SEARCH

SEARCH BY CITATION

Keywords:

  • extratropical cyclones;
  • stratospheric ozone;
  • Southern Hemisphere

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

This study explores the impact of Antarctic stratospheric ozone depletion on extratropical cyclones. Output from the Community Atmosphere Model is combined with a Lagrangian cyclone-tracking algorithm to identify the response of Southern Hemisphere extratropical cyclones to ozone and greenhouse gas forcings over the period 1960–2000. Stratospheric ozone depletion induces a significant poleward shift in cyclone frequency over the Southern Ocean, but has minimal influence on cyclone intensity and lifetime. The response of the cyclones to late 20th century greenhouse gas increases has similar characteristics, but falls within the range of natural variability in the model.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

Extratropical cyclones not only govern many of the observed weather and precipitation patterns at mid- and high latitudes but also directly contribute to the heat, momentum, and moisture fluxes that play a central role in the general circulation of the atmosphere. Hence, it is important to understand how the frequency and intensity of the cyclones respond to anthropogenic forcings in the climate system. Numerous studies have investigated the effect on extratropical cyclones of projected increases in greenhouse gases over the 21st century using a wide range of methodologies (Geng and Sugi, 2003; Bengtsson et al., 2006; Lambert and Fyfe, 2006; Bengtsson et al., 2009; Ulbrich et al., 2009; Booth et al., 2012; Ulbrich et al., 2013), including recent analyses of the global climate model projections from phase 5 of the Coupled Model Intercomparison Project (CMIP5) (Chang et al., 2012; Mizuta, 2012; Zappa et al., 2013). One of the more robust findings in these studies is a poleward shift in the zonal-mean frequency of extratropical cyclones, particularly in the Southern Hemisphere (SH) (Kushner et al., 2001; Yin, 2005; Bengtsson et al., 2006; Chang et al., 2012). Changes in the intensity of the cyclones are, however, less clear and appear to reflect competing influences predicted in increasing greenhouse gas scenarios. The predicted decrease in the equator-to-pole surface temperature gradient and increase in tropospheric static stability might be expected to weaken the cyclones (Geng and Sugi, 2003), whereas the predicted increase in tropospheric moisture content and the equator-to-pole upper tropospheric temperature gradient might be expected to strengthen the cyclones (Booth et al., 2012; Mizuta, 2012).

Another key anthropogenic forcing in the SH is the depletion of Antarctic stratospheric ozone associated with chlorofluorocarbons. The emergence of the Antarctic ozone hole over the last two decades of the 20th century has strengthened the SH stratospheric polar vortex during austral spring and contributed to a poleward shift in the SH mid-latitude jet at tropospheric levels during austral summer. The linkage between the Antarctic ozone hole and changes in the SH tropospheric circulation has been documented both in observations and in climate model experiments (Thompson et al., 2011 for a recent review). In fact, modelling evidence suggests that the SH summertime tropospheric circulation changes associated with the ozone hole have been larger than those associated with increasing greenhouse gases over the second half of the 20th century (McLandress et al., 2011; Polvani et al., 2011a). Over the first half of the 21st century, the SH summertime tropospheric circulation changes associated with the recovery of the ozone hole are projected to counteract those associated with increasing greenhouse gases (Perlwitz et al., 2008; Son et al., 2008; Arblaster et al., 2011; Polvani et al., 2011b).

