Close interactions between the Antarctic cyclone budget and large-scale atmospheric circulation



[1] For the first time, we quantify relationships between the Southern Ocean cyclones and large-scale atmospheric variability indices: the Southern Annular Mode (SAM), the El Niño–Southern Oscillation (ENSO), the Zonal Wave 3 (ZW3) pattern, and the semiannual oscillation (SAO). Using the ERA-Interim 1979–2011 results, we identify cyclones south of 40°S and calculate monthly sectoral cyclone budgets of the Southern Ocean, defined as cyclogenesis minus cyclolysis plus net movement of cyclones into each sector. The SAM index has a strong connection with cyclones across all sectors. Positive SAM values are related to decreased eastward and increased southward movement of cyclones, resulting in higher cyclone densities along the Antarctic coast. The ENSO index shows strong associations with the cyclone behavior in the Amundsen-Ross Seas, whereas other regions are less sensitive to it. The ZW3 index has a stronger association with the meridional movement of cyclones than other indices.

1 Introduction

[2] The Antarctic climate system includes components that are quantitatively poorly known but simultaneously highly important for the global climate. For example, the mass balance of the Antarctic ice sheet and therefore its contribution to the present and future global sea level change is highly uncertain [King et al., 2012]. The Antarctic sea ice extent and its recent slightly increasing trend, arising from compensating regional trends [Holland and Kwok, 2012], is yet another phenomenon that is not fully understood and is not properly modeled [Landrum et al., 2012]. Moreover, better understanding of heat and freshwater budgets of the Southern Ocean is essential for the formation of deep water masses, which affect the global ocean circulation. Common to these processes is that they all are strongly affected by the wind field. Winds carry heat and moisture from lower latitudes to the Antarctic and hence moderate the heat and freshwater budgets of the ocean [Simmonds and Keay, 2000] as well as the heat and mass budgets of the continental ice sheet [Tietäväinen and Vihma, 2008], ice shelves, and sea ice.

[3] The point of view that local winds in the Antarctic are directly impacted by large-scale circulation is, however, simplified. The essential link between the large-scale circulation and local winds is provided by cyclones, which strongly affect the local wind field, while interacting with the large-scale circulation throughout their evolution [Fogt et al., 2012]. Over a flat ocean surface, the relationship between cyclones and local winds is fairly straightforward [see, e g., Holton, 2004, chapter 5], while the relationship between cyclones and the large-scale circulation is far more complex. The large-scale atmospheric circulation in the Antarctic is often characterized by indices such as the Southern Annular Mode (SAM), the Southern Oscillation Index (SOI) which characterizes the strength of El Niño–Southern Oscillation (ENSO), Zonal Wave 3 (ZW3), and the semiannual oscillation (SAO). These indices are interrelated—for example, SAM has strong connections with ENSO and ZW3 [Fogt et al., 2012].

[4] Numerous studies have addressed the relationships between the atmospheric large-scale circulation and Antarctic temperature, precipitation, and sea ice conditions [Liu et al., 2004; Raphael, 2007; Stammerjohn et al., 2008]. In addition, several studies have addressed the impact of local winds on the evaporation and heat flux from the ocean [e.g., Large and Yeager, 2008] and on the drift of Antarctic sea ice [e.g., Vihma et al., 1996]. Recently, Holland and Kwok [2012] summarized links between large-scale circulation and local winds, also showing that trends in sea ice drift can be explained by trends in local winds and that the Antarctic sea ice extent is strongly affected by the sea ice drift. The relationship between the large-scale circulation and cyclones has been addressed in several studies. In particular, it was determined that (1) periods of positive SAM are related to southward shift of the Antarctic Cyclone Track and a higher cyclone occurrence in the Weddell, Amundsen, and Bellingshausen Seas [Lubin et al., 2008; Pezza et al., 2012], (2) La Niña is related to mesocyclogenesis in the Amundsen Sea, whereas El Niño is related to mesocyclogenesis in the Indian Ocean Sector of the Southern Ocean [Claud et al., 2009], and (3) the optimal locations for cyclogenesis over the open ocean nearby the sea ice edge are associated with the ZW3 pattern [Yuan et al., 1999]. Building on these results, we present a comprehensive analysis of the relationship between cyclones and large-scale circulation indices, including the assessment of relative importance of these indices.

[5] We hypothesize that cyclones drive the changes in local winds around Antarctica, simultaneously interacting with the large-scale atmospheric circulation. Accordingly, our objective is to quantify the relationship between cyclones and large-scale circulation indices. We present a novel approach of calculating sectoral cyclone budgets and relate these budgets to four circulation indices: SAM, ZW3, SAO, and SOI. We demonstrate the importance of the SOI on the cyclone budgets and the two-way interaction between the other indices (SAM, SAO, and ZW3) and the cyclone budgets.

