Global Atmosphere – The Antarctic Ozone Hole

Ozone (O₃) is a vitally important component of the Earth's atmosphere. About 90% of atmospheric ozone resides in the stratosphere where it filters out harmful ultraviolet radiation and stops wavelengths shorter than about 300 nm from reaching the surface.¹  The remaining 10% resides in the troposphere where it is a photochemical pollutant.²  Concern over the depletion of the stratospheric ozone layer first focussed on the detrimental effects of increased surface UV on humans, animals and plants. As ozone also absorbs in the infrared (IR), the stratospheric ozone layer also plays a key role in the Earth's climate balance. This role of ozone in climate has become increasingly apparent over the past decade or so and is a second principal motivation for ongoing research into stratospheric ozone.

Ozone in the stratosphere is produced naturally by the photolysis of O₂ molecules by short wavelength UV radiation (λ<242 nm).³ 

Atomic oxygen rapidly recombines with an O₂ molecule to form O₃, and O+O₃ are treated together in stratospheric chemistry as the ‘odd oxygen’ family. While there is only one source of ozone, there are many reactions which remove it via catalytic cycles (see Section 3.2). These cycles can comprise species which are present naturally in the stratosphere (e.g. hydrogen and nitrogen compounds) or species which are largely present through human activities. Overall there is a dynamic equilibrium which maintains the balance of stratospheric ozone through competing loss and destruction cycles. Production by sunlight is favoured at lower latitudes while loss can occur at higher latitudes following transport of ozone via the stratospheric Brewer–Dobson circulation. Given that the natural production of ozone varies only slightly (e.g. over the 11 year solar cycle)^4,5  human perturbations to the ozone layer have occurred by changes to the rate of catalytic loss. As described below, the ultimate cause of the Antarctic ozone hole has been the large increase in stratospheric bromine and chlorine in the period from 1960 onwards due to increased use of gases such as chlorofluorocarbons (CFCs).

Many thousands of research papers have been published on the topic of stratospheric ozone depletion and the Antarctic hole over the past three decades. A number of comprehensive reviews have already been published. Solomon⁶  reviewed stratospheric ozone depletion in general, with a concise summary of the science of the Antarctic ozone hole as of 1999. More recently Müller and coauthors⁷  have provided a comprehensive summary of stratospheric ozone in a Royal Society of Chemistry publication. That book has separate detailed chapters dealing with varied aspects relevant to the chemistry and dynamics of the Antarctic ozone hole. The state of knowledge in stratospheric ozone research is assessed every 4 years in the World Meteorological Organisation/United Nations Environment Programme (WMO/UNEP) assessments.⁸  These publications provide a detailed summary of the most recent literature in the field. An extensive account of the activities of policy makers involved in the signing of the Montreal Protocol is given in Anderson and Sarma.⁹ 

The aim of this review is to provide a concise summary of the Antarctic ozone hole from its discovery, the determination of its cause, the policy action taken to solve the problem and the outlook for the future. The aim is to provide the reader with an up-to-date overview of the subject and he or she is referred to the more detailed reviews cited above for further information. Section 2 discusses observations of Antarctic ozone during the period of the springtime depletion. The processes which lead to this depletion are described in Section 3. Although not the focus of this review, Section 4 briefly gives a summary of ozone depletion which has occurred at other latitudes, and in particular the Arctic. Section 5 summarises the Montreal Protocol, the international agreement to limit the emission of ozone-depleting substances (ODSs). Finally, Section 6 describes some areas of current research in stratospheric ozone and discusses what may happen to the ozone hole in the future.

The Antarctic ozone hole is one of the best known geophysical phenomena among both scientists and the general public. False-colour satellite images of total column ozone at the end of Austral spring are a classic and ubiquitous image in environmental science. An example is given in Figure 1 for October 9, 2006. By this day in the season the ozone column over the Antarctic had been reduced to <90 Dobson Units (DU) with the area below 220 DU, which is commonly taken to indicate the size of the hole, extending over the entire Antarctic continent.

