1. Introduction
Extratropical cyclones, which form in mid- and high latitudes, carry heat, moisture, and angular momentum between lower latitudes and the poles. These events cause gales, precipitation, and temperature variations and are an important component of the climate system (Nakamura and Shimpo 2004; Wang et al. 2006). Extratropical cyclones also directly contribute to the general circulation of the atmosphere (Grise et al. 2014). Studies focusing on extratropical cyclone frequency and intensity in the Southern Hemisphere are relatively scarce, and understanding cyclone climatology and its response to climate change is vital for the Southern Hemisphere as well as Earth.
Unlike tropical cyclones, for which the best track data are available from the International Best Track Archive for Climate Stewardship (IBTrACS) (e.g., Wang et al. 2020; Kishtawal et al. 2012) that is based on the observational record, extratropical cyclones do not have observational data. As a result, detection and tracking algorithms have been widely used in extratropical cyclone research. In classical fluid dynamics, the storm track is detected from the Eulerian perspective, while from the Lagrangian perspective, the storm track is not a fixed point in a fluid space but the motion of particle in a fluid (Blackmon 1976; Trenberth 1981, 1991; Priestley et al. 2020). The Lagrangian method is more suitable for tracking a system with a certain center and for obtaining the location, duration, intensity, and size. Scientists have employed sea level pressure, potential height, and relative vorticity in the lower troposphere to objectively identify and track cyclones since the mid-1990s (Jones and Simmonds 1993; Hodges 1994; Simmonds and Keay 2000; Fyfe 2003; Hodges et al. 2003). Jones and Simmonds (1993) utilized daily data during the period 1975–89 from the Australian Bureau of Meteorology to analyze the climatology of extratropical cyclones in the Southern Hemisphere and revealed two clear midlatitude bands of track density that originate in the Tasman Sea. For large-scale systems, the minimum value of local sea level pressure can be used as an indicator to locate cyclones. However, small-scale systems can be missed, and prior studies (Hoskins and Hodges 2002, 2005; Mohanty et al. 2010) have noted that employing relative vorticity could help to remedy this problem. Hoskins and Hodges (2002, 2005) applied objective methods utilizing relative vorticity at 250 and 850 hPa (ζ250 and ζ850, respectively), potential temperature θ, and mean surface level pressure (MSLP) to identify and track cyclones from 1958 to 2002, with a focus on the spatial distribution, genesis, and lysis in the Southern Hemisphere. A storm path was generally found to be more asymmetric and spiral in winter. Their works also noted that vorticity-tracking methods can identify more small systems than pressure-tracking methods, while the latter are more easily affected by changes in the background field. Vorticity-tracking methods were adopted in our work, but filtering is also required to avoid the impacts of large-scale background and small-scale noise in the atmosphere (Gramcianinov et al. 2019).
We calculated the extratropical cyclones for different seasons and determined that winter is the most active cyclone season in the Southern Hemisphere. Therefore, the current study focuses on austral winter cyclones. In recent years, in addition to the study of cyclone spatial distribution, some studies have also examined the long-term trend characteristics of cyclones. Qin et al. (2017) established cyclone datasets within the area delineated by three Chinese Antarctic scientific investigation stations from 1979 to 2013. The changes in cyclones at the different scientific stations were not consistent. The interannual changes in cyclones at a single station are of great importance to the Antarctic expeditions, while more work is needed to understand the interannual change in cyclones in the whole Southern Hemisphere. Wei and Qin (2016) studied the climatological characteristics of cyclones in the Southern Ocean and found that the number of cyclones showed an increasing trend. These regional studies help us to understand the local variations and effects of cyclones, but the long-term changes in cyclones, tracked by observational datasets, in the whole Southern Hemisphere are not yet clear. Coupled models have been used to discuss the response of midlatitude cyclones to greenhouse warming (Lambert and Fyfe 2006; Butler et al. 2010; Mendes et al. 2010; Shaw et al. 2016; Mindlin et al. 2020). The results demonstrated that the total number of events decreased significantly and the number of strong events increased, with a tendency to move poleward. However, the explanation remains unclear. Several factors ranging from upper-level jets to low-level jets play critical roles in cyclonic activities, as does the special growth of sea ice (Simmonds and Wu 1993; Butler et al. 2010; Simmonds and Li 2021).
It is critical to verify the conclusions given by the multiple models and reanalysis datasets. Our present work shows an obvious decreasing trend in the frequency of austral winter cyclones but identifies a significant increasing trend in cyclone intensity. In contrast, we have not found the poleward shift shown in previous research. The purpose of this study is to explore the possible mechanisms for the long-term trends in the frequency and intensity of austral winter cyclones.
The remainder of this paper is organized as follows: the next section introduces the datasets and methods used. The cyclone climatology characteristics and the long-term trends are presented in section 3. The possible mechanisms of the cyclone changes are discussed in sections 4 and 5, and conclusions are provided in section 6.
2. Data and methods
a. Datasets
The 6-hourly 2.5° × 2.5° gridded reanalysis data from the National Centers for Environmental Prediction–National Center for Atmospheric Research (NCEP–NCAR) Reanalysis-1 (Kalnay et al. 1996) for 70 austral winters [including June–August (JJA)] from 1948 to 2017 are used in this study. Other monthly mean physical fields are also taken from this dataset. Uncertainties exist in the NCEP–NCAR Reanalysis-1 dataset, and the observational data in the SH are rather scarce, especially in the mid- to high latitudes. For the purpose of diversity, the data derived from the European Centre for Medium-Range Weather Forecasts Reanalysis (ERA5) (Hersbach et al. 2023; Bell et al. 2020) and the Japan Meteorological Agency (JMA) Japanese 55-year Reanalysis (JRA-55) (Onogi et al. 2007; G. Chen et al. 2014; Kobayashi et al. 2015) are interpolated to the same resolution and are compared with the cyclone-tracking results of the NCEP–NCAR Reanalysis-1 dataset. Detailed information about the reliability of the three datasets can be found in the appendix. To analyze the influence that sea ice has on extratropical cyclones in the Southern Hemisphere, the sea ice concentration is taken from Centennial In Situ Observation-Based Estimates of SST, version 2 (COBE-SST2; Hirahara et al. 2014).
b. TRACK algorithm
In this study, cyclone detection and tracking methods are carried out using the automated tracking scheme based on the 850-hPa relative vorticity developed by Hodges (Hodges 1994, 1995). Planetary background waves with wavenumbers of less than 5 are filtered (Hodges 1999; Mailier et al. 2006). This algorithm is widely used in studies of the climatological characteristics of extratropical cyclones and polar depressions. Here, cyclones that formed south of 20°S and initially moved eastward are considered extratropical cyclones. As previous studies have noted, it takes several days for an extratropical cyclone to go through its entire life cycle (Mailier et al. 2006). Cyclones with lifetimes of more than 2 days and that cover more than 1000 km are studied here.
