Abstract (eng)
In recent decades, the substantial rise in anthropogenic greenhouse gas (GHG) emissions has become a worldwide concern. Enhancing our understanding of these emissions is crucial to assist policymakers in implementing effective mitigation strategies. A powerful tool to constrain GHG emissions is inverse modeling, where atmospheric measurements are used along with an atmospheric transport model to optimize an a priori estimate of these emissions. This thesis focuses on developing the inverse modeling framework and applying it to determine the global distribution of sulfur hexafluoride (SF6) emissions, the GHG with the highest known global warming potential. In the first part of this thesis, I investigate the uncertainties of atmospheric inversions, when utilizing Lagrangian Particle Dispersion Models (LPDMs) to model the atmospheric transport between emissions and measurements. In this approach, a large number of virtual particles is released from the measurement sites and followed backwards in time to establish a relationship between the measurements and emission sources within a chosen simulation period. As this simulation period is limited due to computational costs, a baseline has to be defined that accounts for all emissions that occur prior to the simulation period, representing a large source of uncertainty. I put a main emphasis on assessing the influence of different baseline methods and different LPDM backward simulation periods on the inversion results. I demonstrate, that commonly employed statistical baseline methods can encounter substantial problems and present the advantages of a Global-Distribution-Based (GDB) approach, that leads to more robust inversion results, accounts for meteorological variability, and allows the inclusion of low-frequency flask measurements in the inversion. I further propose to employ relatively long backward simulation periods, beyond 5–10 days, as this can improve the performance of the LPDM and the inversion. In the second part of this thesis, I employ a global inversion setup that is based on the methodological findings of the first part, to globally determine the distribution of regionally resolved SF6 emissions between 2005 and 2021. My findings show a substantial decline in U.S. SF6 emissions, indicating the positive effects of national regulation measures. I also find a decreasing emission trend in the EU, with a substantial drop after 2017, likely a result of the EU’s 2014 F-gas regulation. Chinese emissions, however, show a strong positive trend, that is even higher than the average global total emission trend. I further demonstrate that national reports to the United Nations Framework Convention on Climate Change underestimated the SF6 emissions in the U.S., EU, and China throughout the whole study period. The aggregation of all the regionally resolved emissions shows a relatively good agreement with total global emissions, however, results are sensitive to the employed a priori emission fields, likely due to the challenges in constraining emissions in regions poorly covered by the observation network. Lastly, monthly inversion results show higher SF6 emissions in summer than in winter in the Northern Hemisphere. This thesis contributes to the development of inverse modeling and globally enhances the knowledge about regionally resolved SF6 emissions. The developed set-up for atmospheric inversions provides various advantages and is suitable for estimating GHG emissions on global, regional, and local scales.