ACPAtmospheric Chemistry and PhysicsACPAtmos. Chem. Phys.1680-7324Copernicus PublicationsGöttingen, Germany10.5194/acp-16-15619-2016Antarctic ozone depletion between 1960 and 1980 in observations and chemistry–climate model simulationsLangematzUlrikeulrike.langematz@met.fu-berlin.deSchmidtFranziskaKunzeMarkushttps://orcid.org/0000-0002-9608-1823BodekerGregory E.https://orcid.org/0000-0003-1094-5852BraesickePeterFreie Universität Berlin, Berlin, GermanyBodeker Scientific, Alexandra, New ZealandKarlsruhe Institute of Technology, Karlsruhe, GermanyUlrike Langematz (ulrike.langematz@met.fu-berlin.de)20December20161624156191562716August20165September20167December20168December2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://acp.copernicus.org/articles/16/15619/2016/acp-16-15619-2016.htmlThe full text article is available as a PDF file from https://acp.copernicus.org/articles/16/15619/2016/acp-16-15619-2016.pdf
The year 1980 has often been used as a benchmark for the return
of Antarctic ozone to conditions assumed to be unaffected by emissions of
ozone-depleting substances (ODSs), implying that anthropogenic ozone
depletion in Antarctica started around 1980. Here, the extent of
anthropogenically driven Antarctic ozone depletion prior to 1980 is examined
using output from transient chemistry–climate model (CCM) simulations from 1960
to 2000 with prescribed changes of ozone-depleting substance
concentrations in conjunction with observations. A regression model is used
to attribute CCM modelled and observed changes in Antarctic total column
ozone to halogen-driven chemistry prior to 1980. Wintertime Antarctic ozone
is strongly affected by dynamical processes that vary in amplitude from year
to year and from model to model. However, when the dynamical and chemical
impacts on ozone are separated, all models consistently show a long-term,
halogen-induced negative trend in Antarctic ozone from 1960 to 1980. The
anthropogenically driven ozone loss from 1960 to 1980 ranges between
26.4 ± 3.4 and 49.8 ± 6.2 % of the total anthropogenic ozone
depletion from 1960 to 2000. An even stronger ozone decline of 56.4 ± 6.8 %
was estimated from ozone observations. This analysis of the
observations and simulations from 17 CCMs clarifies that while the return of
Antarctic ozone to 1980 values remains a valid milestone, achieving that
milestone is not indicative of full recovery of the Antarctic ozone layer
from the effects of ODSs.
Introduction
Over the past few decades, the evolution of Antarctic stratospheric ozone
has been dominated by chemical depletion due to anthropogenic sources of
active chlorine (Cl) and bromine (Br) (e.g. WMO, 1990, and references
therein). This secular ozone change has been modulated by ozone variations
on interannual timescales caused by dynamically induced temperature
fluctuations (e.g. Huck et al., 2005; Newman et al., 2006). To attribute
changes in stratospheric ozone to depletion by halogens, equivalent
effective stratospheric chlorine (EESC) is used as an indicator of the
chemical effects of ozone-depleting substances (ODSs) (Daniel et al.,
1995). EESC is derived from measurements of ground-based halocarbon
concentrations, taking into account their transport times into the
stratosphere and their conversion into reactive chlorine (Cly) and
bromine (Bry) (e.g. Solomon and Wuebbles, 1995; Montzka and Fraser,
2003; Newman et al., 2006).
As a result of the Montreal Protocol on Substances that Deplete the Ozone
Layer and its subsequent amendments and adjustments that regulated the
production and consumption of halocarbons, EESC reached maximum
concentrations around 1996 in midlatitudes and around 2000 over polar
regions. The 4-year difference is due to longer transport times to the polar
stratosphere (see Figs. 1–22 in Carpenter and Reimann, 2014).
Since EESC is expected to continue to decline, stratospheric ozone is
expected to recover from the influence of ODSs. To provide policy-relevant
statements on expected ozone recovery, the return of ozone concentrations to
levels typical of 1980 has frequently been used as a benchmark (e.g.
Bodeker and Waugh, 2007). By comparing ozone concentrations projected by
models with their simulated 1980 values, dates of return to 1980 levels can
be identified. For Antarctica, October mean total column ozone is projected
to reach 1980 values between 2046 and 2057, about 5 years earlier than Cly
will return to 1980 values (Eyring et al., 2010; Bekki and Bodeker, 2011).
There is evidence that the return of ozone to pre-1980 values is not
equivalent to a full recovery of ozone from the effects of ODSs.
Calculations of EESC (e.g. Montzka and Fraser, 2003; Clerbaux and Cunnold,
2007) show a clear upward trend in EESC before 1980, with relatively small
increases in the 1960s followed by steeper increases in the 1970s. By 1980,
polar stratospheric EESC had reached about 50 % of its peak level around 2000,
indicating that considerable Antarctic chemical ozone loss should have
occurred before 1980 (Carpenter and Reimann, 2014).
There are indications from both ground-based total column ozone measurements
(Farman et al., 1985) and Antarctic ozonesonde stations (Solomon et al.,
2005) of ozone depletion prior to 1980, suggesting an early effect of ODSs.