To better represent the cyclone activities, we computed the densities of cyclone genesis, lysis, and trajectories per 5° × 5° grid cell (considered as a spherical cap) per month (Gramcianinov et al. 2019). The genesis density and lysis density are calculated from the first and the last observation, respectively, for each cyclone. The track density is the number of cyclone trajectories passing each grid cell.
3. Characteristics and the long-term trend of extratropical cyclones
a. Number of extratropical cyclones
The track density of cyclones can indicate cyclone-active regions and tell us whether a suitable cyclone growth environment exists. We calculated the cyclone number of each season from 1948 to 2017 and found that the quantities peaked in winter (102 per month), were lowest in summer (79 per month), and roughly remained stable during the other two seasons. These seasonal characteristics are similar to those presented in a previous study (Hoskins and Hodges 2005). Because cyclones are most active in winter, and winter cyclones had the largest frequency variation over the past 70 years, our research is focused on the austral winter cyclones.
Figure 1 shows the time series of cyclone numbers in austral winters. A long-term downward trend can be clearly observed from 1948 to 2017, and the overall trend (with a slope of −0.57 yr−1) is significant at the 99% confidence level based on the Student’s t test. There were more cyclones in the 1950s and 1960s and fewer after the 1980s. Meanwhile, both interannual and interdecadal fluctuations appear in the time series. The orange line is the 9-yr running average, illustrating variations with longer time scales. Here, we mainly focus on the significant decreasing trend, which encourages us to conduct further investigations on the possible reasons.
Time series of extratropical cyclone numbers in austral winters (JJA) from 1948 to 2017. The orange line indicates the interdecadal variation by a 9-yr running average. The blue line represents the long-term trend of the whole time series.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
b. Genesis and lysis density
The climatology of the cyclogenesis and cyclolysis densities in austral winters is shown in Figs. 2a and 2b. Using the same method as Gramcianinov et al. (2019), we calculated the densities with a unit area equivalent to a spherical cap with a 5° × 5° grid cell in units of number per unit area for each season. Austral winter cyclones mainly form between 30° and 50°S, at approximately 45°S. There exists zonal asymmetry in the track density on both sides of the meridian connecting Australia and the tip of South America. In addition, two main cyclogenesis centers appear in South America: one is at 45°S, where cyclogenesis is associated mainly with systems decaying on the windward slope of the Andes and reviving on the leeward slope (Gan and Rao 1991; Hoskins and Hodges 2005). We can see that the downstream genesis density (Fig. 2a) is greater than the upstream lysis density (Fig. 2b), indicating that there are other factors conducive to the formation of cyclones, probably the upper-level jet and favorable terrain. The other main genesis cyclone centers, located at 30°S in South America, are not associated with upstream decay but with low-level baroclinicity (Hoskins and Hodges 2005). Other active cyclogenesis centers during the study period range from those at 120°–150°E around the Antarctic continent to those in southwest Australia and New Zealand.
Winter-season average (a) genesis and (b) lysis density of extratropical cyclones (number per 5° × 5° grid cell per month). For this and subsequent figures that feature the hemispheric map, the outermost zonal circle is 20°S, and dashed lines start from 30° to 75°S, with an interval of 15°.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
We analyzed the genesis locations of 21 396 cyclones during the overall 70 years of austral winters and discovered that the average cyclogenesis latitude was 47°S. In addition, the annually averaged latitude showed a long-term southward trend, indicating that more cyclones formed over higher latitudes over time. We further investigated the trend of cyclogenesis density (Fig. 3) and found that regional increasing trend genesis centers still existed at lower latitudes, which means that the southward-moving trend is not spatially uniform. The two genesis centers in South America have witnessed opposite changes in cyclogenesis density, an increasing trend in the northern part and a decreasing trend in the southern part. The decreasing trend in the southern part and its downstream areas may affect the cyclonic activity in the Atlantic. A significantly downward trend in cyclogenesis density could also be observed in New Zealand and the low-latitude region of the central Pacific. At the same time, there are more systems formed in the midlatitudes of the Pacific. In general, although we see a significant reduction in the number of cyclones, there are more systems in some places due to regional variability, and the various factors causing these different trends will be discussed in the following sections.
Trend of cyclogenesis density over 1948–2017 (per year). Stippling denotes areas that are statistically significant at the 95% confidence level.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
c. Intensity and lifetime
The tracking method used here is based on the relative vorticity at 850 hPa. During the development of a single cyclone, the absolute value of the relative vorticity generally increases to the maximum and then decreases. We regard the maximum absolute value of relative vorticity in the lifetime of a cyclone as the intensity of that cyclone. The average maximum intensity of austral winter cyclones is 6.2 × 10−5 s−1. As shown in Fig. 4a, the long-term cyclone intensity shows a significant increasing trend, which is opposite to the cyclone frequency. Less austral winter cyclones might occur in the future if the climate continues to warm, but they will be of greater intensity. This hypothesis is verified by possible mechanisms in the next section.
(a) Time series of the winter-season average maximum intensity (10−5 s−1) of extratropical cyclones. The intensity represents the maximum of the absolute value of relative vorticity during the whole lifetime of the cyclone. (b) Time series of the winter-season average lifetime (h) of extratropical cyclones. (c) The proportions of strong, medium, and weak cyclones. The strong cyclones (denoted by the red line) are defined as cyclones with a maximum intensity exceeding 8 × 10−5 s−1, the maximum intensity of weak cyclones (denoted by the black line) should be less than 4 × 10−5 s−1, and the remaining cyclones are defined as medium cyclones (denoted by the blue line).
