O&GProfile : Automated attribution of GHGSat point source methane emissions detections to O&G infrastructures for site emissions profile analysis (Permian)

Methane emissions are the second leading cause of global warming. Because of the near-term warming potential of atmospheric methane, reducing its emissions will be essential to achieve the UNFCCC climate objectives. Reducing methane emissions from oil and gas operations is among the most cost-effective and efficient actions governments can take to meet global climate goals. In the "net zero emissions by 2050" scenario, methane emissions decline rapidly in the coming years, reaching this reduction potential by 2030. This is primarily a result of the rapid deployment of emission reduction measures and technologies, leading to the elimination of all technically avoidable methane emissions within this decade.
Current regulations are based on methane emissions figures from regional and national inventories. However, it has been shown repeatedly in the literature that these have a strong tendency to underestimate actual emission levels. In order to be able to move towards adapted and personalized regulations, it is necessary to aim at new methods -other than the current too generalized inventories- improving the current assessment of emission trends at the level of the whole O&G supply chain. Emission profiles should be estimated by operator, by type of O&G site, or by site infrastructures to design and to implement specific regulations.
In order to establish the most complete emission profiles possible - taking into account the current lack of continuous coverage - a maximum of measurements at various scales (satellite, UAVS, group) of methane emissions for a site or infrastructure should be collected by operator/site/infrastructure.
We therefore propose in this paper O&GProfile,an automatic method  to associate GHGSat methane plume detections to detections from other satellite and airbone campaign already tagged by oil and gas sites type (Gathering & boosting, processing, production) and their respective operators. O&GProfile is based on the use of unsupervised machine learning methods for classification purposes,with automated correction of the classification results. By association of GHGsat data to tagged CarbonMapper and Gao data (method mame), our method provides an informed GHGSat dataset necessary to study profile of emissions by site and operators. The study of these site emissions profiles over the period 2020-2021 in the Delaware and Midland basins in the Permian also allows us to connect GHGSat detections to the Gao and carbon mapper detections.