ACCMIP-Ozone-RF-v1_22July1740BST_RD

1Tropospheric 2radiative 3Model ozone changes, attribution to emissions and forcing in the Atmospheric Chemistry and Climate Inter-comparison Project (ACCMIP) 5D.S Stevenson1, P.J Young2,3, V Naik4, J.-F Lamarque5, D.T Shindell6, R Skeie7, 6S Dalsoren7, G Myhre7, T Berntsen7, G.A Folberth8, S.T Rumbold8, W.J 7Collins8, I.A MacKenzie1, R.M Doherty1, G Zeng9, T van Noije10, A Strunk10, D 8Bergmann11, P Cameron-Smith11, D Plummer12, S.A Strode13, L Horowitz14, Y.H 9Lee6, S Szopa15, K Sudo16, T Nagashima17, B Josse18, I Cionni19, M Righi20, V 10Eyring20, K.W Bowman21, O Wild22 11 12[1]{School of GeoSciences, The University of Edinburgh, Edinburgh, United Kingdom} 13[2]{Chemical Sciences Division, NOAA Earth System Research Laboratory, Boulder, 14Colorado, USA} 15[3]{Cooperative Institute for Research in Environmental Sciences, University of Colorado, 16Boulder, Colorado, USA} 17[4]{UCAR/NOAA Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey, USA} 18[5]{National Center for Atmospheric Research, Boulder, Colorado, USA} 19[6]{NASA Goddard Institute for Space Studies, New York, New York, USA} 20[7]{CICERO, Center for International Climate and Environmental Research-Oslo, Oslo, 21Norway} 22[8]{Met Office Hadley Centre, Exeter, UK} 23[9]{National Institute of Water and Atmospheric Research, Lauder, New Zealand} 24[10]{Royal Netherlands Meteorological Institute, De Bilt, Netherlands} 25[11]{Lawrence Livermore National Laboratory, Livermore, California, USA} 26[12]{Canadian Centre for Climate Modeling and Analysis, Environment Canada, Victoria, 27British Columbia, Canada} 1 1[13]{NASA Goddard Space Flight Centre, Greenbelt, Maryland, USA} 2[14]{NOAA Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey, USA} 3[15]{Laboratoire des Sciences du Climat et de l’Environment, Gif-sur-Yvette, France} 4[16]{Department of Earth and Environmental Science, Graduate School of Environmental 5Studies, Nagoya University, Nagoya, Japan} 6[17]{National Institute for Environmental Studies, Tsukuba-shi, Ibaraki, Japan}? 7[18]{GAME/CNRM, Météo-France, CNRS Centre National de Recherches 8Météorologiques, Toulouse, France} 9[19]{Agenzia Nazionale per le Nuove Tecnologie, l'energia e lo Sviluppo Economico 10Sostenibile (ENEA), Bologna, Italy} 11[20]{Deutsches Zentrum für Luft- und Raumfahrt (DLR), Institut für Physik der Atmosphäre, 12Oberpfaffenhofen, Germany} 13[21]{NASA Jet Propulsion Laboratory, Pasadena, California, USA} 14[22]{Lancaster Environment Centre, University of Lancaster, Lancaster, UK} 15 16Correspondence to: D S Stevenson (David.S.Stevenson@ed.ac.uk) 17 18Abstract 19Ozone (O3) from seventeen atmospheric chemistry models taking part in the ACCMIP 20(Atmospheric Chemistry and Climate Model Intercomparison Project) has been used to 21calculate tropospheric O3 radiative forcings (RFs) We calculate a value for the 1750 to 2010 22tropospheric O3 RF of 0.40 W m-2 The model range of pre-industrial to present-day changes 23in O3 produces a spread in RFs of ±17% Three different radiation schemes were used – we 24find differences in RFs between schemes (for the same ozone fields) of about ±10% Applying 25two different tropopause definitions we find differences in RFs of ±3% Given additional 26(unquantified) uncertainties associated with emissions, climate-chemistry interactions and 27land-use change, we estimate an overall uncertainty of ±30% for the tropospheric O RF 28Experiments carried out by a subset of six models find that the tropospheric O RF can be 29attributed to increased emissions of CH4 (46%), NOx (30%), CO (15%) and NMVOCs (9%) 1Normalising RFs to changes in tropospheric column O 3, we find a global mean normalised RF 2of 0.042 W m-2 DU-1 Future O3 RFs (W m-2) for the Representative Concentration Pathway 3(RCP) scenarios in 2030 (2100) are: RCP2.6: 0.31 (0.16); RCP4.5: 0.38 (0.26); RCP6.0: 0.33 4(0.24); and RCP8.5: 0.42 (0.56) Models show some coherent responses of O to climate 5change: decreases in the tropical lower troposphere, associated with increases in water 6vapour; and increases in the sub-tropical to mid-latitude upper troposphere, associated with 7increases in lightning and stratosphere-to-troposphere transport 91 Introduction 10Estimates of many aspects of Earth’s past atmospheric composition can be derived from 11analyses of air trapped in bubbles during ice formation (Wolff, 2011) However, the 12greenhouse gas ozone (O3) is too reactive to be preserved in ice Direct