Physics-informed and AI-supported Methane Plume Point-source Identification

Methane, a potent greenhouse gas, has nearly tripled in atmospheric concentration since preindustrial times, significantly contributing to climate change. Point sources must be measured and identified to quantify emissions and inform mitigation efforts accurately. Traditional methods face challenges such as cloud cover interference and their dependency on local wind speed data and simulations to recognize plume patterns, leading to uncertainties in emission estimates. Our research introduces a novel Physics-informed and AI-supported Methane Point-source Identification (PAMPI) system, merging physical models with advanced machine learning algorithms. This approach addresses limitations by incorporating cloud removal schemes, detailed plume integral models accounting for wind shearing and atmospheric stratification, and physics-informed regularization techniques. By leveraging high-resolution 2-D plume imagery from airborne and satellite platforms, this framework significantly enhances the reliability of methane source-point quantification, providing near-real-time data crucial for climate policy and mitigation. The PAMPI system represents a critical advancement in environmental monitoring, equipping policymakers and global entities with precise tools to achieve methane reduction targets in alignment with international climate agreements and pledges, supporting broader environmental transparency and ultimately fostering improved social, health, and equity outcomes.