Austin et al. (2010) and SPARC CCMVal (2010) found pre-1980 Antarctic ozone
depletion in CCMVal-2 simulations. Some of the CCMs simulated an increasing
Antarctic ozone mass deficit before 1980, and later return dates of Antarctic
ozone to 1960 compared to 1980. However, these studies did not provide any
detail how much of the long-term ozone loss to 2000 had already occurred
before 1980. In a more recent single-model CCM study using specified
dynamics, Shepherd et al. (2014) found that 40 % of the long-term
non-volcanic ozone loss occurred before 1980.
The purpose of this study is to quantify the extent to which Antarctic ozone
was already affected by chemical ozone depletion in the period before 1980.
The results provide a context for the utility of the return to 1980 levels
as a benchmark of the degree of recovery from the effects of ODSs. A
multiple linear regression model is used to attribute simulated ozone
changes in a set of 17 multidecadal CCM simulations to changes in EESC.
Importantly, nonlinear dependence of ozone on EESC is included in the model
so that a false positive does not arise, as would be the case if pure linear
dependence of ozone depletion on EESC had been assumed, i.e. a regression
model with only linear dependence of ozone on EESC applied to an ozone
signal that is constant from 1960 to 1980 and declines after 1980 would
suggest EESC-induced ozone loss prior to 1980. The CCM simulations were
carried out in phase 2 of the SPARC (stratosphere–troposphere processes and
their role in climate) chemistry–climate model validation (CCMVal-2)
initiative (SPARC CCMVal, 2010). These simulations are well suited for this
study for two reasons: first, all models used emissions originating from the
same prescribed scenario according to the CCMVal recommendations for REF-B1
simulations (Eyring et al., 2008). This ensures that all models simulate the
same surface mixing ratios of halocarbons (WMO, 2007) which then determine
long-term stratospheric ozone depletion. Second, the participating CCMs, due
to their different horizontal and vertical resolutions, as well as varying
degrees of implemented physical parameterizations, develop different
dynamical variability and trends, in particular during winter and spring. As
a result, the combination of identical prescribed ODS emissions and quite
different dynamical variability simulated in the CCMs provides a rich set of
simulations that can aid understanding of the observations. A detectable and
consistent EESC-induced Antarctic ozone depletion prior to 1980 throughout
the models would inform the use of 1980 as a benchmark for the definition of
polar ozone recovery from ODSs.
The application of the least-squares regression to separate the effects of
halogens and temperature fluctuations on ozone changes is described together
with the CCM output in Sect. 2. A comparison of the halogen-induced
Antarctic ozone depletion in the models between 1960 and 2000 is compared
with the derived ozone loss prior to 1980 in Sect. 3. A summary of the
results, and conclusions, follow in Sect. 4.
Method, models, and observations
To attribute CCM-simulated Antarctic total column ozone changes to changes
in stratospheric halogen loading and temperature variability, a multiple
linear regression model is fitted to the total column ozone time series as
O3=aESC2+bESC+cT′+d+ε,
where a, b, c, and d are fit coefficients derived by least-squares fitting
of the equation to modelled or measured September to November total column
ozone (O3). Note that in contrast to the real atmosphere, where EESC is
derived from surface emissions of halocarbons, we use here equivalent
stratospheric chlorine (ESC) at 50 hPa, i.e. the actual Cly and
Bry levels provided by the chemistry codes of the models, using a
60-fold efficacy for ozone destruction by bromine (Daniel et al., 1999).
T′ is the temperature anomaly at 100 hPa (with the climatological annual cycle
subtracted). A constant offset is included, leading to the fit coefficient d, and
ε is the residual. All variables are averaged over
Antarctica (60–90∘ S) and over the Southern Hemisphere late
winter/spring (September to November, SON). The regression model includes
basis functions that account for the longer-term effects of halogens on
ozone depletion (ESC2 and ESC) as well as ozone variations due to
year-to-year variations in midlatitude planetary wave activity (represented
by the corresponding Antarctic mean temperature anomalies T′; e.g. Newman
et al., 2004; Huck et al., 2005). The quadratic term (ESC2) allows
for nonlinear dependence of ozone depletion on ESC related to the catalytic
ClO destruction cycle (Jiang et al., 1996). Moreover, it ensures that the
regression model implicitly allows for constant ozone from 1960 to 1980.
Tests with regression models including additional basis functions, such as
the quasi-biennial oscillation (QBO), solar and volcanic effects, following the regression model used in
SPARC CCMVal (2010), resulted generally in larger unexplained residuals of
the ozone time series than the simple model with a quadratic EESC fit used
here (not shown). This indicates that temperature anomalies in the lower
Antarctic stratosphere explain much (if not all) of the wintertime dynamical
variability in total column ozone. Once the fit coefficients a, b, c, and
d are determined for each of the CCMs, the degree of halogen-induced
ozone depletion in any year, for each of the CCMs, can be derived using
O3=aESC2+bESC.
Monthly mean total column ozone, temperature, and Cly and Bry
concentrations from 17 CCMVal-2 simulations (see Table 1) were analysed for
the period 1960 to 2000 using the regression model. All simulations used
observed transient forcings of ODSs, as well as greenhouse gas and sea surface
temperatures/sea-ice concentrations as prescribed for the REF-B1 scenario by
the CCMVal-2 initiative (Eyring et al., 2008). Some models within the set
included some sources of natural variability such as background and volcanic
aerosol, solar variability, the QBO, and ozone
and aerosol precursors, while others did not. The inclusion or exclusion of
these factors was found to have no effect on the results of our study.
