Tropospheric ozone changes and ozone sensitivity from present-day to future under shared socio-economic pathways

. Tropospheric ozone is important to future air quality and climate. We investigate ozone changes and ozone sensitivity to changing emissions in the context of climate change from the present day (2004–2014) to the future (2045–2055) under a range of shared socio-economic pathways (SSPs). We apply the United Kingdom Earth System Model, UKESM1, with an extended chemistry scheme including more reactive volatile organic compounds (VOCs) to quantify ozone burdens as well as ozone sensitivities globally and regionally based on nitrogen oxide ( NO x ) and VOC concentrations. We show that the 5 tropospheric ozone burden increases by 4 % under a development pathway with higher NO x and VOC emissions (SSP3-7.0), but decreases by 7 % under the same pathway if NO x and VOC emissions are reduced (SSP3-7.0-lowNTCF) and by 5 % if atmospheric methane ( CH 4 ) concentrations are reduced (SSP3-7.0-lowCH4). Global mean surface ozone concentrations are reduced by 3–5 ppb under SSP3-7.0-lowNTCF and by 2–3 ppb under SSP3-7.0-lowCH4. However, surface ozone changes vary substantially by season in high-emission regions under future pathways, with decreased ozone concentrations in summer and 10 increased ozone concentrations in winter when NO x emissions are reduced. VOC-limited areas are more extensive in winter (7 %) than in summer (3 %) across the globe. North America, Europe and East Asia are the dominant VOC-limited regions in the present day but North America and Europe become more NO x -limited in the future mainly due to reductions in NO x emissions. The impacts of VOC emissions on ozone sensitivity are limited in North America and Europe because reduced anthropogenic VOC emissions are offset by higher biogenic VOC emissions. Ozone sensitivity is not greatly inﬂuenced by changing CH 4 15 concentrations. South Asia becomes the dominant VOC-limited region under future pathways. We highlight that reductions in NO x emissions are required to transform ozone production from VOC- to NO x -limitation, but that these lead to increased ozone concentrations in high-emission regions, and hence emission controls on VOC and CH 4 are also necessary.


