1
Pre-industrial to end 21st century projections of
2
tropospheric ozone from the Atmospheric Chemistry and
3
Climate Model Intercomparison Project (ACCMIP)
4
5
P. J. Young1,2, A. T. Archibald3, K. Bowman4, J.-F. Lamarque5, V. Naik6, D. S.
6
Stevenson7, S. Tilmes5, A. Voulgarakis8, O. Wild9, D. Bergmann10, P. Cameron-
7
Smith10, I. Cionni11, W. J. Collins12, S. Dalsoren13, R. Doherty7, V. Eyring14, G.
8
Faluvegi15, G. Folberth12, L. W. Horowitz6, B. Josse16, Y. Lee8, I. McKenzie7, T.
9
Nagashima17, D. Plummer18, M. Righi14, S. Rumbold12, R. Skeie13, D. T.
10
Shindell15, S. Strode19, K. Sudo20, S. Szopa21 and G. Zeng22
11
[1] Cooperative Institute for Research in the Environmental Sciences, University of Colorado-
12
Boulder, Boulder, Colorado, USA
13
[2] Chemical Sciences Division, NOAA Earth System Research Laboratory, Boulder,
14
Colorado, USA
15
[3] Centre for Atmospheric Science, University of Cambridge, UK
16
[4] NASA Jet Propulsion Laboratory, Pasadena, California, USA
17
[5] National Center for Atmospheric Research, Boulder, Colorado, USA.
18
[6] NOAA Geophysical Fluid Dynamics Laboratory, Princeton, New Jersey, USA.
19
[7] School of Geosciences, University of Edinburgh, Edinburgh, UK.
20
[8] Department of Physics, Imperial College, London, UK
21
[9] Lancaster Environment Centre, University of Lancaster, Lancaster, UK.
22
[10] Lawrence Livermore National Laboratory, Livermore, California, USA.
23
[11] ENEA, Bologna, Italy
24
[12] Hadley Centre for Climate Prediction, Met Office, Exeter, UK.
25
[13] CICERO, Center for International Climate and Environmental Research-Oslo, Oslo,
26
Norway.
1
1
[14] Deutsches Zentrum fur Luft- und Raumfahrt, Institut für Physik der Atmosphäre,
2
Oberpfaffenhofen, Germany
3
[15] NASA Goddard Institute for Space Studies, New York City, New York, USA.
4
[16] Meteo-France, CNRM/GMGEC/CARMA, Toulouse, France.
5
[17] Frontier Research Center for Global Change, Japan Marine Science and Technology
6
Center, Yokohama, Japan.
7
[18] Canadian Centre for Climate Modeling and Analysis, Environment Canada, Victoria,
8
British Columbia, Canada.
9
[19] NASA Goddard Space Flight Center, Greenbelt, Maryland, USA.
10
[20] Department of Earth and Environmental Science, Graduate School of Environmental
11
Studies, Nagoya University, Nagoya, Japan
12
[21] Laboratoire des Sciences du Climat et de l’Environnement, Gif-sur-Yvette, France.
13
[22] National Institute of Water and Atmospheric Research, Lauder, New Zealand.
14
15
Correspondence to: P. J. Young (paul.j.young@noaa.gov)
16
17
2
1
Abstract
2
Modelled present day tropospheric ozone and its changes between 1850 and 2100 are
3
considered, analysing time slices from the 3 chemical transport models and 12 chemistry
4
climate models that participated in the Atmospheric Chemistry and Climate Model
5
Intercomparison Project (ACCMIP). The mean model generally performs well against present
6
day satellite and in situ observations. The seasonal cycle is well captured, except compared to
7
some locations in the tropical upper troposphere. Observations are consistent in suggesting a
8
high bias for the mean model in the Northern Hemisphere (NH) and a low bias in the
9
Southern Hemisphere (SH), which is also true for the majority of models. However, a range
10
of global mean tropospheric ozone column estimates from satellite data encompasses 2/3 of
11
the models. Compared to the present day, the mean tropospheric ozone burden for 1850 time
12
slice is ~30% lower, with the largest contribution to the change coming from the NH
13
extratropics. The mean burden increases between 1980 and 2000 (4%), although, based on the
14
spread of the model results, this is not significant. Future changes in tropospheric ozone were
15
considered for 2030 and 2100 time slices, using different projections of climate and ozone
16
precursor emissions from four Representative Concentration Pathways (RCPs). Compared to
17
2000, the relative changes for the tropospheric ozone burden in 2030 (2100) for the different
18
RCPs are: -5% (-22%) for RCP2.6, 3% (-8%) for RCP4.5, 0% (-9%) for RCP6.0, and 5%
19
(15%) for RCP8.0. Based on the inter-model spread in the response, these changes are
20
significant except in 2030 for RCP6.0. The decreases apparent for most RCPs are due to
21
reductions in precursor emissions, but the increase in ozone for RCP8.5 is in spite of
22
reductions in nitrogen oxide emissions and can be attributed to the very large increase
23
(~doubling) in methane. Individual models with high ozone burdens and concentrations for
24
the present day also have high ozone levels for the other time slices, but the modelled burden
25
is not correlated with the magnitude of the ozone change. Furthermore, the modelled changes
26
in ozone burden for the different models are not correlated across time slices: i.e. there is no
27
consistent outlier. A unified approach to ozone budget specifications will help future studies
28
diagnose the drivers of the ozone changes and inter-model differences more clearly.
29
30
1
Introduction
31
The Atmospheric Chemistry and Climate Model Intercomparison Project (ACCMIP) is
32
designed to complement the climate model simulations being conducted for the Coupled
3
1
Model Intercomparison Project (CMIP), Phase 5 (e.g. Taylor et al., 2011), and both will
2
inform the Intergovernmental Panel on Climate Change (IPCC) Fifth Assessment Report
3
(AR5). A primary goal of ACCMIP is to use its ensemble of tropospheric chemistry-climate
4
models to investigate the evolution and distribution of short-lived, chemically-active climate
5
forcing agents for a range of scenarios, a topic that is not to be investigated in detail as part of
6
CMIP5. Ozone in the troposphere is one such short-lived, chemically-active forcing agent,
7
and, as it is both a pollutant and greenhouse gas, it straddles research communities concerned
8
with air quality and climate. This study is concerned with documenting the evolution and
9
distribution of tropospheric ozone in the ACCMIP models, detailing the large-scale ozone
10
changes since the pre-industrial period, through to the end of the 21st century, with a focus on
11
where the projected changes from the ensemble are “robust”.
12
(Paragraph on importance of ozone and where it comes from)
13
(Paragraph on other modelling studies that have looked at ozone changes. Importance of
14
different emissions here…although same as HTAP)
15
(Paragraph on what this study will add….and what is missing).
16
This study complements parallel investigation of the ACCMIP ensemble, such as the analysis
17
of past (Naik et al., 2012) and future (Voulgarikis et al., 2012) trends in OH and methane
18
lifetime, and analysis of the radiative impact of short-lived climate forcing agents (Shindell et
19
al., 2012), in particular including analyses focussed on the radiative forcing of tropospheric
20
ozone (Bowman et al., 2012; Stevenson et al., 2012). This study is also complementary to an
21
investigation of tropospheric and stratospheric ozone in the CMIP5 models (Eyring et al.,
22
2012). Ongoing analyses of the ACCMIP ensemble are investigating the impacts of surface
23
ozone changes on crop growth and human health.
