Several COP28 announcements could have a transformative impact on the nascent methane detection sector, and especially for those companies that rely on artificial intelligence and other advanced technologies to identify leaks.
The climate talks saw the unveiling of a $255 million global flaring and methane reduction fund and a United States Environmental Protection Agency rule aiming to curb air pollutants. Both developments are designed to slash emissions of methane, which has 84 times the global warming potential of carbon dioxide over a 20-year timeframe and is responsible for almost a third of the emissions-induced increase in global temperatures since the start of the industrial era.
“This last week has resulted in some watershed moments for methane action,” said Riley Duren, CEO of the California non-profit Carbon Mapper.
Using aerial remote sensing technology perfected at NASA’s Jet Propulsion Laboratory in Pasadena, Carbon Mapper identifies methane leaks of more than 100 kilograms per hour and reports them to U.S. regulatory bodies. This top-down approach to detection has been used to identify more than 11,000 methane plumes, details of which are published on a public portal.
These include several hundred methane super leaks on U.S. soil, around 100 of which have resulted in voluntary mitigation on the part of the emitter. However, Carbon Mapper’s technology can only identify large methane leaks, and is dependent on aircraft or satellite coverage of a given area.
To plug the gaps in this kind of detection, other startups are resorting to bottom-up calculations that use AI and machine learning to calculate the likely emissions from visible exhaust plumes.
One such company, Orbital Sidekick (or OSK), is expecting a tenfold increase in revenues next year following the COP methane announcements, CEO Dan Katz told Latitude Media. Some of OSK’s largest clients are oil and gas pipeline and well pad operators, who will be facing increased pressure post-COP to get methane emissions in check.
“OSK has seen increased demand for its AI/ML-enabled energy infrastructure monitoring services in the second half of 2023,” Katz said. “This has also correlated with significant funding interest to grow and scale our satellite constellation, to service more customers.”
OSK operates a hyperspectral intelligence platform dubbed SIGMA, which it says allows its clients to “meet and exceed regulatory compliance, prevent and detect oil and methane leaks, and assess the overall risk profile of their assets,” he added.
The company is aiming to expand operations in 2024, with a view to offering daily monitoring of energy infrastructure anywhere in the world.
Vít Růžička, a machine learning researcher at the University of Oxford, said the efficacy of the AI technology will improve with increased commercialization. (Růžička is collaborating on methane detection with the consulting and research company Trillium Technologies.)
“AI automation requires further validation by human experts for the moment,” he said. “This kind of workflow will really come into its own as more hyperspectral instruments are flown.”
Analyzing hyperspectral data “is particularly labor intensive,” Růžička added. “That is when AI will be hugely valuable — even just a pre-screening of the huge quantities of incoming data can help.”
Furthermore, automated systems may be able to trigger follow-up verifications of areas of interest, and by doing so collect more evidence that Růžička said “can help bulletproof the manual analysis.”
In November, the analyst firm Wood Mackenzie said methane remains a significant challenge for the oil and gas industry.
“Tracking methane emissions has been a huge challenge,” said Elena Belletti, Wood Mackenzie’s global head of carbon research, at the time. “Current satellite technology is not yet able to detect smaller leaks, and airborne or on-the-ground sensors present significant trade-offs on coverage and comparability between different sites.”.
Duren at Carbon Mapper said the COP announcements would likely help close this information gap.
“What we’ve seen to date has been voluntary,” he said. “What’s been missing are stronger incentives to increase participation.”