measurements of 13tropospheric ozone concentrations prior to the 1970s are also extremely limited (Volz and 14Kley, 1988; Staehelin et al., 1994), and most early measurements used relatively crude 15techniques, such as Schӧnbein papers, that are subject to contamination from compounds 16other than ozone (Pavelin et al., 1999) Only in the last few decades have observation 17networks and analytical methods developed sufficiently to allow a global picture of ozone’s 18distribution in the troposphere to emerge (Fishman et al., 1990; Logan, 1999; Oltmans et al., 192006; Thouret et al., 2006) Despite this paucity of early observations, tropospheric ozone is 20thought to have increased substantially since the pre-industrial era; this is largely based on 21model studies Ozone photochemistry in the troposphere is relatively well understood 22(Crutzen, 1974; Derwent et al., 1996), and anthropogenic (including biomass burning) 23emissions of ozone precursors (methane (CH4,), nitrogen oxides (NOx), carbon monoxide 24(CO), non-methane volatile organic compounds (NMVOCs)) have changed (generally risen) 25dramatically since 1850 (Lamarque et al., 2010) Increasingly sophisticated models of 26atmospheric chemistry, driven by emission estimates, and sometimes coupled to climate 27models, have been used to simulate the rise of ozone since industrialisation (Hough and 28Derwent, 1990; Crutzen and Zimmerman, 1991; Berntsen et al., 1997; Wang and Jacob, 1998; 29Gauss et al., 2006) 30Although the rise of anthropogenic emissions has been the main driver of ozone change, 31several other factors may also have contributed Natural sources of ozone precursor emissions 32(e.g., wetland CH4, soil and lightning NOx, biogenic VOCs) show significant variability and 1have probably also changed since 1850, but these changes are highly uncertain (Arneth et al., 22010) Downwards transport of ozone from the stratosphere is also an important source of 3tropospheric ozone (Stohl et al., 2003; Hsu and Prather, 2009); this source may have been 4affected by stratospheric ozone depletion, and its magnitude is forecast to increasechange in 5the future, via acceleration of the Brewer-Dobson circulation (Hegglin and Shepherd, 2009), 6although significant changes have not yet been observed (Engel et al., 2009) Ozone’s 7removal, via chemical, physical and biological processes is also subject to variability and 8change Increases in absolute humidity (driven by warming), changes in ozone’s distribution, 9and changes in HOx (OH+HO2), have all tended to increase chemical destruction of ozone 10(Stevenson et al., 2006; Isaksen et al., 2009) Dry deposition of ozone at the surface, and to 11vegetation in particular, has been influenced by land-use change, but also changes in climate 12and CO2 abundance (Sanderson et al., 2007; Sitch et al., 2007; Fowler et al., 2009; Andersson 13and Engardt, 2010, Wu et al 2012) Fluctuations in these natural sources and sinks are driven 14by climate variability; climate change and land-use change and may also have contributed 15towards long-term trends in ozone (ref needed) 16Ozone is a radiatively active gas, and interacts with both solar and terrestrial radiation; 17changes in the atmospheric distribution of ozone affect upwards and downwards fluxes of 18radiation We use the concept of radiative forcing (RF) (e.g., as defined by Forster et al., 192007) to quantify the impacts of ozone changes on Earth’s radiation budget.; Sspecifically in 20this paper we follow the IPCC approach for the forthcoming th assessment? and use 21stratospherically adjusted RFs at the tropopause Previous estimates of O RF (e.g., Gauss et 22al., 2006) span the range 0.25-0.65 Wm-2, with a central value of 0.35 Wm-2 (Forster et al., 232007) Skeie et al (2011) recently estimated a value of 0.44 W m -2, with an uncertainty of 24±30%, using one of the models we also use in this study Cionni et al (2011) calculated O 25RFs for the IGAC/SPARC ozone database, and found a value of 0.23 W m -2, using an earlier 26version of the radiation scheme used here We show here that an updated version of the 27radiation scheme with the same ozone field finds an equivalent value of 0.32 W m -2, and this 28value is considered more accurate The tropospheric part of the IGAC/SPARC ozone database 29was constructed from early ACCMIP integrations from two of the seventeen models used here 30(GISS-E2-R and NCAR-CAM3.5) Consequently, the multi-model mean results presented 31here