2 Methods and Data

[6] The cyclone tracking was carried out by using the University of Melbourne Automatic Cyclone Tracking Scheme [Murray and Simmonds, 1991] and based on the mean sea level pressure fields derived from the ERA-Interim reanalysis product [Dee et al., 2011] from 1979 to 2011. ERA-Interim arguably provides the most realistic estimates for the precipitation [Bromwich et al., 2011a] and cyclones [Hodges et al., 2011] in the Antarctic. The cyclone tracking scheme provided locations for each individual cyclone from the time of its generation (cyclogenesis) to the time of its decay (cyclolysis). Using this information and dividing the area south of 40°S into 270 sectors (each of them 4° in latitude and 20° in longitude), we computed cyclone movements and cyclone densities (number of cyclones per sector) as well as cyclogenesis, cyclolysis, and cyclone movement rates. We also developed a concept of a sectoral cyclone budget (cyclogenesis minus cyclolysis plus net movement of cyclones into a sector).

[7] The SAM index was based on Marshall [2003] (available from, while the SOI (available from was based on the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR) I reanalysis data [Kalnay et al., 1996]. The ZW3 index (available from was computed as in Raphael [2007] and based on the NCEP/NCAR I 500 hPa geopotential height. The SAO index was computed as the second harmonics of the zonally averaged sea level pressure difference between 50°S and 65°S following van Loon [1967]. We performed all calculations based on monthly indices and cyclone budgets.

3 Results and Discussion

[8] Upon completing a comprehensive assessment of sectoral cyclone budgets, we confirm and expand upon previously published results [e.g., Simmonds and Keay, 2000]. In summary, Southern Ocean cyclones primarily form at lower latitudes (north of 65°S) and dissipate closer to the Antarctic continent. Throughout their lifetime, cyclones generally track around the continent, forming the Antarctic Circumpolar Trough. The net zonal movement is directed eastward, while the net meridional movement is directed toward the south. Regionally, cyclogenesis is especially prominent on the leeside and south of South America (the Weddell Sea), south of Australia, and southeast of New Zealand, and also in the Antarctic coastal region at 110°E–170°E (please refer to Figure S1 in the supporting information). The latter region is also quite noticeable as maxima in cyclone density and net southward and net eastward transports. On the contrary, the Weddell Sea experiences the smallest net zonal and meridional transports, suggesting large variability in cyclone track directions.

[9] Relationships between the cyclone budget components and SAM are stronger than for the other circulation indices analyzed in this study, indicating the importance of SAM as the leading mode of variability in the Southern Hemisphere (note the different scales in Figures 1-3). The SAO (not shown) has even weaker influence on cyclone behavior.

Figure 1.

Differences in number of cyclones per sector (4° in latitude × 20° in longitude) of (a) cyclone track density, (b) southward movement, (c) eastward movement, (d) cyclogenesis, and (e) cyclolysis between two composites of SAM. The composite differences are between 20 monthly maximum values (5% of all months) and 20 monthly minimum values of SAM from 1979 to 2011. Borders of sectors are plotted in Figure 1a.

Figure 2.

Composite differences in number of cyclones of (a) cyclone track density, (b) southward movement, and (c) eastward movement based on SOI. The composite differences are between 20 monthly maximum values and 20 monthly minimum values of SOI from 1979 to 2011.

Figure 3.

(a–c) As in Figure 2 but for ZW3. (d) A redrawn Figure 1 by Raphael [2007] showing composite differences in the NCEP/NCAR I 500 hPa geopotential height field from 1958 to 2005 based on the phase of ZW3.

[10] During positive SAM events, the number of cyclones increases near the Antarctic continent because of increased southward cyclone motion and cyclogenesis in the south (Figures 1a, 1b, and 1d). The southward cyclone motion increases particularly in the Antarctic coastal sector at 110°E–170°E associated with a large increase in the eastward motion and cyclogenesis. This is one of the most prominent cyclogenesis sectors in the Southern Hemisphere: The increased southward movement of synoptic-scale cyclones is related to the interaction with coastal barrier and katabatic winds, which promotes the secondary development of cyclones [Bromwich et al., 2011b].

[11] More cyclones are generated over the Antarctic Peninsula when the SAM index becomes positive (Figure 1d), which is consistent with Lubin et al. [2008]. This is a sector of stronger westerly winds during positive SAM events which, along with warm-air advection, increase the lee cyclogenesis. The generated cyclones move south-eastward (Figures 1b and 1c) and dissipate in the southern Weddell Sea or further east in the Indian Ocean sector (Figure 1e). Simultaneously, as SAM changes from negative to positive, more cyclones enter the southern Indian Ocean sector, the southern Ross Sea, and the southern Amundsen Sea from the west (Figure 1c). Because more dissipating cyclones enter these sectors, the cyclolysis increases (Figure 1d).