Introduction
Ozone (O 3 ) is a chemically reactive component in the atmosphere that is produced from natural and anthropogenic sources. impacts on human health, ecosystems and climate change (Lefohn et al., 2018;Zhang et al., 2019;Agathokleous et al., 2020).
1 O 3 concentrations are largely governed by the magnitudes of O 3 precursor emissions, transport, deposition and transport from the stratosphere. O 3 exerts a positive radiative forcing on climate forcing (Stevenson et al., 2013;O'Connor et al., 2021; 25 Thornhill et al., 2021a), and changes in climate in turn influence ozone (Fiore et al., 2012;Doherty et al., 2013). Climate change can alter natural emissions of biogenic VOCs (BVOC), lightning NO x and CH 4 , along with temperature, humidity, convection and clouds, which further influence O 3 concentrations (Thornhill et al., 2021b). The interactions between air quality and climate play an important role in the coupled Earth system, and we focus on the impacts of future emissions in the context of climate change on tropospheric O 3 in this study. 30 The tropospheric O 3 burden is controlled by the amount of O 3 production, O 3 destruction, O 3 deposition and the O 3 transport from the stratosphere (Lelieveld and Dentener, 2000;Wild, 2007). From pre-industrial times to the present day, the tropospheric O 3 burden has increased from approximately 240 Tg to 350 Tg mainly due to substantial increases in anthropogenic O 3 precursor emissions (Lamarque et al., 2010;Young et al., 2013;Griffiths et al., 2021). However, regional surface O 3 changes between the pre-industrial and present day vary substantially due to different regional emission changes (Turnock 35 et al., 2020) and to differences in O 3 sensitivity to NO x and VOC emissions. In recent decades, there has been a decrease in surface O 3 concentrations in North America and Europe due to emission controls (Lefohn et al., 2008;Colette et al., 2016).
In contrast, increases in surface O 3 levels are observed in South Asia and East Asia due to industrialization, urbanization and social development (Akimoto, 2003;Ohara et al., 2007). Furthermore, while emission controls have been implemented across industrial regions of China in recent years, these have focused on emissions of NO x and particulate matter, and have led to 40 increased O 3 pollution in some places Silver et al., 2018).
It is important to investigate O 3 sensitivity to understand how O 3 chemical regimes might change in different parts of the world, and to guide suitable emission control strategies. However, few studies have focused on O 3 sensitivity from a global perspective, or on how this might change in the future. O 3 sensitivity, typically characterized by a NO x -or VOC-limited O 3 production regime, is dependent on the relative abundances of NO x and VOC concentrations (Sillman, 1995(Sillman, , 1999, and 45 determines the extent and effectiveness of different emission control strategies. VOC-limited regimes typically occur in highly urbanised regions with high NO x concentrations in which decreases in NO x emissions increase O 3 concentrations, and O 3 production increases with higher VOC emissions. In contrast, changes in NO x concentrations dominate O 3 changes in NO xlimited regimes such that decreases in NO x emissions decrease O 3 concentrations, and O 3 concentrations are less sensitive to VOC emissions. O 3 sensitivity can be assessed by quantifying the ratio between NO x and VOC concentrations and we apply 50 this approach in this study to present-day and future conditions. The shared socio-economic pathways (SSPs) are future emission and climate scenarios accounting for future social, economic and environmental developments van Vuuren et al., 2014). The SSPs represent a range of levels of policy strength (weak, medium and strong) to control emissions of near-term climate forcers (NTCFs) that include tropospheric O 3 , O 3 precursors and aerosols (Rao et al., 2017). Our study is based on simulations using historical and future SSPs emis-55 sions and climate undertaken as part of the Aerosol Chemistry Model Intercomparison Project (AerChemMIP; Collins et al., 2017) and the wider Coupled-Model Intercomparison Project Phase 6 (CMIP6; Eyring et al., 2016). The aim of AerChemMIP is to quantify the effects of chemistry and aerosols on air quality and climate in CMIP6 by conducting historical and future experiments using chemistry-climate models with specified climate and emission trajectories.

60
Model development and application are described in Sect. 2 along with descriptions of the emission and climate scenarios used.
We compare and evaluate the present-day tropospheric O 3 burden and surface O 3 concentrations with two different chemistry schemes in Sect. 3. We then investigate the seasonal, daytime and nighttime differences in O 3 changes in the future compared to present-day for different regions in Sect. 4. Analysis of O 3 concentrations and production is used to quantify O 3 sensitivity and to explain contrasting regional O 3 changes in Sect. 5. We then show the changes in O 3 sensitivity between different seasons 65 and scenarios in Sect. 6 and present our conclusions in Sect. 7.

Model description, development and application
We use version 1 of the United Kingdom Earth System Model, UKESM1 (Sellar et al., 2019), to reproduce present-day (ES) processes have been coupled (Sellar et al., 2019). Atmospheric composition is modelled using a state-of-the-art chemistry and aerosol module, the United Kingdom Chemistry and Aerosol model (UKCA; Morgenstern et al., 2009;O'Connor et al., 2014). UKCA includes a stratosphere-troposphere gas-phase chemistry scheme (StratTrop; Archibald et al., 2020a) coupled 75 to the aerosol scheme GLOMAP-mode (Mann et al., 2010;Mulcahy et al., 2020). The model resolution is N96L85 in the atmosphere, with 1.875°in longitude by 1.25°in latitude, 85 terrain-following hybrid height layers and a model top at 85 km.
While the UKESM1 configuration for CMIP6 used the UKCA StratTrop mechanism, this study also uses an extended gasphase chemistry scheme that incorporates more reactive VOC species to permit a more realistic representation of O 3 production in polluted environments. The extended chemistry scheme (denoted as Ext_StratTrop hereafter) is based on the StratTrop 80 scheme and includes oxidation of the additional chemical components propene (C 3 H 6 ), butane (C 4 H 10 ) and toluene (C 7 H 8 ) to represent alkenes, alkanes and aromatic VOC classes, as described in Liu et al. (2021). The extended chemistry scheme includes 101 species, 244 bimolecular reactions, 26 uni-and termolecular reactions, 70 photolytic reactions, 5 heterogeneous reactions and 3 aqueous phase reactions for the sulphur cycle.
The atmosphere-only configuration of UKESM1 is used with prescribed sea surface temperatures and sea ice to show the 85 transient impacts of emissions under present-day and future climates. These are prescribed using monthly-mean time-evolving fields from the fully coupled UKESM1. Greenhouse gas concentrations are prescribed as in historical and future simulations conducted by UKESM1 as part of CMIP6 (Meinshausen et al., 2017(Meinshausen et al., , 2020.