The Antarctic ozone hole was discovered in the mid 1980s when a sharp downward trend in total ozone column over Halley was observed by Farman, Gardiner and Shanklin.¹¹  Figure 2 shows the updated time series of Halley column ozone which dates back to the 1950s. Between 1960 and the mid-1990s the October monthly mean decreased by over 50%. Similar behaviour is seen in the satellite data of the October minimum column ozone and the ozone hole area. Since the mid 1990s the size and depth of the ozone hole, as determined by various metrics, has not increased. In recent years the observations in Figure 2 indicate a slight decrease in size with larger variability. This is consistent with recent decreases in ODSs (see Section 5) but it is not yet possible to determine with certainty that Antarctic ozone has started to recover¹²  (see Section 6).

The ozone hole develops after the end of Austral polar night and the rapid ozone loss occurs when sunlight returns to the polar region. Figure 3a shows the time series of daily column ozone values at Halley for 1957–1973 and 1990–2009. This shows that in contrast to the earlier period, in 1990–2009 column ozone decreased strongly from late August through to early October. In late winter at Halley, temperatures at 100 hPa (∼15 km) are cold (see Figure 3b) but increase through to December as sunlight warms the polar stratosphere. The delay in the warming in 1990–2009 compared to 1957–1973 is due to the role that ozone plays in absorbing sunlight and warming the stratosphere. The decreased ozone in the latter period delays the warming of the stratosphere – an example of ozone chemistry-climate feedbacks (see Section 6).

The depletion of ozone is not spread uniformly throughout the column but occurs in the lower stratosphere. Indeed, Figure 4 shows that in the Antarctic ozone hole by early October there has been complete ozone loss between about 13 and 21 km. Ozone at higher altitudes is not significantly depleted. The restriction of the rapid ozone loss to the lower stratosphere provides a natural limit to the depth the ozone hole and explains the levelling off in the observations of column ozone in Figure 2.

In the mid 1980s the appearance of the Antarctic ozone hole came as a complete surprise to the atmospheric science community. Based on the early work on CFCs in the 1970s,^14,15  it was expected that any ozone depletion due to these compounds would be modest and would occur in the upper stratosphere at around 40 km. The large loss of ozone in the high latitude lower stratosphere was not predicted by chemical models being used at that time.¹⁶  Because of the surprise occurrence of the ozone hole numerous different theories involving different chemical and dynamical processes were proposed to explain it. A number of measurement campaigns were conducted starting with ground-based observations in 1986 and a larger aircraft-based campaign in 1987. These campaigns rapidly established that the cause of the depletion was chlorine (and bromine)-catalysed ozone loss.¹⁷  However, the occurrence of high levels of ozone-destroying reactive chlorine in the Antarctic spring was not expected and new heterogeneous chemical processes needed to be invoked to explain this.¹⁸  Overall, the annual occurrence of the ozone hole is due to a remarkable combination of dynamical and chemical processes which occur in the Antarctic winter/spring lower stratosphere. These processes are described in this section. It was very fortunate for society that the first large warning that we had for global ozone depletion by anthropogenic emissions occurred in this specific region of the atmosphere, rather than on a global scale.

In the 24 hour darkness of polar night, the Antarctic lower stratosphere cools and a strong westerly circulation forms around the pole – the so-called polar vortex (see Figure 5). This vortex isolates the high latitudes in a ‘containment vessel’ for the subsequent chemistry of ozone hole. Temperatures inside the wintertime Antarctic vortex can fall to around 190 K or less, making it the coldest location in the stratosphere outside of the tropical tropopause. The polar vortex persists until late spring, when warming due to solar heating causes the winter westerly circulation to revert to summer easterlies as the vortex breaks down.

The stratosphere is very dry due to the very cold temperatures (∼190 K) which occur at the tropical tropopause and cause water in air parcels to condense out as they are transported upwards from the troposphere. Therefore, extremely cold temperatures are needed for clouds to form in the stratosphere. For ice clouds these temperatures will only occur in the polar regions or just above the cold tropopause at lower latitudes. Historically, sightings of ‘mother of pearl’ clouds in the polar regions have been reported since the 1880s²⁰  but their role in atmospheric chemistry only became apparent after the observation of the Antarctic ozone hole.