While the impact of Antarctic stratospheric ozone depletion on large-scale tropospheric circulation changes is by now well established, it is unknown whether ozone depletion affects the characteristics of individual extratropical cyclones, and whether its effects are distinct from those associated with increasing greenhouse gases. Several studies have estimated long-term trends in SH extratropical cyclones using reanalysis data sets (Simmonds and Keay, 2000; Wang et al., 2006; Wang et al., 2013), but large uncertainties remain because of temporal discontinuities in the observations assimilated into the reanalyses. Thus, in this study, we choose to examine the response of SH extratropical cyclones to stratospheric ozone depletion using a set of carefully designed climate model experiments, in which observed levels of ozone and greenhouse gases are independently prescribed during the pre-ozone hole (1960) and ozone hole (2000) periods. We identify individual cyclones in the model using a Lagrangian cyclone tracking algorithm and document how the characteristics of the cyclones respond to the ozone and greenhouse gas forcings. The results demonstrate that, during SH summer, the ozone depletion induces a robust poleward shift in the frequency of the cyclones but has little impact on their intensity: this is similar to, but larger than, the response of the cyclones to increasing greenhouse gases in the late 20th century.

2. Methodology

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

The results of this study are based on climate model experiments with the National Center for Atmospheric Research (NCAR) Community Atmosphere Model version 3 (CAM3) (Collins et al., 2006). The model is integrated with a standard T42 horizontal resolution (approximately 2.8° × 2.8° latitude-longitude grid), 26 hybrid vertical levels, and a model top at 2.2 hPa. Sea surface temperatures and sea ice concentrations are prescribed using the Hadley Centre dataset (Rayner et al., 2003). Further details of the model configuration are described in Polvani et al. (2011a).

Here, we analyse 20 years of 6-hourly model output from the four experiments of Polvani et al. (2011a): (1) a control run in which ozone and greenhouse gases are set to 1960 levels (REF1960), (2) a run in which ozone is perturbed to 2000 levels (thus creating an ozone hole over Antarctica in the springtime) and greenhouse gases remain at 1960 levels (OZONE2000), (3) a run in which greenhouse gases are perturbed to 2000 levels and ozone remains at 1960 levels (GHG2000), and (4) a run in which both ozone and greenhouse gases are set to 2000 levels (BOTH2000). The ozone concentrations are taken from Cionni et al. (2011), and the greenhouse gas levels are taken from the historical forcings prescribed in CMIP3 models, as in Polvani et al. (2011a). (Nearly identical ozone and greenhouse gas concentrations were prescribed in the historical simulations of the CMIP5 models.) The sea surface temperatures and sea ice concentrations are specified using the mean of the 17-year period surrounding the date of the greenhouse gas forcing (i.e. 1952–1968 or 1992–2008). Throughout this paper, we use the term response to denote the difference of each perturbed run from the REF1960 run.

To identify extratropical cyclones in the model runs, we use the Lagrangian cyclone-tracking algorithm developed by Hodges (1994, 1995, 1999). Following Grise et al. (2013), the algorithm is applied to the 6-hourly relative vorticity field (filtered to wavenumbers 5–42) at the model level nearest σ = 0.85. Details of the algorithm are provided in the above references. In brief, the algorithm minimizes a cost function to obtain the smoothest possible set of cyclone tracks that meet a minimum vorticity threshold of 1.0 × 10−5 s−1 over the domain 25°S–90°S. The tracks are subject to constraints on smoothness and maximum displacement between time steps. The track database is then filtered for cyclones with a minimum lifetime of 2 days, a minimum track length of 1000 km, and a minimum peak vorticity of 3.0 × 10−5 s−1. Developing cyclones are classified as systems with a growth rate exceeding 2.0 × 10−5 s−1 day−1.

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

Prior to examining the response of SH extratropical cyclones to the ozone and greenhouse gas forcings, we first confirm the ability of our model to capture the climatological features of the cyclones. The left column of Figure 1 shows the observed summertime climatology of SH extratropical cyclones over the period 1992–2011 as derived from the European Centre for Medium-Range Weather Forecasts (ECMWF) Interim reanalysis data set (ERA-Interim) (Dee et al., 2011). The cyclones are most frequent in a circumpolar band over the Southern Ocean centred near 53°S (Figure 1, top left). The cyclone frequency is approximately zonally symmetric, with localized maxima occurring over the southern Atlantic Ocean, southeastern Indian Ocean, and southeastern Pacific Ocean. The mean intensity of the cyclones is largest east of South America, in the southern Indian Ocean, and east of New Zealand (Figure 1, middle left). The cyclones primarily develop over strong meridional sea surface temperature gradients in the mid-latitude Southern Ocean and near the eastern coastlines of South America, Australia, and New Zealand (Figure 1, bottom left). The cyclones primarily decay over the high-latitude Southern Ocean near the coastline of Antarctica (not shown). Overall, the observed climatology derived here compares well with those shown in previous studies (Hoskins and Hodges, 2005).

