Access to this full-text is provided by Springer Nature.
Content available from Nature Communications
This content is subject to copyright. Terms and conditions apply.
ARTICLE
Delay in recovery of the Antarctic ozone hole
from unexpected CFC-11 emissions
S.S. Dhomse 1,2, W. Feng 1,3, S.A. Montzka 4, R. Hossaini 5, J. Keeble 6,7, J.A. Pyle 6,7, J.S. Daniel4&
M.P. Chipperfield 1,2*
The Antarctic ozone hole is decreasing in size but this recovery will be affected by atmo-
spheric variability and any unexpected changes in chlorinated source gas emissions. Here,
using model simulations, we show that the ozone hole will largely cease to occur by 2065
given compliance with the Montreal Protocol. If the unusual meteorology of 2002 is repe-
ated, an ozone-hole-free-year could occur as soon as the early 2020s by some metrics. The
recently discovered increase in CFC-11 emissions of ~ 13 Gg yr−1may delay recovery. So far
the impact on ozone is small, but if these emissions indicate production for foam use much
more CFC-11 may be leaked in the future. Assuming such production over 10 years, dis-
appearance of the ozone hole will be delayed by a few years, although there are significant
uncertainties. Continued, substantial future CFC-11 emissions of 67 Gg yr−1would delay
Antarctic ozone recovery by well over a decade.
https://doi.org/10.1038/s41467-019-13717-x OPEN
1School of Earth and Environment, University of Leeds, Leeds LS2 9JT, UK. 2National Centre for Earth Observation (NCEO), University of Leeds, Leeds LS2
9JT, UK. 3National Centre for Atmospheric Science (NCAS), University of Leeds, Leeds LS2 9JT, UK. 4Earth System Research Laboratory, Global Monitoring
Division, National Oceanic and Atmospheric Administration (NOAA), Boulder, USA. 5Lancaster Environment Centre, Lancaster University, Lancaster, UK.
6Department of Chemistry, University of Cambridge, Cambridge CB2 1EW, UK. 7National Centre for Atmospheric Science (NCAS), University of Cambridge,
Cambridge CB2 1EW, UK. *email: M.Chipperfield@leeds.ac.uk
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 1
1234567890():,;
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Depletion of the stratospheric ozone layer by chlorine and
bromine species has been a major environmental issue
since the early 1970s1,2. Following controls on the pro-
duction of the long-lived halocarbons that transport chlorine and
bromine to the stratosphere, atmospheric concentrations of most
of them are now decreasing3, and the ozone layer is expected to
recover over the course of this century4. Decreases in the stra-
tospheric loading of chlorine and bromine have been observed5,6,
and there are signs of a consequent increase in ozone in the upper
stratosphere and the total column7,8.
The most significant signal of anthropogenic ozone depletion
occurs in the Antarctic in spring—the so-called Antarctic ozone
hole. The size of the hole is typically quantified using a range of
metrics including minimum column ozone, area contained within
a particular value of column ozone and ozone mass deficit4. From
these metrics, it is clear that the hole has also stopped increasing
in size and there are signs of recovery9–12. However, even with
full compliance with the Montreal Protocol, it is expected that the
Antarctic ozone hole will persist for many decades into the future.
The most recent comprehensive assessment of ozone return
dates was performed by Dhomse et al.13, using results from the
chemistry–climate modelling initiative (CCMI) as input to the
2018 WMO Assessment4. They used results from 20 coupled
chemistry–climate models (CCMs) to make a best estimate of the
dates at which future ozone levels would return to their 1980
values for polar, mid-latitude and tropical regions. Based on this
they reported that October column ozone in Antarctic
(60oS–90oS) would return to 1980 values in 2060 (with a 1σ
uncertainty of 2055–2066) and for March in the Arctic
(60oN–90oN) by 2034 (2025–2043). The much earlier return date
in the Arctic is due to the smaller depletion and large variability,
in conjunction with climate change, which was estimated to have
only a small (2-year) effect on the much larger Antarctic loss.
This metric of return to a 1980 value is a straightforward concept
but it is difficult to estimate and needs to be interpreted with
caution. In order to derive values from CCMs, Dhomse et al. (and
similar earlier studies14) employed significant smoothing and
averaging of the modelled ozone. As a result, the influence of
interannual dynamical variability is not included in the uncer-
tainty ranges given above. In addition, in regions where ozone
values return asymptotically to the 1980 reference value, a small
change in ozone can lead to a large change in return date. Indeed,
until and unless ozone values return to the 1980 values then
recovery could be deemed not to have happened.
In addition to uncertainties in how to quantify recovery of the
ozone layer, a number of factors pose a threat to its expected
timescale. Montzka et al.15 recently reported that since the mid-
2000s, atmospheric CFC-11 has not been declining as expected,
and this discrepancy became particularly striking after 2010, the
year production of CFC-11 was reportedly phased out. Their
results suggest new emissions16 that linked to unreported pro-
duction have occurred in recent years. Based on current atmo-
spheric abundances, CFC-11 still contributes about one-quarter15
of anthropogenic chlorine (about one-fifth of all chlorine3)
reaching the stratosphere, so these results could have significant
implications for recovery of the ozone layer. Moreover, because
nearly all produced CFC-11 eventually escapes to the atmosphere,
the impact of this apparent renewed use of CFC-11 on strato-
spheric ozone ultimately depends on the total amount of new,
unreported production, which is currently unknown. If the
detected unexpected emissions arise from CFC-11 produced for
an emissive use, no large increases in CFC-11 production or
banks would be implied, and one would expect the emissions to
diminish rapidly if use were terminated. However, because past
CFC-11 use was primarily for blowing closed-cell insulating foam
that retained most of the CFC17, the detected magnitude of the
new, unexpected emissions could imply much larger CFC-11
production quantities and therefore large future emissions.
Recent studies have also suggested that increasing atmospheric
emissions of very short-lived substance species (VSLS), which are
not controlled by the Montreal Protocol, may cause a delay to
ozone layer recovery. Hossaini et al.18 reported increasing emis-
sions of dichloromethane (CH
2
Cl
2
) based on atmospheric
observations from the NOAA global surface network. Using a
CCM to investigate different assumptions of the future evolution
of CH
2
Cl
2
, they estimated a delay in the return of Antarctic ozone
to 1980 levels of 5 years for constant future CH
2
Cl
2
concentra-
tions, compared with zero atmospheric CH
2
Cl
2
, and up to 30
years for an extreme sensitivity study of constantly increasing
CH
2
Cl
2
concentrations. However, those timescales for ozone
recovery need to be interpreted with caution as the likelihood for
these scenarios to be realised is unknown and the timescales
depend on the slow convergence of ozone to a reference recovery
baseline. Recently, Fang et al.19 reported increases in the atmo-
spheric abundance of the VSLS chloroform (CHCl
3
). They also
used the model results of Hossaini et al. to estimate the impact of
sustained CHCl
3
growth on ozone recovery and thereby derived
significant delays in 1980 return dates, but these values will have
the same caveat of a small change in ozone, causing a large dif-
ference in return date, as discussed above. These results need to
be reassessed for the impact of realistic amounts of chlorinated
VSLS in the context of other changes to chlorine source gases and
for a wider range of ozone-hole recovery metrics.
Therefore, although ozone recovery is underway, there is
uncertainty in how it will progress in the future and how it should
be reported. The date for the atmosphere to return to a specified
state does not take account of variability in that pathway or on
the impact of other transient factors before the final return date.
For the Antarctic ozone hole in particular, there are other mea-
sures of its size, which may give a different perspective on
recovery.
In this paper, we use a detailed atmospheric chemical transport
model (CTM) to investigate the impact of meteorological varia-
bility and non-compliant or uncontrolled chlorine source gas
emissions on polar ozone recovery. We quantify the persistence
of the Antarctic ozone hole by a range of different metrics, which
we show need to be interpreted carefully. We are interested in
how long these metrics suggest an ozone hole will occur and how
soon, given meteorological variability, we may experience a year
without a hole. We also investigate how much longer the Arctic
may be susceptible to large ozone loss based on recent extreme
meteorology observed in the year 2011 (with extensive occurrence
of low temperatures that are conducive to ozone loss). For both
cases, we explore the impact on this recovery of the unexpected
increase in emissions of CFC-11, under different scenario
assumptions. We compare this with the impact of uncontrolled
chlorinated VSLS that also enhance polar ozone loss through the
increase in stratospheric chlorine. Given the large uncertainties in
the CFC-11 emissions related to their source, their main appli-
cation and future trends, our aim is to test example scenarios and
quantify how the impact on ozone varies with the timing and
magnitude of the emissions. Our 3D model results can therefore
be scaled to assess the impact on column ozone of other total
CFC-11 emissions.