We now know that Polar Stratospheric Clouds (PSCs) can be composed of a number of particle types. In the lower stratosphere there exists a ubiquitous layer of liquid binary sulfate aerosols (H₂SO₄–H₂O). At cold temperatures (around 195 K) these aerosols grow and take up HNO₃ to form supercooled liquid ternary solutions (STS, H₂SO₄–H₂O–HNO₃).²¹  At temperatures below about 195 K the formation of solid nitric acid trihydrate particles^22,23  becomes thermodynamically possible, although some degree of supersaturation is required. Finally, below the ice point (∼188 K) solid ice particles can form. These liquid and solid particles provide surfaces for heterogeneous reactions,¹⁸  the most important of which are:

The discovery that heterogeneous reactions can occur on the surface of PSCs was a key step in explaining the cause of the ozone hole. In particular, reactions (2)–(6) cause the conversion of stable reservoir chlorine species (HCl and ClONO₂) into more active forms. Solid PSC particles (NAT, ice) can grow large enough to sediment from the stratosphere and cause permanent removal of HNO₃ (denitrification)²³  and H₂O (dehydration).

PSCs in the atmosphere can be detected by a variety of in situ and remote observations.²⁴  In recent years such observations have been revolutionised by the launch in 2006 of the CALIPSO satellite with the CALIOP lidar onboard.^25,26  This active lidar technique has revealed the full three-dimensional structure of PSCs and can be used to classify the different type and composition of the particles. The instrument is still performing well and has accumulated a long time series of measurements. Figure 6 shows CALIPSO observations of the daily PSC area in the 8 Antarctic winters from 2006–2013. PSCs tend to occur from May to late September between about 12 and 25 km. There is very little interannual variability in the PSC occurrence, reflecting the regular nature of the Antarctic polar vortex. Figure 7 shows the mean distribution of four types of PSC over these 8 winters. Pure Supercooled Ternary Solution (STS) tends to be favoured early in the season with a smaller maximum in late September. The larger ‘denitrifying’ NAT particles occur preferentially at lower altitudes, but are always present as a mixture with liquid STS particles. Smaller NAT particles are also present in a mixture with STS and maybe solid ice.

The role of heterogenous PSC chemistry is illustrated in Figure 8 with satellite observations of chemical species in the Antarctic vortex of 2005.²⁷  In the middle of winter, at altitudes below about 25 km, PSCs cause the almost complete conversion of HCl and ClONO₂ to active forms. This active chlorine is revealed as ClO when sunlight returns to the polar region in September. The formation of PSCs also removes HNO₃ from the gas phase (which is measured by the satellite) either by condensation into PSCs or through sedimentation of solid particles (denitrification).

In the early days of research into atmospheric chemistry it was a mystery how the observed concentrations of stratospheric ozone could be maintained based on production from solar ultraviolet radiation and loss through the direct chemical reaction of O+O₃→2O₂. Although ozone is present at levels of 1–10 parts per million by volume (ppmv) in the stratosphere, this is still a factor of 100 or more greater than the concentration of chemical species which are able to destroy it. The answer to this puzzle lies in the fact that reactive radical chemical species, present in concentrations much lower than ozone, can destroy ozone through catalytic cycles. An example is:

Note how the Cl and ClO species are reformed in this cycle and that the overall stoichiometry is that same as the direct reaction between O+O₃. In this way, chemical species containing hydrogen, nitrogen, chlorine and bromine are able to destroy significant amounts of stratospheric ozone.

However, during the 1980s our understanding of catalytic stratospheric ozone destruction cycles was far from complete. Cycles of the form given above, which catalyse the reaction O+O₃→2O₂, are very slow in the lower stratosphere because of the low concentration of atomic oxygen there. Observations (see Section 2) showed that the Antarctic ozone depletion occurred in this region on a timescale of just a few weeks. This problem was resolved by the discovery of cycles involving ClO and BrO radicals which catalyse the reaction 2O₃→3O₂, and therefore do not depend on atomic oxygen. There are two main cycles which are responsible for the rapid loss in the Antarctic ozone hole. The first is based on the formation of the ClO dimer:²⁸ 

The second involves the interaction of chlorine and bromine:²⁹ 

A minor variant of this second cycle also occurs with the initial formation of BrCl (BrO+ClO→BrCl+O₂ followed by BrCl photolysis). The ability of these cycles to quantitatively reproduce observed ozone loss is demonstrated in Section 3.4.

