Abstract/Summary

Tropospheric ozone (O3) is the third most important greenhouse gas (GHG) and as an air pollutant, constitutes one of the major contributors to ~7 million premature deaths (WHO, 2019) annually worldwide and also to ecosystem damage. Furthermore, ozone determines hydroxyl (OH) radical levels, which serves to breakdown various pollutants and GHGs. Despite its significance, regional and seasonal variations in tropospheric ozone are poorly understood. Ozone in the troposphere is formed from various precursor species (e.g. CO, NOx and VOCs), which have both natural and anthropogenic sources, as well as transported down from the stratosphere. The relative importance of each influence has been the subject of intense debate in recent decades. Using a combination of in situ (ozonesonde) and satellite (GOME-2A and OMI) observations, in conjunction with state-of-the-art chemistry climate models (EMAC and CMAM), comprehensive quantification of the regional and seasonal variability in tropospheric ozone is here provided, including recent changes and source attribution using the models’ stratospheric-tagged ozone tracers. This combined approach is necessary to overcome the shortcomings of individual datasets. The realism of each dataset is first rigorously evaluated and limitations identified. A significant source of retrieval error is identified in the satellite measurements, which will serve to retrospectively improve existing datasets. Using both models, it is quantified that the stratospheric influence is larger than found in previous studies; exceeding 50 % near the surface during winter in the extratropics. Finally, the influence of midwinter Sudden Stratospheric Warmings (SSWs) in relation to upper troposphere/lower stratosphere (UTLS) composition and stratosphere-troposphere exchange (STE) of ozone is presented, as a potential source of interannual variability. Approximately half of all SSWs result in significant, prolonged perturbations in UTLS composition, with implications for STE of ozone at lead times of ~50 days. The radiative and air quality implications are subsequently discussed.