Results
Chlorine scenarios. Figures 1and 2show estimates of CFC-11
emissions and the corresponding mean global atmospheric mix-
ing ratios from a box model (see the ‘Methods' section). The 2018
WMO Assessment20 baseline mixing ratio scenario uses a com-
bined atmospheric observation record up to the beginning of
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
2NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
2017 and then future projections. We use this mixing ratio sce-
nario to infer emissions for past years, and obtain good agree-
ment with other estimates15. Considering future mixing ratios,
the simple WMO scenario with constant future emissions of
CFC-11 of 67 Gg yr−1(the average calculated top–down emis-
sions over 2002–2016, scenario S_CFC11_67) produces a much
slower decrease in CFC-11 than the baseline scenario, with the
global mixing ratio dropping only to 170 ppt in 2080. We have
also constructed an additional CFC-11 emission scenario for the
recent past based on a constant release fraction from the bank
since 200215, which is typically assumed in the creation of future
scenarios20 (S_NoIE; no increased emissions). This scenario
indicates the path that would have been expected without these
post-2010 emissions from unreported production and also earlier
emission changes from 2002, which caused a stabilisation of
emissions (Fig. 1a). Therefore, it represents an upper limit of the
impact of the recent emission changes. Note that with the box
model result, we implicitly assume that the recent variations in
the CFC-11 decay rate are all due to emissions. Montzka et al.15
noted that some of the observed variations could be due to
influences of atmospheric dynamics, which would imply a smaller
emission increase than we have assumed.
To construct alternative future scenarios, we begin with the
estimate of new emissions due to unreported production of 13 Gg
yr−1(based on ref. 15). We relate this to production by using an
estimate of the ratio between rapid initial emission and
accumulation in the bank. We then assume a yearly fractional
release (leakage) rate for this bank. Finally, we assume a timescale
for future rampdown of the unreported production (see
Methods). For these 4 parameters, we perform sensitivity tests
with the box model to investigate the impact on CFC-11 (Figs. 1
and 2). The accumulated CFC-11 emissions scale directly with the
key parameters of initial emissions, emission ratio and timescale
for rampdown (Supplementary Fig. 1). For the ranges assumed
for these values the impact on the return dates is around ±1 year
(see Supplementary Results 2). The results are relatively
insensitive to the future fractional release as the CFC-11 is
eventually emitted to the atmosphere in any case. Given the
ongoing international concern about this issue, we anticipate that
the unreported production (and associated initial emissions) will
likely stop in the near future, so the main uncertainty for ozone
recovery would be the cumulative magnitude of post-2010
unreported production and how it was used (i.e. the emission
ratio). The scenario (S_CFC11_B) that assumes production for
the non-emissive (foam) use that is phased out over 10 years
maintains the peak in emissions of 80 Gg yr−1for a short period
and an additional 15 ppt of CFC-11 (45-ppt Cl) in mid–late
century (Fig. 2).
Antarctic ozone hole. Different metrics are used to assess the size
of the Antarctic ozone hole from observations and model
experiments, including September and October mean column
ozone from 60oSto90
oS; ozone-hole area; ozone mass deficit;
minimum total column ozone4. For all metrics, the observations
show the increasing size of the hole from the time of the first data
point shown in 1980 until the early 2000s (Fig. 3). Subsequently
the metrics show a peak in the size of the hole followed by sug-
gestions of a decrease from around the mid-2000s. Figure 3also
shows the results from a range of model simulations (Table 1).
The control model run CNTL, with time-dependent meteorology,
agrees well with the observations for the various metrics, showing
Year
Sensitivity to fractional release
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
100
80
60
40
20
0
2000 2020 2040 2060 2080
2000 2020 2040 2060 2080
2000 2020 2040 2060 2080
Year
2000 2020 2040 2060 2080
Emissions (Gg/yr) Emissions (Gg/yr)
×3 – ×9
3.5 +/– 1 %/year
5–20 years
Sensitivity to rampdown
Sensitivity to emission ratio
Sensitivity to initial emissions
13 +/– 5 Gg/yr WMO (2018)
WMO 67 Gg/yr
WMO 67 Gg/yr
WMO 67 Gg/yr
WMO (2018)
WMO (2018)
WMO (2018)
Montzka et al (2018)
WMO 67 Gg/yr (S_CFC11_67)
Foam (S_CFC11_B)
No increased emissions (S_NoIE)
b
d
a
c
Fig. 1 Past and potential future emissions of CFC-11. Estimated emissions of CFC-11 for the past derived from atmospheric measurements and for future
scenarios with different assumptions. aFive scenarios including the World Meteorological Organisation (WMO) (2018)20 baseline scenario (solid black
line) and an assumption of constant 67 Gg yr−1emissions (dashed line, S_CFC11_67). Also shown is a scenario with decreasing emissions from an estimate
of the 2002 bank (dotted line, S_NoIE). Emissions based on box model simulations for unreported CFC-11 production for foam use are shown in green
(solid line, S_CFC11_B). The green shading indicates the sensitivity range for S_CFC11_B for initial emissions ranging from 8 to 18 Gg/yr (see Methods for
assumptions). bSimilar to panel a, for WMO (2018) baseline and S_CFC11_67 scenarios with green shading showing sensitivity of scenario S_CFC11_B for
fractional release ranging from 2.5 to 4.5%/year. cSimilar to panel b, and showing sensitivity to emission ratios from ×3 to ×9. dSimilar to panel b, and
showing sensitivity to rampdown of new production between 5 and 20 years.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x ARTICLE
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 3
Content courtesy of Springer Nature, terms of use apply. Rights reserved
that the CTM gives a good quantitative simulation of polar ozone
loss, which is driven by chlorine and bromine chemistry. Run
R2000 is similar to run CNTL but uses repeating 2000 meteor-
ology and can be extended into the future to investigate the
impact of decreasing halogens. This run shows the characteristic
signal of ozone recovery with the dates for the different metrics to
return to 1980 values ranging from 2063 to 2067 (Table 2). Run
fODS, with fixed surface mixing ratios of ozone-depleting sub-
stances (ODS), provides a baseline for ozone changes due solely
to changes in N
2
O, CH
4
and, in the past, dynamics and aerosol
(see Methods). In the future, run fODS also uses repeating 2000
meteorology. The net chemical effect of increasing N
2
O and CH
4
causes a gradual decrease in the mean September/October col-
umn ozone values, which, in itself, extends the return dates of
these metrics compared with the direct effect of halogen
decreases21,22.