For comparison with observations, the same regression model was applied to a
1979 to 2000 database of total ozone column measurements derived by
combining measurements from multiple space-based instruments corrected for
offsets and drifts against the ground-based Dobson and Brewer
spectrophotometer networks (Bodeker et al., 2005). The data set combines
total column ozone measurements from Total Ozone Mapping Spectrometer (TOMS)
instruments, the Global Ozone Monitoring Experiment (GOME), solar
backscatter ultra-violet (SBUV) instruments, and the Ozone Monitoring
Instrument (OMI). A monthly mean, polar cap mean (60–90∘ S), total
column ozone time series was calculated from this database. For the
pre-satellite era (before 1979), total column ozone measurements from Brewer
and Dobson spectrophotometers at four Antarctic stations (Faraday
(previously Argentine Islands) at 65.3∘ S since 1957; Halley at
73.5∘ S since 1957; Showa at 69∘ S since 1961; and the South
Pole at 90∘ S since 1961) were used to estimate Antarctic mean
(60–90∘ S) total column ozone. First, the monthly mean
time series of total column ozone measurements at the Argentine Islands/Faraday
was combined with the time series of measurements from Showa to create a
single time series representative of ozone changes on the periphery of the
continent. Systematic differences between the Argentine Islands/Faraday and
Showa, arising primarily from their different locations, were accounted for
by averaging differences between temporally coincident monthly means
(Argentine Islands/Faraday 12.58 DU higher than Showa on average). Whether
Argentine Islands/Faraday time series is corrected against Showa or vice versa is
irrelevant as the combined Argentine Islands/Faraday and Showa time series
is simply used as a predictor in a regression model and is therefore
insensitive to their absolute value. Monthly means were calculated from
daily data where measurements made outside of the circumpolar vortex
(diagnosed from 550 K potential vorticity fields) were excluded from the
calculation, and were corrected in each year for temporal sampling bias. The
three resultant location-specific monthly mean time series were then used as
basis functions in a regression model which was trained on available polar
cap mean total column ozone obtained from the observational database
described above. Once trained, the regression model can be used to generate
estimates of monthly mean polar cap total column ozone from available
monthly means at the ground-based measurement sites. Different forms of the
regression model were constructed depending on which location time series
had missing data. This approach generates robust estimates in the
pre-satellite era in a way that introduces as little additional information
as possible, errs on the side of underestimating the variability rather
than overestimating the variability, avoids spatial interpolation, and
avoids the use of ancillary data such as output from CCMs. Antarctic mean
temperatures were derived from NCEP/NCAR reanalyses (Kalnay et al., 1996).
The EESC time series was taken from Newman et al. (2007a), assuming a mean
transport time of 5 years. Uncertainty values of the total column ozone and
further derived quantities have been calculated by applying the standard
formulae for error propagation to determine the uncertainties on the
regression model fit coefficients.
List of participating CCMs.
ModelGroupAMTRAC3GFDL, USACAM3.5NCAR, USACCSRNIESNIES, JapanCMAMUniv. of Toronto and York Univ., CanadaCNRM-ACMMétéo-France, FranceEMACMPI Chemistry Mainz, GermanyEMAC-FUBFreie Universität Berlin, GermanyGEOSCCMNASA/GSFC, USALMDZreproIPSL, FranceMRIMeteorological Research Institute, JapanNiwaSOCOLNIWA, New ZealandSOCOLPMOD/WRC and ETHZ, SwitzerlandULAQUniversity of L'Aquila, ItalyUMSLIMCATUniversity of Leeds, UKUMUKCA-METOMetOffice, UKUMUKCA-UCAMUniversity of Cambridge, UKWACCMNCAR, USAHalogen-induced Antarctic ozone loss
Figure 1 shows the time series of the SON average ESC at 50 hPa over
Antarctica from 1960 to 2000, simulated by the CCMs used in this study. Also
included are the EESC time series derived from measurements of halogen-containing
substances following the method of Newman et al. (2007a) for mean
transit times of 4 and 5 years. To facilitate comparison between different
CCMs, the ESC values of the individual CCMs have been adjusted to a common
baseline by subtracting their individual 1960 values and then adding the
multi-model mean value for 1960. The same adjustment was applied to the EESC
time series. All models show a slow ESC increase in the 1960s and 1970s,
followed by a steeper increase until the mid to late 1990s. Except for one
CCM, which simulates an ESC increase of 3.7 ppb between 1960 and 2000, the
majority of the models simulate an ESC increase of about 2.8 ppb from 1960
to 2000. This ensemble shows good agreement with the EESC time series that
is based on a 5-year transport time representative for polar latitudes
(e.g. Newman et al., 2007a). A few models show a flattening of Antarctic ESC
in the second half of the 1990s, and reach a smaller average ESC increase of
about 2.5 ppb. The evolution of polar ESC in these models is more similar to
the EESC time series with a 4-year transport time to middle and higher
latitudes. This indicates that these models have a faster transport of
constituents towards polar latitudes than observed. It is evident that
elevated ESC abundances appeared in the Antarctic lower stratosphere before 1980.
The simulated increase of about 0.9 ppb from 1960 to 1980 corresponds
to one-third of the increase between 1960 and 2000.
