C-CORE (CA)
Methane (CH4) is a significant anthropogenic greenhouse gas, second only to carbon dioxide (CO2) in contributing to global warming. It constitutes approximately 25% of the warming effect since pre-industrial times. Traditional methods for monitoring methane emissions rely on bottom-up approaches, calculating emissions by multiplying activity levels with emission factors. Satellites offer a more efficient alternative, providing near-real-time information on national emissions by sector. This data serves as a baseline for establishing methane reduction goals and allows ongoing monitoring to assess progress. Unlike bottom-up inventories, which may have latencies of a few years, satellites enable the documentation of rapid changes in emissions.
The “AI4CH4” project aims to revolutionize methane atmospheric monitoring using advanced state-of-the-art earth observation data and technology. With a clear focus on innovation, the project aims to address a critical gap in the current methane emission monitoring landscape. As such, the project’s key objective is to develop an advanced end-to-end AI model for automatic plume detection and quantification, surpassing the limitations of conventional physics-based techniques (e.g., high uncertainty, slow processing speed, and dependency to ancillary data, such as wind information). By leveraging the vast datasets from the harmonious synergy of Sentinel data and creating a comprehensive benchmark dataset of plumes, the AI4CH4 project promises automated and accurate identification and quantification of methane emissions based on the integration of advanced deep learning models. This cutting-edge project will advance our collaborations with GHGSat and Harvard University and holds the potential to significantly advance our understanding of methane emissions at various temporal and spatial scales, contributing to global transparency and support for more effective climate change mitigation efforts for a greener, more sustainable future.