Results from runs R2002, R2009 and R2010 illustrate the
impact of different meteorology on ozone recovery. In reality,
interannual variability will change the meteorology from year to
year and give rise to a variable signal in, for example, column
ozone, seen in runs CNTL and fODS from 1980 to 2016. The
background pink line in Fig. 3from 2018 onwards shows results
from the extension of run CNTL with 20-year repeating
meteorology. In September, the meteorology for 2002 stands
out as extreme (Fig. 3a), while October shows a wider range of
meteorology (Fig. 3b). Keeble et al.23 used a 7-member ensemble
of CCM integrations to investigate the range of meteorologically
driven annually averaged Antarctic recovery dates. They found
the earliest recovery of annual mean ozone to values above those
of 1980 at around 2040 and final recovery (after which date ozone
values were always above the 1980 value) in 2060, with an
ensemble spread as large as about 15 years. Further analysis of the
CCM data gives final recovery dates for October monthly mean
column ozone values, averaged from 90oSto60
oS, in the late
2070s, with an ensemble spread ~20 years. This range in return
dates is similar to the CTM behaviour for October when run with
different meteorologies (Fig. 3b). CTMs that repeat the
meteorology of a particular year can complement CCMs by
giving a clear signal of the impact of the different meteorologies
on the ozone hole as chlorine and bromine levels decline. The
disturbed meteorology of run R2002 clearly leads to the smallest
ozone hole by all metrics and the earliest returns to 1980 values:
this is as early as 2021 for September column ozone, 2031 for
October mean column ozone, 2033 for the ozone mass deficit,
2035 for minimum column ozone and 2041 for the ozone-hole
area (Table 2). Clearly different metrics can give rise to different
return dates due to the different timing of low temperatures
(Supplementary Figure 2), and the timing of vortex split and
reformation during that winter24,25. Runs R2009 and R2010 give
ozone return dates that range from 2060–2066 to 2052–2085,
respectively. The range of values is much larger for the 2010
meteorology due to the shift of the ozone-hole timing to later in
the spring, causing smaller loss in September and larger ozone
loss in October (see Supplementary Fig. 2). These results show
that using the October return date in assessment studies13 is not
such a clear measure of recovery due to the large interannual
variations in vortex conditions during that month11. Further-
more, the free-running climate models may capture this
variability to different extents, increasing the uncertainty of the
multi-model mean return date.
Year
Mixing ratio (ppt) Mixing ratio (ppt)
Sensitivity to fractional release
Sensitivity to rampdownSensitivity to emission ratio
Sensitivity to initial emissions
WMO (2018)
WMO 67 Gg/yr (S_CFC11_67)
Foam (S_CFC11_B)
No increased emissions (S_NoIE)
WMO (2018)
WMO 67 Gg/yr (S_CFC11_67)
Foam (S_CFC11_B)
WMO (2018)
WMO 67 Gg/yr (S_CFC11_67)
Foam (S_CFC11_B)
WMO (2018)
WMO 67 Gg/yr (S_CFC11_67)
Foam (S_CFC11_B)
3.5 +/– 1 %/year
×3 – ×9 5–20 years
13 +/– 5 Gg/yr
250
200
150
100
250
200
150
100
250
200
150
100
250
200
150
100
1980 2000 2020 2040 20601980 2000 2020 2040 2060
1980 2000 2020 2040 2060
Year
1980 2000 2020 2040 2060 20802080
2080
2080
b
d
a
c
Fig. 2 Past and potential future concentrations of CFC-11. As Fig. 1but for CFC-11 volume-mixing ratio (ppt). aFour scenarios including the World
Meteorological Organisation (WMO) (2018)20 baseline scenario (solid black line) and an assumption of constant 67 Gg yr−1emissions (dashed line,
S_CFC11_67). Also shown is a scenario with decreasing emissions from an estimate of the 2002 bank (dotted line, S_NoIE). Emissions based on box model
simulations for unreported CFC-11 production for foam use are shown in green (solid line, S_CFC11_B). The green shading indicates the sensitivity range for
S_CFC11_B for initial emissions ranging from 8 to 18 Gg/yr (see Methods for assumptions). bSimilar to panel a, for WMO (2018) baseline and
S_CFC11_67 scenarios with green shading showing sensitivity of scenario S_CFC11_B for fractional release ranging from 2.5 to 4.5%/year. cSimilar to panel
b, and showing sensitivity to emission ratios from ×3 to ×9. dSimilar to panel b, and showing sensitivity to rampdown of new production between 5 and
20 years.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
4NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Arctic ozone depletion. Springtime ozone depletion in the Arctic
is smaller with larger interannual variability compared with the
Antarctic26,27. Due to this large interannual dynamical variability,
it is difficult to determine robust trends in Arctic ozone depletion
and recovery, and unlike the Antarctic, there are few relevant
metrics with which to assess this. Of interest in the Arctic is
predicting how long the region may be susceptible to large che-
mical ozone depletion as chlorine and bromine levels decline.
CCMs predict13 that mean March Arctic ozone levels will return
to 1980 values around 2034 (Fig. 4b), due to the large impact of
dynamics, although they may not capture years of large ozone
depletion under extreme Arctic meteorology. In recent decades,
the year with the conditions most conducive to large chemical
ozone loss was winter 2010/2011, when very low temperatures
gave rise to large ozone depletion28. Figure 4shows the mean
Arctic column ozone in February and March for CTM simula-
tions with a range of meteorology and chlorine scenarios.
Simulation R2000 shows that Arctic chemical ozone loss follows
the time variation of chlorine and bromine with a return to 1980
values only around 2080, i.e. much later than the CCMs. For
2010/2011 meteorology, the model predicts only a small increase
in mean March column ozone from around 350 DU in March
2010 to around 400 DU by 2100. This is still considerably below
the March 1980 baseline of around 450 DU. Hence, whenever
years with extremely cold stratospheric conditions occur in this
century, the Arctic would be susceptible to ozone depletion,
driven by both dynamics and chemical loss29.
CFC-11 emissions from unreported production.Figure3also
includes the results of simulations that consider different CFC-11
scenarios. The results of run R2000_NoIE, which includes neither
the impact of additional inferred post-2010 emissions15,16,northe
impact of stabilisation of emissions from 2002 to 2010, are very
similar to run R2000 (lines essentially overlapping). This shows that
the impact of the additional unexpected emissions to date has likely
CNTL R2010
R2010_CFC11_B
R2002
R2009
fODS
NASA
R2000
R2000_NoIE
R2000_CFC11_B
R2000_CFC11_67
R2000_NoVSLS
R2000_CFC12
Total column ozone [DU]Total column ozone [DU]
Total column ozone [DU]
Mass deficit [million tonnse]
Ozone mass deficit [21Sep–13Oct]Ozone hole area [07Sep–13Oct]
TCO minimum [21Sep–16Oct]
Antarctic [90S-60S] Sep Antarctic [90S-60S] Oct
Year
Area [million km2]
ab
cd
e
1960 1980 2000 2020 2040 2060 2080
Year
1960 1980 2000 2020 2040 2060 2080
40
30
20
10
0
400
350
300
250
30
20
10
0
200
450
400
350
300
250
200
CCMI-1
300
250
200
150
100
Fig. 3 Antarctic ozone and metrics quantifying ozone loss as a function of meteorology and additional CFC-11 emissions. Mean column ozone (DU)
averaged from 90oSto60
oS for aSeptember and bOctober from TOMCAT simulations CNTL (control), fODS (fixed ozone-depleting substances), R2000,
R2002, R2009, R2010 (repeating meteorology from 2000, 2002/2003, 2009/2010 and 2010/2011, respectively), R2000_NoIE (no increased CFC-11
emissions), R2000_NoVSLS (no chlorinated very short-lived substances), R2000_CFC11_67 (with constant CFC-11 emissions of 67 Gg yr−1),
R2000_CFC11_B and R2010_CFC11_B (with additional CFC-11 emissions from box model for 2000 and 2010/2011 meteorology, respectively) (see legend)
from 1960 to 2090. Panel balso shows mean (±1σcyan shading) chemistry-climate modelling initiative (CCMI) results13. Estimates of the size of the
Antarctic ozone hole using carea contained within the 220 DU contour (×106km2) (averaged September 7–October 13), dozone mass deficit (×106
tonnes) (averaged September 21–October 13) and eminimum column ozone (between September 21 and October 16). All panels also show observations
(black line) from NASA Solar Backscatter Ultraviolet (SBUV) instrument (a,b)orhttps://ozonewatch.gsfc.nasa.gov/statistics/annual_data.htm (c–e). The
coloured dots on the fODS line (panels a,b) show the years used for simulations R2000, R2002, R2009 and R2010. The pink line in the background in all
panels from 2018 to 2090 shows the results of the continuation of run CNTL with 20-year repeating meteorology.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x ARTICLE
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 5
Content courtesy of Springer Nature, terms of use apply. Rights reserved
been very small and highlights the effectiveness of the atmospheric
monitoring system for detecting small changes. If these continued
emissions are related to production for an immediately emissive use
(e.g. a solvent) then the overall impact on ozone may also be small.