image

Figure 1. December–February (DJF) climatology of Southern Hemisphere extratropical cyclones for (left column) 1992–2011 ERA-Interim reanalysis and (right column) the BOTH2000 model integration. Cyclones are identified using the relative vorticity field at the model level nearest σ = 0.85, the Hodges (1994) algorithm, and the filtering criteria of Grise et al. (2013). Cyclones are assigned to a 500-km radius region surrounding the vorticity maximum. (Top) Cyclone frequency. (Middle) Cyclone intensity (for grid points where a cyclone is present for at least 50 of the 60 months). (Bottom) Developing cyclone frequency. The shading interval is (top) 0.6 cyclones month−1, (middle) 1.25 × 10−6 s−1, and (bottom) 0.3 cyclones month−1. The thin white lines highlight the (top) 7.2  cyclones month−1 and (middle) 4.5 × 10−5 s−1 contours, and the thick white lines highlight the (middle) 4.75 × 10−5 s−1 contour. In the bottom panels, the thin contours display the observed DJF-mean sea surface temperatures prescribed in the BOTH2000 model integration (contour interval: 1K).

Download figure to PowerPoint

The right column of Figure 1 shows the summertime climatology of SH extratropical cyclones from our model. Here, we use the BOTH2000 integration to best match the period from the reanalysis. Overall, the model reproduces the observed climatology (Figure 1, left) with reasonable accuracy. Recall that both datasets were truncated to the same horizontal resolution (wavenumbers 5–42) for the cyclone tracking. The most notable difference between the reanalysis and model results is the underestimate in the frequency and intensity of cyclones by the model. This underestimate is attributable, at least in part, to the lower native resolution of the model (Jung et al., 2006); similar underestimates in SH cyclone intensity during DJF are also found in CMIP5 models (Chang et al., 2012). Additional differences between the reanalysis and model results include: (1) an equatorward bias in the model cyclone frequency near New Zealand, (2) excessive cyclones in the model near South Africa, and (3) missing cyclone development in the model south of Australia (Figure 1, bottom). Despite these limitations, we submit that CAM3 provides a satisfactory representation of SH extratropical cyclones for our experiments.

To isolate the impact of ozone forcing on SH extratropical cyclones, we compare the summertime climatology of the cyclones from the REF1960 and OZONE2000 model integrations (Figure 2). Recall that ozone is the only forcing that differs between these two experiments; greenhouse gases, sea surface temperatures, and sea ice concentrations are fixed at 1960 levels in both experiments. Overall, the climatologies from the two runs are similar to those shown in Figure 1, but Antarctic stratospheric ozone depletion leads to several notable differences in SH cyclones (Figure 2, bottom row). First, the latitude where extratropical cyclones are most frequent shifts ∼2° poleward (Figure 2, left, white lines). Second, the frequency of cyclones increases over the high-latitude Southern Ocean (Figure 2, bottom left). Third, there is some suggestion that the intensity of cyclones tracking eastward from South America increases (Figure 2, bottom right). Finally, the frequency of developing cyclones weakly increases southwest of South America and decreases near South Africa, southeastern Australia, and New Zealand (not shown).

image

Figure 2. DJF climatology of Southern Hemisphere extratropical cyclones for the (top row) REF1960 and (middle row) OZONE2000 model integrations. The left column displays the cyclone frequency (as in the top row of Figure 1). The thin white lines denote the latitude of the maximum at each longitude for the (dashed) REF1960 and (solid) OZONE2000 integrations. The right column displays the cyclone intensity (as in the middle row of Figure 1). The bottom row shows the difference between the OZONE2000 and REF1960 panels. Regions that are 95% significant via a Student's t-test are denoted by thick bolded outlines. The shading interval in the bottom row is (left) 0.1 cyclones month−1 and (right) 5.0 × 10−7 s−1.