24
(Outline the paper)
25
4
1
2
Models, simulations and analysis details
2
Here we provide brief details of the ACCMIP models and simulations, together with some
3
details on the analysis performed in this study. Lamarque et al. (2012) provide a more
4
complete description of the models and further details for the simulations. Note that the
5
analysis here does not include the ACCMIP model NCAR-CAM5.1.
6
2.1
7
Table 1 summarises the models, scenarios and their time periods analysed in this study (the
8
tropospheric ozone burdens are discussed in later sections of the text). For this study, we used
9
the output from 15 models, although not all of the models provided output for every scenario
ACCMIP models
10
and period, as indicated by “–” in Table 1.
11
Most of the ACCMIP models are climate models with atmospheric chemistry modules, run in
12
atmosphere-only mode; i.e. the models are driven by sea-surface temperature (SST) and sea-
13
ice boundary conditions. GISS-E2-R uniquely was run as a fully-coupled ocean-atmosphere
14
climate model, although the closely-related GISS-E2-R-TOMAS model was run with SSTs
15
and sea-ice prescribed. CICERO-OsloCTM2, MOCAGE and STOC-HadAM3 are chemical
16
transport models (CTMs), with MOCAGE and STOC-HadAM3 using meteorological fields
17
from an appropriate simulation of a climate model, and CICERO-OsloCTM2 using
18
meteorological fields from a single year of a reanalysis dataset.
19
The model chemical schemes vary greatly in their complexity (e.g. as measured by the
20
number of species and reactions), particularly in the range of non-methane VOCs (NMVOCs)
21
that they simulate. Complexity ranges from the simplified and parameterized schemes of
22
CMAM (no NMVOCs) and CESM-CAM-superfast (isoprene as the only NMVOC), to the
23
intermediate schemes of HadGEM2 and UM-CAM (include ≤ C3-alkanes), to the more
24
complex schemes of the other models, which include the more reactive, chiefly anthropogenic
25
NMVOCs (e.g. higher alkanes, alkenes and aromatic species), as well as lumped
26
monoterpenes. Some representation of stratospheric chemistry is included in most models,
27
with
28
LMDzORINCA, STOC-HadAM3 and UM-CAM. Ozone concentrations calculated by the
29
model chemistry scheme are used online in the radiation code for all models, except the
30
CTMs and UM-CAM. CMAM, GEOSCCM, LMDzORINCA, MOCAGE and UM-CAM did
31
not calculated online aerosol concentrations.
the
exception
of
CESM-CAM-superfast,
CICERO-OsloCTM2,
HadGEM2,
5
1
2.2
Scenarios and time slices
2
The ACCMIP simulations broadly correspond with the CMIP5 scenarios (Taylor et al., 2011).
3
Historical scenario (hereafter Hist) simulations cover the preindustrial period to the present
4
day, while a range of Representative Concentration Pathways (RCPs) (van Vuuren et al.,
5
2011) cover 21st century projections. These latter scenarios are named for their nominal
6
radiative forcing level (2100 compared to 1750), such that RCP2.6 corresponds to 2.6 Wm-2,
7
RCP4.5 to 4.5 Wm-2, RCP6.0 to 6.0 Wm-2 and RCP8.5 to 8.5 Wm-2. Ozone precursor
8
emissions from anthropogenic and biomass burning sources were taken from those compiled
9
by Lamarque et al. (2010) for the Hist simulations, whereas emissions for the RCP
10
simulations are described by Lamarque et al. (2012) (see also Lamarque et al., 2011; van
11
Vuuren et al., 2011). Natural emissions, such as CO and VOCs from vegetation and oceans,
12
and NOx from soils and lighting, were determined by each model group. The emissions used
13
by the individual models are discussed further in Sect. 3.
14
With the exception of GISS-E2-R and LMDzORINCA, each model conducted a set of time
15
slice simulations for each scenario. In this study, we analyse output from the 1850, 1980 and
16
2000 time slices from the Hist scenario, and the 2030 and 2100 time slices for the RCPs, to
17
provide near-term and longer-term perspectives. Except for the CTMs and GISS-E2-R, each
18
model used climatological SSTs and sea-ice boundary conditions from coupled ocean-
19
atmosphere CMIP5 simulations of a closely related climate model, averaged for the 10 years
20
about each time slice (e.g. 2026-2035 for the 2030 time slice). CICERO-OsloCTM2 used the
21
same meteorology for each simulation, whereas MOCAGE and STOC-HadAM3 were run
22
similarly to the CCMs, except using decadal-mean meteorological fields from a climate model
23
running the appropriate time slice and scenario, rather than directly using the SSTs and sea-
24
ice to drive an atmosphere model. The number of years that the ACCMIP models simulated
25
for each time slice mostly varied between 4 and 12 years for each model, although CICERO-
26
OsloCTM2 only simulated a single year. However, as the boundary conditions were constant
27
for each year of a given time slice, “interannual” variability is generally small (see Sect. 4).
28
GISS-E2-R and LMDzORINCA both conducted transient simulations, and the data analysed
29
in this study were averaged for the decade about each time slice (e.g. 1976-1985 for the 1980
30
time slice), with some minor exceptions as noted in Table 1.
6
1
2.3
Analyses: Tropopause definition and statistical definitions
2
(Sub)-section here on common gridding, tropopause definition and statistical tests. Will define
3
what is meant by “significant” and “uncertainty”. Something about sensitivity of tropopause
4
definition?
5
6
3
Emissions: differences and similarities between models
7
While the goal of ACCMIP was for models to match each other’s ozone precursor inputs as
8
closely as possible, differences in model parameterisations and complexity means that some
9
model diversity is unavoidable. In particular, natural emissions were not prescribed as part of
10
the experiment design and their differing treatment between models broadens the range of
11
ozone precursors.
12
Figure 1 shows the range of ozone precursors in the ACCMIP models, presenting box-
13
whisker plots for each scenario and time slice, for (a) the tropospheric methane burden, (b)
14
CO emissions, (c) total NOx emissions, (d) lightning NOx emissions (separated out from (c)),
15
and (e) VOC emissions. Each box represents the inter-quartile range (IQR) of the emissions
16
from the different models, with the mean (dot) and median (line) also indicated. The whiskers
17
indicate the full range of the given emission (or burden) that lies outside the IQR. The number
18
of models that constitute the spread of the data is different for different simulations – see
19
Table 1.
20
Figure 1a shows that the tropospheric methane burden is generally well constrained, with the
21
IQR 3-5.5% of the mean burden. The close agreement is due to all models except
22
LMDzORINCA having used prescribed methane surface concentrations for the Hist
23
simulations, and only GISS-E2-R and LMDzORINCA not prescribing concentrations for the
24
RCP simulations. The methane burden approximately doubles from 1850 to 2000, but 2100
25
burdens are 30%, 10% and 2.5% lower than 2000, for the RCP2.6, RCP4.5 and RCP6.0
26
scenarios respectively. For RCP8.5, by 2100 the methane burden has doubled again compared
27
to 2000. For each model, the troposphere is defined using ozone data from the Hist 1850
28
simulation, masking out regions where ozone exceeds 150 ppbv (e.g. Stevenson et al., 2006).