The potential impact from continued emissions is greater if they
have arisen from much larger quantities of production for non-
emissive use (e.g. foam). Run R2000_CFC11_B includes additional
CFC-11 emissions based on the assumption that post-2010 use of
CFC-11 was for non-emissive foam blowing (implying recent
production that was around × 6 larger (see Methods) than the
unexpected emission increase), and steady elimination of this pro-
duction over the next 10 years. In this scenario, ozone loss is
enhanced, causing a delay in the recovery of ozone of about 2 years
no matter which metric is considered (Table 2). The simple
WMO scenario of constant CFC-11 emissions of 67 Gg yr−1
(S_CFC11_67) gives a different CFC-11 time dependence and
much larger impact compared with the S_CFC11_B scenario. With
large, constant emissions, the CFC-11 mixing ratio is maintained at
higher levels late in this century (Fig. 2) and the ozone return to
1980 values (run R2000_CFC11_67) is delayed by around 18 years
(Table 2, Supplementary Figure 6), in agreement with 2D model
estimates with a less detailed treatment of polar processes4.How-
ever, we would note that this scenario of constant future CFC-11
emissionsasusedinWMO2018
4is likely unrealistic.
Chlorinated VSLS also affect polar ozone recovery through
similar chemical processes. The impact of additional CFC-11
emissions can be compared with the results from run
R2000_NoVSLS, which shows that if the stratospheric injection
of chlorinated VSLS decreases to zero (from 2016 value of 114
pptv30) the ozone return dates are brought forward by about 7
years (Table 2), in broad agreement with the delay estimated by
Hossaini et al.18 who only considered CH
2
Cl
2
. This shows the
leverage that chlorine from uncontrolled VSLS also exerts on
polar ozone recovery. While the continued growth sensitivity
scenario of Hossaini et al.18 seems unlikely, further increases in
short-lived chorine would delay ozone recovery in proportion to
the increase in chlorine delivered to the stratosphere by these
gases. Conversely, if the loading were to decrease from the
present-day values ozone recovery would occur earlier. It should
be noted that the impact on Antarctic ozone owing to changes in
chlorine from CFC-11 (or CFC-12 or CCl
4
; see the ‘Discussion'
section) or short-lived gases (e.g. CH
2
Cl
2
) is similar, and depends
primarily on the amount of chlorine delivered to the stratosphere.
The large mean age-of-air in the polar lower stratosphere
(~5 years31) results in a large fractional conversion of most
major organic chlorine (and bromine) source gases (whether long
or short lived) to inorganic Cl
y
in air transported to this region32.
The ozone-hole metrics in Fig. 3can be transformed into a
relative extent of ozone recovery by defining 0% recovery as the
metric value at maximum depletion and 100% recovery as the
1980 value. This is demonstrated in Fig. 5for run R2000,
R2000_NoVSLS, R2000_CFC11_B and R2000_CFC11_67. This
approach means that the extent of recovery at any time can be
compared on a relative scale (see Table 3for these values in 2050).
Presentation of the results in this way avoids the issue of return
dates being strongly affected by the shape of the ozone recovery
trajectory, and that, under some circumstances, the atmosphere
may not return to 1980 values at all. All simulations will give a
numerical value for the extent of recovery at a given date and the
scale of recovery is simply defined in terms of past ozone
concentrations (e.g. 1980 levels and maximum depletion).
The comparison date can be chosen as one that best suits the
timescale of atmospheric processes and policy decisions. In 2050,
the ozone mass deficit is 86% towards return to the 1980 value in
R2000_NoVSLS, 76% in R2000 and 72% in R2000_CFC11_B but
only 62% in run R2000_CFC11_67.
Table 1 Details of 3D model simulations.
Model simulation
CNTL fODS R2000_NoIE R2000 R2002 R2009 R2010 R2000_CFC11_67 R2000_CFC11_B +R2000_CFC12_67 R2010_CFC11_B R2000_NoVSLS
Purpose and
description
Control Fixed ODS No increased
emissions.
CFC-11 emissions
estimated by decay of
2002 bank only
Impact of meteorology.
Simulations with repeating annual
meteorology for years given below
Impact of additional CFC-11 (and CFC-12) emissions.
Simulations with CFC-11 scenario estimated from box model (CFC11_B) or
from WMO (2018) at constant 67 Gg yr−1
Impact of VSLS.
Simulation with
zero
chlorinated VSLS
Time period 1980–2090 1960–2090 2000–2018 1955–2080 2018–2018–2018–2018–2018–2018–2018–2018–
Meteorology Varying Varying 2000 2000 2002/
2003
2009/
2010
2010/
2011
2000 2000 2000 2010/2011 2000
CFC-11
emissions
WMO 1960 vmr Box model S_NoIE WMO WMO WMO WMO 67 Gg yr−1Box model 67 Gg yr−1Box model WMO
Cl VSLS Yes –– ––––– – – – No
+Also includes additional constant CFC-12 emissions of 59 Gg yr−1from 2018 onwards.
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
6NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Dependence on halogen loading and CFC-11 emissions. The
dependence of ozone recovery on the decline of halogens can help
explain the difference of our Antarctic return dates with the CCM
estimate of Dhomse et al.13 of 2060 (2055–2066) and the impact
of different CFC-11 and VSLS loadings (Table 1). Supplementary
Fig. 3 shows temporal changes in both ozone and equivalent
chlorine in the Antarctic diagnosed at 50 hPa in September and
October. The post-2000 increase in ozone is driven by the
decrease in equivalent chlorine. For equivalent chlorine, the only
variation between the simulations is caused by differences in
CFC-11 and VSLS, but ozone variability is also affected by
meteorology. Supplementary Figure 4 compares Cl
y
and total
chlorine from our CTM runs with CCM results13. For Cl
y
the
individual CCMs show large variability and often underestimate
observed values based on microwave limb sounder (MLS)
data33,34. The Cl
y
values from the CTM runs (which also include
VSLS) are at the upper end of the CCM range and agree well with
MLS. Apart from a few outliers, the total chlorine loading from
the CCMs is less variable, indicating that although most CCMs
have an overall realistic halogen loading for the past, the parti-
tioning into Cl
y
in the Antarctic vortex is underestimated. In the
future, the total chlorine (and Cl
y
) loading in our CTM simula-
tions is larger than the CCMs. This is partly due to the VSLS
included in the CTM and partly due to the use of the updated
WMO (2018) ODS scenarios (Supplementary Fig. 5), which delay
the return of equivalent chlorine to 1980 values by about 4 years
compared with the 2011 WMO Assessment35 scenario used
in CCMI.
Ozone return dates for the total column and at 50 hPa from the
individual CCMs tend to correlate with Cl
y
return dates13. This is
shown in Supplementary Fig. 7 along with our CTM results for
varying meteorology and chlorine loading. Comparing the results
Table 2 Dates for return to 1980 values for different ozone-hole metrics for 3D model simulations.
Ozone-hole metric Model simulation+
R2000 R2002 R2009 R2010 R2000_CFC11_67 R2000_CFC11_B R2010_CFC11_B R2000_NoVSLS
Minimum column ozone 2063 2035 2062 2057 2078 2065 2058 2057
Hole area 2067 2041 2061 2056 2085 2069 2059 2061
Mass deficit 2067 2033 2061 2059 2083 2069 2060 2060
September mean
ozone column
(90oS–60oS)
2067 2021 2060 2052 2084 2069 2053 2060
October mean ozone
column (90oS–60oS)*
2067 2031 2066 2085 2085 2069 2087 2060
*For comparison the multi-model mean results from Dhomse et al.13 (MMM1S) for this metric are 2060 (2055–2066). +Simulation R2000_CFC12_67 includes additional constant CFC-12 emissions of
59 Gg yr−1from 2018 onwards. Recovery to 1980 values does not occur in this run by 2100, so the run is not included in the table.
