How do transport processes affect inter-annual variability and trends in lower stratospheric ozone over the recent decade?

Recent observations show a significant decrease of lower stratospheric (LS) ozone (O3) concentrations in the tropics and mid-latitudes (60°N-60°S) between 1998-2016, stating that O3 concentrations have declined by ~2 DU since 1998. The cause of the negative trend in observed LS O3 is not identified yet. A possible explanation is the speed up of the Brewer-Dobson circulation in a changing climate that influences stratospheric distribution of O3. It is known from Chemistry-climate model (CCM) studies that enhanced tropical upwelling drives LS tropical O3 to decrease. However, how mid-latitude O3 is affected by changes in the circulation, and which processes dominate (advection by the residual circulation versus horizontal mixing) is not entirely clarified yet. In this study we analyse the inter-annual variability as well as trends over the recent decades in LS O3 in CCMs and seek to attribute O3 variability and trends to transport processes. To do so, we use free-running simulations with the EMAC model performed for the climate chemistry model inter-comparison project (CCMI-1), and contrast EMAC results with results from other CCMI-1 models. We find a wide spread in LS O3 trends over the recent decades (1998-2010/2016) in the CCMs. The different trends in CCMs indicate that inter-annual variability in O3 likely contributes to trends over this time period, and thus could lead to the observed trend. We further investigate in how far O3 variability and trends are determined by transport processes. To do so we analyse the inter-annual correlation between local O3 and mean age of air (measure of stratospheric transport). All CCMs show high positive correlations throughout the lower tropical to mid-latitude stratosphere, confirming that O3 is strongly dynamically determined there. To gain a better understanding of the transport processes that control the LS O3 variability and trends, we separate stratospheric transport to the process of residual circulation and mixing. Results indicate that tropical upwelling influences tropical O3 trends, whereas mixing influences mid-latitude LS O3 variability and trends. Overall, our results imply that the recently reported trends in LS O3 are very likely caused by trends in the stratospheric circulation, which are influenced by inter-annual variability over the analysed period