29
(This to go in Sect. 2.3)
30
The general trends in CO (Fig. 1b) and total NOx (Fig. 1c) emissions are similar to one
31
another, increasing from 1850 to 2000 for the Hist scenario, decreasing thereafter for all the
7
1
RCPs, except for the 5% higher NOx emissions for the 2030 time slice of RCP8.5, compared
2
to Hist 2000. NOx emissions show a stronger increase over the 20th century, with the mean
3
trebling between 1850 and 2000, whereas CO emissions slightly more than double. For all
4
RCPs, by 2100 CO and NOx emissions are lower by 30-45% compared to 2000.
5
Both CO and NOx emissions show a greater degree of variation between the models than the
6
methane burden. For a given time slice, the IQRs vary between approximately 10-30% of the
7
corresponding mean emission, whereas the full range (maximum minus minimum emission)
8
is between 20-100% of the mean emission. The spread is due to both the varying natural
9
emissions (NOx from soils and lightning; CO from oceans and vegetation), as well as the less
10
complex models including extra CO emissions as a surrogate for missing NMVOCs (e.g.
11
CMAM, HadGEM2).
12
An idea of the variation in natural emissions can be seen from Fig. 1d, which shows the
13
spread in the lightning NOx source (LNOx) for the models. Parameterisations for LNOx are
14
generally dependent on cloud top heights and convective mass fluxes (e.g. Price and Rind,
15
1992; Allen and Pickering, 2002), which likely show large variability between models,
16
accounting for spread. The IQRs are generally 40-55% of the mean emission, and the full
17
range is 90-170% of the mean. The maximum emissions come from the MIROC-CHEM,
18
whose LNOx was erroneously high (by 60%) in the ACCMIP simulations. The minimum
19
emissions (< 2 Tg N / yr) are generally from the erroneously low HadGEM2 LNO x, although
20
emissions are low for CMAM for the RCP simulations. Our knowledge of LNOx is generally
21
poor, but (excluding HadGEM2 and MIROC-CHEM) LNOx for the Hist 2000 simulation
22
ranges between 3.8-7.7 Tg N/yr, within the range of 5±3 Tg N a-1 estimated by Schumann and
23
Huntrieser (2007) for a range of LNOx parameterisations.
24
Possible changes in lightning activity with climate change (e.g. Williams, 2009 and refs.
25
therein) has been recognised as potentially important for LNOx and the subsequent impact on
26
tropospheric composition (e.g. Price and Rind, 1994; Hauglustaine et al., 2005; Fiore et al.,
27
2006; Schumann and Huntrieser, 2007; Zeng et al., 2008), even if only the spatial distribution
28
changes (Stevenson et al., 2005). An increase in LNOx from 2000 to 2100 (RCP8.5) is
29
generally robust across the ACCMIP models, and ranges in magnitude from 10-75%. CMAM
30
is an outlier in this case, with 45% lower emissions for RCP8.5 2100 compared to Hist 2000.
31
CMAM is also the only model using an LNOx parameterisation based on the study of Allen
32
and Pickering (2002). Jacobson and Streets (2009) also modelled lower LNOx in a warmer
8
1
climate, using a different parameterisation again. Clearly further study is required into the
2
implications of the use of different parameterisations for LNOx, and the different sensitivities
3
across models.
4
Finally for emissions, Fig. 1e shows the VOC emissions used in the ACCMIP models. As
5
with LNOx, the emissions cover a wide range and there is no clear trend in the mean across
6
the simulations. Many of the reasons for the differences are familiar from the above
7
discussion, particularly the fact that some models include more VOCs than others. Another
8
reason for the spread in VOC emissions comes from treatment of VOCs of biogenic origin,
9
particularly isoprene, which likely dominates the total NMVOC emissions (e.g. Guenther et
10
al., 1995). EMAC, GISS-E2-R and STOC-HadAM3 simulations were the only ones to include
11
climate-sensitive isoprene emissions, and these are the only models with a positive change in
12
VOC emissions between Hist 2000 and RCP8.5 2100, arising from the positive temperature
13
dependence of isoprene emission algorithms (e.g. Guenther et al., 2006; Arneth et al., 2007).
14
Arneth et al. (2011) noted that the isoprene emission computed from a given algorithm is
15
highly sensitive to the input meteorological data and vegetation boundary conditions, giving
16
further cause for variation in VOC emissions between models. For the rest of the ACCMIP
17
ensemble, if they included isoprene chemistry, constant present day isoprene emissions were
18
used for all simulations, and their individual VOC emission trends broadly resemble those of
19
CO and NOx.
20
21
22
4
Present day ozone distribution and model-observation comparison will
shorten
23
This section presents the distribution of tropospheric ozone (surface, column and zonal mean)
24
as simulated by the ACCMIP models for the Hist 2000 simulation. The ozone concentrations
25
and columns output from this simulation are then used to evaluate the models performance
26
against observational datasets, from both satellites and ozonesondes. As with much of the
27
discussion in subsequent sections, the emphasis is largely on the distribution and performance
28
of the ACCMIP ensemble mean, describing the spread of model results with statistical
29
metrics, although highlighting any particular patterns seen for individual models. The
30
distributions of ozone and data on model performance for all the individual models can be
31
found in the supplementary material.
9
1
To calculate the tropospheric ozone column, the troposphere was defined as where the ozone
2
concentration is less than or equal to 150 ppbv in the Hist 1850 time slice, applied on a per
3
model basis, and varying by month. This definition is used throughout the study to give a
4
consistent troposphere, and we use the 1850 time slice to avoid issues with different degrees
5
of stratospheric ozone depletion across the ensemble. (moving to sect 2.3)
6
4.1
7
Figure 2 shows the ensemble mean, annual mean distribution of ozone, presenting the zonal
8
mean, tropospheric column and surface ozone concentrations, as well as their inter-model
9
variability. The variability is indicated both by the standard deviation and the coefficient of
Zonal mean, tropospheric column and surface ozone from Hist 2000
10
variation, i.e. the standard deviation expressed as a percentage of the mean concentration.
11
The general patterns of the ozone distribution in Figs. 2a, 2d and 2g are consistent with those
12
reported from satellite (Fishman et al., 1990; Ziemke et al., 2011) and ozonesonde (Logan,
13
1999; Thompson et al., 2003) measurements, as well as the multi-model data shown by
14
Stevenson et al. (2006). Convective lifting of low-ozone air masses coupled with lofting of
15
ozone precursors (Lawrence et al., 2003; Doherty et al., 2005) results in the characteristic
16
tropical zonal mean profile in Fig. 2a. The hemispheric asymmetry in mid-tropospheric ozone
17
concentrations reflects the larger input of stratospheric ozone in the NH, due to the stronger
18
Brewer-Dobson circulation there (Rosenlof, 1995), as well as the larger emissions of ozone
19
precursors (Lamarque et al., 2010). While both the tropospheric column (Fig. 2d) and surface
20
concentrations (Fig. 2g) also show higher ozone levels over source regions, these plots also
21
indicate enhanced concentrations downwind of the source regions, due to transport of ozone,
22
ozone precursors, and “reservoir” species, such as PAN (Moxim et al., 1996; Fiore et al.,
23
2009). Figure 2d also shows the “wave-1” pattern in the tropical tropospheric ozone column
24
(Thompson et al., 2003), with a minimum in ozone over the Pacific Ocean and maximum over
25
the Atlantic. Surface ozone concentrations are also very low over the equatorial Pacific
26
Ocean.