Total column ozone [DU]
Year
ab
Year
1960
550
Arctic [February] Arctic [March]
500
450
400
350
550
500
450
400
350
1980 2000 2020 2040 2060 2080 1960 1980 2000 2020 2040 2060 2080
CNTL
R2010
R2010_CFC11_B
R2002
R2009
fODS
NASA SBUV
CCMI-1
R2000
R2000_NoIE
R2000_CFC11_B
R2000_CFC11_67
R2000_NoVSLS
R2000_CFC12
Fig. 4 Past and simulated future Arctic ozone showing the influence of meteorology and additional CFC-11 emissions. Mean column ozone (DU)
averaged from 90oNto60
oN for aFebruary and bMarch from TOMCAT simulations CNTL (control), fODS (fixed ozone-depleting substances), R2000,
R2002, R2009, R2010 (repeating meteorology from 2000, 2002/2003, 2009/2010, and 2010/2011, respectively), R2000_NoIE (no increased CFC-11
emissions), R2000_NoVSLS (no chlorinated very short-lived substances), R2000_CFC11_67 (with constant CFC-11 emissions of 67 Gg yr−1),
R2000_CFC11_B and R2010_CFC11_B (with additional CFC-11 emissions from box model for 2000 and 2010/2011 meteorology, respectively) (see legend)
from 1960 to 2090. Panel balso shows mean (±1σshading) chemistry-climate modelling initiative (CCMI) results from Dhomse et al.13 and observations
from NASA Solar Backscatter Ultraviolet (SBUV) instrument (black line). The coloured dots on the fODS line show the years used for simulations R2000,
R2002, R2009 and R2010. The pink line in the background from 2018 to 2090 shows results of the continuation of run CNTL with 20-year repeating
meteorology.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x ARTICLE
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 7
Content courtesy of Springer Nature, terms of use apply. Rights reserved
from R2002 with R2010 shows the large impact of meteorology
on the ozone return date. However, for a given meteorology, the
variation of ozone return date appears to vary almost linearly
with Cl
y
(e.g. R2000_NoVSLS, R2000 and R2000_CFC11_B) and
to have a similar dependence (2000 vs. 2010 meteorology). Based
on the different chlorine loadings in the simulations, the delay in
ozone return date (Table 2) is around 10 years for an additional
150 pptv Cl. Thus, the impact of other CFC-11 scenarios on
Antarctic ozone can be estimated without the expense of
rerunning the full 3D model, as already noted for VSLS18. The
simple dependence of ozone return on Cl
y
(whatever its source) is
expected because of the role that chlorine plays in polar ozone
loss cycles through the ClO +ClO and ClO +BrO catalytic loss
cycles36 and the small impact of climate change on Antarctic
ozone recovery13. The same dependency of ozone return on Cl
y
is
likely to exist for the individual CCMs with realistic polar
chemistry shown in Supplementary Figs. 4 and 7.
There is a compact, near-linear correlation between mean
Antarctic column ozone depletion and the accumulated equiva-
lent CFC-11 emissions (Fig. 6a, Supplementary Figs. 8 and 9),
despite the different time evolutions of emissions in runs
R2000_CFC11_B and R2000_CFC11_67, and the inclusion of
CFC-12 emissions in run R2000_CFC12_67. This near-linear
relationship can be explained by the behaviour of the ClO +ClO
and ClO +BrO catalytic cycles for the relatively small perturba-
tion of chlorine produced in the simulations37,38 (see Supple-
mentary Results 3). As expected, the runs with larger overall
chlorine emissions give larger ozone depletion, but in the next few
Year
1980 2000 2020
O3 mass defi.
October
September
100
80
60
40
20
0
d
b
a
O3 hole area
c
TCO minimum
e
2040 2060 2080
Year
1980 2000 2020 2040 2060 2080
Ozone recovery (%)
100
80
60
40
20
0
Ozone recovery (%)
100
80
60
40
20
0
Ozone recovery (%)
R2000
R2000_CFC11_B
R2000_CFC11_67
R2000_NoVSLS
Fig. 5 Extent of recovery for Antarctic ozone hole showing the influence of varying chlorine loading. Extent of recovery (%) for the metrics of
aSeptember mean column ozone (90oS–60oS), bOctober mean column ozone, carea contained within the 220 DU contour (averaged September
7–October 13), dozone mass deficit (averaged September 21–October 13) and eminimum column ozone (between September 21 and October 16) from
TOMCAT simulations R2000 (2000 meteorology), R2000_CFC11_B (with additional CFC-11 emissions from box model), R2000_CFC11_67 (with constant
CFC-11 emissions of 67 Gg yr−1) and R2000_NoVSLS (no chlorinated very short-lived substances) (see legend) from 1980 to 2080. For the metrics 0%
recovery is defined as the maximum depletion (which occurs around 1998) and 100% recovery is defined as return to the 1980 value. The blue horizontal
and vertical lines indicate 100% recovery and 2050, respectively.
Table 3 Percentage recovery (return to 1980 values) by 2050 from four model simulations with different chlorine loading for
various ozone-hole metrics.
Ozone-hole metric Model simulation
R2000 (%) R2000_NoVSLS (%) R2000_CFC11_B (%) R2000_CFC11_67 (%)
Minimum column ozone 61 75 57 49
Hole area 58 72 54 44
Mass deficit 76 86 72 65
September mean ozone column (90oS–60oS) 69 79 67 61
October mean ozone column (90oS–60oS) 60 72 56 49
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
8NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
decades the emission history does not play a role. For run
R2000_CFC11_B, the emissions have decreased strongly by
around 2080, and CFC-11 emitted early in the run has been
removed from the atmosphere, hence the breakdown in the
correlation on this longer timescale. The runs with fixed
emissions also show a smaller deviation from linearity late in
the century when some early emissions have been removed from
the atmosphere. The slope of Fig. 6a (0.6 DU/100 Gg CFC-11)
provides a means for estimating the Antarctic ozone impact of
different emission scenarios. The additional CFC emissions
increase the ozone mass deficit (Fig. 6b) and delay recovery.
Despite the ongoing ozone depletion shown, runs
R2000_CFC11_B and R2000_CFC11_67 are still estimated to
have returned to 1980 values by 2069 and 2083, respectively
(Table 2).
Discussion
Recovery of the ozone layer is happening but any metric used to
quantify its timescale or extent needs to be chosen with caution.
For the Antarctic ozone hole, metrics related to the areal extent,
ozone mass deficit or minimum column ozone appear more
robust across different meteorological situations. In contrast,
October mean column ozone can produce different year-to-year
variations compared with other metrics due to meteorological
variability in that month. This suggests a problem in using that
metric to quantify ozone recovery from CCM simulations. Also,
using return-to-1980-value dates can cause apparently large
changes in recovery for small changes in column ozone18.A
clearer picture of recovery under different scenarios can be
obtained by estimating the degree of recovery achieved by a
certain date. This can be designed to avoid over-emphasising the
long tail in the recovery process when the ozone concentration is
changing only slowly, but is very close to pre-ozone-hole values.
Whatever metric is used to estimate the recovery of the mean
ozone layer, atmospheric variability will cause variations from year
to year. It is very likely that the first year without an ozone hole (by
the usual metrics) will occur well before the mean return date. For a
year with disturbed meteorology like that observed in 2002, then, by
some metrics, the ozone hole may temporarily recover to 1980
conditions even with the enhanced halogen loadings of the 2020s.
Clearly, an early year without an ozone hole (i.e. in the next decade
or so) does not indicate full recovery from the effects of ODSs.
Renewed production and emission of CFC-11 will delay the
recovery of the ozone layer. For the Antarctic ozone hole, there is
a clear link between the additional amount of chlorine injected
into the stratosphere, the additional polar ozone loss and the
delay to recovery. However, even if the renewed production so far
is for closed-cell foam use, immediate effective measures to stop
this could imply a delay of just a few years. Should this renewed
production be allowed to continue the impact will be corre-
spondingly more severe, and for the Antarctic ozone hole, can be
estimated directly from the chlorine-loading enhancement. The
estimates derived here for the large WMO (2018) emission sce-
nario of constant 67 Gg yr−1of a 18-year delay for an extra
~70-ppt CFC-11 are substantial but still do not change the overall
trajectory of recovery. Note that for the Antarctic, the expected
impact of climate change is only a small (2-year) advance of the
1980 return date13. In comparison, the potential impact of
additional CFC-11 emissions can be large.