Atmospheric numerical models are based on our best current understanding of the relevant physics and chemistry expressed in a mathematical form. In order to be realistic, models need a representation of all of the important chemical, physical and dynamical processes; a model cannot contain unknown processes. The atmospheric models available in the 1970s and 1980s famously failed to predict the occurrence of the Antarctic ozone hole. The lack of representation of heterogeneous PSC chemistry alone would have caused this failure, but the catalytic ClO dimer cycle was also unknown and not contained in the models. Models do now include a treatment of these and other processes relevant to the Antarctic ozone hole. Nevertheless, we should not forget the fact that our understanding of atmospheric science is incomplete and our best current models can only be as good as our current understanding.

Numerical models which are now used to simulate the Antarctic ozone hole can be summarised as:

• Chemical Box or Trajectory models. These models have a detailed representation of chemistry which occurs in a single air parcel. They contain all of the relevant gas-phase and heterogeneous chemical reactions and they are computationally cheap. They can be used to integrate the chemistry at a single point or along an air mass trajectory.

• Off-Line Three-Dimensional Chemical Transport Models (3-D CTMs). These models simulate the global atmospheric domain in an array of grid-boxes. They also contain a detailed description of stratospheric chemistry. Off-line models are driven by external meteorological fields of winds and temperature³⁰  which makes them computationally efficient and well suited to interpretation of past observations.

• Coupled Chemistry-Climate Models (CCMs). These 3-D models have similar chemical schemes to CTMs but calculate their own winds and temperatures. They are therefore much more computationally expensive but these are the only models which can be used for future predictions. These models will be able to capture the coupled interaction of atmospheric chemistry and climate: for example they can simulate the impact of changing ozone concentrations on atmospheric temperature (e.g. a feedback illustrated in Figure 3).

Chemical box models which employ the observed levels of ClO (1–2 ppbv) and BrO (5–10 pptv) in the Antarctic polar vortex are able to quantitatively reproduce the observed rates of ozone loss, giving us confidence in our understanding of the cycles described above. An example is given in Figure 9 from the work of Frieler et al.³¹  With updated information on the kinetics of reaction (10) their model was able to reproduce the rates of O₃ loss derived from a Lagrangian analysis of Antarctic (and Arctic) ozonesonde observations.³² 

Figure 8 also includes results from a state-of-the-art 3-D CTM, TOMCAT/SLIMCAT. This model has a detailed treatment of polar ozone chemistry including heterogeneous chemistry on PSCs and the relevant catalytic loss cycles.³³  The model generally captures the depletion of HCl in midwinter and the production of ClO. The model also captures the decrease in HNO₃ in midwinter below about 21 km (denitrification) and the lower stratosphere ozone loss in September. A comparison of column O₃ from the same CTM for early October 2006 is shown in Figure 1. This clearly shows that the model is able to reproduce the ozone hole. Although there are some differences in detail in the comparisons in Figure 8, for example ClONO₂, the comparisons show that we now generally have a good quantitative understanding of polar ozone depletion gained through the intensive research efforts since the mid 1980s. Nonetheless, we should still be careful not to repeat the mistake of the late 1970s of assuming that our models are complete and accurate in all situations.

The Antarctic ozone hole is the largest manifestation of human impact on the ozone layer, and indeed one of the most dramatic impacts on the whole global environment. Ozone depletion has also been observed at other latitudes, though to a smaller extent.

The dynamical and chemical processes described in Section 3 also occur in the Arctic, but to smaller extent. The difference in the behaviour of the two polar regions in winter and spring is driven by the difference in surface orography and its impact on stratospheric temperatures and dynamics. Northern mid-high latitudes are much more mountainous than the southern mid-latitudes which are largely ocean. This orography causes the Arctic winter polar vortex to be smaller and more disturbed. As a result winter polar temperatures in the Arctic are warmer than the Antarctic and also much more variable. Some ‘warm’ Arctic winters may see no PSC activity at all and therefore essentially zero ozone loss. In contrast, ‘cold’ Arctic winters may experience relatively large ozone depletion, but confined to a much smaller vortex than in the south.³⁴ 

One of the largest episodes of Arctic ozone depletion occurred in the recent winter of 2010–2011.³⁵  Temperatures in this winter were unusually cold which led to extensive PSC activity and activation of chlorine in the Arctic polar vortex. Extensive ozone loss then ensued (see Figure 10). Although the Arctic loss was very large, this was not unexpected given the meteorology. Models which had been developed to model the large loss in the Antarctic (e.g. Figure 1) could reproduce the Arctic loss of 2010–2011 using the same chemistry.³⁶  Therefore, this winter turned out as a corroboration of our understanding. This recent winter did, however, show the ongoing potential for large ozone loss despite the reductions in stratospheric chlorine and bromine that have started to happen.