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
Duration is also a key factor in describing a cyclone. In our present work, only systems lasting longer than 48 h are considered, and the extratropical cyclones in austral winter last approximately 4.5 days (108 h) on average. The lifetime of cyclones generally shortened over the 70 winters, with a slope of −0.165 h yr−1 (Fig. 4b). Specifically, it mainly shortened from the late 1940s to the middle of the 1980s and slightly increased after the late 1980s. Notably, the lifespan of cyclones is proportional to the maximum intensity (with a significant correlation coefficient of 0.45), which implies that stronger cyclones tend to last longer. However, from the perspective of long-term variability, the trends of the average maximum intensity and lifespan are anticorrelated, and the further reason for the inversely varied intensity has been discussed, as shown in Fig. 4c. The number of strong cyclones, whose maximum intensity exceeds 8 × 10−5 s−1, accounts for 18% of all cyclones and has increased significantly in recent decades, which explains why the average maximum intensity (Fig. 4a) exhibits a dramatic increasing trend.
d. Track density
The climatology of the track density of extratropical cyclones in austral winters is shown in Fig. 5, which is calculated by the same rules as in the genesis density figures (Fig. 3). The density in the mid- and high latitudes exhibits a major spiral distribution from the Atlantic Ocean through the Indian Ocean to the Antarctic Peninsula (60°W), with a clearly asymmetric storm track path. The appearance of such a spiral is considered to be related to a stationary Rossby wave forced by the asymmetries of convective heating in the north equator (Inatsu and Hoskins 2004). There are active movements of cyclones between 45°and 60°S in the Atlantic and Indian Oceans, as well as west of the Baleny Islands near the Antarctic continent, which coincide with the cyclogenesis centers shown in Fig. 2a.
Winter-season average track density (number per 5° × 5° grid cell per month).
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
The track density depends on the number and lifespan of cyclones. Given that the number and duration of cyclones in the NCEP–NCAR dataset are both lower, it is no surprise to see a significant decreasing trend in the track density in most areas (Fig. 6). The cyclone track density decreases significantly in the mid- and high latitudes of the Atlantic and Indian Oceans, as well as in the lower latitudes of the Pacific. Around the Antarctic continent, including the Amundsen Sea, Ross Sea, and Weddell Sea, there is a decreasing trend. There are also some cyclonic movements in some other regions, including the midlatitudes of the Pacific Ocean (45°–60°S), southern Australia, the Tasman Sea, and the northern cyclogenesis center at 30°S in South America. It is interesting to note that there are some shifting polar signals in the South Pacific, but none in the South Atlantic and Indian Oceans.
Trend of cyclone track density during 1948–2017 (per year). Stippling denotes statistical significance at the 95% confidence level.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
Some studies have stated that the capacity to describe cyclones in the Southern Hemisphere varies among the different datasets (Hodges et al. 2003; Wang et al. 2006; Hodges et al. 2011; Befort et al. 2016); therefore, we compared the descriptions of cyclone climatology and variability in three datasets, NCEP–NCAR, ERA5, and JRA-55 (see Fig. A1 in the appendix). Although the three datasets cover different time periods, the number of cyclones exhibits a common downward trend from 1958 to 2017 in all three datasets. From the spatial distribution of track density, although the latter two datasets tracked more cyclones, the spatial distribution showed the same spiral shape. As far as the trend of track density is concerned, the results of JRA-55 and NCEP–NCAR are rather similar, and there are some different patterns shown in ERA5, mainly in the area from southern South America to the Antarctic Peninsula and the southern part of the African continent. Nevertheless, a significant reduction in extratropical cyclones can be clearly observed in the south Indian Ocean (the area of 30°–120°E, 30°–60°S) in all cases.
The decrease in cyclone activity shown here is accompanied by a decrease in weak cyclones and an increase in strong cyclones. Here, we try to explain why the track densities of cyclones have decreased significantly in some regions, while other regions have shown the opposite trend.
4. Reasons for the long-term trend of extratropical cyclone activity
In most parts of the Southern Hemisphere, especially the south Indian Ocean, there has been a sharp reduction in track density, as observed by NCEP–NCAR and confirmed by the results of the ERA5 and JRA-55 datasets (shown in Fig. A2). At the same time, however, extratropical cyclone activities have increased throughout the region west of the La Plata Plain, south of Australia and the Tasman Sea, in the South Pacific at 45°–60°S. It is believed that cyclogenesis is related to both the low-level baroclinicity and the waveguide effect of the upper jet (Hodges 2005). We will discuss their variations and how they affect the long-term trends of cyclone activities in this section.
a. Low-level baroclinicity
Dynamic meteorology suggests that synoptic-scale disturbances are frequently forced by baroclinic instability, which largely explains why differentially heated rotating planets spontaneously generate transient eddies. As reported in the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (IPCC AR5), the climate system has warmed unequivocally in recent decades. However, the rates of warming and the gradients of air temperature vary by region. According to thermal wind theory, the meridional temperature gradient is proportional to the vertical wind shear of the zonal wind, and an increase in the vertical wind shear would further cause instability. We use the maximum Eady growth rate to describe mid- to high-latitude baroclinic instability (Eady 1949; Lindzen and Farrell 1980; Simmonds and Lim 2009), calculated as 0.31(|f|/N)(∂V/∂Z), where f is the Coriolis parameter, N is the Brunt–Väisälä frequency [where N2 = (g/θ)(∂θ/∂z), g is acceleration due to gravity, θ is the potential temperature, and z is the vertical coordinate], and ∂V/∂Z is the monthly average wind shear between 700 and 850 hPa during JJA. As shown in Fig. 7a, the active regions of the atmosphere are located in south Australia and the temperate Pacific at the same latitude, from South America to the high latitude of the south Indian Ocean and near the Antarctic continent. The maximum Eady growth rate is matched with the mean state of the active cyclogenesis regions, with the exception of the polar area. From the perspective of the long-term trend of low-level baroclinicity, there is significant weakening from the end of the South American continent to the ocean to the south of the African continent and near New Zealand, and decreasing trends in cyclone track densities are observed in areas located downstream of these areas (especially over the Indian Ocean) (Fig. 7b). The regional downward trend of low-level baroclinicity may partly explain the changes in cyclone frequency (as well as genesis changes). The regional increasing trend of low-level baroclinicity explains why the cyclone intensity increased. For example, fewer cyclones will be formed in the weakened baroclinicity midlatitude regions, and they move forward to the east and higher latitudes. If some of the cyclones enter the area where baroclinicity is increasing (e.g., the spiral increasing band along 60°S), the intensities of these cyclones will be strengthened, which is the reason why fewer but more intense cyclones will occur in the future.