The relationship between Antarctic ozone recovery, emissions
and polar Cl
y
loading can be used to apply our results to per-
turbations of other source gases. In the aged air of the polar lower
stratosphere, most major source gases are nearly completely
converted to inorganic forms. Therefore, if CFC-12 is being co-
produced along with the additional CFC-11 and is largely being
contained in a bank, a process that we have not explicitly con-
sidered, the impact on Antarctic ozone recovery will scale with
chlorine loading supplied by this additional CFC-12 as it is
eventually emitted. Minimum co-production rates of CFC-12 in
the most common industrial process for producing CFC-11 are
~30%, so this additional chlorine might be substantial39. Simi-
larly, the polar impact of any additional emissions of CCl
4
, which
is not decreasing as rapidly as expected in the atmosphere40–42,
will depend on the resulting increase in stratospheric chlorine.
2050
2077 2083
ab
2069
0 2000 4000 6000 8000
2050
2050
2050
dTCO [Sep 21–Oct 13] (DU)
dOMD [Sep 21–Oct 13] (million tonnes)
Accumulated CFC-11 emissions (Gg)
02000
0
–10
Difference in column ozone Difference in ozone mass deficit
7
2020
2040
2060
2080
2100
6
5
4
3
2
1
0
–20
–30
–40
4000 6000 8000
Accumulated CFC-11 emissions (Gg)
2050
2050
2050
2050
R2000_CFC11_B
R2000_CFC11_Ex
R2000_CFC11_67
R2000_CFC12_67
Fig. 6 Antarctic ozone depletion versus accumulated chlorine emissions. a The mean column ozone difference (DU) between run R2000 (2000
meteorology) and runs R2000_CFC11_B (with additional CFC-11 emissions from box model, circle), R2000_CFC11_67 (with constant CFC-11 emissions of
67 Gg yr−1, diamond) and R2000_CFC12_67 (with constant CFC-11 emissions of 67 Gg yr−1and CFC-12 emissions of 59 Gg yr−1,+symbol) in regions
60oS–90oS for the period September 21–October 13 (corresponding to the time period for the ozone mass deficit metric in Fig. 3) is plotted against
accumulated additional equivalent CFC-11 emissions (Gg). Also shown are the results for simulation R2000_CFC11_Ex (star, see Supplementary
Information). The colour shading indicates the year for each data point; the points for 2050 are plotted in black. The slope of best-fit line through the 2050
data points is 0.6 DU/100 Gg CFC-11. bDifference in estimated ozone mass deficit (million tons) versus accumulated equivalent CFC-11 emissions for the
same simulations as panel (a).
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x ARTICLE
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 9
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Emissions of VSLS are also a source of stratospheric chlorine.
Previous sensitivity studies18 and our simulation R2000_NoVSLS
demonstrate the potential for VSLS to influence timescales for
ozone recovery through changing stratospheric chlorine loading.
However, the degree to which such potential is realised and the
overall significance of VSLS with respect to ozone will depend on
both the magnitude and trend of their future emissions. For
CH
2
Cl
2
, the most abundant chlorinated VSLS, global emissions
increased by a factor of ~2 between 2000 (~500 Gg yr−1) and
2016 (~1000 Gg yr−1), as derived from NOAA data4. While such
growth may have acted to offset the rate of upper stratospheric
HCl decline, by ~15% since the mid-2000s30,firm conclusions on
any future impacts require more accurate estimates of likely
future emission changes. For VSLS with both mixed emissive and
non-emissive applications, notably CH
2
Cl
2
, such scenarios could
be developed using production information and market analyses.
The Montreal Protocol is rightly seen as a seminal interna-
tional agreement that has successfully led to decreasing levels of
atmospheric chlorine and bromine and early signs of ozone
recovery. For the Antarctic ozone hole, this recovery is best
measured by metrics related to the extent of ozone loss in Sep-
tember rather than October and by the decrease in depletion by a
certain date, rather than return to the 1980 value. Previous his-
toric use of October metrics was related to datasets used to dis-
cover the ozone hole2, but, with our detailed knowledge of
the processes involved, we can now use different metrics to
measure recovery (e.g. ref. 11). In this paper, we have shown that
three decades since it was ratified, its continued success does face
some challenges from recent unreported CFC-11 production.
However, with swift action to curb this production and any other,
the long-term success of the protocol will be ensured.
Methods
Model configuration and experimental design. We have used the TOMCAT/
SLIMCAT offline three-dimensional (3D) CTM to calculate the impact of chlorine
scenarios and meteorological variability on stratospheric ozone43. The model has
been widely used in previous studies44,45 and simulates stratospheric ozone well.
The model includes a detailed treatment of stratospheric chemistry, including a full
description of processes related to polar ozone depletion. The model is forced by
ERA-Interim reanalyses provided by the European Centre for Medium-Range
Weather Forecasts (ECMWF)46. The model was integrated in series of experiments
at a horizontal resolution of 2.8o× 2.8owith 32 levels from the surface to ~60 km.
The runs were forced by observed and predicted surface mixing ratios of long-lived
source gases from 1955 to 2100. The upper troposphere mixing ratios of short-lived
chlorine species were specified using estimates from 2000 to 201630, with constant
values before and after this period.
We drive the future model simulations with repeating meteorological analyses
from previous years but with time-dependent ODS concentrations, which for the
control run are taken from the baseline A1 scenario of WMO (2018)4. The surface
CH
4
and N
2
O scenarios are taken from the Special Report on Emissions (SRES)
scenario A1b (see Supplementary Information of Dhomse et al.13). This approach
ignores the impact of climate change on stratospheric temperatures and circulation,
which will become increasingly important as the simulations progress. Climate
change certainly has an important impact on ozone recovery in the upper
stratosphere where cooling acts to increase ozone and adds to the effect of
decreasing chlorine4. However, in this paper, we focus on polar lower stratospheric
ozone loss where the impact of temperature trends is less important13. Moreover,
by using this approach, we ensure that the model results have realistic polar
meteorological conditions that are important for accurate simulation of ozone loss.
By running the model with different repeating analysis years we aim to span the
effects of climate change on polar vortex dynamics.
Model simulation R2000 was integrated from 1955 to 2080 using repeating 2000
meteorology and time-dependent ODS concentrations (see Table 1). Control
simulation CNTL was initialised from run R2000 in 1980 and integrated with varying
meteorology until 2018. This simulation gives the most realistic representation of the
atmosphere over the past 4 decades. Run CNTL was continued to 2080 using a cycle
of 20 years of repeating meteorology from 1999 to 2018 (i.e. 1999 meteorology in
model years 2019, 2039, 2059 and 2079). A series of future runs with repeating annual
meteorology were initialised in 2018 from run R2000: R2002 (May 2002–April 2003),
R2009 (May 2009–April 2010) and R2010 (May 2010–April 2011). These runs used
repeating meteorology from May to April to avoid discontinuities in the polar winter/
spring of either hemisphere. Run R2000_NoIE is the same as R2000 but without the
effect of the post-2002 stabilisation and post-2010 increase in CFC-11 emissions (box
model scenario S_NoIE in Fig. 1). Runs R2000_CFC11_B and R2010_CFC11_B were
the same as R2000 and R2010, respectively, but with additional emissions of CFC-11
using box model scenario S_CFC11_B (Figs. 1and 2). Run R2000_CFC11_67
was the same as R2000 but with constant CFC-11 emissions of 67 Gg yr−1
(scenario S_CFC11_67, Figs. 1and 2). Run R2000_CFC12_67 was the same as
R2000_CFC11_67 but with constant CFC-12 emissions of 59 Gg yr−1(corresponding
to equal numbers of molecules of CFC-11 and CFC-12 emitted). This is used as a
sensitivity run to examine how the co-emission of longer-lived CFC-12 (lifetime 102
years47) will affect the model projections. Run R2000_NoVSLS was the same as
R2000 but with post-2018 emissions of short-lived chlorine species set to zero
(corresponding to 114 ppt less chlorine in the stratosphere). Finally, we performed a
simulation fODS that was identical to R2000 but with halogenated ODS values
constant at 1960 values. The length of the sensitivity runs varied depending on the
rate of recovery to 1980 ozone values.