Observed ozone loss
Figure 2 shows the observed Antarctic total column ozone time series between 1960
and 2000 (black line, left panel). The observations include estimates
of polar cap total column ozone, based on the four sites listed earlier, until 1978,
and then satellite-based measurements thereafter. Year-to-year
variations in SON mean Antarctic stratospheric ozone are apparent and
fluctuate around a longer-term downward trend. The regression model fit to
the observations (Eq. 1; blue line in Fig. 2a) reproduces
this behaviour well, explaining ∼ 91 % of the variance in
the observations. The residuals are normally distributed (Fig. 2b), adding
confidence to the robustness of the regression model fit.
Evolution of Antarctic SON average ESC (in ppb) in the CCMs between 1960
and 2000, adjusted to a common baseline of 1960. Black lines show EESC (in ppb),
provided by Newman et al. (2007) for mean transit times of 4 (solid) or 5 (dashed) years.
(a) Antarctic mean, SON mean total column ozone from
observations (black line), from the full regression model (blue line), and from
the regression model including the EESC term only (red line). (b) Histogram
of the residuals.
The Antarctic SON mean total column ozone evolution attributable to changes
in EESC, as derived using Eq. (2), is shown by the red line in Fig. 2.
The regression analysis quantifies an EESC-induced Antarctic total column
ozone loss from 1960 to 1980 of 76.7 ± 3.1 DU, and from 1960 to 2000
of 136.0 ± 10.9 DU. Slightly more than half (56.4 ± 6.8 %) of
the Antarctic ozone loss caused by EESC in the period from 1960 to 2000
occurred already before 1980. While the EESC-induced ozone change leads to a
continuous downward trend of Antarctic column ozone through the 1960s and 1970s,
the ozone time series itself shows a flatter evolution over the 2
decades. This can be explained by a warming of the southern polar lower
stratosphere in that period, associated with noticeable dynamically induced
year-to-year temperature variations (Labitzke and Kunze, 2005; Newman et
al., 2007b). Antarctic winters with strong and cold polar vortices were
alternating with others that developed dynamically disturbed polar vortices.
The high total column ozone in Antarctic spring 1968 (Fig. 2a), for
example, was associated with a weak and warm polar vortex that broke down in
early spring (see Fig. 6 in Langematz and Kunze, 2006), while the low ozone
column amounts in the springs 1966 and 1969 were connected to strong and
cold vortices with late breakdown dates.
Ozone loss in CCMs
As for the observations, simulated halogen-induced ozone losses in the CCMs
were derived using the statistical model of Sect. 2. The blue line in
Fig. 3a shows, for example, the results of fitting the full
regression model (Eq. 1) to EMAC-FUB total column ozone (black line).
The regression model, which explains 96 % of the variance in the CCM-simulated
ozone signal, reproduces the slow downward trend well along with the
year-to-year variability in the simulated Antarctic ozone. For all other
CCMs (except for one outlier), similar regression results were obtained with
the regression model explaining between 93 and 99 % of the variance in the CCM signals.
By applying Eq. (2), the ESC-induced Antarctic ozone loss in the CCMs was
calculated. In EMAC-FUB, the regression analysis suggests an ESC-induced
Antarctic ozone loss from 1960 to 1980 of 39.8 ± 1.0 DU, and from 1960
to 2000 of 89.4 ± 4.2 DU (red line in Fig. 3a). A total of 44.5 ± 3.2 %
of the ozone loss caused by ESC in the period from 1960 to 2000 occurred before 1980.
(a) Total column ozone from the EMAC-FUB CCM (black line),
from the full regression model (blue line), and from the regression model
including the ESC term only (red line). (b) Histogram of the residuals.
Table 2 lists the ESC-induced SON Antarctic ozone losses for all CCMs used
in this study, together with the observational estimate. The ESC-only terms
in the regression model have been used to separately diagnose ozone losses
for the periods 1960–1980 and 1960–2000; the percentage contribution of
the 1960–1980 ozone loss to the ozone loss over the whole period 1960–2000 is
given in the rightmost column. As expected, all CCMs simulate ESC-induced
Antarctic ozone depletion in late winter and spring between 1960 and 2000,
ranging from 53.9 ± 4.1 to 182.0 ± 15.5 DU. The observed
EESC-induced ozone decline over that period amounts to 136.0 ± 10.9 DU.
Antarctic total ozone column depletion (in DU, September–October–November
average) between 1960 and 2000 due to ESC in REF-B1 CCM simulations and due to
EESC in observations, adjusted to a common baseline (1960 mean of CCMs).
Antarctic temperature in 100 hPa (in ∘C, September–October–November
average) between 1960 and 2000 in REF-B1 CCM simulations (coloured lines) and
observations (black). Temperatures have been constructed by combining the offset
and linear trend coefficients of a regression model similar to Eq. (1) applied
piecewise to two periods from 1960 to 1980 and 1980 to 2000.
Halogen-induced Antarctic ozone depletion in DU in CCM simulations and
observations for the periods 1960 to 1980 and 1960 to 2000. Uncertainties of 1σ
were derived by applying error propagation rules. The rightmost column indicates
changes from 1960 to 1980 as percentage of the changes from 1960 to 2000.
The temporal evolution of SON Antarctic mean total column ozone from 1960
to 2000 for the individual CCMs and observations is illustrated in Fig. 4.