(a) Winter-season average of the maximum Eady growth rate (day−1). The maximum Eady growth rate is calculated as 0.31(|f|/N)(∂V/∂Z). The blue contours show the variance, with an interval of 0.015 day−2. (b) Trend of the maximum Eady growth rate (per day per year). Stippling denotes significance at the 95% confidence level. Areas in Antarctica with elevations greater than 1500 m have been masked.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
It is worth noting that in the south Indian Ocean, the enhancement of low-level baroclinicity is not matched by the significant reduction in cyclone track densities and the minimal reduction in cyclone numbers. We speculate that another mechanism is responsible for the change in cyclone activities in this region. Strong westerly winds are considered a key influencing factor, and their effects will be discussed in the next subsection.
b. Upper-level jet
Before the arrival of winter in the Southern Hemisphere, the upper jet undergoes a transition from an annular mode to a spiral pattern (shown in Fig. 8a). The north branch is the subtropical jet, and the south branch is the polar-front jet (Bals-Elsholz et al. 2001; Williams et al. 2007; Lachmy and Harnik 2014). The subtropical jet diminished, while the polar-front jet strengthened, according to an investigation into the fluctuating trend of relative vorticity and zonal wind speed. Previous studies have noted that the poleward shift of the subtropical jet stream serves as a waveguide and is an important driver of cyclone movement (Yin 2005; Zhang et al. 2012; L. Chen et al. 2014). As illustrated in Fig. 2, the high-level airflow plays a crucial role in how cyclones decay upstream and reform downstream when they hit mountains, such as the Andes Mountains in South America. Moreover, secondary circulation and vorticity advection accompanied by the upper jet stream in the outflow and inlet favor the emergence and development of weather systems. We discovered active cyclonic activities in the equivalent lower layer, including the Southern Ocean and the southern Indian Ocean. In addition to the mean state, the high correlation coefficient between cyclone frequency and upper jet intensity can also be confirmed by the long-term trend. Figure 8b shows the spatially opposite changes in the zonal wind trends on the northern and southern sides of the subtropical jet. The subtropical jet exit area corresponds to the reverse change in cyclone activity in the South Pacific, as well as in the Indian Ocean. Therefore, in these areas, the upper-level jet explains the increase in cyclone activity patterns better than can be explained by the low-level baroclinicity.
(a) Winter-season climatology (shadings; m·s−1) and variance (contour; interval is 10 m2·s−2 from 20 to 30 m2·s−2) of the zonal wind at 200 hPa from 1948 to 2017. (b) Trend of the zonal wind (shadings; m·s−1). Stippling denotes significance at the 95% confidence level.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
Over the last 70 years, the upper-level zonal wind has shown an annular variation. Specifically, westerly jet anomalies can be seen north of 45°S and south of 70°S, and easterly jet anomalies can be seen between 45° and 70°S in austral winters. The decreasing trend of cyclone numbers can be found at 20°–45°S and south of 70°S, while a very weak decreasing trend can be seen between 45° and 70°S (Table 1). The weakening zonal wind north of 45°S and south of 70°S is consistent with the reduction in the cyclone numbers; however, the intensification of the polar-front jet in the mid- and high latitudes (45°–70°S) is antiphased to the variation in cyclone activity, and the possible reason will be discussed in the next section.
Trend of the cyclone numbers and track density between three latitudinal zones during 1948–2017 (number per year). The two asterisks indicate that the correlation coefficients are significant at the 95% confidence level.
In general, the significantly decreasing cyclone track densities and dramatically decreasing cyclone numbers are closely related to the weakened upper-level troposphere jet stream and low-level baroclinicity. The change in baroclinicity can basically explain the regional change in cyclonic activities, especially the frequency and intensity changes. However, in places like the entrance and exit zones of upper-level subtropical jets, the jets play a more significant role in explaining changes in cyclone frequency.
5. Discussion
As mentioned above, low-level baroclinicity and upper-level jets are two major factors that influence extratropical cyclone activities. In fact, both of the factors are partly influenced by the meridional temperature gradients. As the climate system warms, cyclones are predicted to form less frequently in the subtropical region and more frequently in the subpolar region, as determined in previous research (Ulbrich et al. 2009). In comparison with their results, ours are consistent within the subtropical regions but inconsistent in the subpolar regions. There are many factors that make the midlatitudes, particularly in the south Indian Ocean, unfavorable for cyclonic activity. In contrast, cyclones are frequent at higher latitudes, especially in the westerly anomaly region of the polar-front jet stream. Contrary to expectations, the track density of extratropical cyclones does not clearly indicate a poleward-shifting trend. Some likely causes and the limitations of our results are discussed below.
