1. Introduction
Antarctic sea ice extent (SIE) has increased by about 1% decade−1 since the introduction of reliable (satellite based) measurements in 1979 (e.g., Turner et al. 2013) and reached its highest observed value in September 2013 (Fetterer et al. 2009). The question of why Antarctic sea ice has increased in a warming world represents one of the most fundamental unsolved mysteries in polar climate science. Previous studies have suggested a number of possible explanations. Bintanja et al. (2013) suggest that freshwater input by Antarctic ice sheet melt has driven the observed sea ice trend, but Swart and Fyfe (2013) argue that the contribution of freshwater forcing is too small to explain the observed sea ice increase. Holland and Kwok (2012) showed that most of the 1992–2010 sea ice changes are forced by changing wind patterns, but it is unclear if such wind trends have been driven by natural variability [as suggested by Polvani and Smith (2013) and Swart and Fyfe (2013)] or external forcings.
Some studies have suggested that the SIE increase can be explained by atmospheric circulation changes associated with the Antarctic ozone hole. The positive correlation between intraseasonal variations in the southern annular mode (SAM) and SIE (Hall and Visbeck 2002; Sen Gupta and England 2006) could lead one to infer that the ozone hole, which has induced a positive SAM trend (e.g., Thompson et al. 2011), would indeed lead to increased Antarctic SIE. In an atmosphere-only climate model Turner et al. (2009) found a deepening of the Amundsen–Bellingshausen Sea low in response to ozone depletion, which they linked to regional features of the observed sea ice trends.
By contrast, more recent studies indicate that the ozone hole cannot explain the observed increase in Antarctic SIE. Sigmond and Fyfe (2010, hereafter SF10) directly simulated the Antarctic sea ice response to ozone depletion using a coupled atmosphere–ocean–sea ice model. Contrary to expectations they found that the ozone hole does not lead to an increase but instead to a year-round decrease in SIE. Their conclusions were consistent with other studies that have suggested that the positive trends in the SIE are unrelated to those in the SAM (Liu et al. 2004; Lefebvre et al. 2004; Simpkins et al. 2012). SF10 found that ozone depletion leads to a positive SAM response in austral summer, which mechanically drives warming of the upper ocean and induces sea ice melt. Because of the large thermal inertia of the ocean, this ocean warming persists throughout the year, causing the sea ice decrease to maximize in austral spring. Other subsequent studies using various National Center for Atmospheric Research (NCAR) climate model versions have confirmed the SF10 result that the ozone hole leads to decreased SIE (Bitz and Polvani 2012; Smith et al. 2012). However, this result has not been fully accepted by the scientific community because it has only been demonstrated for two models. Here we establish the robustness of this result between different models through the analysis of transient simulations of climate models participating in phases 3 and 5 of the Coupled Model Intercomparison Project (CMIP3 and CMIP5).
2. CMIP3 results
We analyze sea ice trends in CMIP3 simulations of the past using observed radiative forcing [twentieth-century climate simulation (20C3M)] and of the future using a moderate radiative forcing scenario (A1B simulation). Because ozone forcing was not constrained in the CMIP3 experimental setup, some CMIP3 models included time-varying stratospheric ozone (i.e., ozone depletion for the past and ozone recovery for the future) while other models were forced with monthly climatological ozone fields that do not change from year to year. This has provided a unique opportunity to derive the effect of ozone depletion and recovery on climate by comparing climate trends in CMIP3 models with and without time-varying stratospheric ozone. The assumption here is that the trend difference caused by the different ozone forcing between the two groups is larger than the difference caused by the fact that both groups are composed of different models that may have different sensitivities to ozone forcing. This assumption was shown to be valid for various climate variables (Son et al. 2009), which encouraged us to repeat the analysis for Antarctic SIE, which is defined as the total area with at least 15% ice cover. For our analysis, we employ all models of Son et al. (2009) for which sea ice variables were available but exclude the Flexible Global Ocean–Atmosphere–Land System Model gridpoint, version 1.0 (FGOALS-g1.0), which suffered from an unusually large sea ice bias. Details of the CMIP3 models can be found in Table 1. The results of this analysis are shown in Fig. 1 and can be summarized as follows:
For the past, the group with ozone depletion shows a statistically significant decrease in SIE, while the sea ice response is not statistically different from zero in the constant ozone forcing group. We thus find that ozone depletion is associated with decreased SIE, which is consistent with SF10 and Bitz and Polvani (2012). It has to be noted though that the sea ice trend in the ozone depletion group is not well separated from that in the constant ozone forcing group as the uncertainty bars (which represent the 95% confidence intervals) overlap strongly.