Absolute values have been adjusted to a common baseline, i.e. the mean total
column ozone in 1960 of all CCMs. It is evident that ozone depletion by
halogens started prior to 1980. All CCMs consistently simulate an
ESC-induced decrease in SON mean Antarctic total column ozone of between
19.9 ± 1.0 and 90.7 ± 3.5 DU from 1960 to 1980. The
ESC-induced ozone loss is, however, masked in those CCMs that reproduce the
observed polar stratospheric warming between 1960 and 1980, and is enhanced
in those CCMs that simulate a cooling in that period. Figure 5 shows the
evolution of the polar cap mean SON mean temperature at 100 hPa between 1960
and 2000, fitted with piecewise linear trends from 1960 to 1980 and from 1980
to 2000. While the Antarctic lower stratosphere temperature
observations showed warming in SON from 1960 to 1980, the CCMs span a wide
range of trends, indicating different temperature trend regimes resulting
from model dynamical variability differing from what happened in reality. In
addition, the absolute temperature values differ between the CCMs by more
than 10 K, which directly affects the potential for chemical ozone loss in
the models. Figure 6 compares linear trends in Antarctic total column ozone
against polar 100 hPa temperatures in the CCMs and observations between 1960
and 1980. CCMs simulating a stronger cooling of the Antarctic lower
stratosphere, and therefore a more stable wintertime polar vortex, show a
stronger total column ozone decline than those models that produce no
cooling or even a warming.
Scatter diagram of the linear trends in the Antarctic total column
ozone (in DU yr-1) and the Antarctic temperature at 100 hPa (in K yr-1)
in September–October–November between 1960 and 1980 for the 17 CCMs of this
study and the observations.
According to the CCM simulations, ESC caused Antarctic ozone depletion (prior
to 1980) of 26.4 ± 3.4 to 49.8 ± 6.2 % of the total depletion
between 1960 and 2000. This estimate is marginally lower than the observed
value of 56.4 ± 6.8 % with only a few CCMs replicating the observed
pre-1980 Antarctic ozone depletion within the uncertainties. Most of the
ESC-induced Antarctic ozone depletion between 1960 and 1980 took place in
the second decade of this period. For the observations, the regression
yields an ozone depletion of -41.9 ± 2.7 DU between 1970 and 1980;
54.6 ± 1.3 % of the ozone loss between 1960 and 1980 took place in
the second decade of this period. With the exception of three CCMs, the models
generally show a stronger ESC-induced ozone loss between 1970 and 1980 of
the order of 60 to 75 % after 1970, in agreement with elevated ESC in the
CCMs after 1970 (Fig. 1).
Discussion and conclusion
Output from CCMVal-2 REF-B1 CCM simulations forced by a realistic transient
scenario of ODSs for the period 1960 to 2000 was used to investigate
anthropogenic ozone depletion over Antarctica prior to 1980. A regression
model was fitted to Antarctic SON vortex average total ozone columns taking
effects of ESC and temperature variations into account. The regression
results with regression coefficients varying from R2= 0.89 to
R2= 0.99 showed that Antarctic ozone levels are dominated by halogen
chemistry and dynamical effects. By evaluating only the ESC terms in the
regression model, we were able to derive ESC-induced ozone depletion for the
periods 1960–1980 and 1960–2000.
The observed decrease in total column ozone between 1960 and 2000 was
reproduced – within its uncertainty range – by seven models (CMAM, LMDZrepro,
UMSLIMCAT, UMUKCA-METO, UMUKCA-UCAM, WACCM and ULAQ). Two of these CCMs
(CMAM, WACCM) obtained the highest ranking in an evaluation of their
photochemistry and transport characteristics performed within the SPARC
CCMVal activity (SPARC CCMVal, 2010) and discussed in Sect. 2 of the
2010 WMO ozone assessment (WMO, 2011), providing confidence in the robustness of
their results. Four CCMs (AMTRAC3, CNRM-ACM, GEOSCCM, MRI) simulated a stronger
ozone decline, and six CCMs (CAM3.5, CCSRNIES, EMAC, EMAC-FUB, NIWA-SOCOL,
SOCOL) underestimated the observed ozone decline. This divergent model
behaviour may be due to the representation of polar ozone chemistry in the
models, their dynamical and transport characteristics, or to a combination
of both. Based on the detailed evaluation performed as part of the SPARC
CCMVal activity (SPARC CCMVal, 2010), we found in our study that the CCMs
that represent the observations well, generally (with one exception) show a
good potential for chlorine activation and (all) a good representation of
chemical ozone depletion in Antarctic spring. CCMs with a stronger ozone
loss than observed (cf. Table 2) partly tend to a slight overestimation of
chemical ozone depletion (AMTRAC3, GEOSCCM). For some CCMs with weaker ozone
decline between 1960 and 2000, a consistent underestimation of chemical ozone
depletion was found (CCSRNIES, EMAC, CAM3.5). Thus, the deviations of some
CCMs from the observed ozone decline can partly be explained by deficiencies
in their polar ozone chemistry. However, in addition, models that
underestimate the observed ozone decline were found to suffer from either a
transport of air into the Antarctic polar vortex that is too fast (SOCOL,
NIWA-SOCOL) or an isolation of the polar vortex from midlatitudes that is too weak
in the lower stratosphere (CAM3.5, CCSRNIES, EMAC, SOCOL, NIWA-SOCOL). Both
effects led to lower ESC concentrations by the end of the 20th century
in these models (cf. Fig. 1), and as a result, an underestimation of the
observed polar ozone decline due to ESC.