are also considered to be a better estimate of atmospheric composition change than the 32IGAC/SPARC database 1Because ozone is a secondary pollutant (it is not directly emitted) it is most useful to 2understand how emissions of its precursors have driven up its concentration Model 3experiments carried out by Shindell et al (2005, 2009) attributed ozone changes to pre4industrial to present-day increases in CH4, NOx and CO/NMVOC emissions between the pre5industrial and present-day periods.; Furthermore these authors found that CH4 emissions were 6responsible for most of the O3 change These emissions also influence the oxidising capacity 7of the atmosphere in general, and affect a range of radiatively active species beyond ozone, 8including methane and secondary aerosols (Shindell et al., 2009) 9In In this paper, we present results from global models participating in tthe Atmospheric 10Chemistry and Climate Model Intercomparison Project (ACCMIP; see 11www.giss.nasa.gov/projects/accmip), Within ACCMIP, multiplea considerable number of 12global models (~17) simulated atmospheric composition between 1850-2100 Lamarque et al 13(2012a) give an overview of ACCMIP whilst Lamarque et al (2012b) present detailed 14descriptions of the participating models Shindell et al (2012) describe total radiative 15forcings, particularly those from aerosols; Lee et al (2012) further focusses on black carbon 16aerosol Young et al (2012) describes the tropospheric ozone results for the pre-industrial, 17present-day and future periodsin detail, including a range of comparisons with observations; 18Bowman et al (2012) focus on comparisons with measurements from TES (Tropospheric 19Emission Spectrometer) Finally, two papers focus on the historical and future evolution of the 20oxidising capacity of the atmosphere (Naik et al., 2012; Voulgarakis et al., 2012) In this 21paper, we estimate tropospheric ozone radiative forcing based on results from global models 22participating in ACCMIP In Section 2, the models used and the experiments they performed 23are described Results of simulated tropospheric? ozone and resulting radiative forcings are 24presented in Section 3; these are discussed and conclusions drawn in Section For reasons of 25space and conciseness, the main text focusses on generalised results (often presented as the 26multi-model mean); specific results from individual models are predominantly presented in 27the extensive Supplementary Material 28 12 22.1 Methods Models employed 3Results from seventeen different models are analysed here (Table 1) Detailed model 4descriptions are provided elsewhere (Lamarque et al., 2012b; Huijnen et al., 2010) All are 5global atmospheric chemistry models, and most are coupled to climate models which provide 6the driving meteorological fields online Climate model output of sea-surface temperatures 7and sea-ice (SST/SI) from prior CMIP-5 runs typically provide the lower boundary 8conditions; well-mixed atmospheric greenhouse gas concentrations are also specified Two 9models (B and Q) are chemistry-transport models, driven by meteorological analyses – these 10provide only a single year’s output for each experiment and were run with the same 11meteorology in each case Models M and O are chemistry-transport models driven by climate 12models, but chemical fields are not passed back to the climate model In all other models the 13chemical fields are regularly passed to the climate model’s radiation scheme: they are fully 14coupled chemistry-climate models (CCM) Models G and H are two versions of GISS, but set 15up in different ways: G has a fully interactive coupled ocean (the only model with this); H 16uses SST/SI and also includes aerosol chemistry Models I and J are two versions of 17HadGEM2: I uses a relatively simple tropospheric chemistry scheme, whereas J has a more 18detailed scheme with several hydrocarbons Several models (C, D, E, F, N?) include detailed 19stratospheric chemistry schemes; tropospheric schemes range from simple methane oxidation 20(A, C?) through models with a basic representation of NMVOCs (G, H, I, P?) to those with 21more detailed hydrocarbon schemes (B, E, F, J, K, L, M, N, O, Q?) In addition, some models 22include interactions between aerosols and gas-phase chemistry (B, F, H, I, J, K, L, N?) 23Models with no stratospheric chemistry handled their upper levels in a variety of different 24ways Model B prescribed a stratospheric ozone influx following SYNOZ (McLinden et al., 252000) Models I, J, O and P all used the IGAC/SPARC ozone climatology (Cionni et al., 262011) to prescribe O3 in the stratosphere In models I and J, ozone is overwritten in all model 27levels which are levels (approximately 3-4 km) above the tropopause Model O used the O 28fields together with vertical winds, to calculate a vertical O flux at 100 hPa, added as a 29source at these levels in regions of descent Model P prescribed O at pressures below 100 hPa 30between 50°S-50°N and pressures below 150 hPa poleward of 50°.Other models did what…? 