Since the pre-industrial era, methane has contributed as much as 0.5°C to global warming. With a global warming potential of 28-34 times that of CO2 over 100 years, and with 80 times the warming power of CO2 in the first 20 years, methane has a large impact on climate change. Measurements from the satellite instrument TROPOMI help inform on the global atmospheric content of methane. We frequently observe methane plumes in TROPOMI data. These plumes come from super-emitters, and those provide a significant but uncertain contribution to global methane emissions, and they are relatively easy to mitigate. The TROPOMI automated plume detection algorithm makes plume detection and emission quantification possible around the world, which is infeasible without automation. However, the current TROPOMI plume emission quantification algorithm makes use of mass balance methods to estimate the methane emissions of point sources. These approaches provide reasonable but highly uncertain estimates. Hence, in this thesis, we aim to improve the accuracy of the automated emission quantification algorithm with the help of more sophisticated techniques based on atmospheric transport models.

First, this study analysed mass balance methods by quantifying emissions from synthetic plumes having known emission rates. Synthetic plumes were generated using the WRF and FLEXPART atmospheric transport models. A classification algorithm was developed to segregate synthetic plumes into different categories based on their geometries to determine challenging plume emission quantification scenarios for the mass balance methods. We found that mass balance methods produce uncertain emission estimates due to several inherent limitations like missing plume pixels, performance under low wind speed conditions, and no utilisation of three-dimensional wind speeds. Next, the atmospheric transport model based plume scaling approach was analysed by quantifying emissions from synthetic plumes. This analysis revealed that the plume scaling approach could overcome several inherent limitations of the mass balance methods and reduce the uncertainty of plume emission quantification by nearly 10%. Finally, the plume scaling approach was applied to TROPOMI plumes. The FLEXPART model was observed to be most suitable for replicating TROPOMI plumes. With the help of results obtained from this research, a decision tree algorithm was developed. This decision tree can choose the most suitable plume emission quantification method between the mass balance methods and the plume scaling approach for a given TROPOMI plume, maintaining a balance between ease of use of the mass balance methods and accuracy of the plume scaling approach.