2.2 Emissions and experiments
Present-day CMIP6 anthropogenic and biomass burning emissions are taken from Hoesly et al. (2018) and van Marle et al. 90 (2017), respectively. Biogenic VOC emissions are calculated interactively within the iBVOC emissions scheme (Pacifico et al., 2011) in the Joint UK Land Environmental Simulator (JULES) land-surface scheme which is coupled to UKCA. Other aspects of the emissions used here are the same as described in Turnock et al. (2020). Anthropogenic emissions are categorised into five sectors (industry, power plants, transport, residences and agriculture) as inputs to the model, with independent diurnal and vertical emission profiles applied for each sector (Bieser et al., 2011;Mailler et al., 2013;Liu et al., 2021).
105 Table 1. Model configurations for present-day and future simulations. "Emissions" refers to emissions of O3 precursors and aerosols. "CH4 conc." refers to prescribed surface CH4 concentrations. "SST/SI" refers to prescribed sea surface temperature and sea ice concentrations.
"Historical" means that the emissions, CH4 concentrations or SST/SI evolve as for the CMIP6 historical simulations, and "Reference" means that they evolve as for SSP3-7.0. "Low" emissions or CH4 concentrations evolve following SSP3-7.0 but with lower emissions or CH4 concentrations.   (Archibald et al., 2020b). The magnitude of the O 3 burden is also consistent with the CMIP6 multi-model mean burden of 356 ± 31 Tg for 2005-2014 (Griffiths et al., 2021). The extended chemistry scheme produces a 5 % higher tropospheric O 3 burden than that of StratTrop (358 Tg), demonstrating a more reactive environment for net O 3 formation throughout the troposphere due 115 to reactive VOCs. This is also reflected in the higher rates of chemical O 3 production (11 %), loss (6 %) and deposition (6 %) with the extended chemistry scheme. However, the higher O 3 production is offset by greater O 3 destruction and by faster O 3 deposition, and hence the mean O 3 chemical lifetime remains very similar at about 22 days, which is consistent with previous multi-model estimates of the mean lifetime of 22.2 ± 2.2 days (Stevenson et al., 2006).  South Asia (Lu et al., 2018). This leads to higher O 3 concentrations in winter than that in summer, consistent with Gao et al. winter is also seen in other heavily populated regions that have high NO x emissions such as North America and Europe.

140
Numerical diffusion of O 3 precursor emissions due to coarse model horizontal resolution may explain the biases (Wild and 7 Prather, 2006;Stock et al., 2014;Fenech et al., 2018), and we note that these seasonal O 3 biases are also evident in other chemistry-climate models (Young et al., 2018;Turnock et al., 2020). Insufficient turbulent mixing in the planetary boundary layer may also contribute to the bias (O'Connor et al., 2014), as accumulation of NO x at the surface leads to greater O 3 production in summer and greater titration by NO in winter. However, we note that a much more comprehensive chemistry This suggests that other biases in the model are primarily responsible for the biases in surface O 3 . We choose to apply the extended chemistry scheme as it permits representation of a more appropriate chemical environment for O 3 production in high-emission areas and is thus more suitable to investigate O 3 sensitivity.