Ozone depletion at mid latitudes (30°S–60°S and 30°N–60°N) is much smaller than in the polar regions and occurs all year round. The largest depletion observed was a reduction of around 8% in the north and 10% in the south.⁸  This loss is also linked to the increase in atmospheric chlorine and bromine through human activities, though the activation of chlorine occurs on sulfate aerosols rather than PSCs. These aerosols are enhanced following large volcanic eruptions which reach the stratosphere. The most recent such eruption was Mt Pinatubo in June 1991 which was responsible for the lowest northern hemisphere mid-latitude ozone columns observed so far.³⁷ 

The ozone layer is naturally much thinner over the tropics (30°S–30°N) compared to mid and especially high latitudes. The observed trend in column ozone in the tropics is very small. Observations in the stratosphere do show evidence of a small downward trend but this is not apparent in the column, possibly due to compensating increases in tropospheric ozone.³⁸  This lack of large trend is consistent with our understanding of the tropics being a source region of ozone which is then transported to higher latitudes. However, the low ozone column and low average daytime solar zenith angle means that the Earth's surface in the tropics experiences large UV indices. Even in the absence of large ozone depletion, humans can experience a much larger UV dose in the tropics than in the Antarctic region in late October under the ozone hole.³⁹ 

As summarized above, stratospheric ozone depletion, especially in the Antarctic ozone hole, is caused by chlorine and bromine radicals which are released from long-lived reservoirs. This section discusses the compounds responsible for the transport of chlorine and bromine to the stratosphere and the international agreement to limit production and release of the most damaging man-made examples.

Chlorine and bromine is carried to the stratosphere in the form of source gases. These are compounds which are emitted at the surface but have long atmospheric lifetimes meaning that they are transported to the stratosphere before they degrade.^40,41  The major natural source gas of stratospheric chlorine is CH₃Cl, with a tropopause mixing ratio of 0.6 ppbv. There are a range of other anthropogenic source gases and the use of these compounds resulted in increases of around a factor 6 in chlorine and a factor 2 in bromine compared to natural levels by the late 1990s.

The major anthropogenic chlorine source gases are the CFCs, CF₂Cl₂ and CFCl₃. Their atmospheric abundances grew along with their production at ∼3% per year during much of the 1970s and 1980s due to their widespread use as aerosol propellants, refrigerants, foam blowing agents, etc. Neither compound reacts in the troposphere, making them particularly useful, but at high enough altitudes they can be broken down by UV radiation and by reaction with O(¹D). For example

Notice that both molecules carry more than one Cl to the stratosphere. The time to reach the sufficient altitudes for destruction leads to very long lifetimes, between 50 and 100 years for these compounds. In effect there is a wide range of compounds which carry ozone-depleting halogens to the stratosphere but they can be classified into a few groups which are summarized in Table 1 with some specific compounds in Table 2.

The Montreal Protocol on Substances that Deplete the Ozone Layer was a seminal international agreement signed in September 1987. In retrospect, it can be viewed as a triumph for the ‘precautionary principle’. The main motivation for the Protocol was the belief, in the 1970s and early 1980s, that CFCs may damage the ozone layer. Recall that at the time atmospheric models predicted only modest ozone depletion from CFCs. Supporters for the Protocol argued that given the very long residence times in the atmosphere, waiting for definite scientific evidence of ozone depletion (as requested by the opponents) was too dangerous. As it turned out, preparations for the Montreal Protocol were well underway when the dramatic losses in the Antarctic were discovered and attributed to chlorine.