Satellite data. To compare with our past model simulations we use various satellite
data products. For total column comparisons, we use the NASA Solar Backscatter
Ultra-Violet SBUV (Version 8.6) merged dataset that is constructed by merging
individual SBUV/SBUV/2 (total and profile ozone) satellite datasets48. SBUV-merged
total ozone data are obtained from https://acd-ext.gsfc.nasa.gov/Data_services/
merged/. Values for Antarctic minimum ozone, ozone-hole area and ozone mass
deficit average d over defined time periods are obtained from https://ozonewatch.gsfc.
nasa.gov/. For gridded total ozone data comparisons we use ESA’sCopernicusCli-
mate Change Service (C3S) data. These data combine ozone total column retrievals
from various UV–nadir, limb and occultation satellite sensors. A key feature of these
data is that climate data records (CDR) and interim-CDR parts of each product are
generated using the same software and algorithms. Total ozone data used here are so-
called level 4 data that combine data from 15 satellite instruments and fill the missing
values using data assimilation system. These data are available from January 1970 to
present at 1.0o×1.0
oresolution and are obtained from https://cds.climate.copernicus.
eu/cdsapp#!/dataset/satellite-ozone?tab=overview
Emissions box model. We have related emissions of CFC-11 to mean atmospheric
mixing ratios using a global 1-box model (Figs. 1and 2). The model assumes a
CFC-11 lifetime of 54.5 years, which was diagnosed from a TOMCAT full
chemistry simulation. First, the model was used to relate the WMO (2018) global
mean surface CFC-11 scenario4to an estimate of annual emissions, assuming that
the surface values correspond to a global mean. This gave good agreement with
independent estimates from a multi-box model15. Our estimated emissions were
then used as a basis for estimating emissions from non-reported production of
CFC-11 since 2010 and for future example sensitivity scenarios, bearing in mind
the large uncertainties related to the new CFC-11 emissions and how they will vary.
We use 13 ± 5 Gg yr−1as the estimate of emissions related to unreported pro-
duction in recent years15, which we relate to an assumed total production and
project forward in time, assuming that the CFC-11 is produced for use in closed-
cell foams. Scenario S_CFC11_B assumes that 15% of CFC-11 produced is released
immediately49 (ratio 5.66:1), followed by 3.5% yr−1(ref. 15) as leakage from the
bank. We assume that through policy action the unreported production will
rampdown to zero over 10 years. Further box model runs were performed to
examine the sensitivity to these assumptions (see panels in Figs. 1and 2). The
estimate of recent emissions was varied from 8 to 18 Gg yr−1. Estimates of the ratio
of production to initial emission are as large as 26 (ref. 49), which seems incon-
sistent with likely applications. Therefore, we adopt the representative range of ×3
to ×9. The fractional release was varied from 2.5 to 4.5% yr−1, which spans the
upper range given in ref. 49. Finally we varied the time for the production ramp-
down from 5 to 20 years. A further future scenario was defined by taking the WMO
(2018) scenario with an additional 67 Gg yr−1of emissions (the average estimated
top–down emissions over 2002–2016, S_CFC11_67). This is likely unrealistic but it
serves as a 3D model sensitivity simulation. Scenario S_CFC11_B leads to around
45-pptv additional chlorine in the atmosphere over the next few decades (Sup-
plementary Fig. 5). The larger emissions of scenario S_CFC11_67 cause the CFC-
11 decay to slow and by 2080 its vmr is around 170 ppt (510 ppt chlorine). A total
of 4 CFC-11 scenarios used in the 3D CTM are shown in Fig. 2a, with one further
described in Supplementary Results 1 and used in Fig. 6.
Data availability
The observational datasets are available from the web links described in the text, i.e.
SBUV-merged total ozone data from https://acd-ext.gsfc.nasa.gov/Data_services/merged/.
Values for Antarctic minimum ozone, ozone-hole area and ozone mass deficit averaged
over defined time periods from https://ozonewatch.gsfc.nasa.gov/; gridded total ozone
data from ESA’s Copernicus Climate Change Service (C3S) from https://cds.climate.
copernicus.eu/cdsapp#!/dataset/satellite-ozone?tab =overview. The TOMCAT model
results are available by emailing the corresponding author and via the web page http://
homepages.see.leeds.ac.uk/~lecmc/ftp/CFC11/.
Code availability
The TOMCAT model is a research tool that is available to NERC-funded researchers in
the United Kingdom and other collaborators who have access to suitable computing
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
10 NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
facilities. The code is not otherwise publicly available. Reasonable requests to use the
model should be made to the corresponding author.
Received: 15 April 2019; Accepted: 20 November 2019;
References
1. Molina, M. J. & Rowland, F. S. Stratospheric sink for chlorofluoromethanes:
chlorine atom-catalysed destruction of ozone. Nature 249, 810–812 (1974).
2. Farman, J. C., Gardiner, B. G. & Shanklin, J. D. Large losses of total ozone in
Antarctica reveal seasonal ClOx/NOx interaction. Nature 315, 207–210
(1985).
3. Engel, A., et al. Update on ozone-depleting substances (ODSs) and other gases
of interest to the Montreal Protocol. Chapter 1 in Scientific Assessment of
Ozone Depletion 2018 (World Meteorological Organization, Geneva,
Switzerland, 2018).
4. World Meteorological Organisation (WMO). Scientific Assessment of Ozone
Depletion: 2018, Global Ozone Research and Monitoring Project—Report No.
58. (2018).
5. Froidevaux, L. et al. Temporal decrease in upper atmospheric chlorine.
Geophys. Res. Lett. 33,8–12 (2006).
6. Kohlhepp, R. et al. Observed and simulated time evolution of HCl, ClONO
2
,
and HF total column abundances. Atmos. Chem. Phys.12, 3527–3557 (2012).
7. Harris, N. R. P. et al. Past changes in the vertical distribution of ozone—Part 3:
analysis and interpretation of trends. Atmos. Chem. Phys. 15, 9965–9982
(2015).
8. Weber, M. et al. Total ozone trends from 1979 to 2016 derived from five
merged observational datasets—the emergence into ozone recovery. Atmos.
Chem. Phys. 18, 2097–2117 (2018).
9. de Laat, A. T. J., van Weele, M. & van der, A. R. J. Onset of stratospheric ozone
recovery in the Antarctic ozone hole in assimilated daily total ozone columns.
J. Geophys. Res. 122, 880–11,899 (2017).
10. Kuttippurath, J., Kumar, P., Nair, P. J. & Pandey, P. C. Emergence of ozone
recovery evidenced by reduction in the occurrence of Antarctic ozone loss
saturation. npj Clim. Atmos. Sci.1, 42 (2018).
11. Solomon, S. et al. Emergence of healing in the Antarctic ozone layer. Science
353, 269–274 (2016).
12. Strahan, S. E. & Douglass, A. R. Decline in Antarctic ozone depletion and
lower stratospheric chlorine determined from Aura Microwave Limb Sounder
observations. Geophys. Res. Lett. 45, 382–390 (2018).
13. Dhomse, S. S. et al. Estimates of ozone return dates from Chemistry-Climate
Model Initiative simulations. Atmos. Chem. Phys. 18, 8409–8438 (2018).
14. Eyring, V. et al. Multi-model assessment of stratospheric ozone return dates
and ozone recovery in CCMVal-2 models. Atmos. Chem. Phys.10, 9451–9472
(2010).
15. Montzka, S. A. et al. An unexpected and persistent increase in emissions of
ozone-depleting CFC-11. Nature 557, 413–417 (2018).
16. Rigby, M. et al. Increase in CFC-11 emissions from eastern China based on
atmospheric observations. Nature 569, 546–550 (2019).
17. McCulloch, A., Ashford, P. & Midgley, P. M. Historic emissions of
fluorotrichloromethane (CFC-11) based on a market survey. Atmos. Environ.
35, 4387–4397 (2001).
18. Hossaini, R. et al. The increasing threat to stratospheric ozone from
dichloromethane. Nat. Commun. 8, 15962 (2017).
19. Fang, X. et al. Rapid increase in ozone-depleting chloroform emissions from
China. Nat. Geosci. 12,89–93 (2019).
