a potentially beneficial use of deep learning and artificial intelligence in limiting the climate impact of planes.
The firm said that by avoiding flight paths where contrails – the streak of white, smoke-like material sometimes seen coming out of planes as they fly through the air – are likely to form, planes can significantly reduce their global warming impact. account for about 35% of all of aviation’s global warming impact. Google explained that contrails can “persist as cirrus clouds for minutes or hours” depending on atmospheric conditions, which traps heat that should have left the earth’s atmosphere.
Google’s proposal is to use AI to see flight paths where contrails may occur, so planes can be advised to avoid them. It has partnered with American Airlines and the Bill Gates-founded sustainable energy organization Breakthrough Energy to process satellite imagery, weather and flight path data, to make the necessary predictions.
Seventy flight tests were done in a six-month period showing that pilots were able to reduce contrails by 54%. Planes were shown to have increased fuel consumption by 2% as it altered its flight, but Google said that, at scale, extra fuel burned across an airline’s flights could be as low as 0.3%. This suggested that, according to Google, airlines can avoid contrails for costs at about $5 to $25 per ton of carbon dioxide emissions avoided.
While the results show promise, the test sample size is still small and the research isn’t yet peer-reviewed as noted by