1Some models allowed natural emissions of ozone precursors to vary with climate (all except 2B and E for lightning NOx; only D,E, G, and O for isoprene); others fixed these sources (Table 32) 42.2 Experiments analysed 5The main experiments analysed here are multi-annual simulations for the 1850s and the 62000s Every model performed these experiments Table shows the model run length for 7each experiment: typically 10 years, but in a few cases longer or shorter Model G ran five 108year ensemble members In most cases, driving climate models simulated climates of the 91850s and 2000s, typically by specifying decadal-mean SST/SI fields (from prior coupled 10ocean-atmosphere climate simulations) and setting well-mixed greenhouse gas concentrations 11at appropriate levels Models B, J and Q ran with the same climate in the 1850s as in their 122000s runs, so only assess how emissions have changed composition; single year experiments 13are thus not unreasonable in these cases 14All models used anthropogenic (including biomass burning) emissions from Lamarque et al 15(2010) This harmonisation of all models to the same source of emissions removes a 16potentially large source of inter-model difference (c.f Gauss et al., 2006) However, as each 17model did not run exactly the same years to represent the 1850s and 2000s (see Table 1), and 18models used a range of values for natural emissions (Table 2) there are still some differences 19between models in the magnitude of the applied change in emissions (see Young et al., 2012, 20Figure 1) These differences are also added to as a result of the by different chemistry 21schemes used in the different models and decisions within each model of how to partition 22NMVOC emissions between individual species and/or direct CO emissions 23Most models ran with prescribed methane concentrations of around 791 ppbv (1850) and 241751 ppbv (2000) (Meinshausen et al., 2011) One model (K) ran with methane emissions that 25varied overfor the historical period; this model and another (G) ran with methane emissions in 26the future 27The experiment set used in this paper includes additional simulations to those described in 28Young et al (2012), Voulgarakis et al (2012) and Lamarque et al (2102) Six of the models 29(Table 1) ran a series of attribution experiments, based on the 2000s simulations In these, 30specific drivers of O3 change (anthropogenic emissions of NOx, CO, NMVOCs, and CH4 31concentrations) were individually reduced to 1850s levels These experiments are closely 1related to previous studies with the GISS model (Shindell et al., 2005, 2009), and allow us to 2attribute CH4 and O3 radiative forcings since the 1850s to these individual drivers 3A subset of ten models (Table 1) ran experiments where they fixed emissions at 2000s levels, 4but applied an 1850s climate These simulations allow us to investigate how climate change 5has contributed to the ozone change since the 1850s Most of these models ran equivalent 6experiments for future climates 7Finally, most models (Table 1) ran additional historical and future simulations, utilizing 8harmonized emissions from the ‘Representative Concentration Pathway’ (RCP) scenarios 9Ozone fields from these experiments are presented in detail by Young et al (2012) – here we 10use future column O3 changes in conjunction with normalised radiative forcings (mW m -2 DU111) (for 1850s-2000s; we assume this normalised forcing does non’t change significantly in 12future) to estimate future ozone radiative forcings 132.3 Radiative forcing calculations 14Ozone fields were inserted into an offline version of the Edwards and Slingo (1996) radiation 15scheme, updated and described in Walters et al (2011) (their Section 3.2) The scheme 16includes gaseous absorption in six bands in the SW and nine bands in the LW The treatment 17of O3 absorption is as described in Zhong et al (2008) The RF calculations use an updated 