Consistent negative Antarctic ozone changes were diagnosed in the CCMs prior
to 1980 as a result of chemical depletion by ESC. This pre-1980
halogen-induced Antarctic ozone depletion amounts to values between
26.4 ± 3.4 and 49.8 ± 6.2 % of the simulated ozone depletion
between 1960 and 2000. Hence, the CCM simulations are consistent with the
observational estimate of a significant EESC-induced ozone decline in 1960–1980,
albeit nearly all CCMs underestimate the observed decline of 56.4 ± 6.8 %,
derived from the NIWA combined total column ozone database. However, note that the two CCMs, ranked highest in the SPARC CCMVal
evaluation of their photochemistry and transport characteristics, CMAM and
WACCM (SPARC CCMVal, 2010), nearly agree with the observed decline between 1960
and 1980 within its uncertainty range.
Apart from the chemical and dynamical performance of the models, their
underestimation of the ozone loss might be related to a sampling bias of the
CCMs that include high latitudes with little ozone depletion in their polar
cap mean, while the observations are made in sunlight. However, this bias
exists in September only. Moreover, the region of the polar cap that is in
perpetual darkness is very small (as a fraction of the whole area poleward
of 60∘ S) and shrinks to zero by the end of September. So, this
effect should be of the order of a few percent. Another potential reason for
the underestimation of the Antarctic ozone decline before 1980 in most CCMs
is the effect on chemical ozone depletion by short-lived bromine compounds
of natural biogenic origin, so-called very short lived substances (VSLSs).
The effects of VSLSs which contribute 20–30 % to the present-day
stratospheric bromine content (WMO, 2011) on Antarctic stratospheric ozone
were not included in the CCMVal-2 simulations. Braesicke et al. (2013) and
Sinnhuber and Meul (2014) showed that taking brominated VSLSs in their CCMs
into account leads to a significant reduction of Antarctic polar ozone. In a
transient REF-B1 simulation using the same FUB-EMAC CCM as included in this
study but with prescribed VSLS sources, Sinnhuber and Meul (2014) found a
reduction of October mean ozone in the lower Antarctic stratosphere of more
than 20 %. However, although constant VSLS emissions were prescribed over
the whole simulation period of 1960–2005, the impact of VSLSs was stronger in
the most recent period after 1980 with enhanced chlorine due to combined
bromine–chlorine catalytic ozone loss cycles. Hence, including the VSLS
effect leads to an enhancement of the 1960–2000 Antarctic ozone depletion,
but reduces the relative change in 1960–1980 compared to the full period.
Further insight is expected from the analysis of the new CCMI (Chemistry-Climate Model Initiative) simulations
that will include the effects of VSLSs.
Our results show that CCM-modelled declines in Antarctic polar cap average
total column ozone from 1960 to 1980 are not intrinsically in disagreement
with observations which show little change in polar cap average ozone over
this period. The apparent discrepancy results from the particular instance
of reality in which Antarctic stratospheric temperatures increased over the
period 1960 to 1980 that significantly offset the EESC-induced depletion of
ozone. In the CCM simulations in which stratospheric warming occurs from 1960
to 1980, similar to observations, no statistically significant changes
in ozone prior to 1980 are observed. These results reiterate that while the
return of Antarctic ozone to 1980 levels remains a valid milestone on the
path to recovery, attaining this milestone cannot be indicative of the full
recovery of Antarctic ozone from the effects of ODSs since appreciable
ODS-induced ozone depletion occurred prior to 1980.
Data availability
The CCM data used in this study are available from the CCMVal-II database at
the British Atmospheric Data Centre (British Atmospheric Data Centre, 2009)
(http://browse.ceda.ac.uk/browse/badc/ccmval/data/CCMVal-2/Reference_Runs/REF-B1/, login required).
The NIWA combined total column ozone database can be obtained from http://www.bodekerscientific.com/data/the-bdbp.
NCEP/NCAR reanalyses are available from http://rda.ucar.edu/datasets/ds090.0/#description.
Acknowledgements
We acknowledge the modelling groups for making
their simulations available for this analysis, the chemistry–climate model
validation (CCMVal) activity of WCRP's (World Climate Research Programme)
SPARC (stratosphere–troposphere processes and their role in climate) project
for organizing and coordinating the model data analysis activity, and the
British Atmospheric Data Centre (BADC) for collecting and archiving the
CCMVal model output. We thank Ted Shepherd for his thoughts on the
methodology underlying this paper. We thank Christian Blume for input to
Fig. 1. U. Langematz was supported by the International Bureau of the Federal
Ministry of Education and Research.
Edited by: M. Dameris
Reviewed by: two anonymous referees
ReferencesAustin, J., Struthers, H., Scinocca, J., Plummer, D. A., Akiyoshi, H., Baumgaertner,
A. J. G., Bekki, S., Bodeker, G. E., Braesicke, P., Brühl, C., Butchart, N.,
Chipperfield, M. P., Cugnet, D., Dameris, M., Dhomse, S., Frith, S., Garny, H.,
Gettelman, A., Hardiman, S. C., Jöckel, P., Kinnison, D., Kubin, A., Lamarque,
J. F., Langematz, U., Mancini, E., Marchand, M., Michou, M., Morgenstern, O.,
Nakamura, T., Nielsen, J. E., Pitari, G., Pyle, J., Rozanov, E., Shepherd, T. G.,
Shibata, K., Smale, D., Teyssèdre, H., and Yamashita, Y.: Chemistry-climate
model simulations of spring Antarctic ozone, J. Geophys. Res., 115, D00M11,
10.1029/2009JD013577, 2010.
