Stratospheric Ozone in the Post-CFC Era

The stratospheric ozone layer protects life on the Earth from the Sun's harmful ultraviolet radiation. In the middle 1970s scientists found that this natural shield could be depleted by human made chlorine and bromine containing chemicals, in particular, chlorofluorocarbons (CFCs), which were widely used as refrigerants, foaming agents, and in many other applications. Stratospheric ozone depletion can greatly increase the rates of skin cancer and cataracts, and can significantly diminish global crop yields. By the mid-1 980s, clear evidence of ozone depletion was found over Antarctica, the so-called, "Antarctic ozone hole." In response to this threat, governments around the world signed an international agreement, the 1987 Montreal Protocol with its subsequent amendments and adjustments, to phase out CFCs and other ozone-depleting substances (ODs). With the compliance with these agreements, it is projected that the ODs amounts in the stratosphere will return to pre-1980 levels in the 2060s. The ODs level, however, is not the only factor affecting ozone recovery. Changes in temperature and transport in the stratosphere, because of greenhouse gases (GHG), could also have significant impacts on ozone recovery. In this study, using simulations from the Goddard Earth Observing System Coupled Chemistry-Climate Model (GEOS CCM), we try to answer an outstanding scientific question: how does climate change affect ozone recovery? The model simulations show that the ODs will recover to 1980 values in the 2060s, but stratospheric ozone in the 2060s differs significantly from that in 1980. In fact, the global total ozone levels in 2060 will be higher than in the 1980s with peak increases of 23 DU (6%) in the extratropics, "a super recovery". On the other hand, the tropical column ozone, however, does not recover to 1980 levels up to the Abstract Vertical and latitudinal changes in the stratospheric ozone in the post-chlorofluorocarbon (CFC) era are investigated using simulations of the recent past and the 21st century with a coupled chemistry-climate model. Model results reveal that, in the 2060s when the stratospheric halogen loading is projected to return to its 1980 values, the extratropical column ozone is significantly higher than that in 1975-1984, but the tropical column ozone does not recover to 1980 values. Upper and lower stratospheric ozone changes in the post-CFC era have very different patterns. Above 15 hPa ozone increases almost latitudinally uniformly by 6 Dobson Unit (DU), whereas below 15 hPa ozone decreases in the tropics by 8 DU and increases in the extratropics by up to 16 DU. The upper stratospheric ozone increase is a photochemical response to greenhouse gas induced strong cooling, and the lower stratospheric ozone changes are consistent with enhanced mean advective transport due to a stronger Brewer-Dobson circulation. The model results suggest that the strengthening of the Brewer-Dobson circulation plays a crucial role in ozone recovery and ozone distributions in the post-CFC era. ozone decrease outweighs the upper stratospheric ozone increase. These results suggest that circulation changes play a larger role in ozone recovery than previously thought.


I Introduction
The stratospheric ozone layer is expected to recover to pre-1980 levels in the middle of the 2lSt century with the projected decline of the stratospheric halogen loading (WMO, 2007).
Coupled chemistry-climate model (CCM) simulations have found that the recovery of the stratospheric ozone and halogen to 1980 levels will not happen at the same time, because ozone recovery is strongly dependant on temperature and transport, which in the middle 21St century, are significant different from those in the 1980s (Eyring et al., 2007). Increased greenhouse gases (GHG) will cool the stratosphere, which will lead to an increase in ozone due to the temperature dependence of the chemical reactions involved in ozone loss (Barnett et al., 1975). Additionally, model simulations find that the Brewer-Dobson circulation will speed up with the increase in GHGs (e.g., Butchart et al., 2006), resulting in more ozone transport to the mid-high latitudes and could advance ozone recovery in the extratropical region (Austin and Wilson, 2006).
An important aspect of ozone recovery is the vertical and latitudinal characteristics of the ozone distribution in the post-chlorofluorocarbon (CFC) era, which has not been investigated in detail by previous work. Changes in temperature and transport have very different effects on ozone abundance in the upper and lower stratosphere because of the ozone photochemical lifetime differences in the two regions (Shepherd, 2008). The acceleration of the Brewer-Dobson circulation will change the latitudinal distribution of ozone by bringing more ozonepoor air into the tropical lower stratosphere and more ozone-rich air into the mid-high latitudes. A detailed examination of the vertical and latitudinal structure of ozone changes will help us to understand how the ozone layer is affected by climate change.
This paper investigates the impacts of climate change on ozone recovery using simulations of the recent past (late 20"' century) and future (2lSt century) from the Goddard Earth Observing System (GEOS) CCM (Pawson et al., 2008). We focus on vertical and latitudinal ozone changes after the stratospheric halogen loading returns to 1980 levels. Decadal differences in the stratospheric ozone between 2060s and 1975-1984 are examined when the halogen amounts are nearly the same during these two decades. These differences are interpreted as mainly caused by climate change and are consistent with an increased Brewer-Dobson circulation in the lower stratosphere and strong cooling in the upper stratosphere.

