Airlines Target Contrails to Cut Their Climate Impact

Aviation's most promising climate lever needs no new aircraft and no new fuel. It can run through flight-planning software airlines already use. In one of the largest operational trials yet run, dispatchers used it on fewer than one flight in ten.

 

The 62 percent reduction figure from American Airlines and Google has been quoted in every write-up of that trial. The number underneath it, how often dispatchers actually used the system, has barely been reported. Which one you look at decides whether contrail avoidance is a real near-term option or another aviation promise with a 2050 delivery date.

 

In this trial, much of the gap was operational rather than atmospheric: flight planning software, dispatcher workload and a set of habits no climate model accounts for. That makes contrail avoidance a different kind of problem from the rest of aviation decarbonisation, and possibly an easier one.

 

What a contrail actually is

 

A jet engine burns kerosene and produces carbon dioxide, water vapour and tiny particles of soot. At cruising altitude the air is far below freezing, and how much moisture it holds changes from one patch of sky to the next. Hot, damp exhaust meets that cold air, the vapour condenses onto the soot and freezes almost instantly. That is the white streak behind the aircraft.

 

Most break up within minutes because the surrounding air is too dry to sustain the ice. The problem is the minority that survive. In air already saturated with moisture, the crystals persist, spread sideways and form a sheet of artificial cirrus lasting hours. It reflects some sunlight, which cools, but absorbs more outgoing heat, which warms. The net effect is warming, strongest at night.

 

One statistic shapes everything that follows: roughly 2 percent of flights produce around 80 percent of annual contrail energy forcing, the warming a contrail delivers over its lifetime.

 

The science got smaller. The opportunity got bigger.

 

The figures have fallen a long way since 2021, when David Lee at Manchester Metropolitan University and colleagues published the assessment that still anchors most coverage. They put contrail-cirrus radiative forcing for 2018 at 111 milliwatts per square metre against aviation's cumulative CO2 forcing of 34, the source of the claim that contrails warm three times as much as aviation CO2. The confidence interval ran from 33 to 189, which rarely survives into headlines.

 

Where it went. In 2024, Roger Teoh's group at Imperial College London rebuilt the estimate using real flight trajectories and arrived at 62.1 for 2019, around 44 percent lower. Marius Bickel and colleagues at the German Aerospace Center then went further, modelling the surface-temperature response rather than the radiation balance. They found contrail cirrus produces far less surface warming per unit of forcing than CO2, with a central efficacy estimate, a ratio rather than a forcing, of 0.21. In the model, rising CO2 burns off low reflective cloud and amplifies its own warming; contrail cirrus produced no such feedback. Lee did not defend his own figure. Speaking to Science last November, he said the revision reversed his paper's message and called the smaller number better knowledge.

 

A complication. Work led by Andreas Petzold at Forschungszentrum Jülich found more than 80 percent of long-lived contrails form inside cloud already there, where their added warming can be smaller and harder to pin down. That is a targeting problem: avoidance systems assume you can tell which contrails matter.

 

The consensus holds regardless. Sebastian Eastham, also at Imperial, points out: contrails warm, and not by a small amount. What changed is size and confidence, not direction. And a smaller effect removable at five to twenty dollars a tonne still beats a larger one costing three hundred to address through fuel.

 

Three climate metrics that are easy to mix up

Radiative forcing (RF) is the raw energy imbalance a contrail creates. Most studies report this one.

Effective radiative forcing (ERF) includes rapid atmospheric adjustments that follow the initial forcing. For contrail cirrus, current estimates put it below classical RF: Lee's contrail ERF was around 57, against an RF of 111.

Global warming potential (GWP) restates everything as CO2-equivalent over a fixed period. EU reporting requires 20, 50 and 100-year versions, which give very different answers.

The error to watch for is comparing one study's ERF against another's RF. Set contrail ERF of 57 against CO2 RF of 34 and you get 1.7 times. Like for like, it is 111 against 34. If a source does not say which measure it used, the number cannot be checked.

 

The trial that proved the physics

 

American Airlines, Google, flight-planning provider Flightkeys and Breakthrough Energy's Contrails.org ran a randomised controlled trial over seventeen weeks in early 2025, across roughly 2,400 eastbound transatlantic flights on twenty-eight city pairs from American's ordinary schedule. Contrail forecasts were built into the software dispatchers already use.

 

Intent to treat. 1,232 flights crossed the warming threshold and formed the eligible group. Across all of them, contrail formation fell 11.6 percent (p=0.011).

 

Per protocol. Among the 112 planned for avoidance and flown that way, observable contrail formation fell 62.0 percent (p<0.001), with roughly 69 percent less estimated warming and no significant fuel difference.

 

The 62 percent result was validated using satellite observations and automated flight attribution rather than a contrail-formation model, with analysts blind until the trial closed. The warming reduction is a climatological estimate. Both carry the same limit, which the authors state plainly: not all contrails are visible from a satellite, so the true reduction is likely smaller. The paper is also not yet peer-reviewed, posted to arXiv in March 2026 and still under review.

 

The 112 that matters more than the 62

 

Of 1,232 eligible flights, 190 were released with a contrail-optimised plan and 112 flew it as planned. Fewer than one in ten. If avoidance is as cheap and effective as that suggests, where did the other 1,120 chances go?

 

What the study establishes. Participation was voluntary, and dispatcher priorities ran safety first, efficiency second, contrails third. The paper also names specific obstacles. Avoidance plans sometimes needed climbs or descents mid-flight, safe and coordinated with air traffic control but which crews preferred not to fly. American also prefers the North Atlantic Organised Track System, and the Flightkeys features that make track adherence easy could not be switched on for contrail flights. Dispatchers hitting that friction often did not attempt avoidance. The interface lacked a vertical profile view, so they could not see why the software wanted a given altitude. Take rates varied widely and fell as workload rose.