18version of the radiation code compared with those presented in Cionni et al (2011), and it is 19found that these updates make substantial differences in the values We recommend that the 20values presented here are used rather than those presented in Cionni et al (2011) 21The offline code was set up so that all input fields except ozone remained fixed (at present22day values) – thus differences between two runs of the radiation code with different ozone 23yield the changes in fluxes of radiation due to ozone change Monthly mean ozone fields were 24interpolated from each model to a common resolution: 5° longitude by 5° latitude, and 64 25hybrid vertical levels The vertical levels were chosen to be compatible with the base 26climatological fields (temperature, humidity, cloud fields) which were taken from a present27day simulation of the HadAM3 model (Pope et al., 2000; Tian and Chipperfield, 2005) 28Values for cloud particle effective radii were taken from the GRAPE dataset (Sayer et al., 292011) 30To calculate an ozone radiative forcing, the code is applied as follows A base calculation of 31radiation fluxes is performed, using multi-annually averaged monthly ozone data from the 11850s, for each column of model atmosphere The radiation calculation is then repeated, 2keeping everything the same, but using a different ozone field (e.g., from the 2000s) The 3change in net radiation at the tropopause between these two calculations gives the 4instantaneous radiative forcing 6By changing the ozone field, heating rates in the stratosphere will have changed If such a 7change were to happen in the real atmosphere, stratospheric temperatures would respond 8quickly (days to months) – much more quickly than the surface-troposphere system, which 9will adjust on multiannual timescales A better estimate of the long-term forcing on the 10surface climate takes into account this short-term response of stratospheric temperatures 11(Forster et al., 2007) Stratospheric temperature adjustment was achieved by first calculating 12stratospheric heating rates for the base atmosphere The stratosphere was assumed to be in 13thermal equilibrium, with dynamical heating exactly balancing the radiative heating 14Furthermore, the dynamics were assumed to remain constant following a perturbation to 15ozone Hence to maintain equilibrium, radiative heating rates must also remain unchanged To 16achieve this, stratospheric temperatures were iteratively adjusted in the perturbed case, until 17stratospheric radiative heating rates returned to their base values This procedure is called the 18fixed dynamical heating approximation (Ramanathan and Dickinson, 1979) Here we report 19annual mean forcings at the tropopause, after stratospheric temperature adjustment 20We make some compareisons theof calculations fromwith the Edwards-Slingo radiation 21scheme to results from similar schemes used in the Norwegian Met Office in from Oslo and at 22the National Center from Atmospheric research (NCAR) in the USA The Oslo radiative 23transfer calculations are performed with a broad band longwave scheme (Myhre and Stordal, 241997) and a model using the discrete ordinate method (Stamnes et al., 1988) for the shortwave 25calculations (see further description in Myhre et al (2011)) In the radiative transfer 26calculations meteorological data from ECMWF are used and stratospheric temperature 27adjustment is included Ozone RFs were also calculated offline using the NCAR Community 28Climate System Model radiative transfer model and allowing for stratospheric temperatures 29to adjust (Conley et al., 2012) We compute the net longwave and shortwave all-sky flux at 30the tropopause (based on a climatology of tropopause pressure from the NCAR/NCEP 31reanalyses) using the same conditions for all parameters except for the ozone distribution 32 13 23.1 Results Pre-industrial (1850s) and present-day (2000s) simulations 33.1.1 Core ACCMIP experiments 43.1.1.1 Ozone distributions and changes 5Figure shows the multi-model mean (MMM) annual zonal mean (AZM) ozone for the 61850s and 2000s All models are included in the MMM, with equal weighting In the 7supplementary material, Figure S1 shows AZM (ppb) for the 1850s and 2000s for all of the 8individual 17 models Figure shows maps of MMM tropospheric column ozone (DU) for 9the 1850s and 2000s Figure S2 is the equivalent for all 17 models In these figures, we use 10the same monthly zonal mean climatological tropopause (hitherto