For the future, the simulated sea ice decrease is smaller in the group with ozone recovery than in the group with constant ozone forcing. In other words, ozone recovery acts to mitigate the future sea ice decrease associated with increased greenhouse gases, which is consistent with Smith et al. (2012). The uncertainty bars of the two groups overlap, but the overlap is smaller than for the past trends.
CMIP3 Models used in this study.
Past and future Antarctic SIE trends in CMIP3 models. For each model the linear trends are computed for 1961–99 in the 20C3M integrations and 2000–49 in the A1B scenario integrations. The mean trend and its 95% confidence interval is shown by the red circle and bar for the group of models with ozone depletion and by the blue square and bar for the group of models with ozone recovery. The mean trend and confidence interval for models without time-varying stratospheric ozone is shown by the green symbols and bars. SIE is defined as the total area with at least 15% ice cover.
Citation: Journal of Climate 27, 3; 10.1175/JCLI-D-13-00590.1
3. CMIP5 results
a. Impact of ozone depletion
In this section, we analyze all CMIP5 models from which the ozone impact can be derived. To isolate the ozone impact, single forcing runs are required, which were available for six CMIP5 models (Table 2). We first consider the response of the atmospheric circulation as quantified by the SAM index. The red bars in Fig. 2 show the historical trend of the December–February (DJF) SAM index in the six CMIP5 models with available ozone attribution runs. Consistent with previous modeling studies (e.g., Son et al. 2009; SF10; McLandress et al. 2011) all models show a statistically significant positive SAM trend (note the inverse scale on the left axis: the error bars represent the 5%–95% confidence interval and are calculated from the time series of the mean of all ensemble members for each model). The mean trend and its confidence internal (shown in Fig. 2 on the right) is calculated from the mean time series averaged over all 27 available CMIP5 ensemble members. As a side result, we find that the poleward jet shift associated with the positive SAM trend is stronger for models with a more equatorward jet position. The correlation between the mean SAM (averaged between 1951 and 2005) and the SAM trend over that period is −0.89. Such a relationship has already been identified in previous studies for the response to increasing greenhouse gases (Kidston and Gerber 2010), but this is the first time that this relationship is reported for the response to ozone depletion. It is consistent with the fluctuation–dissipation theory and suggests that current climate models, which tend to have an equatorward bias in the climatological jet position, may overestimate the SAM response to ozone variations.
Number of ensemble members available for historical CMIP5 simulations that combine all anthropogenic and natural forcings and for single forcing simulations of ozone, greenhouse gas, and anthropogenic aerosols.
Ozone induced trends in the SAM (DJF) and Antarctic SIE (annual mean) in CMIP5 models. Red bars show the SAM trend (inverted scale on the left axis), purple bars show the SIE trend (scale on right axis), and the black error bars show the 5%–95% confidence intervals as calculated from the ensemble mean time series. Trends are shown for 1951–2005 and all CMIP5 models with available ozone attribution simulations. The SAM is defined as the zonal mean sea level pressure difference between 40° and 65°S. Here and in subsequent plots, the mean trend and its error bar (labeled MEAN) are calculated from the time series averaged over all ensemble members and models.
Citation: Journal of Climate 27, 3; 10.1175/JCLI-D-13-00590.1
Since all CMIP5 models considered in this study reproduce the observed positive correlation between intraseasonal variations in the SAM and SIE (not shown), the positive SAM response suggests that SIE might increase in response to the ozone forcing. On the other hand, previous studies have shown that the positive SAM response does not lead to increased but instead to decreased SIE. To investigate the robustness of this result, we now turn to the sea ice response in the CMIP5 ozone attribution runs. The key result of this study is depicted by the purple bars in Fig. 2. They show that the annual mean Antarctic SIE trends in the ozone attribution runs are significantly negative for all six models. In other words, all CMIP5 models consistently show that the positive SAM associated with ozone depletion induces decreased Antarctic SIE, confirming the surprising conclusions of previous single model studies. The SIE response calculated from the time series averaged over all 27 CMIP5 ensemble members is −0.13 ± 0.02 × 106 km2 decade−1, which, over the entire 1951–2005 period, is equivalent to −0.69 ± 0.12 × 106 km2 (95% confidence interval). This value is quite similar to that found with time-slice simulations in SF10 (−0.55 × 106 km2) and comparable to the values found in Bitz and Polvani (2012) (from −0.77 to −0.85 × 106 km2) and Smith et al. (2012) (−0.47 × 106 km2).