27
There is generally good agreement between the models for the zonal mean profile of ozone.
28
Figure 2c shows that the standard deviation is less than 20% of the mean throughout much of
29
the troposphere, with exception of some lower troposphere regions and the upper troposphere.
30
The spatial patterns of the spread in surface ozone concentrations in Figs. 2h and 2i suggest
31
that much of the lower troposphere variability is over regions with large anthropogenic,
10
1
pyrogenic or biogenic emissions, where both the absolute and relative uncertainty is largest.
2
In anthropogenic and biomass burning source regions, much of the model diversity could
3
reflect the spread in VOC composition (low vs. high reactivity species), which means
4
different ozone production efficiencies (e.g. Russell et al., 1995). For tropical Africa and
5
South America, large variations are apparent over isoprene source regions, which reflects
6
differences in the total emission (some models have no isoprene), as well as differences in
7
isoprene chemistry (Archibald et al., 2010).
8
Large model variation is also found for the high latitude SH, chiefly for the tropospheric
9
column but also the surface ozone concentrations. Ozone levels in the SH are generally low,
10
but there is relatively large diversity in the overhead stratospheric ozone column (standard
11
deviation for the total ozone column is 10-15% of the mean; not shown). This results in a
12
spread in the stratospheric input as well as potentially some impacts through changes in
13
photolysis rates, for those models with photolysis schemes that use the model-calculated
14
ozone column (e.g. Fuglestvedt et al., 1994; see also Voulgarikis et al. 2012). There is less
15
uncertainty in the tropospheric column in the NH, coupled with less spread in between models
16
for the total ozone column. However, larger uncertainty at NH high latitudes could be related
17
to transport and chemistry related to NH emitting regions. As was also found by Stevenson et
18
al. (2006), there are large relative uncertainties over the equatorial Pacific Ocean, but the
19
concentrations are very low.
20
Figure 3a shows the annual mean tropospheric ozone burden for all the models, as well as the
21
ACCMIP mean. Values for the tropospheric burden for all models and scenarios can also be
22
found in Table 1. The mean burden is 337 ± 23 Tg, very close to the 336 ± 27 Tg reported for
23
a subset of the ACCENT models (Stevenson et al. 2006), and the 335 ± 10 Tg estimated from
24
measurement climatologies by (Wild, 2007). The error bars for the individual models in Fig.
25
3a indicate the uncertainty in the ozone burden, as represented by the standard deviation of
26
the range of burdens computed for individual years of the time slice. The uncertainty is small,
27
and the standard deviations are less than 2% of the burden.
28
There is a significant correlation (r = 0.67) between the modelled ozone burden and the total
29
VOC emissions; the models are arranged in order of increasing VOC emissions in Fig. 3a. In
30
the absence of ozone budget data and information on the VOC make-up of the models, it is
31
difficult to rationalise this correlation satisfactorily. However, Wild (2007) demonstrated
32
increased VOC emissions lead to an increased ozone burden.
11
1
Figure 3b indicates the distribution of the mean ozone burden throughout the troposphere,
2
using the regions defined by Lawrence et al. (2001) (that they used describe the distribution of
3
OH) to give a mass-weighted view of the zonal ozone distribution. The hemispheric
4
asymmetry in ozone is apparent from Fig. 3b, which shows that 57.5% of the ozone mass is in
5
the NH. The NH extratropics has 60% more ozone than the SH extratropics, but the NH
6
tropics has only slightly more ozone (~3%) than the SH tropics. The greatest burdens are
7
found in the extratropical upper troposphere, reflecting the importance of stratosphere-to-
8
troposphere transport of ozone. Relative to the rest of the globe, substantial burdens are found
9
in the comparatively more polluted NH lower troposphere, as well as the tropical upper
10
troposphere. This latter region is impacted by convective transport of ozone precursors and
11
lightning emissions (Jacob et al., 1996). The standard deviation in the fractional distribution
12
of ozone is also in Fig. 3b, showing that the model uncertainty in distribution of ozone mass is
13
largest in the NH extratropics and SH upper troposphere, consistent with the results in Fig. 2.
14
4.2
15
Figure 4 compares the ACCMIP mean, median and individually modelled ozone
16
concentrations from the Hist 2000 simulation against ozonesonde data, in the same manner as
17
Stevenson et al. (2006) (their Fig. 2). Ozonesonde measurements are taken from datasets
18
compiled by Logan (1999) (representative of 1980-1993) and Thompson et al. (2003)
19
(representative of 1997 onwards), and consist of a total of 48 stations, split 5, 15, 10 and 18
20
between the SH extratropics, SH tropics, NH tropics and NH extratropics respectively. In
21
addition, Fig. 4 shows satellite-derived ozone concentrations retrieved from the Tropospheric
22
Emission Spectrometer (TES), where a 2005-2010 climatology of TES data were put on the
23
same grid as the ACCMIP models and sampled at the ozonesonde locations. Figure 4 also
24
shows the ACCENT model mean (Stevenson et al. 2006), to place the ACCMIP results in
25
context of recent multi-model comparisons. The correlation and mean normalised bias error
26
(MNBE) are shown for mean model from the ACCMIP and ACCENT ensembles, relative to
27
the ozonesonde observations.
28
In terms of matching the magnitude of the concentrations, the ACCMIP mean and median
29
models are within the standard deviation of the observations for all locations and altitudes.
30
Compared to the mean observations, the largest relative errors are found for the NH
31
extratropics, where the mean model overestimates the concentrations, and SH tropics, where
32
the mean model underestimates the concentrations. The individual model biases in these
Comparison to ozonesondes and satellite data
12
1
locations are significantly correlated with total VOC emissions (r = 0.57; i.e. more VOC
2
emissions give a more positive, or less negative, bias), although several other chemical and
3
transport factors likely play a role. However, the mean model captures the annual cycle in
4
ozone concentrations extremely well in most locations (as measured by the correlation
5
coefficient), suggesting that, broadly speaking, the seasonality in circulation patterns,
6
stratosphere-to-troposphere and natural emissions (chiefly biomass burning in the tropics, and
7
isoprene in the NH extratropics) is captured well. The statistics for the NH tropical mid and
8
upper troposphere suggest that the seasonality is less well modelled, although we note that, 1)
9
the observed-model correlation is significant (r > 0.58), 2) there is considerable interannual
10
variability in the upper troposphere, and 3) the bias and correlation are improved compared to
11
the ACCENT mean. Compared to ACCENT, the correlation is improved with the ACCMIP
12
mean model for most locations, and the bias for some locations, although the two mean
13
models are not significantly different (double check!).