referred to as MASKZMT) 11for all models, based on the PVU definition applied to NCEP/NCAR reanalysis data (Cionni 12et al., 2011) We also calculate O3 changes and radiative forcing results using a different 13tropopause definition (1850s O3 = 150 ppbv; hitherto referred to as MASK150) to test how 14sensitive O3 and RF results are to this choice Table compares global mean tropospheric 15column O3 changes using both definitions for all models Evaluation of simulated present-day 16ozone fields against a variety of observational data sets can be found elsewhere (Young et al., 172012) Present-day ozone distributions of AZM and tropospheric columns are similar to those 18presented in Stevenson et al (2006) from the ACCENT PhotoComp model intercomparison 19Figure shows the multi-model mean change in AZMzonal annual mean ozone (ppb) and the 20change in tropospheric column ozone (DU) for MASKZMT between 2000 and 1850 Figure 21S3 is the equivalent for all 17 models Ozone generally increases throughout the troposphere, 22most strongly in the Northern Hemisphere sub-tropical upper troposphere (Figure 3) This 23mainly reflects the industrialised latitudes where emissions are concentrated, and the fact that 24the ozone lifetime is longer in the upper troposphere Decreases in ozone are seen in the high 25latitudes of the Southern Hemisphere (SH), to varying degrees in many models (Figures 3, 26S3a) This reflects downwards transport of ozone depleted air from the stratosphere, and is 27especially pronounced in models M, G and H This effect is strong enough in several models 28(especially M, G, H) to produce decreases in column ozone in high SH latitudes (Figure S3b) 10 2From the emissions and the atmospheric lifetime of CH4 we calculate the corresponding 3atmospheric source of CO2 as a function of time For the NMVOCs we have assumed an 4average carbon content of 80% by mass We then calculate the resulting development in the 5CO2 concentrations using the impulse response function for CO2 (IRF 1) given in IPCC 6(2007) The change in the mixing ratio of CO2 (XCO2(t)) is given by 10The contribution to atmospheric CO2 is given in figure 2a 11 12The radiative forcing due to the CO2 from the fossil/anthropogenic emissions of methane, 13CO and NMVOCs are calculated by the simple parameterization given in IPCC (2001) 14 15 16 17The calculations are done with X = 278 ppm The radiative forcings are shown in figure 2b 18In 2010 the RFs are 18, 87 and 33 mWm -2 for emissions of methane, CO and NMVOC 19respectively 20 21 11 45 1Figure Contribution from fossil/anthropogenic emissions of CH4, CO and NMVOCs to 2atmospheric CO2 and radiative forcing 46 1Figures 3Figure S1: Annual zonal mean ozone (ppb) for the 1850s and 2000s For scale see Figure 1 47 2Figure S1 (continued) 48 2Figure S2: Annual mean tropospheric column O3 (DU), for the 1850s and 2000s, using the 3MASKZMT For scale see Figure 49 2Figure S2 (continued) 50 2Figure S3: Changes (2000s-1850s) in annual zonal mean O3 (ppbv) and tropospheric column 3O3 (both masked using MASKZMT) See Figure for scales 51 2Figure S3 (continued) 52 2Figure S4a: Total tropospheric O3 RFs for all models, masked using MASKZMT 53 2Figure S4b: Annual mean tropospheric O3 total, SW, and LW RFs for the IGAC/SPARC 3ozone dataset, as used in Cionni et al (2011) Compare to multi-model mean in Figure 4; also 4compare to Cionni et al.’s Figure 15a, which shows the total RF calculated with an earlier 5version of the E-S radiation code, and finds a total RF 27% lower 54 3Figure S5: Normalised total tropospheric? RFs for MASKZMT See Figure for scale 55 2Figure S6: An example of one model’s results (model B) for the attribution experiments Left 3hand side shows contributions to changes in zonal annual mean ozone, right hand side shows 4contributions to change in annual mean tropospheric ozone column Referring to the 5experiment numbers in Table 3, rows from the top show: experiment #1-#0 (all components); 6#1-#2 (CH4); #1-#3 (NOx); #1-#4 (CO); #1-#5 (NMVOC) 56 2Figure S7: Radiative forcings for model B, from the attribution experiments Left-hand plot 3shows total 2000s-1850s (#1-#0); middle shows the CH4 component (#1-#2); and the right4hand plot shows the NOx component (#1-#3) The CO and NMVOC components are 5significantly less (38 and 37 mWm-2, respectively) and are not shown 57 3Figure S8: As Figure 6, but showing the impact of climate change on O3 from 1850s to the 4(a) 2030s; (b) 2100s 58 2Figure S8 continued 59