In 1985, a treaty called the Convention for the Protection of the Ozone Layer was signed by twenty nations in Vienna. The signing nations agreed to take appropriate measures to protect the ozone layer from human activities. The Vienna Convention supported research, exchange of information, and future protocols. In response to growing concern, the Montreal Protocol was signed in 1987 and ratified in 1989. The Protocol established legally binding controls for developed and developing nations on the production and consumption of halogen source gases known to cause ozone depletion. As the scientific basis of ozone depletion became more certain in subsequent years and substitutes and alternatives became available for the principal halogen source gases, the Protocol was strengthened with Amendments and Adjustments. These added new controlled substances, accelerated existing control measures, and scheduled phaseouts of the production of certain gases. The initial Protocol called for only a slowing of CFC and halon production. The 1990 London Amendments to the Protocol called for a phaseout of the production of the most damaging ozone-depleting substances in developed nations by 2000 and in developing nations by 2010. The 1992 Copenhagen Amendments accelerated the date of the phaseout to 1996 in developed nations. Further controls on ozone-depleting substances were agreed upon in later meetings in Vienna (1995), Montreal (1997), Beijing (1999), and Montreal (2007). Figure 11 illustrates the stratospheric chlorine loading that would have occurred without the Montreal Protocol and under different amendments and adjustments. Recall that the Antarctic ozone hole first became detectable in the mid 1980s when the relative equivalent effective stratospheric chlorine (EESC) loading was 1 on the scale in the figure. Without the Montreal Protocol, EESC would have increased to about 4 by 2020 at which point severe ozone depletion would have become widespread.

The 1987 provisions of the Montreal Protocol would have only slowed the approach to the large effective chlorine values by one or more decades in the 21st century. Not until the 1992 Copenhagen Amendments and Adjustments did the Protocol show a decrease in future effective stratospheric chlorine values. Now, with full compliance to the Montreal Protocol and its Amendments and Adjustments, the emissions of the major human-produced ozone-depleting gases will ultimately be phased out and the effective stratospheric chlorine value will slowly decay to reach pre-“ozone-hole” values in the mid 21st century. This sets the timescale for the future disappearance of the Antarctic ozone hole, if other atmospheric variables remain the same (see Section 6).

The Montreal Protocol provides for the transitional use of hydrochlorofluorocarbons (HCFCs) as substitute compounds for principal halogen source gases such as CFC-12. HCFCs differ chemically from CFC source gases in that they contain H atoms in addition to the Cl and F. HCFCs are used for refrigeration, blowing foams, and as solvents, which were primary uses of CFCs. HCFCs are 1 to 10% as effective as CFC-12 in depleting stratospheric ozone because they are partially chemically removed in the troposphere by reaction with OH (see Ozone Depletion Potentials or ODPs in Table 2). In contrast, CFCs and other halogen source gases are chemically inert in the troposphere and hence reach the stratosphere without incurring any removal. Because HCFCs still contribute to the halogen abundance in the stratosphere, the Montreal Protocol requires the production and consumption of HCFCs to end in developed and developing nations by 2040.

The observed behaviour and future expected evolution of a range of chlorine and bromine source gases is shown in Figure 12. The atmospheric concentration of the most abundant CFCs has already started to decrease, slightly earlier for CFC-11 as it has an atmospheric lifetime around 50 years compared to 100 years for CFC-12. The concentration of solvents such as methyl chloroform and carbon tetrachloride is also decreasing strongly. Under the provision of the Montreal Protocol the concentration of HCFCs is still increasing, but this too will decrease in the near future. In summary, the overall atmospheric chlorine and bromine loading peaked around 1998 and has been decreasing since. The behaviour of the gases shown in Figure 12 is generally consistent with our expectations based on the Montreal Protocol and gives us confidence that the agreement is working. At present one gas whose atmospheric behaviour is not fully understood is carbon tetrachloride. It is decreasing in the atmosphere but at a rate slower than that expected based on reported emission reductions.⁴² 

The meeting to discuss the final wording of the Montreal Protocol was held in September 1987, coincidentally at the same time as the first major aircraft campaign was in the Antarctic to determine the cause of the newly discovered hole.¹⁷  The agreement would then come into force when it was ratified by at least 11 Parties representing at least two thirds of the 1986 estimated global consumption of regulated species. This happened on January 1, 1989 when 29 counties, representing 83% of global consumption, signed up.

There has been near universal participation in the agreements to protect the ozone layer: 183 countries ratified the treaty and only 11 did not. Key innovations in the Montreal Protocol process which helped ensure this were: (i) having modest initial control measures in the first instance, which allowed countries to sign up, along with the mechanism for strengthening the agreement if later evidence suggests this is necessary, and (ii) the provision for financial/technical assistance for developing countries to help implement the Protocol.