Model Simulations
This study uses past and future climate simulations from GEOSCCM Version 1, which is based on the GEOS-4 General Circulation Model. It includes a comprehensive stratospheric chemistry scheme that is coupled with the physical processes through the radiation code. The model has a horizontal resolution of 2" x 2.5" and 55 vertical levels with a model top at 0.01 hPa. A detailed description of GEOSCCM Version 1 is given by Pawson et al. (2008).
A number of time-slice and time-dependent simulations of the recent past (195 1-2005) and future (200 1-2099) have been performed using the GEOSCCM Version 1. Model results have been evaluated and analyzed to address several different scientific questions (Eyring et al., 2006(Eyring et al., , 2007Pawson et al., 2008;Perlwitz et al., 2008). Overall, model simulations of the recent past agree reasonably well with observations in stratospheric dynamics, transport, and ozone distribution and depletion (Eyring et al., 2006(Eyring et al., ,2007Pawson et al., 2008).
For this study, two sets of simulations for the period 1975-2070 are used. The first set is forced with standard scenarios of greenhouse gases (GHG) and ozone-depleting substances (ODS) (Eyring et al., 2006). Sea-surface temperature and sea-ice amounts (SST) are born integrations of the NCAR Community Climate System Model. The external forcings used for the second simulation differ from those in the first one in that, 1. the halogen concentrations are fixed at 1960 levels, and 2. SST data for the late 2oth century are from observations. These two sets of simulations are referred to as FREF (future reference runs) and C160 (fixed 1960 halogen amount simulations).