14
Except for the NH Tropics at 250 hPa, the TES data are positively biased compared to the
15
ozonesondes, although, taking the interannual variability into account, we note that they are
16
not notably (significantly?) different (especially as the uncertainty for the TES measurements
17
are not included). This positive bias means that, compared to the ozonesonde data, the
18
ACCMIP mean model bias against TES is improved for the NH extratropics, about the same
19
for the NH tropics (opposite in sign), but worsened for the SH. Changes in correlation are
20
more marginal.
21
Figure 5 makes a similar comparison to ozonesonde data, this time using the compilation of
22
Tilmes et al. (2011). This dataset mostly consists of the same station data described by Logan
23
(1999) and Thompson et al. (2003), but updated to cover 1995-2007, and aggregated into 12
24
regions that exhibit similar ozone concentration characteristics (see the top panel of Fig. 5 and
25
Tilmes et al. (2011)). The figure presents the mean, median and spread of the MNBE for the
26
individual ACCMIP models (cf. Fig. 1), showing that the full range of performance
27
encompasses positive and negative biases for each region and altitude.
28
The information in Fig. 5 is consistent with that in Fig. 4, but with more longitudinal
29
information. For instance, we see that the negative bias in the SH tropical ozone is driven by
30
the less favourable comparison of the model mean with the sites in the Atlantic/Africa region
31
(dark green), and the sign of the bias is consistent across more than 75% of the models. A
32
positive bias is apparent in all the NH extratropical regions in the low and mid-troposphere,
13
1
and again is shared by the majority of the models. Figure 5 also shows low biases for the high
2
latitude regions in the upper troposphere/lower stratosphere (a region is not shown in Fig. 4).
3
In contrast, comparison of the ACCMIP mean total ozone column against satellite
4
measurements from the merged Total Ozone Mapping Spectrometer/solar backscatter
5
ultraviolet (TOMS/SBUV) data (Stolarski and Frith, 2006) suggests that the models
6
overestimate the total ozone column by around 5% at high latitudes (not shown). Validation
7
of stratospheric ozone in these models is beyond the scope of this study, but this would help
8
resolve whether ozonesonde-model comparisons at higher altitudes are consistent with the
9
satellite data.
10
Tropospheric ozone columns are available from a combination of the Ozone Monitoring
11
Instrument (OMI) and Microwave Limb Sounder (MLS) data. Figure 6 compares the
12
ACCMIP mean tropospheric ozone column (Fig. 2d) against the OMI-MLS climatology
13
derived by Ziemke et al. (2011), covering October 2004 to December 2010. Many of the
14
differences between the ACCMIP mean model and OMI (Fig. 6c) are broadly consistent with
15
the comparison against ozonesonde data. The mean model appears to overestimate the column
16
across the NH mid-latitudes, and underestimate the column over tropical oceans and for all
17
regions poleward of approximately 30°S, although the underestimate is stronger than
18
suggested by the ozonesonde data. The negative bias over the equatorial Pacific in Fig. 6c is
19
not consistent with the ozonesonde comparison in Fig. 5, which suggests a neutral or positive
20
bias for the mean model. However, this region is poorly represented by ozonesonde
21
measurements.
22
Table 2 has further information on the comparison against the OMI data for the ACCMIP
23
mean and individual models, showing the modelled column and bias for the global (60°S-
24
60°N) annual average tropospheric ozone column and the spatial correlation coefficient. The
25
global mean OMI tropospheric ozone column is 31.1 DU, compared to 31.5 DU for the mean
26
model. Data from the TES instrument provide an additional observationally based estimate of
27
29.8-32.8 DU, a range encompassing 2/3 of the models. GFDL-AM3 and EMAC
28
overestimate the column by the greatest amount. Compared to mean model, EMAC has
29
anomalously high ozone in the tropical mid and upper troposphere, whereas GFDL-AM3 has
30
anomalously high ozone in the lower troposphere, and SH upper troposphere. The most
31
negative biases are seen for HadGEM2 and CESM-CAM-superfast, which both have
32
anomalously low ozone throughout most of the free troposphere. The spatial correlation
14
1
between OMI and the models is generally very good, as is also evident from a qualitative
2
comparison of Figs. 6a and 6b. MOCAGE shows the poorest comparison for this metric, but
3
this model does not have as well-defined features of tropical ozone (e.g. the equatorial Pacific
4
minimum) compared to the others (see supplementary material).
5
Table 3 divides up the OMI-model comparison into similar latitude bands as Fig. 4. Biases in
6
the mean column for a given latitude band are well correlated with those for the ozonesondes,
7
at any of the pressure levels (r ≥ 0.75), providing a generally consistent picture of the model
8
evaluation between the OMI column and ozonesonde data. Moreover, correlations between
9
the biases at different latitude bands is strong for the two tropical regions (r = 0.88),
10
suggesting that similar processes are operating in the regions, even if the sign of the bias is
11
different between them. Correlations are significant between the global mean biases and those
12
for the different latitude bands, but more weakly so for the extratropical regions – i.e. a given
13
model bias in the NH or SH is not as good a predictor of its global bias, compared to the
14
model’s bias in the tropics.
15
Overall, Figs. 4, 5 and 6 and Tables 3 and 4 suggest that the mean model performs well
16
against ozone observations, and is within interannual variability. Compared to the mean
17
observations, there appears to be a systematic high bias in for the NH extratropics and NH
18
tropics, with, respectively, 14 and 11 out of the 15 models exceeding the observed
19
tropospheric column in these regions. Conversely, there is systematic low bias in the SH
20
extratropics and SH tropics, with, respectively, 14 and 10 out of the 15 models under-
21
predicting the observed column. For the NH, this could suggest an issue with the
22
anthropogenic and biomass burning emissions of ozone precursors, whereas the SH bias may
23
point to a pervasive deficiency in modelling the sources and chemistry related to natural
24
emissions, or to biases in processes related to stratospheric ozone and its transport to the
25
troposphere. (Relate to CO biases perhaps? Vaishali?)
26
27
5
Tropospheric ozone from 1850 to 2100
28
In this section we consider the changes in tropospheric ozone projected by the ACCMIP
29
models for the past (1850 and 1980), as well as for the near (2030) and more distant (2100)
30
future, using the range of RCPs. While we are mainly concerned with discussing the mean
31
result and its significance, expressed in terms of the model spread (see Sect 2.3), we also
32
consider the range of model ozone changes and note any unusual patterns for a given model.
15
1
This section presents the global-scale changes, and then more regional changes, before
2
considering the drivers of the change. Need to add comparison to Kawase et al. (2011) and
3
Lamarque et al. Clim. Dyn.
4
5.1
5
Figure 7a shows the annual average tropospheric ozone burden for the ACCMIP models for
6
all the simulations and time slices considered. Figure 7b shows the difference in the
7
tropospheric ozone burden compared to the Hist 2000 simulation. Both panels in Fig. 7 show
8
the IQR, mean, median and spread of the burden, or the difference in the burden, similarly to
9
Figs.1 and 5. Data for individual models can be found in Table 2.
Global-scale changes: Tropospheric ozone burden
10
The evolution of the mean tropospheric burden in Fig. 7a shows a 25% increase between 1850
11
and 1980, and a 29% increase between 1850 and 2000; the burden increases 4% between
12
1980 and 2000. As might be expected, the future projections vary with the scenario.