20. Carpenter, L. et al. Scenarios and information for policy makers Chapter 6 in
Scientific Assessment of Ozone Depletion: 2018. Chapter 6 in Scientific
Assessment of Ozone Depletion (2018).
21. Revell, L. E., Bodeker, G. E., Huck, P. E., Williamson, B. E. & Rozanov, E. The
sensitivity of stratospheric ozone changes through the 21st century to N
2
O
and CH
4
.Atmos. Chem. Phys. 12, 11309–11317 (2012).
22. Butler, A. H. et al. Diverse policy implications for future ozone and surface
UV in a changing climate. Environ. Res. Lett. 11,2–9 (2016).
23. Keeble, J., Brown, H., Abraham, N. L., Harris, N. R. P. & Pyle, J. A. On ozone
trend detection: using coupled chemistry-climate simulations to investigate
early signs of total column ozone recovery. Atmos. Chem. Phys. 18, 7625–7637
(2018).
24. Newman, P. A. & Nash, E. R. The unusual southern hemisphere stratosphere
winter of 2002. J. Atmos. Sci. 62, 614–628 (2005).
25. Simmons, A. et al. ECMWF analyses and forecasts of stratospheric winter
polar vortex breakup: September 2002 in the southern hemisphere and related
events. J. Atmos. Sci. 62, 668–689 (2005).
26. Chipperfield, M. P. & Jones, R. L. Relative influences of atmospheric chemistry
and transport on Arctic ozone trends. Nature 400, 551–554 (1999).
27. Rex, M. et al. Arctic ozone loss and climate change. Geophys. Res. Lett. 31,
L04116 (2004).
28. Manney, G. L. et al. Unprecedented Arctic ozone loss in 2011. Nature 478,
469–475 (2011).
29. Bednarz, E. M. et al. Future Arctic ozone recovery: the importance of
chemistry and dynamics. Atmos. Chem. Phys. 16, 12159–12176 (2016).
30. Hossaini, R. et al. Recent trends in stratospheric chlorine from very short‐lived
substances. J. Geophys. Res. 124, 2318–2335 (2019).
31. Waugh, D. W. & Hall, T. M. Age of stratospheric air: theory, observations, and
models. Rev. Geophys.40, 1010 (2002).
32. Newman, P. A., Daniel, J. S., Waugh, D. W. & Nash, E. R. A new formulation
of equivalent effective stratospheric chlorine (EESC). Atmos. Chem. Phys. 7,
4537–4552 (2007).
33. Waters, J. W. et al. The Earth observing system microwave limb sounder (EOS
MLS) on the aura satellite. IEEE Trans. Geosci. Remote 44, 1075–1092 (2006).
34. Livesey, N. J. et al. Version 4.2x Level 2 Data Quality and Description
Document. (2018).
35. World Meteorological Organization (WMO). Scientific Assessment of Ozone
Depletion: 2010, Global Ozone Research and Monitoring Project—Report No.
52. (2011).
36. Solomon, S. Stratospheric ozone depletion: a review of concepts and history.
Rev. Geophys. 37, 275–316 (1999).
37. Searle, K. R., Chipperfield, M. P., Bekki, S. & Pyle, J. A. The impact of spatial
averaging on calculated polar ozone loss: 2. Theoretical analysis. J. Geophys.
Res.103, 25409–25416 (1998).
38. Fernandez, R. P., Kinnison, D. E., Lamarque, J. F., Tilmes, S. & Saiz-Lopez, A.
Impact of biogenic very short-lived bromine on the Antarctic ozone hole
during the 21st century. Atmos. Chem. Phys. 17, 1673–1688 (2017).
39. Technology and Economic Assessment Panel Progress Report (Volume 3).
(2018).
40. Liang, Q. et al. Constraining the carbon tetrachloride (CCl
4
) budget using its
global trend and inter-hemispheric gradient. Geophys. Res. Lett. 41, 5307–5315
(2014).
41. Lunt, M. F. et al. Continued emissions of the ozone-depleting substance
carbon tetrachloride from Eastern Asia. Geophys. Res. Lett. 45, 11,423–11,430
(2018).
42. Sherry, D., McCulloch, A., Liang, Q., Reimann, S. & Newman, P. A. Current
sources of carbon tetrachloride (CCl
4
) in our atmosphere. Environ. Res. Lett.
13, 024004 (2018).
43. Chipperfield, M. P. New version of the TOMCAT/SLIMCAT off-line chemical
transport model: Intercomparison of stratospheric tracer experiments. Q. J. R.
Meteorol. Soc. 132, 1179–1203 (2006).
44. Chipperfield, M. P. et al. Quantifying the ozone and ultraviolet benefits
already achieved by the Montreal Protocol. Nat. Commun. 6, 7233 (2015).
45. Dhomse, S., Chipperfield, M. P., Feng, W. & Haigh, J. D. Solar response in
tropical stratospheric ozone: a 3-D chemical transport model study using ERA
reanalyses. Atmos. Chem. Phys.11, 12773–12786 (2011).
46. Dee, D. P. et al. The ERA-Interim reanalysis: configuration and performance
of the data assimilation system. Q. J. R. Meteorol. Soc. 137, 553–597 (2011).
47. Chipperfield, M. P. et al. Multimodel estimates of atmospheric lifetimes of
long‐lived ozone‐depleting substances: present and future. J. Geophys. Res.
119, 2555–2573 (2014).
48. Frith, S. M. et al. Recent changes in total column ozone based on the SBUV
Version 8.6 Merged Ozone Data Set. J. Geophys. Res. 119, 9735–9751 (2014).
49. Ashford, P., Clodic, D., McCulloch, A. & Kuijpers, L. Emission profiles from
the foam and refrigeration sectors comparison with atmospheric
concentrations. Part 1: methodology and data. Int. J. Refrig. 27, 687–700
(2004).
Acknowledgements
The TOMCAT modelling work was supported by the UK Natural Environment Research
Council (NERC) through the SISLAC project (NE/R001782/1) and performed on the
Archer HPC machine. We thank ECMWF for providing the ERA-Interim reanalyses. We
acknowledge use of the publicly available C3S and SBUV data. RH is supported by a
NERC Independent Research Fellowship (NE/N014375/1). J.K. and J.A.P. received
funding from the European Community’s Seventh Framework Programme (FP7/
2007–2013) under Grant agreement no. 603557 (StratoClim).
Author contributions
M.P.C. conceived the idea and initiated the study in discussion with S.D. M.P.C., S.D. and
W.F. performed and analysed the model runs. The figures were prepared by S.D. and
M.P.C. S.A.M. and J.S.D. provided guidance in scenario development. R.H., J.A.P. and
J.K. provided comments on the model simulations. M.P.C. wrote the paper and included
the comments from all of the coauthors.
Competing interests
The authors declare no competing interests.
NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x ARTICLE
NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications 11
Content courtesy of Springer Nature, terms of use apply. Rights reserved
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/s41467-
019-13717-x.
Correspondence and requests for materials should be addressed to M.P.C.
Peer review information Nature Communications thanks Donald Wuebbles and the
other, anonymous, reviewer(s) for their contribution to the peer review of this work. Peer
reviewer reports are available
Reprints and permission information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in
published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons
Attribution 4.0 International License, which permits use, sharing,
adaptation, distribution and reproduction in any medium or format, as long as you give
appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons license, and indicate if changes were made. The images or other third party
material in this article are included in the article’s Creative Commons license, unless
indicated otherwise in a credit line to the material. If material is not included in the
article’s Creative Commons license and your intended use is not permitted by statutory
regulation or exceeds the permitted use, you will need to obtain permission directly from
the copyright holder. To view a copy of this license, visit http://creativecommons.org/
licenses/by/4.0/.
© The Author(s) 2019
ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-019-13717-x
12 NATURE COMMUNICATIONS | (2019) 10:5781 | https://doi.org/10.1038/s41467-019-13717-x | www.nature.com/naturecommunications
Content courtesy of Springer Nature, terms of use apply. Rights reserved
1.
2.
3.
4.
5.
6.
Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:
use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
use bots or other automated methods to access the content or redirect messages
override any security feature or exclusionary protocol; or
share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at
onlineservice@springernature.com
Available via license: CC BY 4.0
Content may be subject to copyright.
