The central test: A climb or descent may require additional fuel and therefore emit more CO2. The trial is designed to measure the full trade-off: actual fuel penalty against the estimated reduction in contrail-caused warming, including whether forecasts correctly identify the small fraction of flights and airspace responsible for a disproportionate share of the effect. It is not yet evidence that the manoeuvres produce a net benefit; that is the experiment’s main question.
Scale, money and partners: Department for Transport support is through the Aerospace Technology Institute programme. The consortium includes Google, NATS, Contrails.org, Imperial College London, the University of Cambridge and the Met Office, which is working as a subcontractor to Contrails.org.
How results are to be checked: Satellite imagery and flight data will be used to see whether the forecast contrail regions and avoidance instructions correspond with observed outcomes. Academic partners provide independent scientific scrutiny; the programme says its methods and results are intended for publication, which matters because contrail prediction, climate metrics and fuel trade-offs remain scientifically and operationally uncertain.
What it builds on: Google and Contrails.org have developed tools that link satellite-detected contrails to individual flights and provide contrail forecasts, while earlier airline demonstrations have tested flight-planning or tactical altitude changes. Blue Skies’ distinction is moving from airline-led, individual-flight experiments to controller-led operations across a busy oceanic corridor.
Potential significance: If validated, the approach could give aviation a relatively near-term way to address part of its non-CO2 climate impact without waiting for fleet replacement or new fuels. It would still be a complement—not a substitute—for cutting CO2, because avoidance cannot remove emissions from fuel burn and cannot eliminate all contrails.
There is a real accounting question, but it should not be overstated as a direct contradiction. Google reports that its data-centre electricity demand rose 27% in 2024 amid growing AI demand, while it says data-centre energy emissions fell 12% through clean-energy procurement and efficiency measures. A contrail-avoidance system could yield a climate benefit far larger than the electricity needed to run its forecasts—but that has not yet been publicly demonstrated with a transparent, end-to-end lifecycle comparison.
The credible standard is therefore to publish: the model-compute energy and associated marginal emissions; observed fuel changes; observed contrail reductions; the climate metric and time horizon used; forecast false positives and missed contrails; and an independent assessment. Until that is available, the claimed climate benefit is promising but unproven.