13
Compared to 2000, the relative changes in 2030 (2100) for the difference RCPs are: -5% (-
14
22%) for RCP2.6, 3% (-8%) for RCP4.5, 0% (-9%) for RCP6.0, and 5% (15%) for RCP8.5.
15
RCP8.5 is the only scenario to show an ozone increase for both time slices, whereas RCP4.5
16
shows an increase in 2030, before decreasing in 2100. The ozone burden for the 2030 time
17
slice of RCP6.0 is essentially unchanged compared to 2000, although it is still higher than
18
1980 (unlike the time slices with negative changes).
19
Figure 7 also shows a large range in the modelled ozone burdens and their differences, with
20
overlapping IQRs between many of the time slices. There is a good, but not perfect,
21
correlation between the modelled ozone burden (or the column bias) for Hist 2000 and that of
22
another time slice. GFDL-AM3 and EMAC consistently have the highest ozone burden, and
23
generally CESM-CAM-superfast and HadGEM2 having the lowest ozone burden, although
24
this is more variable (see Table 1). However, there is neither a correlation between the
25
modelled ozone burden and a given burden change, nor between the changes in ozone burden
26
for any two periods. This suggests that a model’s ozone bias is not related to its ozone change,
27
at least at these larger scales.
28
The significance of the burden change with respect to Hist 2000 can also be assessed, using
29
the inter-model spread of the differences to calculate the 95% confidence interval for the
30
mean difference (Sect 2.3). This analysis suggests that the all the changes in the ozone burden
31
are significantly different from zero at the 5% level, except for between Hist 1980 and Hist
16
1
2000, and Hist 2000 and RCP4.5 2030. The latter result could be anticipated from Fig. 7b, as
2
this is the only time slice where the models do not agree on the sign of the change. We again
3
note that “significance” here does not mean that the change is significant with respect to
4
interannual variability, merely a measure of whether the models agree on a change.
5
5.2
6
Figure 8 shows the mean model regional ozone burden changes relative to Hist 2000 for the
7
Hist 1850 and RCP 2100 simulations, dividing up the troposphere in the same manner as Fig.
8
3b. The figure also indicates the fraction that each region contributes to the overall ozone
9
change, i.e. highlighting asymmetries in the change. From Fig. 7, we see that the overall
10
ozone burden change is negative for these cases, except RCP8.5 2100. Based on the spread of
11
model results, all of the regional burden changes are significantly different from zero.
12
For Hist 1850 and RCP2.6 2100 the burden change is negative for all regions, with the largest
13
contribution to the change coming from the lower ozone precursor emissions in the NH
14
extratropics compared to Hist 2000. Unlike for the other RCPs, stratospheric ozone recovery
15
(e.g. Eyring et al., 2010) does not result in increased tropospheric ozone in the SH upper
16
troposphere, despite a 30% increase in the total column ozone (not shown). The SH
17
extratropics makes a small contribution to the overall change for in both the Hist 1850 and
18
RCP2.6 2100 case.
19
The overall decrease for RCP4.5 2100 is half as much as that for RCP2.6, but is still largely
20
dominated by the decrease in precursor emissions in the NH extratropics, with some
21
contribution from the NH tropical lower troposphere. This overall decrease is countered by a
22
relatively large increase in the SH upper troposphere, likely related to ozone recovery. The
23
magnitude and patterns of absolute ozone changes are similar for RCP6.0, although the
24
tropical upper troposphere makes more of a contribution to the overall change than in
25
RCP4.5, in both absolute and relative terms. For RCP8.5, ozone increases everywhere,
26
although the largest contribution is from the 500 to 250 hPa band.
27
Figures 9, 10 and 11 present information on the annual mean spatial patterns of ozone
28
concentration changes of all the time slices, respectively showing the percentage change in
29
zonal mean ozone, the tropospheric ozone column and surface ozone, relative to Hist 2000.
30
Greyed-out areas indicate where the absolute change in ozone is not significantly different
31
from zero, based on the spread of the model results.
Regional-scale changes: Burdens, columns and concentrations
17
1
Concentrations for Hist 1850 are less than Hist 2000 everywhere except the stratosphere,
2
showing the impact of ozone depletion. Relative decreases are largest for the lower
3
troposphere, exceeding 40% for the column (Fig. 10a) and surface (Fig. 11a) for NH mid-
4
latitudes, and particularly for East Asia and the USA. [Still drafting]
5
5.3
6
Figure 11 and links with Oliver’s HTAP linearisation.
NOx and methane as drivers of the change
7
8
6
Isolating climate- and emission-driven changes leave for Stevenson paper?
9
The results in Sect. 4 present the sum of emission and climate changes on tropospheric ozone,
10
but, as stated in the Introduction, several previous studies have identified separate impacts of
11
climate and emissions changes. A subset of models within ACCMIP performed an additional
12
sensitivity simulation to examine the role of climate and emissions changes between 1850 and
13
2000, completing a simulation with Hist 2000 emissions and Hist 1850 climate. Figure XX
14
shows the result of the…… Emissions dominate. Climate small and negative. Link with
15
Naik/Lamarque climate changes?
16
Also show Em2000Cl2100. Larger climate change.
17
18
7
Discussion
19
[Originally in Sect 3.] We note that while using the multi-model mean enhances robust
20
features across the models, the fact that models are often closely related to one another
21
(Masson and Knutti, 2011) means that we do not really have as many independent realisations
22
as there are models. For example, a particular convective scheme, isoprene mechanism or dry
23
deposition module, which may be shared by several models, could have a disproportionate
24
influence on the ensemble mean, as each model is weighted equally. However, such
25
relationships between chemistry-climate models have yet to be fully documented and
26
analysed.
27
28
Why differences? Chemistry, lightning emissions, BVOC emissions, complexity, model top
29
etc. (More on processes will come in post-IPCC paper). Links with OH and CH4, AQ etc.
18
1
Some recommendations…
2
3
8
Summary and conclusions
4
Would be nice to have consistent budget data....
5
6
Acknowledgements
7
ACCMIP is organized under the auspices of Atmospheric Chemistry and Climate (AC&C), a
8
project of International Global Atmospheric Chemistry (IGAC) and Stratospheric Processes
9
And their Role in Climate (SPARC) under the International Geosphere-Biosphere Project
10
(IGBP) and World Climate Research Program (WCRP). The authors are grateful to the British
11
Atmospheric Data Centre (BADC), which is part of the NERC National Centre for
12
Atmospheric Science (NCAS), for collecting and archiving the ACCMIP data. The work of
13
DB and PC was funded by the U.S. Dept. of Energy (BER), performed under the auspices of
14
LLNL under Contract DE-AC52-07NA27344, and used the supercomputing resources of
15
NERSC under contract No. DE-AC02-05CH11231. The CESM project is supported by the
16
National Science Foundation and the Office of Science (BER) of the U. S. Department of
17
Energy. The National Center for Atmospheric Research is operated by the University
18
Corporation for Atmospheric Research under sponsorship of the National Science Foundation.
19
GZ acknowledges NIWA HPCF facility and funding from New Zealand Ministry of Science
20
and Innovation.