First, let us discuss the possible mechanisms explaining why the long-term high-latitude cyclone numbers undergo a downward trend. We found that more than 50% of cyclones emerging between 20° and 45°S eventually move to the regions south of 45°S. Fewer cyclones form between 45° and 70°S. The decrease in cyclone track density over the mid- and high latitudes can also be partly attributed to this phenomenon. Furthermore, seasonal sea ice is present around Antarctica between 55° and 75°S, and it may play a dominant role in the polar region. The changes in the edge of sea ice are believed to greatly affect the cyclone activities around Antarctica (Zhang et al. 2004; Inoue et al. 2012). Existing studies on Arctic cyclones and their interactions with sea ice indicate that the loss of sea ice during autumn results in enhanced moisture availability, increased local baroclinicity, and changes in vertical stability that favor cyclogenesis (Koyama et al. 2017; Valkonen et al. 2021). We used the sea ice concentration, defined as the percentage of sea ice area per unit area (Liu et al. 2015; Lecomte et al. 2017), to reveal the impact that autumn sea ice melting and formation has on winter cyclones. The sea ice concentration over Antarctica during the autumn seasons has shown a clear increasing trend except in areas around the Amundsen Sea and Bellingshausen Sea. The correlation coefficients between the number of winter cyclones between 45° and 70°S and the autumn sea ice concentration in the same year are also calculated. Figure 9 shows that in most parts around Antarctica, the increase in sea ice concentration is accompanied by a decrease in cyclones. The formation of sea ice cools the surface and low-level atmosphere; thus, the vertical thermal structure of the atmosphere becomes more stable. Additionally, as the underlying surface, sea ice is not conducive to cyclogenesis; therefore, fewer cyclones form on the ice. An increased concentration of polar sea ice with expanded ice edges and varying vertical stability structures acts to suppress cyclones over most of the oceans around Antarctica. This is an important reason why extratropical cyclone track density shows a decreasing trend around the polar region. We thus suspect that the force driving polar cyclogenesis is not the lower-atmospheric baroclinicity or the upper jet but the surface static stability.
Correlation coefficients between the number of winter cyclones at 45°–70°S and the monthly average sea ice concentration in austral autumns during 1948–2017. Stippling denotes statistical significance at the 95% confidence level. The outermost zonal circle is 45°S, and dashed lines start from 60°S, with an interval of 15°.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
Second, we would also like to bring attention to some limitations associated with our findings. The quality of reanalysis datasets varies, and the reanalysis data for the Southern Hemisphere before the satellite era (1979) may not be reliable. We provide further investigations in the appendix. First, we use the same tracking algorithm to obtain cyclones from ERA5 (1950–2017) and JRA-55 (1958–2017), and then we compare their results with the results of NCEP–NCAR Reanalysis-1 data. The asymmetric spiral shape of the spatial distribution of track density trends is the same among the three datasets, although their magnitudes vary (appendix Fig. A2). Although all data have been interpolated to the same grid resolution, they obviously suffer from different systematic errors embedded in the data: different datasets use different methods to collect the observational data included in their catalog, adopt different models to generate the estimated ensemble data before the satellite era, give different initial and boundary conditions, and use various assimilation schemes. For all these reasons, we cannot elaborate on the differences in their tracking results.
However, this research focuses on the consistent decreasing trends of cyclone track density in the south Indian Ocean across all three datasets, rather than the different characteristics. Let us note the fact that both long-term trends and abrupt changes in the statistical characteristics of tracked cyclones within a dataset have nothing to do with the specific resolution of the datasets (since that has remained fixed during the whole period). Furthermore, there is no reason to believe that the systematic errors (including the errors related to the estimated data for the presatellite era) embedded in the three datasets are correlated in the sense of affecting the cyclone-tracking results in similar ways. Therefore, if we still observe common long-term trends and regional change patterns across the datasets, it is a strong signal suggesting the robustness of our findings: even though the different datasets contain different systematic errors, we still observe the same regional trends in cyclone activities. In other words, the statistical likelihood of making false statements decreases if our findings are consistent across all datasets. We believe that each reanalysis dataset is internally consistent throughout the whole period. From this point of view, the physical mechanisms we found from the long-term reanalysis data are reliable.
6. Conclusions
This study investigated austral winter extratropical cyclones in the Southern Hemisphere based on NCEP–NCAR reanalysis data. The main conclusions are as follows:
The average cyclogenesis latitude was 47°S, and the annually averaged latitude showed a long-term southward-moving trend. The average lifetime of austral winter cyclones is approximately 4.5 days, and the average maximum relative vorticity of cyclones is 6.15 × 10−5 s−1.
A downward trend in annual cyclone frequency and an upward trend in intensity were detected for the 1948–2017 period. However, strong cyclones occurred more frequently, while the frequency of weak, short-lived cyclones decreased. This indicates an increasing risk of extreme events in the future.
A significant reduction in extratropical cyclone activities can be clearly observed in the south Indian Ocean, which is mainly due to the weakened upper-level subtropical jet, especially in the entrance and exit areas at midlatitudes (20°–45°S). The regional increasing trend of low-level baroclinicity illustrates why the intensity of cyclones has increased. The baroclinic variations can also be used to explain the changes in cyclone activities at approximately 45°S in South America. However, cyclones forming near the Antarctica region can be greatly influenced by the instability forced by sea ice.
In this study, we sought to shed additional light on the mechanisms underlying the decreasing trend of cyclone track density and the increasing trend of cyclone intensity. The low-level baroclinicity and high-level jet are considered to be the main factors. Furthermore, sea ice in the Antarctic region plays a critical role in mediating polar cyclones.
Acknowledgments.
We thank three anonymous reviewers for their insightful comments and constructive suggestions. We are grateful to Kevin Hodges for his invaluable help. This work is jointly supported by the Strategic Project of Chinese Academy of Science (Grant XDA19070402) and NSFC Grant 41605054.
Data availability statement.
Multilevel horizontal wind (u wind and υ wind) and surface temperature are from NCEP–NCAR Reanalysis-1 datasets, which can be found online (https://psl.noaa.gov/data/gridded/data.ncep.reanalysis.pressure.html). The sea ice concentration is taken from COBE-SST2 data products from the Physical Sciences Laboratory (https://psl.noaa.gov/data/gridded/data.cobe2.html).
APPENDIX
Comparison of Cyclones from Three Reanalysis Datasets
There is no observational dataset for extratropical cyclones. To analyze the climatological characteristics and long-term variations in extratropical cyclones, we must employ automatic cyclone-tracking algorithms and rely on reanalysis datasets. Therefore, the quality of different reanalysis datasets should be assessed. Previous findings indicate that there are some discrepancies, both in climatology and variability, between the cyclones recognized by different datasets (Wang et al. 2006; Hodges et al. 2011; Befort et al. 2016). The number of cyclones obtained by the TRACK algorithm based on data from NCEP–NCAR Reanalysis-1, ERA5, and JRA-55 is selected here to make a comparison of the climatological features and long-term trends (Fig. A1).