Bekki, S. and Bodeker, G. E. (Coordinating Lead Authors), Bais, A. F., Butchart,
N., Eyring, V., Fahey, D. W., Kinnison, D. E., Langematz, U., Mayer, B., Portmann,
R. W., Rozanov, E., Braesicke, P., Charlton-Perez, A. J., Chubarova, N. E.,
Cionni, I., Diaz, S. B., Gillett, N. P., Giorgetta, M. A., Komala, N., Lefèvre,
F., McLandress, C., Perlwitz, J., Peter, T., and Shibata, K: Future ozone and
its impact on surface UV, in: chap. 3 in Scientific Assessment of Ozone
Depletion: 2010, Global Ozone Research and Monitoring Project – Report No. 52,
World Meteorological Organization, Geneva, Switzerland, 516 pp., 2011.
Bodeker, G. E., and Waugh, D. W. (Lead Authors), Akiyoshi, H., Braesicke, P.,
Eyring, V., Fahey, D. W., Manzini, E., Newchurch, M. J., Portmann, R. W., Robock,
A., Shine, K. P., Steinbrecht, W., and Weatherhead, E. C.: The Ozone Layer in
the 21st Century, in: chap. 6 in Scientific Assessment of Ozone Depletion: 2006,
Global Ozone Research and Monitoring Project – Report No. 50, World Meteorological
Organization, Geneva, Switzerland, 2007.Bodeker, G. E., Shiona, H., and Eskes, H.: Indicators of Antarctic ozone depletion,
Atmos. Chem. Phys., 5, 2603–2615, 10.5194/acp-5-2603-2005, 2005.Braesicke, P., Keeble, J., Yang, X., Stiller, G., Kellmann, S., Abraham, N.
L., Archibald, A., Telford, P., and Pyle, J. A.: Circulation anomalies in the
Southern Hemisphere and ozone changes, Atmos. Chem. Phys., 13, 10677–10688,
10.5194/acp-13-10677-2013, 2013.
Carpenter, L. J. and Reimann, S. (Lead Authors), Burkholder, J. B., Clerbaux, C.,
Hall, B. D., Hossaini, R., Laube, J. C., and Yvon-Lewis, S. A.: Ozone-Depleting
Substances (ODSs) and Other Gases of Interest to the Montreal Protocol, in:
chap 1 in Scientific Assessment of Ozone Depletion: 2014, Global Ozone Research
and Monitoring Project – Report No. 55, World Meteorological Organization,
Geneva, Switzerland, 2014.
Clerbaux, C. and Cunnold, D.: Long-lived compounds, in: Scientific assessment
of ozone depletion: 2006, chap. 1, WMO, Geneva, Switzerland, 2007.
Daniel, J. S., Solomon, S., and Albritton, D. L.: On the evaluation of halocarbon
radiative forcing and global warming potentials, J. Geophys. Res., 100, 1271–1285, 1995.
Daniel, J. S., Solomon, S., Portman, R. W., and Garcia, R. R.: Stratospheric
ozone destruction: The importance of bromine relative to chlorine, J. Geophys.
Res., 104, 23871–23880, 1999.Eyring, V., Butchart, N., Waugh, D., Akiyoshi, H., Austin, J., Bekki, S., Bodeker,
G., Boville, B., Brühl, C., Chipperfield, M. P., Cordero, E., Dameris, M.,
Deushi, M., Fioletov, V. E., Frith, S., Garcia, R., Gettelman, A., Giorgetta, M.,
Grewe, V., Jourdain, L., Kinnison, D., Mancini, E., Manzini, E., March, M.,
Marsh, D., Nagashima, T., Newman, P. A., Nielsen, J., Pawson, S., Pitari, G.,
Plummer, D., Rozanov, E., Schraner, M., Shepherd, T. G., Shibata, K., Stolarski,
R., Struthers, H., Tian, W., and Yoshiki, M.: Assessment of temperature, trace
species and ozone in chemistry-climate model simulations of the recent past,
J. Geophys. Res., 111, D22308, 10.1029/2006JD007327, 2006.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, Atmos. Chem. Phys., 10,
9451–9472, 10.5194/acp-10-9451-2010, 2010.
Eyring, V., Chipperfield, M. P., Giorgetta, M. A., Kinnison, D. E., Manzini, E.,
Matthes, K., Newman, P. A., Pawson, S. Shepherd, T. G., and Waugh, D. W.:
Overview of the New CCMVal Reference and Sensitivity Simulations in Support of
Upcoming Ozone and Climate Assessments and the Planned SPARC CCMVal, SPARC
Newsletter No. 30, SPARC IPO, Toronto, Canada, 20–26, 2008.Farman, J., Gardiner, B., and Shanklin, J.: Large losses of total ozone in
Antarctica reveal seasonal ClOx/NOx interaction, Nature, 315, 207–210, 1985.Huck, P. E., McDonald, A. J., Bodeker, G. E., and Struthers, H.: Interannual
variability in Antarctic ozone depletion controlled by planetary waves and
polar temperature, Geophys. Res. Lett., 32, L13819, 10.1029/2005GL022943, 2005.