Download figure to PowerPoint

Figure 3 decomposes the results from Figure 2 (bottom row) into three terciles (i.e. approximately equal-sized populations of cyclones) based on the maximum intensity of each cyclone: weak cyclones (3.0 × 10−5 s−1–4.3 × 10−5 s−1), moderate cyclones (4.3 × 10−5 s−1–5.9 × 10−5 s−1), and strong cyclones (greater than 5.9 × 10−5 s−1). The results reveal that the response of the SH extratropical cyclones to the ozone forcing does not arise from a single category of cyclones. Weak, moderate, and strong cyclones all experience a poleward shift in frequency and a slight intensification east of South America. The most significant response in cyclone frequency arises from the strongest cyclones (Figure 3, bottom left), whereas the most significant response in cyclone intensity over the South Atlantic Ocean arises largely from weak to moderate cyclones (Figure 3, right).

image

Figure 3. As in the bottom row of Figure 2, but separated into (top row) cyclones of weak maximum intensity (3.0 × 10−5 s−1–4.3 × 10−5 s−1), (middle row) cyclones of moderate maximum intensity (4.3 × 10−5 s−1–5.9 × 10−5 s−1), and (bottom row) cyclones of strong maximum intensity (greater than 5.9 × 10−5 s−1). Cyclone intensity is shown for grid points where a cyclone is present for at least 30 of the 60 months.

Download figure to PowerPoint

The quantitative response of the SH summertime cyclones to the ozone forcing is summarized in Table 1. The most robust response of the cyclones is the significant decrease in frequency over the 35°S–45°S latitude band and increase in frequency over the 55°S–65°S latitude band (Figure 2). The strongest cyclones also experience a significant increase in frequency over the 55°S–65°S latitude band (Figure 3). The poleward shift in cyclone frequency closely mirrors the poleward shift in the SH mid-latitude tropospheric jet (cf. Figure 4 of Polvani et al., 2011a). Because the translation speed of cyclones closely follows the lower tropospheric mean flow (Hoskins and Hodges, 2005), the cyclones at higher latitudes experience a significant increase in translation speed, and the cyclones at lower latitudes experience a significant decrease in translation speed. The ozone forcing has little effect on the lifetime of the cyclones, so the cyclones at high latitudes travel on average 250 km farther in the OZONE2000 run.

Table 1. December–February extratropical cyclone characteristics in selected latitude bands. In each cell, the leftmost value lists the mean value from the REF1960 integration, and the subsequent italicized values list the response to the ozone forcing, greenhouse gas forcing, and both forcings. Responses that are 95% significant are listed in bold type. The bottom row of cells is limited to cyclones of strong maximum intensity (greater than 5.9 × 10−5 s−1)
 35°S–45°S45°S–55°S55°S–65°S
Frequency (cyclones month1)42.07, −2.26, −0.95, −2.6655.45, −0.61, +0.55, +0.2939.73, +4.02, +0.83, +2.87
Intensity (105 s1)3.65,+0.00, −0.04, −0.054.05, +0.04, +0.03, +0.023.92, +0.09, +0.04, +0.07
Lifetime (days)5.33, −0.03, +0.05, −0.055.37, −0.03, −0.01, −0.115.75, −0.12, +0.03, −0.16
Track length (km)5903, +45.5, +132, −33.95917, +249.4, +80.5, +142.45714, +253.5, +157.4, +233.1
Speed of movement (ms1)16.37, −0.43, −0.09, −0.6216.71, +1.25, +0.35, +1.2010.45, +1.46, +0.53, +1.47
Strong cyclone frequency (cyclones month1)16.13, −1.04, −0.47, −1.7422.58, +0.17, +0.54, +0.0316.82, +2.74, +1.22, +2.12
image