Time series of extratropical cyclone numbers of NCEP–NCAR (1948–2017), ERA5 (1950–2017), and JRA-55 (1958–2017) over the study periods. The numbers at the upper-right corner of the graph represent long-term trends in the study periods. Two asterisks denote that the correlation coefficients are significant at the 95% confidence level.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
We conducted some statistical analyses on annual cyclone numbers for the same period of 1958–2017 (Fig. A1) and found that the correlation coefficient between NCEP–NCAR and ERA5 is 0.29, 0.27 for NCEP–NCAR and JRA-55, and 0.43 for ERA5 and JRA-55. Based on Student’s t test, all three correlation coefficients are significant at the 95% confidence level, suggesting that significant positive correlations exist among the three datasets. If we calculate the 9-yr running average of the cyclone time series, the correlation coefficient between any two datasets is further improved. More specifically, although more cyclones have been tracked by ERA5 and JRA-55 than by NCEP–NCAR, all of them have experienced a decreasing trend over the whole period, with slopes of −0.25 for JRA-55, −0.19 for ERA5, and −0.57 for NCEP–NCAR.
In addition to the number of cyclones, the characteristics of cyclones among the three datasets, for example, the average lifetime, track length, and location of the cyclogenesis, are shown in Table A1. The cyclones obtained by NCEP–NCAR are relatively inactive, which is also indicated by shorter life cycles and weaker intensities. There also exists a common long-term trend among the cyclones. The results from both NCEP–NCAR and JRA-55 suggest that the lifetime of a generic cyclone has been shortened during the last 70 years, while the result from ERA5 shows the opposite trend. The location of cyclogenesis is another key factor. All statistics in various datasets are consistent, both in terms of the climatology features and long-term trends. The cyclones’ mean meridional positions exhibited significant poleward movement.
Average intensity (the absolute value of relative vorticity; 10−5 s−1), lifetime (h), and location of extratropical cyclones identified by three datasets in the Southern Hemisphere. The values in parentheses are the long-term trends of the corresponding variables covering different periods (1948–2017 for NCEP–NCAR; 1950–2017 for ERA5; 1958–2017 for JRA-55). The two asterisks indicate that the correlation coefficient values are significant at the 95% confidence level.
Based on the three datasets, we contrasted the patterns of cyclone track densities, with a focus on the mean state and variance during the same period of 1958–2017 (Fig. A2). Although their results differ quantitatively, all of them exhibit the same asymmetric spiral shape. The regions of East Antarctica, northern South America, and Australia all share similarities in the track density patterns (Figs. A2d–f), especially for the south Indian Ocean.
(a)–(c) Climatology and (d)–(f) long-term trends of extratropical cyclone track density in winters during 1958–2017 for (left) NCEP–NCAR; (center) ERA5; and (right) JRA-55. The unit of track density is the number of cyclones per 5° radius spherical cap per month. Stippling in (d)–(f) denotes statistical significance at the 95% confidence level. The outermost zonal circle is 20°S, and dashed lines start from 30° to 75°S, with an interval of 15°.
Citation: Journal of Applied Meteorology and Climatology 62, 8; 10.1175/JAMC-D-22-0061.1
In general, the ERA5 and JRA-55 datasets tracked more and more durable cyclone systems, but the long-term trends of cyclone frequency in the south Indian Ocean are consistent. That is, although there are differences in the number of systems and their intensities between various datasets, this does not greatly impair the validity of the conclusions reported in this article.
REFERENCES
Bals-Elsholz, T. M., E. H. Atallah, L. F. Bosart, T. A. Wasula, M. J. Cempa, and A. R. Lupo, 2001: The wintertime Southern Hemisphere split jet: Structure, variability, and evolution. J. Climate, 14, 4191–4215, https://doi.org/10.1175/1520-0442(2001)014<4191:TWSHSJ>2.0.CO;2.
Befort, D. J., S. Wild, T. Kruschke, U. Ulbrich, and G. C. Leckebusch, 2016: Different long-term trends of extra-tropical cyclones and windstorms in ERA-20C and NOAA-20CR reanalyses. Atmos. Sci. Lett., 17, 586–595, https://doi.org/10.1002/asl.694.
Bell, B., and Coauthors, 2020: ERA5 hourly data on pressure levels from 1950 to 1978 (preliminary version). Copernicus Climate Change Service Climate Data Store, accessed 30 June 2021, https://cds.climate.copernicus.eu/cdsapp#!/dataset/reanalysis-era5-pressure-levels-preliminary-back-extension?tab=overview.
Blackmon, M. L., 1976: A climatological spectral study of the 500 mb geopotential height of the Northern Hemisphere. J. Atmos. Sci., 33, 1607–1623, https://doi.org/10.1175/1520-0469(1976)033<1607:ACSSOT>2.0.CO;2.
Butler, A. H., D. W. J. Thompson, and R. Heikes, 2010: The steady-state atmospheric circulation response to climate change–like thermal forcings in a simple general circulation model. J. Climate, 23, 3474–3496, https://doi.org/10.1175/2010JCLI3228.1.
Chen, G., T. Iwasaki, H. Qin, and W. Sha, 2014: Evaluation of the warm-season diurnal variability over East Asia in recent reanalyses JRA-55, ERA-Interim, NCEP CFSR, and NASA MERRA. J. Climate, 27, 5517–5537, https://doi.org/10.1175/JCLI-D-14-00005.1.
Chen, L., B. Tan, N. G. Kvamstø, and O. M. Johannessen, 2014: Wintertime cyclone/anticyclone activity over China and its relation to upper tropospheric jets. Tellus, 66A, 21889, https://doi.org/10.3402/tellusa.v66.21889.
Eady, E. T., 1949: Long waves and cyclone waves. Tellus, 1, 33–52, https://doi.org/10.3402/tellusa.v1i3.8507.
Fyfe, J. C., 2003: Extratropical Southern Hemisphere cyclones: Harbingers of climate change? J. Climate, 16, 2802–2805, https://doi.org/10.1175/1520-0442(2003)016<2802:ESHCHO>2.0.CO;2.
