Impact of transport model resolution on the estimate of Swiss synthetic greenhouse gases emissions by inverse modelling

Atmospheric inverse modelling is a ’top-down’ emission estimation method, which utilises numerical models to estimate emissions from observed and simulated concentrations of atmospheric compounds. Inverse emission modelling can be applied for the support of emission inventories and emission reporting, which are usually based on ’bottom-up’ methods. The latter employ activity data and emission factors for the relevant processes. Depending on the emitting process, both may be afflicted by large uncertainties, especially when spatially-resolved emissions are considered on sub-national scales. Inverse modelling offers an alternative tool to emission estimation, validation and optimization of emission inventories. It is widely used by the scientific community for different atmospheric compounds and from global to the facility scale.Atmospheric inversions can be carried out by combining source sensitivities simulated by atmospheric transport models, observations, and an inversion framework. Here, we focus on emissions of synthetic greenhouse gases (GHG) in the Swiss domain. 'Bottom-up' estimates of these emissions are connected to large uncertainties in the leakage rates of these compounds from various applications (e.g., refrigeration, foam blowing). Globally, synthetic GHGs account for a considerable fraction of the total anthropogenic radiative forcing (~10%), and their future environmental impact depends on the replacement of compounds with long lifetimes by compounds with short lifetimes and minimal global warming potential (GWP). In Switzerland, synthetic GHGs contribute about 3.5% to national total GHG emissions according to bottom-up reporting.Newly available synthetic gases observations, collected as part of the Swiss project IHALOME (Innovation in Halocarbon Measurements and Emission Validation), from the Swiss Plateau at the Beromünster and Sottens tall towers, allow us to localise and quantify the emissions in Switzerland and in the neighboring countries. We apply the Lagrangian Particle Dispersion Model (LPDM) FLEXPART, driven by meteorological fields of the Numerical Weather Prediction (NWP) model COSMO, at two different spatial resolutions (7 km x 7 km and 1 km x 1 km). During the last decade, FLEXPART-COSMO was successfully operated at 7 km x 7 km spatial resolution to estimate Swiss emissions of methane and nitrous oxide. Reliable simulations at 1 km x 1 km resolution were recently established and required an update of FLEXPART-COSMO's turbulence scheme.Inversion results for the most important (by emissions) synthetic GHGs (HFCs and SF6) are presented. Special attention is given to comparisons between inversions for different transport model resolutions and the question if the high resolution simulations are able to enhance the capability of the inversion method to localise emissions. Additionally, the sensitivity of the inversions to different a priori emission fields is presented. Finally, the sensitivity of the inversion towards covariance parameters, either obtained from maximum likelihood optimisation or from expert judgment, is examined. Inversions with the high resolution model amplify the emission differences between the Swiss Plateau and the high altitude regions in the Alps by both increasing the emissions in the big cities and decreasing the emissions in the high altitude regions. At the same time, no significant difference in total national emissions is observed between high and low resolution model inversions.