21
22
References
23
24
25
Allen, D. J., and Pickering, K. E.: Evaluation of lightning flash rate parameterizations for use
in a global chemical transport model, Journal of Geophysical Research, 107, 4711,
10.1029/2002JD002066, 2002.
26
27
28
Archibald, A. T., Jenkin, M. E., and Shallcross, D. E.: An isoprene mechanism
intercomparison, Atmospheric Environment, 44, 5356-5364, 10.1016/j.atmosenv.2009.09.016
2010.
29
30
31
32
Arneth, A., Niinemets, Ü., Pressley, S., Back, J., Hari, P., Karl, T., Noe, S., Prentice, I. C.,
Serca, D., Hickler, T., Wolf, A., and Smith, B.: Process-based estimates of terrestrial
ecosystem isoprene emissions: incorporating the effects of a direct CO2-isoprene interaction,
Atmospheric Chemistry And Physics, 7, 31-53, 2007.
19
1
2
3
4
Arneth, A., Schurgers, G., Lathière, J., Duhl, T., Beerling, D. J., Hewitt, C. N., Martin, M.,
and Guenther, A.: Global terrestrial isoprene emission models: sensitivity to variability in
climate and vegetation, Atmospheric Chemistry And Physics, 11, 8037-8052, 10.5194/acp11-8037-2011, 2011.
5
6
7
Doherty, R. M., Stevenson, D. S., Collins, W. J., and Sanderson, M. G.: Influence of
convective transport on tropospheric ozone and its precursors in a chemistry-climate model,
Atmospheric Chemistry And Physics, 5, 3205-3218, 2005.
8
9
10
11
12
13
14
15
Eyring, V., Cionni, I., Bodeker, G. E., Charlton-Perez, A. J., Kinnison, D. E., Scinocca, J. F.,
Waugh, D. W., Akiyoshi, H., Bekki, S., Chipperfield, M. P., Dameris, M., Dhomse, S., Frith,
S. M., Garny, H., Gettelman, A., Kubin, A., Langematz, U., Mancini, E., Marchand, M.,
Nakamura, T., Oman, L. D., Pawson, S., Pitari, G., Plummer, D. A., Rozanov, E., Shepherd,
T. G., Shibata, K., Tian, W., Braesicke, P., Hardiman, S. C., Lamarque, J. F., Morgenstern,
O., Pyle, J. A., Smale, D., and Yamashita, Y.: Multi-model assessment of stratospheric ozone
return dates and ozone recovery in CCMVal-2 models, Atmospheric Chemistry And Physics,
10, 9451-9472, 10.5194/acp-10-9451-2010, 2010.
16
17
18
Fiore, A. M., Horowitz, L. W., Dlugokencky, E. J., and West, J. J.: Impact of meteorology
and emissions on methane trends, 1990–2004, Geophysical Research Letters, 33, L12809,
10.1029/2006GL026199, 2006.
19
20
21
22
23
24
25
26
27
28
Fiore, A. M., Dentener, F. J., Wild, J. O. F., Cuvelier, C., Schultz, M. G., Hess, P., Textor, C.,
Schulz, M., Doherty, R. M., Horowitz, L. W., MacKenzie, I. A., Sanderson, M. G., Shindell,
D. T., Stevenson, D. S., Szopa, S., van Dingenen, R., Zeng, G., Atherton, C. S., Bergmann, D.
J., Bey, I., Carmichael, G. R., Collins, W. J., Duncan, B. N., Faluvegi, G., Folberth, G. A.,
Gauss, M., Gong, S., Hauglustaine, D., Holloway, T., Isaksen, I. S. A., Jacob, D. J., Jonson, J.
E., Kaminski, J. W., Keating, T. J., Lupu, A., Marmer, E., Montanaro, V., Park, R. J., Pitari,
G., Pringle, K. J., Pyle, J. A., Schroeder, S., Vivanco, M. G., Wind, P., Wojcik, G., Wu, S.,
and Zuber, A.: Multimodel estimates of intercontinental source-receptor relationships for
ozone pollution, Journal of Geophysical Research, 114, D04301, doi:10.1029/2008JD010816,
2009.
29
30
31
Fishman, J., Watson, C. E., Larsen, J. C., and Logan, J. A.: Distribution of Tropospheric
Ozone Determined From Satellite Data, Journal of Geophysical Research, 95, 3599-3617,
10.1029/JD095iD04p03599, 1990.
32
33
34
35
Fuglestvedt, J. S., Jonson, J. E., and Isaksen, I. S. A.: Effects of reductions in stratospheric
ozone on tropospheric chemistry through changes in photolysis rates, Tellus Series B,
Chemical and Physical Meteorology, 46, 172-192, 10.1034/j.1600-0889.1992.t01-3-00001.xi1, 1994.
36
37
38
39
Guenther, A. B., Hewitt, C. N., Erickson, D., Fall, R., Geron, C., Graedel, T. E., Harley, P.,
Klinger, L., Lerdau, M. T., McKay, W. A., Pierce, T. E., Scoles, B., Steinbrecher, R.,
Tallaamraju, R., Taylor, J., and Zimmerman, P. R.: A global model of natural volatile organic
compound emissions, Journal of Geophysical Research, 100, 8873-8892, 1995.
40
41
42
Guenther, A. B., Karl, T., Harley, P., Wiedinmyer, C., Palmer, P. I., and Geron, C.: Estimates
of global terrestrial isoprene emissions using MEGAN (Model of Emissions of Gases and
Aerosols from Nature), Atmospheric Chemistry And Physics, 6, 3181-3210, 2006.
43
44
45
Hauglustaine, D. A., Lathièire, J. A., Szopa, S., and Folberth, G. A.: Future tropospheric
ozone simulated with a climate-chemistry-biosphere model, Geophysical Research Letters,
32, L24807, 10.1029/2005GL024031, 2005.
20
1
2
3
4
5
Jacob, D. J., Heikes, E. G., Fan, S. M., Logan, J. A., Mauzerall, D. L., Bradshaw, J. D., Singh,
H. B., Gregory, G. L., Talbot, R. W., Blake, D. R., and Sachse, G. W.: Origin of ozone and
NOx in the tropical troposphere: A photochemical analysis of aircraft observations over the
South Atlantic basin, Journal of Geophysical Research, 101, 24235-24250,
10.1029/96JD00336, 1996.
6
7
8
Jacobson, M. Z., and Streets, D. G.: Influence of future anthropogenic emissions on climate,
natural emissions, and air quality, Journal of Geophysical Research, 114, D08118,
10.1029/2008JD011476, 2009.
9
10
11
12
13
14
Lamarque, J.-F., Bond, T. C., Eyring, V., Granier, C., Heil, A., Klimont, Z., Lee, D., Liousse,
C., Mieville, A., Owen, B., Schultz, M. G., Shindell, D., Smith, S. J., Stehfest, E., van
Aardenne, J., Cooper, O. R., Kainuma, M., Mahowald, N., McConnell, J. R., Naik, V., Riahi,
K., and van Vuuren, D. P.: Historical (1850–2000) gridded anthropogenic and biomass
burning emissions of reactive gases and aerosols: methodology and application, Atmospheric
Chemistry And Physics, 10, 7017-7039, 10.5194/acp-10-7017-2010, 2010.
