-11, HCFC-22, HFC-134a and PFC-218 all follow one rule: add 90 to the number. The three digits you get are the number of carbon, hydrogen and fluorine atoms. Every other bond on the carbons holds chlorine.
The molecules on the Spectra tab
Name
Formula
Family
Lifetime
RE (W m⁻² ppb⁻¹)
GWP-100
Where it comes from
RE is the radiative efficiency, GWP-100 the global warming potential over 100 years. Values are from IPCC AR6 Table 7.SM.7. They are defined in the glossary below.