Gan, M. A., and V. B. Rao, 1991: Surface cyclogenesis over South America. Mon. Wea. Rev., 119, 1293–1302, https://doi.org/10.1175/1520-0493(1991)119<1293:SCOSA>2.0.CO;2.
Gramcianinov, C. B., K. I. Hodges, and R. D. Camargo, 2019: The properties and genesis environments of South Atlantic cyclones. Climate Dyn., 53, 4115–4140, https://doi.org/10.1007/s00382-019-04778-1.
Grise, K. M., S.-W. Son, G. J. P. Correa, and L. M. Polvani, 2014: The response of extratropical cyclones in the Southern Hemisphere to stratospheric ozone depletion in the 20th century. Atmos. Sci. Lett., 15, 29–36, https://doi.org/10.1002/asl2.458.
Hersbach, H., and Coauthors, 2023: ERA5 hourly data on single levels from 1940 to present. Copernicus Climate Change Service Climate Data Store, accessed 30 June 2021, https://doi.org/10.24381/cds.adbb2d47.
Hirahara, S., M. Ishii, and Y. Fukuda, 2014: Centennial-scale sea surface temperature analysis and its uncertainty. J. Climate, 27, 57–75, https://doi.org/10.1175/JCLI-D-12-00837.1.
Hodges, K. I., 1994: A general method for tracking analysis and its application to meteorological data. Mon. Wea. Rev., 122, 2573–2586, https://doi.org/10.1175/1520-0493(1994)122<2573:AGMFTA>2.0.CO;2.
Hodges, K. I., 1995: Feature tracking on the unit sphere. Mon. Wea. Rev., 123, 3458–3465, https://doi.org/10.1175/1520-0493(1995)123<3458:FTOTUS>2.0.CO;2.
Hodges, K. I., 1999: Adaptive constraints for feature tracking. Mon. Wea. Rev., 127, 1362–1373, https://doi.org/10.1175/1520-0493(1999)127<1362:ACFFT>2.0.CO;2.
Hodges, K. I., B. J. Hoskins, J. Boyle, and C. Thorncroft, 2003: A comparison of recent reanalysis datasets using objective feature tracking: Storm tracks and tropical easterly waves. Mon. Wea. Rev., 131, 2012–2037, https://doi.org/10.1175/1520-0493(2003)131<2012:ACORRD>2.0.CO;2.
Hodges, K. I., R. W. Lee, and L. Bengtsson, 2011: A comparison of extratropical cyclones in recent reanalyses ERA-Interim, NASA MERRA, NCEP CFSR, and JRA-25. J. Climate, 24, 4888–4906, https://doi.org/10.1175/2011JCLI4097.1.
Hoskins, B. J., and K. I. Hodges, 2002: New perspectives on the Northern Hemisphere winter storm tracks. J. Atmos. Sci., 59, 1041–1061, https://doi.org/10.1175/1520-0469(2002)059<1041:NPOTNH>2.0.CO;2.
Hoskins, B. J., and K. I. Hodges, 2005: A new perspective on Southern Hemisphere storm tracks. J. Climate, 18, 4108–4129, https://doi.org/10.1175/JCLI3570.1.
Inatsu, M., and B. J. Hoskins, 2004: The zonal asymmetry of the Southern Hemisphere winter storm track. J. Climate, 17, 4882–4892, https://doi.org/10.1175/JCLI-3232.1.
Inoue, J., M. E. Hori, and K. Takaya, 2012: The role of Barents Sea ice in the wintertime cyclone track and emergence of a warm-Arctic cold-Siberian anomaly. J. Climate, 25, 2561–2568, https://doi.org/10.1175/JCLI-D-11-00449.1.
Jones, D. A., and I. Simmonds, 1993: A climatology of Southern Hemisphere extratropical cyclones. Climate Dyn., 9, 131–145, https://doi.org/10.1007/BF00209750.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-Year Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437–471, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
Kishtawal, C. M., N. Jaiswal, R. Singh, and D. Niyogi, 2012: Tropical cyclone intensification trends during satellite era (1986–2010). Geophys. Res. Lett., 39, L10810, https://doi.org/10.1029/2012GL051700.
Kobayashi, S., and Coauthors, 2015: The JRA-55 Reanalysis: General specifications and basic characteristics. J. Meteor. Soc. Japan, 93, 5–48, https://doi.org/10.2151/jmsj.2015-001.
Koyama, T., J. Stroeve, J. Cassano, and A. Crawford, 2017: Sea ice loss and Arctic cyclone activity from 1979 to 2014. J. Climate, 30, 4735–4754, https://doi.org/10.1175/JCLI-D-16-0542.1.
Lachmy, O., and N. Harnik, 2014: The transition to a subtropical jet regime and its maintenance. J. Atmos. Sci., 71, 1389–1409, https://doi.org/10.1175/JAS-D-13-0125.1.
Lambert, S. J., and J. C. Fyfe, 2006: Changes in winter cyclone frequencies and strengths simulated in enhanced greenhouse warming experiments: Results from the models participating in the IPCC diagnostic exercise. Climate Dyn., 26, 713–728, https://doi.org/10.1007/s00382-006-0110-3.
Lecomte, O., H. Goosse, T. Fichefet, C. de Lavergne, A. Barthélemy, and V. Zunz, 2017: Vertical ocean heat redistribution sustaining sea-ice concentration trends in the Ross Sea. Nat. Commun., 8, 258, https://doi.org/10.1038/s41467-017-00347-4.
Lindzen, R. S., and B. Farrell, 1980: A simple approximate result for the maximum growth rate of baroclinic instabilities. J. Atmos. Sci., 37, 1648–1654, https://doi.org/10.1175/1520-0469(1980)037<1648:ASARFT>2.0.CO;2.
Liu, T., Y. Liu, X. Huang, and Z. Wang, 2015: Fully constrained least squares for Antarctic sea ice concentration estimation utilizing passive microwave data. Geosci. Remote Sens. Lett., 12, 2291–2295, https://doi.org/10.1109/LGRS.2015.2471849.