The ozone forcing in the CMIP5 runs combine tropospheric and stratospheric forcing. Although tropospheric ozone forcing is thought to affect the NH circulation (Allen et al. 2012), additional simulations with the CanESM2 indicate that the effect of tropospheric ozone forcing on the simulated SAM and Antarctic SIE trends is small. Over the 1951–2005 period, CanESM2 simulations with only stratospheric ozone forcing exhibited a DJF SAM trend of 0.32 ± 0.27 hPa decade−1 and an annual mean Antarctic SIE trend of −1.26 ± 0.46 × 106 km2. These trends are statistically indistinguishable from the trends due to the combined stratospheric and tropospheric ozone forcing depicted in Fig. 2 (0.44 ± 0.30 hPa decade−1 and −0.94 ± 0.50 × 106 km2 decade−1, respectively). In other words, the results of these additional simulations indicate that the SAM and SIE trends in the CMIP5 ozone attribution simulations are primarily driven by stratospheric ozone depletion.
Figure 3 shows the seasonality of the SAM and SIE responses. We find that, consistent with previous modeling studies, the SAM trend in most models peaks in DJF. Statistically significant SAM responses are also found for March–May (MAM) in the L’Institut Pierre-Simon Laplace (IPSL) and the two Goddard Institute for Space Studies (GISS) models, in June–August (JJA) for the Commonwealth Scientific and Industrial Research Organisation (CSIRO) model, and for September–November (SON) in the two GISS models. Although the response of the atmospheric circulation (as quantified by the SAM) generally peaks in DJF, most models show negative SIE trends in all seasons. In fact, the SIE response tends to peak in SON, the season with the largest climatological SIE, which is consistent with the findings of SF10.
Ozone induced SAM and Antarctic SIE trends in CMIP5 models as a function of season and their 5%–95% confidence intervals.
Citation: Journal of Climate 27, 3; 10.1175/JCLI-D-13-00590.1
b. Impact of ozone depletion in the context of other forcings
To place the SAM and SIE responses to ozone depletion in the context of other historical forcings, we analyze historical simulations that combine all anthropogenic and natural forcings (ALL) and historical simulations in which only greenhouse gases (GHG) or anthropogenic aerosols (AntrAer) are varied in time (Table 2). Figure 4 (top) shows that in response to all historical forcings, all CMIP5 models considered here show a positive and statistically significant SAM trend in DJF. The residual SAM trends are statistically indistinguishable from zero, which indicates linear additivity of the SAM responses to the historical forcings. In half of the models (CCSM4 and the two GISS models) ozone depletion is the dominant driver of the historical SAM trend, which is consistent with the general scientific consensus (e.g., Thompson et al. 2011). However, it is interesting to note that in the other half of the models the contribution of the GHG forcing is similar to or even larger than that of ozone depletion. Even though this finding may seem to contradict previous literature, we note that some other studies (Staten et al. 2011; Hardiman et al. 2012; Fyfe et al. 2012) have also reported on a more prominent role of historical GHG forcing in driving the historical SAM trend in DJF. This is an interesting finding that is left for further investigation. Finally, we note that in four of the five models with AntrAer simulations, historical trends in anthropogenic aerosols act to decrease the SAM, which is consistent with Gillett et al. (2013).
Historical DJF SAM and annual mean Antarctic SIE trends due to all combined anthropogenic and natural forcings, time-varying ozone, greenhouse gas, and anthropogenic aerosol forcing and the residual trend. The residual trend and its uncertainty is calculated from the time series of the difference between ALL and the sum of the O3, GHG, and AntrAer simulations.
Citation: Journal of Climate 27, 3; 10.1175/JCLI-D-13-00590.1
Figure 4 (bottom) shows that all CMIP5 models simulate a statistically significant decrease of Antarctic SIE in response to all historical forcings. The SIE response in this subset of CMIP5 models is consistent with that found in the entire CMIP5 archive (Turner et al. 2013). The single forcing simulations show that in most models (all except the two GISS models) the dominant driver of the historical sea ice trend is the GHG forcing, with ozone depletion playing a secondary role. This is not surprising as GHG forcing presumably combines dynamical (SAM) and radiative forcing changes, which both facilitate decreased SIE. Finally we note that, in response to historical anthropogenic aerosol changes, four of the five models show increased SIE. This SIE increase is consistent with the SAM decrease found in response to anthropogenic aerosols in those models.