15
16
17
18
Lamarque, J.-F., Kyle, G. P., Meinshausen, M., Riahi, K., Smith, S. J., Vuuren, D. P., Conley,
A. J., and Vitt, F.: Global and regional evolution of short-lived radiatively-active gases and
aerosols in the Representative Concentration Pathways, Climatic Change, 109, 191-212,
10.1007/s10584-011-0155-0, 2011.
19
20
Lawrence, M. G., Jöckel, P., and von Kuhlmann, R.: What does the global mean OH
concentration tell us?, Atmospheric Chemistry And Physics, 1, 37-49, 2001.
21
22
23
Lawrence, M. G., von Kuhlmann, R., Salzmann, M., and Rasch, P. J.: The balance of effects
of deep convective mixing on tropospheric ozone, Geophysical Research Letters, 30, 1940,
10.1029/2003GL017644, 2003.
24
25
26
Logan, J. A.: An analysis of ozonesonde data for the troposphere: Recommendations for
testing 3-D models and development of a gridded climatology for tropospheric ozone, Journal
of Geophysical Research, 104, 16,115–-116,150, 1999.
27
28
Masson, D., and Knutti, R.: Climate model genealogy, Geophysical Research Letters, 38,
L08703, 2011.
29
30
31
Moxim, W. J., Levy II, H., and Kasibhatla, P. S.: Simulated global tropospheric PAN: Its
transport and impact on NO$_x$, Journal of Geophysical Research, 101, 12,621–612,638,
1996.
32
33
Price, C., and Rind, D. H.: A simple lightning parameterization for calculating global
lightning distributions, Journal of Geophysical Research, 97, 9919-9933, 1992.
34
35
Price, C., and Rind, D. H.: Possible implications of global climate change on global lightning
distributions and frequencies, Journal of Geophysical Research, 99, 10823-10831, 1994.
36
37
Rosenlof, K. H.: Seasonal cycle of the residual mean meridional circulation in the
stratosphere, Journal of Geophysical Research, 100, 5173-5191, 10.1029/94JD03122, 1995.
38
39
40
Russell, A., Milford, J., Bergin, M. S., McBride, S., McNair, L., Yang, Y., Stockwell, W. R.,
and Croes, B.: Urban Ozone Control and Atmospheric Reactivity of Organic Gases, Science,
269, 491-495, 10.1126/science.269.5223.491, 1995.
41
42
Schumann, U., and Huntrieser, H.: The global lightning-induced nitrogen oxides source,
Atmospheric Chemistry And Physics, 7, 3823-3907, 2007.
21
1
2
3
Stevenson, D. S., Doherty, R. M., Sanderson, M. G., Johnson, C. E., Collins, W. J., and
Derwent, R. G.: Impacts of climate change and variability on tropospheric ozone and its
precursors, Faraday Discussions, 130, 41-57, 10.1039/b417412g, 2005.
4
5
6
7
8
9
10
11
12
Stevenson, D. S., Dentener, F. J., Schultz, M. G., Ellingsen, K., van Noije, T. P. C., Wild, J.
O. F., Zeng, G., Amann, M., Atherton, M., Bell, N., Bergmann, D. J., Bey, I., Bulter, T.,
Cofala, J., Collins, W. J., Derwent, R. G., Doherty, R. M., Drevet, J., Eskes, H. J., Fiore, A.
M., Gauss, M., Hauglustaine, D. A., Horowitz, L. W., Isaksen, I. S. A., Krol, M. C.,
Lamarque, J.-F., Lawrence, M. G., Montanaro, V., Müller, J.-F., Pitari, G., Prather, M. J.,
Pyle, J. A., Rast, S., Rodriguez, J. M., Sanderson, M. G., Savage, N. H., Shindell, D. T.,
Strahan, S. E., Sudo, K., and Szopa, S.: Multimodel ensemble simulations of present-day and
near-future tropospheric ozone, Journal of Geophysical Research, 111, D08301,
10.1029/2005JD006338, 2006.
13
14
15
Stolarski, R. S., and Frith, S. M.: Search for evidence of trend slow-down in the long-term
TOMS/SBUV total ozone data record: the importance of instrument drift uncertainty,
Atmospheric Chemistry And Physics, 6, 4057-4065, 10.5194/acp-6-4057-2006, 2006.
16
17
18
Taylor, K. E., Stouffer, R. J., and Meehl, G. A.: An Overview of CMIP5 and the Experiment
Design, Bulletin of the American Meteorological Society, 111007060139000,
10.1175/BAMS-D-11-00094.1, 2011.
19
20
21
22
23
Thompson, A. M., Witte, J. C., Oltmans, S. J., Schmidlin, F. J., Logan, J. A., Fujiwara, M.,
Kirchoff, V. W. J. H., Posny, F., Coetzee, G. J. R., Hoegger, B., Kawakami, S., Ogawa, T.,
Fortuin, J. P. F., and Kelder, H. M.: Southern Hemisphere Additional Ozonesondes
(SHADOZ) 1998–-2000 tropical ozone climatology 2. Tropospheric variability and the zonal
wave-one, Journal of Geophysical Research, 108, 8241, 10.1029/2002JD002241, 2003.
24
25
26
27
Tilmes, S., Lamarque, J. F., Emmons, L. K., Conley, A., Schultz, M. G., Saunois, M.,
Thouret, V., Thompson, A. M., Oltmans, S. J., Johnson, B., and Tarasick, D.: Ozonesonde
climatology between 1995 and 2009: description, evaluation and applications, Atmospheric
Chemistry and Physics Discussions, 11, 28747-28796, 10.5194/acpd-11-28747-2011, 2011.
28
29
30
31
van Vuuren, D. P., Edmonds, J., Kainuma, M., Riahi, K., Thomson, A., Hibbard, K., Hurtt, G.
C., Kram, T., Krey, V., Lamarque, J.-F., Masui, T., Meinshausen, M., Nakicenovic, N., Smith,
S. J., and Rose, S. K.: The representative concentration pathways: an overview, Climatic
Change, 109, 5-31, 10.1007/s10584-011-0148-z, 2011.
32
33
Wild, J. O. F.: Modelling the global tropospheric ozone budget: exploring the variability in
current models, Atmospheric Chemistry And Physics, 7, 2643-2660, 2007.
34
35
Williams, E. R.: The global electrical circuit: A review, Atmospheric Research, 91, 140-152,
10.1016/j.atmosres.2008.05.018 2009.
36
37
38
Zeng, G., Pyle, J. A., and Young, P. J.: Impact of climate change on tropospheric ozone and
its global budgets, Atmospheric Chemistry And Physics, 8, 369-387, 10.5194/acp-8-3692008, 2008.
39
40
41
42
Ziemke, J. R., Chandra, S., Labow, G. J., Bhartia, P. K., Froidevaux, L., and Witte, J. C.: A
global climatology of tropospheric and stratospheric ozone derived from Aura OMI and MLS
measurements, Atmospheric Chemistry And Physics, 11, 9237-9251, 10.5194/acp-11-92372011, 2011.
22