 

What it does not establish. The paper does not say how much of the shortfall belongs to each cause, and warns the take rate reflects this trial's design and may not generalise.

 

One barrier was not scientific uncertainty or fuel cost, but a software feature that could not be switched on in time. So the miss rate is not a verdict on the technology, though nobody has shown the gap closes when the workflow is mandatory or tied into air traffic control. The physics worked; the system around it mostly did not. An airline citing 62 percent while flying the plan on 7.8 percent of flights is not delivering 62 percent.

 

Europe moves from measuring to paying

 

Stage one: mandatory measurement. Since 1 January 2025, under Implementing Regulation (EU) 2024/2493, airlines in the EU Emissions Trading System have had to monitor and report non-CO2 effects including contrails. Coverage runs to intra-EEA flights plus departures to Switzerland and the UK, widening to all flights from 2027. No allowances must be surrendered: airlines count, but do not pay.

 

Stage two: a price signal, proposed. On 17 July 2026 the Commission published COM(2026) 616, revising the aviation ETS. For the first time it targets contrails directly: up to three million allowances through 2033, capped at 420,000 a year, for carriers building approved contrail forecasting into flight planning.

 

Two cautions. This is only at the start of the legislative procedure and can still be amended. And IATA's response has been hostile, though not on contrail grounds: Willie Walsh attacked the geographical extension as "repeating a historic error." Transport & Environment wants the contrail measures strengthened.

 

The case for caution

 

The strategy depends on forecasting humidity at altitude, among the hardest things to predict, and a wrong forecast burns certain, long-lived CO2 to avoid a contrail that would never have formed. Modelling puts the fuel penalty at 0.1 to 0.2 percent across an airline's flights, with climate benefit far exceeding the extra warming, but the trial could not confirm it: real-world fuel use varies too much to isolate a penalty that small. Petzold's finding compounds the problem, since some manoeuvres will target contrails doing little damage.

 

American has been measured about the pace, and Jill Blickstein, a co-author on the trial, frames the work as still answering scientific questions. Defensible for a carrier that would wear the reputational cost if the numbers moved. It also requires nothing.

 

What it adds up to by 2050

 

A Cambridge study led by Jessie Smith, published in Nature Communications in January 2026, projects that without avoidance, aviation adds 0.040 K of CO2 warming and 0.054 K of contrail warming by 2050, together taking 19 percent of the remaining budget to two degrees. Avoidance phased in between 2035 and 2045 could recover around 9 percent of that budget; delay a decade and it falls to roughly 2 percent. The headline assumes very high effectiveness, so treat it as an upper bound. What matters is the decay rate: 0.6 percent of the budget lost per year of delay.

 

The ICCT's Aviation Vision 2050, published September 2025, put contrail avoidance at 42.5 percent of avoided warming in 2050 under its most ambitious scenario, at five to twenty dollars per tonne against more than three hundred for sustainable fuel. These are separate studies using different measures, and they do not add up. The ICCT also part-funded the American trial, worth knowing when reading its advocacy.

 

"Governments aren't yet prioritizing the low hanging fruit of contrail avoidance." Xinyi Sola Zheng, International Council on Clean Transportation, September 2025.

 

Where this hits corporate reporting

 

Business air travel falls under Scope 3 Category 6, but its non-CO2 effects are not required in the core inventory and remain an area of developing guidance. GHG Protocol is reconsidering how aviation radiative forcing is accounted for, with the consolidated GHG Protocol-ISO standard now due in Q4 2028. For aviation-sector targets, the Science Based Targets initiative requires formulations to state explicitly that they exclude non-CO2 factors. A substantial part of a flight's climate impact does not appear in the number most organisations disclose.

 

The 2025 UK conversion factors showed how unstable the ground can be, cutting air travel emission factors hard, with long-haul falling around 40 percent, driven by post-COVID load factor recovery and updated aircraft data, not anything airlines did. A company could fly more, report less, and call it progress without misstating a thing.

 

Five questions for your airline

1. Can your flight planning system generate contrail-avoidance routings?

2. What proportion of eligible flights actually fly the plan? The 7.8 percent question.

3. Is avoidance a dispatcher option or the default? Voluntary participation produced single-digit adoption above.

4. How is performance verified? Satellite attribution is the standard the American trial set.

5. What are you reporting under the EU non-CO2 regime, and will you share it?

 

Few travel buyers ask any of this. They press carriers on sustainable fuel blending, expensive and supply-constrained, while giving far less attention to something deployable with today's aircraft.

 

The software feature that stopped dispatchers avoiding contrails over the North Atlantic can be switched on. Forecasts will improve. What is missing is a reason for airlines to bother, and the EU is now proposing a direct financial incentive.

 

An industry that has spent two decades explaining that decarbonisation is unaffordable now has a comparatively cheap option in front of it, and is still deciding whether to pick it up.

 

Sources: Lee et al. 2021, Atmospheric Environment, for the 2018 forcing baseline. Teoh et al. 2024, Atmospheric Chemistry and Physics, for the revised 2019 estimate. Bickel et al. 2025, Journal of Climate, for contrail efficacy. Petzold et al. 2025, Nature Communications, on contrails forming in existing cirrus. Sankar et al., arXiv preprint under review at JECATS, for the American Airlines trial. Smith et al. 2026, Nature Communications, for the 2050 modelling. ICCT Aviation Vision 2050, September 2025. Regulation: Implementing Regulation (EU) 2024/2493 and COM(2026) 616. UK factors: DESNZ 2025 conversion factors methodology paper. Lee and Eastham quoted via Science, November 2025.