Although the MP was ratified quickly, initially there was not universal support. In particular, major developing countries (especially India and China) were dissatisfied, arguing that they had done little to cause the problem but were being required to impose the same cuts in consumption. Moreover, their economies were less able to cover the costs of switching to alternative substances. For this reason the first meeting to amend the Protocol (London 1990) set up the Multilateral Fund. This is a fund which is paid into by developed countries and then used to support the compliance of developing (so-called “Article 5”) countries. An Executive Committee of 14 members (seven each from developed and developing countries) award the money in response to requests to help fund the transition from ozone-depleting substance use to alternatives in developing countries.

The chlorine and bromine source gases listed in Table 1 are also very efficient greenhouse gases (GHGs). These molecules have strong absorption features in the IR and have fairly long atmospheric lifetimes from a few years to around a century (see Table 2). Indeed, their Global Warming Potential (GWP) is much larger than for CO₂ over the time frame of their atmospheric lifetime, as they have low abundances and atmospheric absorption is not saturated at the different wavelengths where these molecules absorb. As pointed out by Velders et al.,⁴³  by leading to the reduction in these gases, the Montreal Protocol has had a very large benefit in reducing climate change. It is estimated⁸  that by 2020 the Montreal Protocol will already have avoided adding the equivalent of 22 Gt of CO₂ to the atmosphere and will have reduced the direct radiative forcing by these compounds by 0.6 Wm⁻². This compares with the estimated actual 1750–2010 radiative forcings of ∼1.68 Wm⁻² for CO₂ emissions, ∼0.97 Wm⁻² for CH₄ emissions, and ∼0.17 Wm⁻² for N₂O emissions.

Past global stratospheric ozone depletion has partially offset some of the positive radiative forcing caused by ODS emissions. However, the net effect has been an increase in radiative forcing overall and these effects should not be separated. The overall climate impacts need to include the direct effect of the GHGs and the indirect effect through ozone depletion.⁴⁴ 

Hydrofluorocarbons (HFCs) are also used as substitute compounds for CFCs and other halogen source gases. HFCs contain no chlorine or bromine so they do not cause ozone depletion. Because of this, HFCs are not controlled by the Montreal Protocol but like all long-lived source gases they are effective GHGs (see Table 2). HFCs therefore fall under classes of gases to be controlled in treaties to prevent climate change.

While the Montreal Protocol appears to be on track to decrease stratospheric chlorine and bromine levels over the rest of this century, there are other factors which may perturb the stratospheric ozone layer, and other important ways in which a changing stratosphere may interact with the Earth system. This section discusses some of those topics which represent active areas in stratospheric research.

In recent years it has become apparent that the stratospheric bromine loading is larger than can be explained by the long-lived source gases of CH₃Br and halons alone. Observations of inorganic degradation products in the stratosphere show around 3–8 pptv additional bromine which can be explained through the transport of brominated very short-lived species (VSLS) to the stratosphere.^45,46  The term VSLS is used for species with an atmospheric lifetime of 6 months or less. The transport of these short-lived species to the stratosphere depends on rapid vertical transport to the upper troposphere, for example in tropical convection systems. The two main brominated species are bromoform (CHBr₃) and dibromomethane (CH₂Br₂) which are naturally emitted from the ocean. At present there is large uncertainty in these surface emissions. Figure 13 shows profiles of bromine calculated from a 3-D model⁴⁷  with four different surface emission datasets for CHBr₃ and CH₂Br₂. The model results show that bromine from VSLS can reach the stratosphere with a predicted range of 4–8 pptv.

Current research into halogenated VSLS therefore combines tropospheric and stratospheric chemistry. VSLS contribute about 5 (3–8) pptv bromine out of a total stratospheric loading of ∼20 pptv, i.e. ∼25%. This bromine from VSLS species reaches the polar lower stratosphere where it contributes to polar ozone loss via the BrO+ClO cycle (see Section 3). Therefore, there is a direct coupling between transport of brominated VSLS in the tropics and polar ozone loss. This contribution of natural bromine to the stratosphere may change in the future either due to changing emissions from the ocean (e.g. through changing temperature) or from changing transport. The latter process was tested in a climate model by Hossaini et al.⁴⁸  who found that the transport of CHBr₃ to the tropical upper troposphere was enhanced under large future increases in GHGs (see Figure 14).

