Posted in | News | Climate Change

Using Hyperspectral Machine Learning Models to Combat Climate Change

Researchers from the University of Oxford have created a tool that uses machine learning and hyperspectral data from Trillium Technologies’ NIO.space to autonomously detect methane plumes on Earth from orbit. This could make it easier to pinpoint methane “super emitters” and allow for more efficient greenhouse gas emission reduction. The journal Nature Scientific Reports has published the findings.

Space satellite orbiting the earth. Elements of this image furnished by NASA.

Image Credit: Andrei Armiagov/Shutterstock.com

While reducing methane emissions is a crucial step in slowing global warming, Net Zero targets primarily concentrate on reducing CO2 emissions. Compared to CO2, methane traps heat 80 times more effectively, but it lasts in the atmosphere for only 7 to 12 years, as opposed to centuries.

Therefore, reducing methane emissions from anthropogenic sources would be a quick and effective way to slow global warming and improve air quality. Reductions in methane emissions that are easily attainable are predicted to prevent warming by almost 0.3 °C over the next 20 years.

However, there are currently very few ways to map methane plumes from aerial data, and the processing phase takes a long time. This is due to the fact that methane gas is transparent to most satellite sensors' spectral ranges as well as the human eye.

To identify methane plumes, even in cases where satellite sensors function within the appropriate spectral range, noise frequently masks the data, necessitating tedious human methods.

These problems are addressed by a novel machine-learning method created by Oxford researchers that finds methane plumes in hyperspectral satellite images. Compared to more popular multispectral satellites, these have smaller bands of detection, which facilitates noise reduction and tuning to the precise methane signature. However, due to the volume of data they generate, processing them without artificial intelligence (AI) becomes difficult.

The model was trained using 167,825 hyperspectral tiles (each representing 1.64 km2) collected by NASA’s aerial sensor AVIRIS over the Four Corners region of the United States.

The algorithm was then applied to data collected by other hyperspectral sensors in orbit, such as NASA’s new hyperspectral sensor EMIT (Earth Surface Mineral Dust Source Investigation mission), which is connected to the International Space Station and provides near-global coverage of the Earth.

All things considered, the model detected huge methane plumes with an accuracy of over 81%, which is 21.5% better than the previous best method. Additionally, compared to the most accurate previous methodology, the method’s false positive detection rate for tile categorization was greatly improved, falling by almost 41.83%.

Both the annotated dataset and the model’s code are publicly available on the GitHub project page in an effort to encourage more studies into methane detection. As part of the NIO.space program, they are currently investigating whether the model could operate directly aboard the satellite, enabling additional satellites to carry out follow-up observations.

Such on-board processing could mean that initially, only priority alerts would need to be sent back to Earth, for instance, a text alert signal with the coordinates of an identified methane source. Additionally, this would allow for a swarm of satellites to collaborate autonomously: an initial weak detection could serve as a tip-off signal for the other satellites in the constellation to focus their imagers on the location of interest.

Vít Růžička, Study Lead Researcher and DPhil Student, Department of Computer Science, University of Oxford

Professor Andrew Markham (Department of Computer Science), supervisor for the research, added, “In the face of climate change, these kinds of techniques allow independent, global validation about the production and leakage of greenhouse gases. This approach could easily be extended to other important pollutants, and building on earlier work, our ambition is to run these approaches on-board the satellites themselves, making instant detection a reality.

The project was carried out as part of the Trillium Technologies initiative Networked Intelligence in Space (NIO.space) and was supported by the European Space Agency (ESA) Φ-lab through the ‘Cognitive Cloud Computing in Space’ (3CS) program.

Journal Reference:

Růžička, V., et al. (2023) Semantic segmentation of methane plumes with hyperspectral machine learning models. Nature Scientific Reports. doi:10.1038/s41598-023-44918-6.

Tell Us What You Think

Do you have a review, update or anything you would like to add to this news story?

Leave your feedback
Your comment type
Submit

While we only use edited and approved content for Azthena answers, it may on occasions provide incorrect responses. Please confirm any data provided with the related suppliers or authors. We do not provide medical advice, if you search for medical information you must always consult a medical professional before acting on any information provided.

Your questions, but not your email details will be shared with OpenAI and retained for 30 days in accordance with their privacy principles.

Please do not ask questions that use sensitive or confidential information.

Read the full Terms & Conditions.