Emission changes
The SSP3-7.0 pathway is characterized by relatively strong emission controls in some parts of the world such as North America and Europe but weaker controls or emission increases elsewhere. Increases in NO x and VOC emissions from anthropogenic and biomass burning sources are seen in Central and South America, North Africa, the Middle East, and South and East 155 Asia (Fig. 3a, c). SSP3-7.0-lowCH4 has the same NTCF emissions as SSP3-7.0 but lower CH 4 emissions that lead to lower CH 4 concentrations ( Table 2). SSP3-7.0-lowNTCF represents strong emission controls across the globe, with reductions in emissions in most major high-emission regions except for South Asia (Fig. 3b, d). Total BVOC emission changes are driven by changes in land-use, vegetation and temperature. under SSP3-7.0-lowCH4 are slightly higher than in the present day partly due to higher hydroxyl radical (OH) concentrations that promote O 3 production. However, these higher O 3 production rates are offset by higher O 3 loss rates, and result in lower O 3 net production under SSP3-7.0 and SSP3-7.0-lowCH4. We find that the O 3 chemical lifetime decreases slightly by 0.4-1.6 days under future pathways partly due to decreased O 3 net production, and partly due to increased O 3 loss associated with higher temperature and humidity in a warmer climate (Young et al., 2018). Changes in O 3 dry deposition rates principally   Fig. 4a-f, along with a comparison between SSP3-7.0 and SSP3-7.0-lowCH4 (Fig. 4g, h). SSP3-7.0 represents less stringent emission control policies, and has slightly higher global mean O 3 concentrations (0.7-0.9 ppb; Fig

O 3 sensitivity in the present day and the future
Non-linearity in chemical O 3 formation can result in differences in the effectiveness of emission control strategies regionally, and may aggravate O 3 pollution issues. We therefore investigate O 3 sensitivity for the present day and the future, as it is important to understand how O 3 concentrations will respond to changing emissions. Ratios of NO x and VOC concentrations provide a useful indicator of regional O 3 sensitivity regimes. Here we quantify the critical NO x /VOC ratio that distinguishes 220 VOC-limited and NO x -limited regimes by examining monthly mean surface O 3 concentrations and net chemical production rates as a function of monthly mean NO x and VOC concentrations, see Fig. 6. Monthly mean data from all months and all scenarios are used to plot the isopleths. Approximate thresholds of monthly mean NO x /VOC ratios for O 3 sensitivity are shown in Fig. 6, ranging from 0.6 to 1 with a central value of 0.8. We hence apply a threshold value of 0.8 to distinguish O 3 sensitivity regimes hereafter. Areas above the threshold represent VOC-limited regimes in which increased NO x emissions 225 reduce O 3 concentrations and O 3 production rates, and areas below the threshold represent NO x -limited regimes. We find the highest O 3 concentrations and O 3 production close to the threshold line, demonstrating that the threshold values are consistent and that the approach is robust.  Fig. 7b, c. We identify the dominant region based on the number of model grid cells with the corresponding NO x and VOC concentrations. This approach reveals differences in regional O 3 sensitivity. We also show the shift in O 3 sensitivity in different regions between the present day (Fig. 7b) and the future (SSP3-7.0-lowNTCF; Fig. 7c) to demonstrate the impacts of decreased NTCF emissions on the evolution of O 3 sensitivity on a regional basis.  surface O 3 levels are typically low and where O 3 production is NO x -limited. In contrast, Europe, North America and East Asia dominate the high-NO x and high-VOC environments in the present day where O 3 levels are high and O 3 production is VOC-limited. Europe and North America have similar VOC concentrations but NO x concentrations are generally higher in Europe than in North America, which results in Europe lying further above the NO x /VOC threshold. This demonstrates that 240 stricter controls on NO x emissions are required for Europe to shift from VOC-limited to NO x -limited regimes than for North America. East Asia dominates VOC-limited regimes due to much higher NO x and VOC concentrations than other regions.
Parts of South Asia and Middle East are also VOC-limited. Major biogenic emission source regions such as South America, South Africa and South East Asia have high VOC concentrations but moderate levels of NO x , and the chemical environment is therefore NO x -limited.

