Research
Deep Learning Maps Methane Emissions from Space
Google Research Bloghigh confidence
Why it matters
- →MAPL-EMIT enhances methane detection accuracy, crucial for climate action.
- →It supports global emission reduction targets by providing precise data.
- →The model's availability on public platforms democratizes access to advanced tools.
Google Research has developed MAPL-EMIT, a deep-learning model that automates the detection and quantification of methane emissions from space. Utilizing NASA's EMIT instrument, the model achieves an 84% recall rate on expert-annotated plumes, significantly improving methane tracking capabilities. This advancement supports global efforts to reduce methane emissions, a potent greenhouse gas, by providing precise data for stakeholders in various sectors. The model and its data are available on platforms like Kaggle and GitHub, facilitating broader scientific and environmental applications.
Read original