Model Results
The WMO Scientific Assessment of Ozone Depletion report (2007) estimates that full recovery of ozone will likely occur when the Equivalent Effective Stratospheric Chlorine (EESC) returns to pre-1980 levels. EESC is a measure to quantify the effects of chlorine and bromine containing halogens on ozone depletion in the stratosphere (Newman et al., 2007).
Evolution of EESC has the largest effect on ozone in the upper stratosphere and polar lower stratosphere. Figure 1 shows that EESC in these regions returns to 1980 values by the early 2060s in FWF. EESC in the tropics and mid-latitudes reach 1980 levels during the 2040s (not shown). Figure 1 also shows that, except in the tropics, the column ozone returns to 1980 values much earlier than the EESC does. The earliest ozone recovery occurs in the Northern Henlisphere (NH) mid-latitudes and the Arctic at about 2025. Recovery of the global, Southern Hemisphere (SH) mid-latitude, and Antarctic column ozone happens between 2040-2045. In the early 2060s when the EESC returns to 1980 values, ozone amounts in these regions are higher than their 1980 values, indicating a super recovery. The tropical ozone is an exception. It is least affected by EESC with a maximum depletion of less than 3%, but it never recovers to 1980 values during the simulations.
The ozone evolution from 1980-2070 in FREF is similar to many other CCM projections (Eyring et al., 2007). The different recovery dates between ozone and EESC strongly suggest that GHG increases have a significant impact on ozone recovery. The focus of this study is the effects of climate change on the vertical and latitudinal ozone changes. In order to separate the contributions to ozone changes from ODs and GHG increase, we examine the decadal differences in ozone between 1975-1984 and 2060-2069. In FREF simulations, EESC has recovered to pre-1980 values in the 2060s. The differences in EESC between 2060-2069 and 1975-1984 are within 10% in regions that are mostly affected by ODs, i.e., the upper stratosphere and polar regions. Therefore we interpret decadal differences in the stratospheric ozone and dynamics between these two decades are mostly due to GHG increase.
Ozone changes between 2060-2069 and 1975-1984   Changes of the lower stratospheric ozone appear to support an enhanced advection. In a simplified view, the stratospheric transport consists of a mean advection by the Brewer-Dobson circulation and horizontal mixing. The horizontal mixing in CCMs is very difficult to be quantitatively diagnosed as an underlying general theory is still missing, but we can assess changes in the mean transport by examining the residual circulation. Figure   suggesting that lower stratospheric ozone changes are controlled mainly by changes in the mean advection. Interestingly, mean transport also shows inter-hemispheric differences in the mid-high latitudes in the lower stratosphere. It is likely that the reduced ozone advection in 50's-70's is responsible for a smaller SH ozone increase in this region. Note that the details of the lower stratosphere changes between the mean transport and ozone abundance are quite different. Overall, changes in ozone are smoother and have a smaller latitudinal gradient compared with changes in the mean transport. These differences are most likely due to horizontal mixing which acts to smooth tracer concentrations across latitude.
Strengthening of the Brewer-Dobson circulation is a robust response to GHG increase in model simulations (Butchart et al., 2006;Li et al., 2008). Butchart and Scaife (2001) found  Fig. 3) are in good agreement with those of the lower stratospheric ozone changes in FREF. These similarities support our argument that lower stratospheric ozone changes can be largely attributed to an enhanced mean ozone transport.
In the upper stratosphere, changes in ozone abundance and mean ozone transport have differing patterns. The mean advection shows increases in the tropics and decreases in the extratropics, whereas ozone displays a nearly latitudinal uniform increase. The opposite changes between ozone concentrations and mean advection in the extratropical upper stratosphere indicate that long-term ozone change is not determined by transport in this region, where ozone is in photochemical equilibrium (KO et al., 1989). One may also notice that the regions of negative advection in the tropics, where mixing is weak, extend to 5 hPa ( Fig. 2f), but ozone decrease is confined below 15 hPa (Fig. 2b). We think this difference suggests a shift from transport control to chemical control of ozone across 10 hPa. We estimated the lifetimes for mean ozone advection Tadv and photochemical ozone loss ~1,,,, and z:~~ = -E, where is the zonal mean ozone loss rate.
[4] m ?Y The chemical loss lifetime decreases very sharply with height. In the tropics (20"s-20W), the chemical loss lifetime is more than 100 times longer than the advective lifetime at 70 hPa, and becomes comparable to the advective lifetime at 20 and 30 hPa (100-200 days), and is 10 times shorter than the advective timescale at 10 hPa. Because of the longer advective timescale compared with the ozone photochemical lifetime above 10 hPa, strengthening of the Brewer-Dobson circulation has little effect on the upper stratospheric ozone.
Another approach to demonstrate ozone changes caused by GHG increase is to use the Cl60 simulation in which the halogen loading is fixed at 1960 levels. Differences between 2060s and 1975-1984 in C160 must be caused by climate change. The decadal differences in the stratospheric ozone, temperature, and mean transport between 2060-2069 and 1975-1984 in C160 are qualitatively very similar to those in FREF. Figure3 displays these C160 decadal differences, and shows that FREF differences (Fig. 2) are very well reproduced in C160. These similarities include: small decrease in the tropical total ozone, significant increase in the extratropical ozone with a larger peak value in the NH than in the SH, almost uniformly 6 DU upper stratospheric ozone increase, enhanced tropical upwelling, and increased extratropical downwelling (except in the Arctic lower stratosphere and a band around 60%). The most notable difference between Figures 2 and 3 is that C160 has a larger ozone increase in the Arctic lower stratosphere (Fig. 3c), which appears to be caused by a stronger downwelling in the NH in C160 (Figs. 3e-f). There are other quantitative differences between Figs. 2 and 3, which are likely caused by different SSTs used in simulation of the recent past, difference in EESC, and model internal variability. The overall qualitative agreement between C160 and FREF results, however, strongly support our interpretations of Fig. 2 as mostly attributed to climate change.

Discussion and Conclusions
The GEOSCCM simulates EESC recovery to 1980 values in the 2060s. The decadal differences in the stratospheric ozone and dynamics between 2060-2069 and 1975-1984  Many CCMs have predicted that extratropical column ozone will be higher than 1980 values when EESC return to pre-1980 levels (Eyring et al., 2007). Previous work has attributed the extratropical ozone "super recovery" to increase in the middle to upper stratospheric ozone due to GHG-induced stratospheric cooling (Eyring et al., 2007). Here we show that, in the GEOSCCM simulations when EESC returns to pre-1980 levels, lower stratospheric ozone increases make significant contribution to the "super recovery" of the extratropical total ozone. In the NH extratropics, increases in the lower stratospheric ozone exceed those in the middle to upper stratosphere. Model results also project that the tropical total column ozone will not recover to 1980 values. In the tropics, the two major mechanisms through which climate change affect the stratospheric ozone, upper stratospheric cooling and strengthening of the Brewer-Dobson circulation, have opposite effects on ozone abundance. The tropical lower stratospheric ozone decrease outweighs the upper stratospheric ozone increase. These results suggest that circulation changes play a larger role in ozone recovery than previously thought.
The model simulated mean ozone transport changes are qualitatively consistent with the lower stratospheric ozone changes, suggesting that changes of ozone advection play a crucial role in ozone recovery. One interesting finding in this study is the inter-hemispheric differences in