4. Summary and discussion
Many recent studies have attempted to explain the increase of Antarctic sea ice since the 1970s. Some have suggested that atmospheric circulation trends driven by the Antarctic ozone hole can explain the sea ice increase, while the results of recent single model studies are inconsistent with this view. In this study, we have analyzed all available CMIP3 and CMIP5 model simulations suitable to address this question and found new modeling evidence that ozone depletion is not associated with increased but instead with decreased Antarctic sea ice extent (SIE).
We caution that our conclusion is based on climate models that are imperfect, especially regarding the simulation of the Antarctic climate (e.g., Turner et al. 2013). However, we note that the negative SIE response to ozone depletion is robustly found for all six CMIP5 models, which each have a distinct Antarctic sea ice climatology and variability as shown by Fig. 2 of Zunz et al. (2013). Hence, our main result does not seem to depend on the quality of the simulation of Antarctic sea ice.
Based on this analysis, it cannot be ruled out that all coupled climate models lack or misrepresent critical physical processes, resulting in unrealistic simulations of the SIE response to stratospheric ozone forcing. However, we do not have reason to believe that is the case. The statistical analysis of Swart and Fyfe (2013) showed that when accounting for internal variability, the average CMIP5 sea ice area trend is statistically consistent with the observed trend, suggesting that the SIE response to external forcings simulated by these models are credible. Other studies have also highlighted the large internal variability, which may play a critical role in explaining the observed increase in Antarctic sea ice cover (Zunz et al. 2013; Polvani and Smith 2013; Mahlstein et al. 2013).
In summary, the CMIP3 and CMIP5 model results presented here provide important confirmation of the surprising conclusions of previous single model studies that processes not linked to stratospheric ozone depletion must be invoked to explain the observed increase in Antarctic sea ice.
Acknowledgments
We thank Bill Merryfield, Neil Swart, and three anonymous reviewers for their insightful comments. We acknowledge the modeling groups, the Program for Climate Model Diagnosis and Intercomparison, and the WCRP Working Group on Coupled Modelling for their roles in making available the WCRP CMIP multimodel datasets. Support of this dataset is provided by the Office of Science, U.S. Department of Energy.
REFERENCES
Allen, R. J., S. C. Sherwood, J. R. Norris, and C. S. Zender, 2012: Recent Northern Hemisphere tropical expansion primarily driven by black carbon and tropospheric ozone. Nature, 485, 350–354, doi:10.1038/nature11097.
Bintanja, R., G. J. van Oldenborgh, S. S. Drijfhout, B. Wouters, and C. A. Katsman, 2013: Important role for ocean warming and increased ice-shelf melt in Antarctic sea-ice expansion. Nat. Geosci., 6, 376–379, doi:10.1038/ngeo1767.
Bitz, C. M., and L. M. Polvani, 2012: Antarctic climate response to stratospheric ozone depletion in a fine resolution ocean climate model. Geophys. Res. Lett., 39, L20705, doi:10.1029/2012GL053393.
Fetterer, F., K. Knowles, W. Meier, and M. Savoie, 2009: Sea ice index. National Snow and Ice Data Center, Boulder, CO, digital media, doi:10.7265/N5QJ7F7W.
Fyfe, J. C., N. P. Gillett, and G. J. Marshall, 2012: Human influence on extratropical Southern Hemisphere summer precipitation. Geophys. Res. Lett., 39, L23711, doi:10.1029/2012GL054199.
Gillett, N. P., J. C. Fyfe, and D. E. Parker, 2013: Attribution of observed sea level pressure trends to greenhouse gas, aerosol, and ozone changes. Geophys. Res. Lett., 40, 2302–2306, doi:10.1002/grl.50500.
Hall, A., and M. Visbeck, 2002: Synchronous variability in the Southern Hemisphere atmosphere, sea ice, and ocean resulting from the annular mode. J. Climate,15, 3043–3057.
Hardiman, S. C., N. Butchart, T. J. Hinton, S. M. Osprey, and L. J. Gray, 2012: The effect of a well-resolved stratosphere on surface climate: Differences between CMIP5 simulations with high and low top versions of the Met Office Climate Model. J. Climate, 25, 7083–7099.