Figure 4. Cyclone frequency response to (top row) ozone forcing, (middle row) greenhouse gas forcing, and (bottom row) both forcings. The left column is for the DJF climatology, and the right column is for the June–August (JJA) climatology. Regions that are 95% significant via a Student's t-test are denoted by thick bolded outlines. The shading interval is 0.1 cyclones month−1.

Download figure to PowerPoint

Apart from regional features that might be model dependent (Figure 2, bottom right), the intensity of the cyclones is only weakly affected by the ozone forcing. In fact, other than a slight increase in translation speed, there is no significant net effect of the ozone forcing on cyclone properties over the entire SH mid-latitudes (35°S–65°S); that is, the response of the cyclones to the ozone forcing is best characterised as a poleward shift. Because the ozone forcing strongly affects the meridional temperature gradient in the upper troposphere and lower stratosphere, our results suggest a larger role for lower tropospheric temperature gradients and moisture in governing cyclone intensity changes in our model.

Having documented the impact of ozone depletion on SH extratropical cyclones, we now compare that response to the cyclones' response to increasing greenhouse gases. Figure 4 shows the responses of the cyclones to the ozone forcing (OZONE2000), greenhouse gas forcing (GHG2000), and both forcings (BOTH2000) for both the austral winter and summer seasons. During the winter months, the responses of the cyclones to all forcings are difficult to interpret and not statistically significant (Figure 4, right column). During the summer months, the response of the cyclones to the ozone forcing is dominant over the 1960–2000 period (Figure 4, left column). There is some suggestion that increasing greenhouse gases also contribute to a poleward shift in cyclone frequency (Figure 4, middle left), as noted in previous studies (Bengtsson et al., 2006; Chang et al., 2012). However, the response of the SH extratropical cyclones to the greenhouse gas forcing is neither as robust nor as pronounced as their response to the ozone forcing. Interestingly, we note that the qualitative responses of the cyclones to the ozone and greenhouse gas forcings have very similar character (Table 1), presumably because both forcings tend to drive the SH mid-latitude westerly jet in the same direction.

4. Summary

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

In this study, we have examined the response of SH extratropical cyclones to stratospheric ozone depletion and increasing greenhouse gases over the second half of the 20th century. Our model experiments have indicated that, over the 1960–2000 period, the response of the summertime extratropical cyclones to stratospheric ozone depletion is similar in structure but larger in magnitude than their response to increasing greenhouse gases. Both forcings are associated with a poleward shift in cyclone frequency over the Southern Ocean, minimal changes in cyclone intensity and duration, and an increase in cyclone translation speed and track length at higher latitudes.

While the greenhouse gas forcing over the latter half of the 20th century is not strong enough to yield a robust response in SH extratropical cyclones in our model, it is expected that the influence of increasing greenhouse gases on SH cyclones will become more robust (Geng and Sugi, 2003; Bengtsson et al., 2006; Chang et al., 2012). However, the recovery of the Antarctic ozone hole in the 21st century is expected to induce oppositely signed changes in the SH tropospheric circulation and cyclones to those discussed here (Perlwitz, 2011). Consequently, any influence of increasing greenhouse gases on SH extratropical cyclones will be most apparent during the austral winter months.

Acknowledgements

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References

We thank J. F. Booth for helpful comments and assistance with the cyclone tracking algorithms. We also thank K. I. Hodges for providing us with his tracking algorithm. K. M. G. and L. M. P. were supported by a grant from the National Science Foundation to Columbia University. S. W. S. was supported by the Korea Polar Research Institute (KOPRI) grant under Project PE13010.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methodology
  5. 3. Results
  6. 4. Summary
  7. Acknowledgements
  8. References