top of page


There has been widespread political consensus, including the Paris Agreement, that ‘2°C’ above pre-industrial levels represent the upper threshold for unsafe and potentially irreversible climate change, with urgent mitigation actions demanded across all sectors to stay below 1.5°C, including aviation. However, the gap between what is necessary to avoid 1.5°C and aviation’s contribution to global warming has uncomfortable implications. 

The strategies pursued to reduce aviation-induced climate impact over the last years have followed a common, win-win pattern for airlines and linked economical agents: they have sought to reduce fuel burnt, claiming that their strategies are climate-friendly due to fuel burnt being proportional to CO2 emissions. This claim is however biased: global aviation emissions contribute to anthropogenic climate change by warming the Earth’s near-surface atmosphere through both carbon dioxide (CO2) and non-CO2 emissions (following Lee et al. 2021). The latter have been practically ignored in the past. This is, in part, because they are not yet fully understood and still linked with medium to high uncertainties. Contrails effects show large uncertainties since they are subject to meteorological, regional, and seasonal variations, yet each being individually characterised by different spatiotemporal scales.


Indeed, under some specific circumstances, aircraft can generate anthropogenic cirrus with cooling effects -negative Radiative Forcing (RF). Thus, together with continuing the decarbonisation, the need for research into contrails (and other non-CO2 effects) and its associated uncertainties to be considered in aviation climate mitigation actions becomes unquestionable.

Non-CO2 emissions associated with air traffic comprise water vapour (H2O), nitrogen oxide (NOx), sulphur oxides and soot. Not all non-CO2 emissions have a direct effect on climate. NOx emissions are not radiatively active themselves, but they are responsible for the chemical production of the greenhouse gas (GHG) ozone (O3) and the destruction of the GHG methane (CH4). Furthermore, induced by non-CO2 emissions, contrails and contrail-cirrus can form and alter the radiation budget (effective radiative forcing -ERF-). It provides state-of-the-art estimates (and their uncertainties) of the global aviation climate impact for individual forcing components. Red bars indicate a warming, while blue bars indicate a cooling of the atmosphere. Comparing the individual forcing components reveals that the largest contributions to the overall positive radiative forcing are due to contrail-cirrus formation and aviation-induced cloudiness (57% of the ERF) and overall, non-CO2 forcings represent about two thirds of global aviation ERF. Nonetheless, the understanding of aviation non-CO2 is still incomplete, and this is reflected by the fact that non-CO2 terms (in particular, those associated to contrails and aviation-induced cloudiness) are also the ones with the largest uncertainties.


We ambition to blend cutting-edge AI techniques, namely deep learning, and climate science with application to the aviation domain, aiming at closing (at least partially) the existing gap in terms of understanding aviation-induced climate impact. Indeed, the possibility of using Artificial Intelligence (AI) techniques is emerging with great force in various fields. In a recent survey published in Nature communications by Vinuesa et al., the authors claim that AI can enable the accomplishment of 134 targets across all the goals established in 2030 Agenda for Sustainable Development. This includes Earth sciences (climate change and meteorological prediction) and the aviation domain.

bottom of page