Sun. Dec 3rd, 2023
Using Machine Learning to Detect Methane Super-Emitter Plumes from Space

The Copernicus Sentinel-5P satellite, equipped with the Tropomi instrument, is revolutionizing our ability to detect methane leaks and monitor their impact on climate change. Methane is a potent greenhouse gas, and reducing its emissions is crucial in the fight against global warming.

Methane has a shorter lifespan in the atmosphere compared to carbon dioxide, but it can retain 30 times more heat over a century. This means that reducing methane emissions can have a significant impact on global warming within a relatively short timeframe.

Methane super-emitters, which release large amounts of methane compared to other sources, are particularly important to address. These super-emitters are often found in industrial facilities such as oil and gas operations, coal mines, and landfills, and fixing them can lead to significant climate gains.

The Tropomi instrument on the Copernicus Sentinel-5P satellite is the only instrument that provides a daily global map of methane concentrations. It measures methane by observing Earth’s atmosphere using shortwave infrared bands, which act as unique fingerprints for methane. This allows for precise detection of methane emissions.

Researchers from SRON Netherlands Institute for Space Research have developed a new algorithm that uses machine learning to automatically identify methane super-emitter plumes in the Sentinel-5P data. The algorithm also calculates the associated emissions based on measured concentrations and wind speeds.

The detection of these methane plumes is a collaborative effort. The information gathered is shared with the International Methane Emissions Observatory and other satellites with higher resolution for further investigation and potential mitigation actions.

In 2021 alone, the algorithm identified 2,974 methane super-emitter plumes. The majority of these plumes were from oil and gas facilities (45%), followed by urban areas (35%) and coal mines (20%).

The findings highlight the significant impact of human-made methane emissions and the potential for easy fixes to reduce these emissions. This approach of using satellite data and machine learning provides a valuable tool in our fight against climate change by targeting and addressing methane super-emitters.