Mailier, P. J., D. B. Stephenson, C. A. T. Ferro, and K. I. Hodges, 2006: Serial clustering of extratropical cyclones. Mon. Wea. Rev., 134, 2224–2240, https://doi.org/10.1175/MWR3160.1.
Mendes, D., E. P. Souza, J. A. Marengo, and M. Mendes, 2010: Climatology of extratropical cyclones over the South American–Southern Oceans sector. Theor. Appl. Climatol., 100, 239–250, https://doi.org/10.1007/s00704-009-0161-6.
Mindlin, J., T. G. Shepherd, C. S. Vera, M. Osman, G. Zappa, R. W. Lee, and K. I. Hodges, 2020: Storyline description of Southern Hemisphere midlatitude circulation and precipitation response to greenhouse gas forcing. Climate Dyn., 54, 4399–4421, https://doi.org/10.1007/s00382-020-05234-1.
Mohanty, U. C., K. K. Osuri, A. Routray, M. Mohapatra, and S. Pattanayak, 2010: Simulation of Bay of Bengal tropical cyclones with WRF Model: Impact of initial and boundary conditions. Mar. Geod., 33, 294–314, https://doi.org/10.1080/01490419.2010.518061.
Nakamura, H., and A. Shimpo, 2004: Seasonal variations in the Southern Hemisphere storm tracks and jet streams as revealed in a reanalysis dataset. J. Climate, 17, 1828–1844, https://doi.org/10.1175/1520-0442(2004)017<1828:SVITSH>2.0.CO;2.
Onogi, K., and Coauthors, 2007: The JRA-25 Reanalysis. J. Meteor. Soc. Japan, 85, 369–432, https://doi.org/10.2151/jmsj.85.369.
Priestley, M. D. K., D. Ackerley, J. L. Catto, K. I. Hodges, R. E. McDonald, and R. W. Lee, 2020: An overview of the extratropical storm tracks in CMIP6 historical simulations. J. Climate, 33, 6315–6343, https://doi.org/10.1175/JCLI-D-19-0928.1.
Qin, T., L. Wei, and C. Li, 2017: The statistic and variance of cyclones enter in scientific investigation station of China in Antarctic. Acta Oceanogr. Sin., 39, 44–60, https://doi.org/10.3969/j.issn.0253-4193.2017.05.005.
Shaw, T. A., and Coauthors, 2016: Storm track processes and the opposing influences of climate change. Nat. Geosci., 9, 656–664, https://doi.org/10.1038/ngeo2783.
Simmonds, I., and X. Wu, 1993: Cyclone behaviour response to changes in winter Southern Hemisphere sea-ice concentration. Quart. J. Roy. Meteor. Soc., 119, 1121–1148, https://doi.org/10.1002/qj.49711951313.
Simmonds, I., and K. Keay, 2000: Variability of Southern Hemisphere extratropical cyclone behavior, 1958–97. J. Climate, 13, 550–561, https://doi.org/10.1175/1520-0442(2000)013<0550:VOSHEC>2.0.CO;2.
Simmonds, I., and E.-P. Lim, 2009: Biases in the calculation of Southern Hemisphere mean baroclinic eddy growth rate. Geophys. Res. Lett., 36, L01707, https://doi.org/10.1029/2008GL036320.
Simmonds, I., and M. Li, 2021: Trends and variability in polar sea ice, global atmospheric circulations, and baroclinicity. Ann. N. Y. Acad. Sci., 1504, 167–186, https://doi.org/10.1111/nyas.14673.
Trenberth, K. E., 1981: Observed Southern Hemisphere eddy statistics at 500 mb: Frequency and spatial dependence. J. Atmos. Sci., 38, 2585–2605, https://doi.org/10.1175/1520-0469(1981)038<2585:OSHESA>2.0.CO;2.
Trenberth, K. E., 1991: Storm tracks in the Southern Hemisphere. J. Atmos. Sci., 48, 2159–2178, https://doi.org/10.1175/1520-0469(1991)048<2159:STITSH>2.0.CO;2.
Ulbrich, U., G. C. Leckebusch, and J. G. Pinto, 2009: Extra-tropical cyclones in the present and future climate: A review. Theor. Appl. Climatol., 96, 117–131, https://doi.org/10.1007/s00704-008-0083-8.
Valkonen, E., J. Cassano, and E. Cassano, 2021: Arctic cyclones and their interactions with the declining sea ice: A recent climatology. J. Geophys. Res. Atmos., 126, e2020JD034366, https://doi.org/10.1029/2020JD034366.
Wang, S., T. Rashid, H. Throp, and R. Toumi, 2020: A shortening of the life cycle of major tropical cyclones. Geophys. Res. Lett., 47, e2020GL088589, https://doi.org/10.1029/2020GL088589.
Wang, X. L., V. R. Swail, and F. W. Zwiers, 2006: Climatology and changes of extratropical cyclone activity: Comparison of ERA-40 with NCEP–NCAR reanalysis for 1958–2001. J. Climate, 19, 3145–3166, https://doi.org/10.1175/JCLI3781.1.
Wei, L., and T. Qin, 2016: Characteristics of cyclone climatology and variability in the Southern Ocean. Acta Oceanol. Sin., 35, 59–67, https://doi.org/10.1007/s13131-016-0913-y.
Williams, L. N., S. Lee, and S.-W. Son, 2007: Dynamics of the Southern Hemisphere spiral jet. J. Atmos. Sci., 64, 548–563, https://doi.org/10.1175/JAS3939.1.
Yin, J. H., 2005: A consistent poleward shift of the storm tracks in simulations of 21st century climate. Geophys. Res. Lett., 32, L18701, https://doi.org/10.1029/2005GL023684.
Zhang, X., J. E. Walsh, J. Zhang, U. S. Bhatt, and M. Ikeda, 2004: Climatology and interannual variability of Arctic cyclone activity, 1948–2002. J. Climate, 17, 2300–2317, https://doi.org/10.1175/1520-0442(2004)017<2300:CAIVOA>2.0.CO;2.
Zhang, Y., Y. Ding, and Q. Li, 2012: Interdecadal variations of extratropical cyclone activities and storm tracks in the Northern Hemisphere (in Chinese). Chin. J. Atmos. Sci., 36, 912–928.