[12] Due to the geographic proximity of the Equatorial Pacific, the SOI has the strongest impact on the cyclone budget in the Amundsen-Ross Seas, with differences in cyclone numbers as high as those of SAM over the same sector (Figures 2a and 1a). Generally, SOI is negative during El Niño and positive during La Niña. During La Niña, the number of cyclones increases in the Amundsen Sea compared to El Niño conditions (Figure 2a), which is consistent with earlier studies [Yuan et al., 1999; Pezza et al., 2012; Fogt et al., 2012]. This is due to cyclones moving increasingly northward in the Ross Sea, increasingly eastward to the northern Amundsen Sea, and increasingly southward in the Amundsen-Bellingshausen Seas (Figures 2b and 2c) during La Niña conditions. Since cyclone motion in the northern Bellingshausen Sea is increased, cyclogenesis increases east of South America during positive SOI events (La Niña, not shown). As a result, more cyclones move south toward the Antarctic Peninsula (Figure 2b), where cyclolysis rates increase as these cyclones decay (not shown).

[13] A comparison of Figure 3d (a copy of Figure 1 by Raphael [2007]) with Figure 3a can be used to identify relationships between the occurrence of cyclones and the areas of geopotential height anomalies due to ZW3. Positive (negative) geopotential anomalies indicate anticyclonic (cyclonic) flow with warm poleward flows on their western (eastern) side and cold equatorward flows on their eastern (western) side. The locations of the positive (negative) geopotential anomalies (Figure 3d) correspond well with the locations of decreased (increased) number of cyclones in Figure 3a. In the Amundsen Sea, in the eastern Weddell Sea, and south of Australia, the number of cyclones increases with ZW3, while in the Ross Sea, in the Bellingshausen Sea, around South America, in the southern Weddell Sea, and at longitudes 40°E to 90°E, the number of cyclones decreases with ZW3.

[14] Consistent with the meridional flow anomalies around ZW3 centers, more cyclones depart to the south in the Amundsen Sea, south of Australia, and in the eastern Weddell Sea, while fewer cyclones enter from north to the southern Ross Sea sector, in the western Weddell Sea, and at longitudes 45°E to 90°E (Figure 3b). The increased southward movement south of Australia is associated with increased cyclogenesis (not shown) and enhanced eastward motion in the Antarctic coastal sector at 130°E–170°E (Figure 3c). The increased number of cyclones moving north in the western Weddell Sea, on the other hand, is associated with the enhanced eastward movement of cyclones (Figure 3c) and increased cyclogenesis (not shown) over the Antarctic Peninsula sector. The cyclolysis increases at the Antarctic coastal locations where the enhanced southward movement of cyclones brings more dissipating cyclones (Figure 3b).

[15] Not surprisingly, the most prominent effect of the ZW3 is seen in the meridional movement of cyclones. In fact, the influence of ZW3 on the meridional movement of cyclones is bigger than the influences of SAM and SOI (compare Figures 1b, 2b, and 3b). The close association between ZW3 and the meridional cyclone movement affects many important processes of the Antarctic climate system: Southward moving cyclones transport heat and moisture over the Antarctic ice sheet [King et al., 2012]; they cause secondary cyclogenesis over the coastal sectors [Bromwich et al., 2011a]; and they modify the sea ice cover [Uotila et al., 2011] and the atmosphere-ocean energy fluxes further affecting the formation of deep water masses [Fahrbach et al., 1995]. From this point of view, ZW3 is a more important large-scale circulation mode than SAM.

[16] A positive SAO index is related to decreased cyclone densities in the northern Amundsen-Bellingshausen Seas, in the Drake Passage, and in the Weddell Sea, resulting from both reduced cyclogenesis and reduced cyclone movement from the west (not shown). Over the coastal sectors of the Amundsen-Ross Seas, however, the number of cyclones increases. During positive SAO, the cyclone tracks in the Amundsen Sea are closer to counterclockwise due to the increased eastward movement along the coast, decreased eastward movement in the northern Amundsen Sea, and increased southward movement in the Ross Sea (not shown). However, the magnitudes of the SAO index influences are smaller than SAM, SOI, and ZW3.

4 Conclusion

[17] We illustrate a close association between the Antarctic cyclone behavior and the large-scale atmospheric circulation indices. The identified close association between the cyclone movement rates and the large-scale atmospheric flow eases the interpretation of the effects of the latter. In this study we particularly emphasize (a) the importance of ZW3 mode and its association with the meridional movement of cyclones, (b) that positive SAM values are related to decreased eastward and increased southward movement of cyclones, resulting in higher cyclone densities along the Antarctic coast, and (c) that the impact of ENSO in the Amundsen-Ross Seas is comparable to that of SAM.

[18] Large-scale atmospheric circulation distinctly influences other components of the climate system, such as the sea ice, land ice, and the oceans, realized through the direct behavior of cyclones. The close associations between the cyclones and the large-scale indices, and fast response times revealed from the monthly analysis both confirm this statement.


[19] The research was funded by the Academy of Finland through the AMICO project (grants 128533 and 263918). The ECMWF is acknowledged for providing us with the ERA-Interim results. We thank two anonymous reviewers for constructive comments on the manuscript.

[20] The Editor thanks two anonymous reviewers for their assistance in evaluating this paper.