As a consequence of the Montreal Protocol (see Section 5.2), it is predicted that the stratospheric EESC will return to 1980 levels by about the year 2050. This is expected to lead to the ‘recovery’ of the ozone layer from the effect of chlorine and bromine-induced depletion. However, the atmosphere of the latter half of this century will be very different to that of 1980. In particular, increasing greenhouse gases will cool the stratosphere and change its chemical composition. A cooler stratosphere will generally have more ozone, through slower gas-phase loss in catalytic cycles,⁴⁹  though a colder polar lower stratosphere could cause more PSCs. Increases in CH₄ will change the H₂O and HO_x (OH+HO₂) budget of the stratosphere. Increases in N₂O will lead to more NOy.⁵⁰  A prediction of the recovery of the ozone layer therefore requires detailed chemistry–climate models which combine all of the known feedbacks.

Figure 15 shows results from a number of CCMs from groups worldwide which have been used to simulate the ozone layer from 1960–2100.^51,52  The models included changing ODSs and GHGs, and model experiments were used to separate these impacts. In the Antarctic lower stratosphere, the region of the ozone hole (see Figure 15d), the models with both ODS and GHG changes show that ozone depletion peaked between 2000 and 2010 and will slowly recover for the rest of this century. The return to 1980 levels will occur by about 2060. Sensitivity experiments with just ODS or just GHG changes reveal how these two forcing terms affect O₃. In the Antarctic lower stratosphere the evolution of ozone is dominated by the change in ODSs and the impact of climate change through increasing GHGs is small. A different behaviour is predicted for the Antarctic upper stratosphere (see Figure 15c). The change in ODS alone would cause a recovery to 1980 levels by 2060. However, stratospheric cooling through increasing GHGs (see Figure 16) is driving an increase in upper stratospheric ozone which accelerates the recovery. Similar results are predicted for the Arctic, but the magnitude of the depletion is smaller.

As ozone is such a radiatively important gas, changes to its distribution in the stratosphere can affect not only the climate of the stratosphere but also the climate at the surface of the Earth. As the Antarctic ozone hole represents the largest perturbation to the ozone layer, with near 100% removal in the lower stratosphere in spring, it would be expected that such impacts would be most strongly seen at southern high latitudes. The cooling of the Antarctic lower stratosphere associated with the ozone hole has caused a delay in the breakdown of the Antarctic polar vortex.⁸  The shift in wind patterns extends to the surface during the spring and summer months and a poleward shift in the tropospheric jet has been observed.⁵³  In fact, the dynamical influence of the ozone hole on surface climate is largely manifested through changes to the structure of the Southern Annular Model (SAM).

As the ozone hole recovers in the future the reverse influences would be expected. Figure 17 shows results from a range of climate models and illustrates the negative impact on the predictions of models which fail to account for stratospheric ozone depletion.⁵⁴  The figure shows the change in DJF zonal mean zonal wind for 2050 compared to 2000 for CCMVal models, with a fully resolved stratosphere and representation of the ozone hole, and different types of models from the IPCC 4th Assessment Report (AR4). The CCMVal models (see Figure 17a) show a decreasing trend in the SH midlatitude jet, a reversal of the observed increasing trend during the past period of growing Antarctic depletion. In stark contrast, AR4 models which do not represent stratospheric ozone changes (see Figure 17d) do not show this behaviour and therefore do not produce accurate forecasts of the surface climate. These results show that climate models need to have a realistic treatment of stratospheric ozone. It should be noted that meteorological agencies are also increasing the altitude of their model top boundaries (well above the stratopause) in order to improve surface weather forecasts.³⁰ 

This article has briefly reviewed the science of the Antarctic ozone hole and the policy action taken to prevent further stratospheric ozone loss. The Montreal Protocol is currently acting to reduce stratospheric chlorine and bromine and should lead to the recovery of the ozone layer during this century. Two-way interactions between ozone and climate will likely affect the rate and extent of this recovery, however. Overall, the story of the Antarctic ozone hole and the Montreal Protocol is a seminal example of world-wide cooperative scientific research and successful policy which has allowed the world to avoid catastrophic destruction of the ozone layer.