Holland, P. R., and R. Kwok, 2012: Wind-driven trends in Antarctic sea-ice drift. Nat. Geosci., 5, 872–875, doi:10.1038/ngeo1627.
Kidston, J., and E. P. Gerber, 2010: Intermodel variability of the poleward shift of the austral jet stream in the CMIP3 integrations linked to biases in 20th century climatology. Geophys. Res. Lett., 37, L09708, doi:10.1029/2010GL042873.
Lefebvre, W., H. Gossee, R. Timmermann, and T. Fichefet, 2004: Influence of the southern annular mode on the sea ice–ocean system. J. Geophys. Res., 109, C09005, doi:10.1029/2004JC002403.
Liu, J., J. A. Curry, and D. G. Martinson, 2004: Interpretation of recent Antarctic sea ice variability. Geophys. Res. Lett., 31, L02205, doi:10.1029/2003GL018732.
Mahlstein, I., P. R. Gent, and S. Solomon, 2013: Historical Antarctic mean sea ice area, sea ice trends, and winds in CMIP5 simulations. J. Geophys. Res. Atmos.,118, 5105–5110, doi:10.1002/jgrd.50443.
McLandress, C., T. G. Shepherd, J. F. Scinocca, D. A. Plummer, M. Sigmond, A. I. Jonsson, and M. C. Reader, 2011: Separating the dynamical effects of climate change and ozone depletion. Part II: Southern Hemisphere troposphere. J. Climate, 24, 1850–1868.
Polvani, L. M., and K. L. Smith, 2013: Can natural variability explain observed Antarctic sea ice trends? New modeling evidence from CMIP5. Geophys. Res. Lett.,40, 3195–3199, doi:10.1002/grl.50578.
Sen Gupta, A., and M. H. England, 2006: Coupled ocean–atmosphere–ice response to variations in the southern annular mode. J. Climate, 19, 4457–4486.
Sigmond, M., and J. C. Fyfe, 2010: Has the ozone hole contributed to increased Antarctic sea ice extent? Geophys. Res. Lett., 37, L18502, doi:10.1029/2010GL044301.
Simpkins, G. R., L. M. Ciasto, D. W. J. Thompson, and M. H. England, 2012: Seasonal relationships between large-scale climate variability and Antarctic sea ice concentration. J. Climate, 25, 5451–5469.
Smith, K. L., L. M. Polvani, and D. R. Marsh, 2012: Mitigation of 21st century Antarctic sea ice loss by stratospheric ozone recovery. Geophys. Res. Lett., 39, L20701, doi:10.1029/2012GL053325.
Son, S.-W., N. F. Tandon, L. M. Polvani, and D. W. Waugh, 2009: Ozone hole and Southern Hemisphere climate change. Geophys. Res. Lett., 36, L15705, doi:10.1029/2009GL038671.
Staten, P. W., J. J. Rutz, T. Reichler, and J. Lu, 2011: Breaking down the tropospheric circulation response by forcing. Climate Dyn., 39, 2361–2375, doi:10.1007/s00382-011-1267-y.
Swart, N. C., and J. C. Fyfe, 2013: The influence of recent Antarctic ice sheet retreat on simulated sea ice area trends. Geophys. Res. Lett., 40, 4328–4332, doi:10.1002/grl.50820.
Thompson, D. W. J., S. Solomon, P. J. Kushner, M. H. England, K. M. Grise, and D. J. Karoly, 2011: Signatures of the Antarctic ozone hole in Southern Hemisphere surface climate change. Nat. Geosci., 4, 741–749, doi:10.1038/ngeo1296.
Turner, J., and Coauthors, 2009: Nonannular atmospheric circulation change induced by stratospheric ozone depletion and its role in the recent increase of Antarctic sea ice extent. Geophys. Res. Lett., 36, L08502, doi:10.1029/2009GL037524.
Turner, J., T. J. Bracegirdle, T. Phillips, G. J. Marshall, and J. S. Hosking, 2013: An initial assessment of Antarctic sea ice extent in the CMIP5 models. J. Climate, 26, 1473–1484.
Zunz, V., H. Goosse, and F. Massonnet, 2013: How does internal variability influence the ability of CMIP5 models to reproduce the recent trend in Southern Ocean sea ice extent? Cryosphere,7, 451–468, doi:10.5194/tc-7-451-2013.
