A new AI-powered satellite will create Google Maps for methane pollution

Google and the Environmental Defense Fund have teamed up to track the elusive emissions—from space.
MethaneSAT concept art above Earth
Methane is very hard to track around the world, but a new satellite project could help address the issue. MethaneSAT LLC

Share

Methane emissions, be it from industrial cattle farming or fossil fuel extraction, are responsible for roughly 30 percent of the Earth’s climate change issues. But despite the massive amounts of methane emissions released into the atmosphere every year, it’s often difficult to track the pollutant—apart from being invisible to the human eye and satellites’ multispectral near-infrared wavelength sensors, methane is also hard to assess due to spectral noise in the atmosphere.

To help tackle this immediate crisis, Google and the Environmental Defense Fund are teaming up for a new project with lofty goals. Announced in a new blog post earlier today, MethaneSAT in a new, AI-enhanced satellite project to better track and quantify the dangerous emissions, with an aim to offer the info to researchers around the world.

Google Earth Image screenshot displaying methane geodata map
EDF’s aerial data, available in Earth Engine, shows both high-emitting point sources as yellow dots, and diffuse area sources as a purple and yellow heat map. MethaneSAT will collect this data with the same technology, at a global scale and with more frequency. Credit: Google

“MethaneSAT is highly sophisticated; it has a unique ability to monitor both high-emitting methane sources and small sources spread over a wide area,” Yael Maguire, Google’s VP and General Manager of Geo Developer & Sustainability, said in a February 14 statement.

[Related: How AI could help scientists spot ‘ultra-emission’ methane plumes faster—from space.]

To handle such a massive endeavor, the EDF developed new algorithmic software with researchers at the Smithsonian Astrophysical Observatory andHarvard University’s School of Engineering and Applied Science and its Center for Astrophysics. Their new supercomputer-powered AI system can calculate methane emissions in specific locations, and subsequently track those pollutants as they spread in the atmosphere. 

MethaneSAT is scheduled to launch aboard a SpaceX Falcon 9 rocket in early March. Once deployed at an altitude of over 350 miles, the satellite will circle the Earth 15 times per day at roughly 1,660 mph. Aside from emission detection duties, Google and EDF intend to harness their AI programs to compile a worldwide map of oil and gas infrastructure systems to hone in what aspects rank as the worst offenders. According to Google, this will function much like how its AI programs interpret satellite imagery for Google Maps. Instead of road names, street signs, and sidewalk markers, however, MethaneSAT will help tag points like oil storage containers.

Google satellite imagery displaying oil wells
The top satellite image shows a map of dots, which are correctly identified as oil well pads. Using our satellite and aerial imagery, we applied AI to detect infrastructure components. Well pads are shown in yellow, oil pump jacks are shown in red, and storage tanks are shown in blue. Credit: Google

“Once we have this complete infrastructure map, we can overlay the MethaneSAT data that shows where methane is coming from,” Maguire said on Wednesday. “When the two maps are lined up, we can see how emissions correspond to specific infrastructure and obtain a far better understanding of the types of sources that generally contribute most to methane leaks.” Datasets like these could prove valuable to watchdogs and experts attempting to rein in oil and gas emission locations that may become more prone to leaks.

All this much-needed information is intended to become available later this year through the official MethaneSAT website, as well as Google Earth Engine, the company’s open-source global environmental monitoring platform. In the very near future, the new emissions data will be able to combine alongside datasets concerning factors like waterways, land cover, and regional borders to better assess where we are as a global community, and what needs to be done in order to stave off climate change’s worst outcomes.