245
The impacts of reductions in NTCF emissions are shown in Fig. 7c. We find that Europe and North America are no longer the most dominant VOC-limited regimes due to decreased NO  Global spatial distributions of annual O 3 sensitivity in the present day and the future are shown in Fig. 9 Europe becomes NO x -limited. However, VOC limited regimes in East Asia, particularly China, are persistent due to projected increases in NO x emissions until 2055 under SSP3-7.0 and SSP3-7.0-lowCH4. We find that reductions in CH 4 concentrations have relatively little influence on O 3 sensitivity over continental regions (Fig. 9b vs 9d). We note that SSP3-7.0-lowNTCF shows the smallest VOC-limited areas across the globe. East Asia is still partly VOC-limited under this scenario, particularly in northern China (Fig. 9c), which indicates that further reductions in NO x emissions are required in addition to those expected 270 in this region in the future (Fig. 3b). conditions, but these account for less than 7 % of the total area of the world in the present day (Table 5). Over 50 % of North America, Europe and East Asia are VOC-limited in winter in the present day. This explains why North America and Europe show O 3 increases in winter (Fig. 4) despite reduced NO x emissions. However, there are fewer VOC-limited regions across the globe under all future pathways (Fig. 10). About 1 % of North America is VOC-limited in summer, and less than 7 % of Europe. In contrast, over 48 % (winter) and 39 % (summer) of South Asia is VOC-limited in future. Slightly more areas are 280 VOC-limited under SSP3-7.0-lowCH4 than under SSP3-7.0 (Fig 10c, d). Overall, reductions in NO x emissions are important to reduce O 3 production in high-emission regions and shift VOC-limited areas to NO x -limitation but this may lead to higher O 3 concentrations in winter without further emission controls on VOC and CH 4 .   [2045][2046][2047][2048][2049][2050][2051][2052][2053][2054][2055]. CMIP6 future scenarios representing 'regional rivalry' development pathways (SSP3-7.0, SSP3-7.0-lowNTCF and SSP3-7.0-lowCH4) are used from the AerChemMIP project. We have examined O 3 changes from the present day to the future and investigated regional O 3 sensitivities to explain contrasting O 3 changes in different seasons.
An extended chemistry scheme incorporating more reactive VOC species is used to permit representation of more active O 3 sensitivity is quantified using monthly mean NO x /VOC concentration ratios to give a broad assessment of regional O 3 sensitivity. The estimated monthly mean NO x /VOC thresholds range from 0.6 to 1.0, and 0.8 is used to distinguish O 3 305 sensitivity regimes. Most VOC limited regimes occur in high-emission regions across the northern hemisphere, such as North America, Europe, the Middle East, South Asia and East Asia. More areas in North America and Europe become NO x -limited under all future pathways due to the projected decrease in NO x emissions. There are more VOC-limited areas in East Asia under SSP3-7.0 and SSP3-7.0-lowCH4 due to the projected increase in NO x emissions, although there are fewer VOC-limited areas in East Asia under SSP3-7.0-lowNTCF. South Asia becomes the dominant region for VOC-limited O 3 production in the 310 future. Projections of regional O 3 sensitivity demonstrate that reductions in NO x emissions are the most important factor to shift VOC-limited regimes to NO x -limitation.
We highlight that O 3 sensitivity varies by season. There are more VOC-limited regimes in winter (7 %) than in summer to reduce surface O 3 concentrations from a globe perspective but will lead to increased O 3 concentrations in some regions, and hence emission controls on VOC and CH 4 are necessary to mitigate regional O 3 pollution during the transition from VOC-to NO x -limitation.
Data availability. The data generated in this study are available upon request.
Author contributions. ZL, RD, OW, FO'C, ST designed the study. ZL set up the model, conducted model simulations and performed the 325 analysis. ZL, RD and OW prepared the paper. All co-authors contributed to reviewing and editing the paper.
Competing interests. The authors declare that they have no conflict of interest.