Jiang, Y. B., Yung, Y. L., and Zurek, R. W.: Decadal evolution of the Antarctic
ozone hole, J. Geophys. Res., 101, 8985–8999, 1996.
Kalnay, E., Kanamitsu, M., Kistler, R., Collin, W., Deaven, D., Gandin, L.,
Iredell, M., Saha, S., White, G., Woollen, J., Zhu, Y., Leetmaa, A., Reynolds,
B., Chelliah, M., Ebisuzaki, W., Higgins, W., Janowiak, J., Mo, K. C., Ropelewski,
C., Wang, J., Jenne, R., and Joseph, D.: The NCEP/NCAR 40-year reanalysis project,
B. Am. Meteorol. Soc., 77, 437–471, 1996.
Labitzke, K. and Kunze, M.: Stratospheric temperatures over the Arctic: comparison
of three data sets, Meteorol. Z., 14, 65–74, 2005.Langematz, U. and Kunze, M.: An update on dynamical changes in the Arctic and
Antarctic stratospheric polar vortices, Clim. Dynam., 27, 647–660, 10.1007/s00382-006-0156-2, 2006.
Montzka, S. A., and Fraser, P. J. (Lead Authors), Butler, J. H., Cunnold, D.,
Daniel, J., Derwent, D., Connell, P., Lal, S., McCulloch, A., Oram, D., Reeves,
C., Sanhueza, E., Steele, P., Velders, G. J. M., Weiss, R. F., and Zander, R.:
Controlled substances and other source gases, in: chap. 1 in Scientific Assessment
of Ozone Depletion: 2002, Global Ozone Research and Monitoring Project – Report
No. 47, World Meteorological Organization, Geneva, Switzerland, 2003.Newman, P. A., Kawa, S. R., and Nash, E. R.: On the size of the Antarctic
ozone hole, Geophys. Res. Lett., 31, L21104, 10.1029/2004GL020596, 2004.Newman, P. A., Nash, E. R., Kawa, S. R., Montzka, S. A., and Schauffler, S. M.:
When will the Antarctic ozone hole recover?, Geophys. Res. Lett., 33, L12814,
10.1029/2005GL025232, 2006.Newman, P. A., Daniel, J. S., Waugh, D. W., and Nash, E. R.: A new formulation
of equivalent effective stratospheric chlorine (EESC), Atmos. Chem. Phys., 7,
4537–4552, 10.5194/acp-7-4537-2007, 2007a.
Newman, P. A., and Rex, M. (Lead Authors), Canziani, J. P. O., Carslaw, K. S.,
Drdla, K., Godin-Bekmann, S., Golden, D. M., Jackman, C. H., Kreher, K., Langematz,
U., Müller, R., Nakane, H., Orsolini, Y. J., Salawitch, R. J., Santee, M. L.,
von Hobe, M., and Yoden, S.: Polar Ozone: Past and Present, in: chap. 4 in
Scientific Assessment of Ozone Depletion: 2006, Global Ozone Research and
Monitoring Project – Report No. 50, World Meteorological Organization, Geneva, Switzerland, 2007b.Shepherd, T. G., Plummer, D. A., Scinocca, J. F., Hegglin, M. I., Fioletov, V.
E., Reader, M. C., Remsberg, E., von Clarmann, T., and Wang, H. J.: Reconciliation
of halogen-induced ozone loss with the total-column ozone record, Nat. Geosci.,
7, 443–449, 10.1038/ngeo2155, 2014.Sinnhuber, B. and Meul, S.: Simulating the impact of emissions of brominated
very short lived substances on past stratospheric ozone trends, Geophys. Res.
Lett., 42, 2449–2456, 10.1002/2014GL062975, 2014.
Solomon, S. and Wuebbles, D. (Lead Authors), Isaksen, I., Kiehl, J., Simon, P.,
and Sze, N.-D.: Ozone depletion potentials, global warming potentials, and future
chlorine/bromine loading, in: chap. 13 in Scientific Assessment of Ozone
Depletion: 1994, Global Ozone Research and Monitoring Project – Report No. 37,
World Meteorological Organization, Geneva, Switzerland, 1995.Solomon, S., Portmann, R. W., Sasaki, T., Hofmann, D. J., and Thompson, D. W. J.:
Four decades of ozonesonde measurements over Antarctica, J. Geophys. Res., 110,
D21311, 10.1029/2005JD005917, 2005.
SPARC CCMVal: SPARC Report on the Evaluation of Chemistry-Climate Models,
edited by: Eyring, V., Shepherd, T. G., and Waugh, D. W., SPARC Report – No. 5,
WCRP-132, WMO/TD-No. 1526, SPARC IPO, Toronto, Canada, 2010.
WMO: Scientific Assessment of Ozone Depletion: 1989, World Meteorological
Organization Global Ozone Research and Monitoring Project, Report No. 50,
Geneva, Switzerland, 1990.
WMO: Scientific Assessment of Ozone Depletion: 2006, World Meteorological
Organization Global Ozone Research and Monitoring Project, Report No. 50,
Geneva, Switzerland, 2007.
WMO: Scientific Assessment of Ozone Depletion: 2010, World Meteorological
Organization Global Ozone Research and Monitoring Project, Report No. 52,
Geneva, Switzerland, 2011.