Statistical Characteristics of Cyclones in the Southern Oceans during the Summer of 2004 - 2007 ABSTRACT Characteristics of cyclones and explosively developing cyclones (or “bombs”) over the Southern Oceans in the summer months (December, January, February) from 2004 to 2007 are analyzed using the Final Analysis (FNL) data issued by the National Centers for Environmental Prediction (NCEP). Statistical results show that both cyclones and explosively developing cyclones frequently develop in January, and most of them occur within the latitude zone 55°S-70°S. Those cyclones gradually approach to the Antarctic Continent from December to February. In most cases, cyclones and bombs move east-southeastward with some exceptions moving northeastwardly. The lifetime of cyclones is around 2-6 days, and the horizontal scale is about 1000 km. For explosive cyclones, however, the lifetime is about 1 week, and the horizontal scale can reach up to 3000 km. Compared with cyclones developed in the northern hemisphere, cyclones over the southern oceans have much higher occurrence frequencies, lower central pressures and larger horizontal scales, which may be caused by the unique geographical environment of the southern hemisphere. Key Words: cyclone, explosive cyclone, southern hemisphere, lifetime, horizontal scale. 1. Introduction Explosive cyclones, referred to as meteorological “bombs,” are usually accompanied with rapid central pressure reduction and extreme intensity. These cyclones are hard to predict and can pose serious threats to human life and property when they approach coastal regions, or intersect with shipping lanes. Explosive cyclones are also considered to be an important contributor to global climate change and energy transportation. Over the last two decades explosive cyclones have received considerable attention from the research community, much research are directed toward the exploration of the mechanisms that trigger rapid pressure reduction and to documenting some of the mean characteristics of NH explosive cyclones. According to the landmark study of Sanders and Gyakum (1980), explosive cyclone events are predominantly maritime and cold-season phenomena. Moreover, in the Northern Hemisphere it has been observed that the maximum frequency of bombs occurs in the westernmost portions of both the Atlantic and Pacific Oceans within, or just north of the warm currents of the Gulf Stream and the Kuroshio, respectively (Sanders and Gyakum 1980; Chen et al. 1992). Chen et al. (1992) showed that in the northwestern Pacific region, explosive cyclones often appeared over the Sea of Japan and over the Kuroshio Current. They also suggested that two peaks of monthly frequency existed in early winter (January) and late winter (March). Yoshida and Asuma (2003) reported the structures and environment of explosively developing extratropical cyclones in the northwestern pacific region. In contrast, there have been only a few studies about cyclones over the southern hemisphere. Sinclair (1995) presented a compilation of rapidly developing cyclones using 7 yr (1980-86) of operational European Centre for Medium-Range Weather Forecasts (ECMWF) analyses. Lim and Simmonds (2002) showed a compilation of Southern Hemisphere explosively developing cyclones which assembled based on the National Centers for Environmental Prediction-Department of Energy reanalysis-2 data over the 21-yr period form 1979-1999. They also discussed the characteristics of bombs in the SH and a comparison with NH events, and pointed out that summer explosive cyclones in the SH are six times more numerous than those in the NH, although there are about twice as many bombs appear in the NH as in the SH every year. Chen et al (2000) found that there were much more explosive cyclones developed during summer time over the southern hemisphere. The purpose of the present paper is to examine the statistical characteristics of the entire cyclones and explosive cyclones developed during the summer over the Southern Ocean. In particular, we describe the geographical and monthly variations and the statistical properties of these events. Data sources are described in section 2. 2. Data and methods This study is based on the Final Analysis (FNL) dataset provided by the National Centers for Environmental Prediction (NCEP). It is a global gridded dataset with a 1o×1o latitude-longitude resolution and 26 vertical levels from the surface to 10 hPa at 6 hour intervals. In this study, we examine the cyclones that occur during three summers, including the following months: December 2004 , January 2005, February 2005, December 2005, January 2006, February 2006, December 2006, January 2007, February 2007. The spatial coverage is between 40°S and 90°S in latitude and between 0° and 360° in longitude (Fig. 1). Synchronous GOES-9 infrared satellite images obtained from Kochi University are also used in this study. Sanders and Gyakum (1980) defined an explosively developing extratropical cyclone as a cyclone that had a central sea level pressure decrease normalized at 60 N over 24 hPa in a day (24h). The cyclone deepening rate (Bergeron) is calculated from the following definition for each cyclone: Cyclone deepening rate =[ pt 6 pt 6 sin( 60) ][ ] t 6 t 6 12hPa sin 2 (1) where t is analysis time in hours, p is the sea level pressure at the cyclone center, and φ is the latitude at the cyclone center. A 12-h pressure change is used in this paper to find an instance of most rapid deepening in a cyclone’s life. An explosively developing cyclone is defined as cyclones having a deepening rate of at least 1 Bergeron. Cyclones that disappear within 24 h after initial appearance are excluded from the analysis. 3. Results 3.1 Larger-scale environment Previous studies show that environmental baroclinicity affects the geographical distribution of explosive cyclones and cyclone structures may change under the influence of atmospheric and geographical environments. The Antarctic continent is often surrounded by a distinct low-level baraclinic zone. Cyclones in the summer with high intensity and low central pressure (lower than 985 hpa) are found over the entire region south of 40oS with averaged moving speeds over 100 km/h. Much of the anticyclones are located in the subtropical area and along the Antarctic coast. Baroclinicity plays an important role in SH explosive cyclone occurrence as in the NH. However, the SH baroclinicity is less affected by the seasonal cycle with a mean circumpolar trough existing (where?) associated with the mean sea level pressure. The maximum propagation speed of migrating cyclones in SH is about twice as fast as those in NH, and the situation is similar for the mean westerly (Qu et al. 2001). From December to February, sea surface temperature increases and land temperature decreases gradually in SH, which result in an increasing temperature difference between ocean and land. [Kuo: Do we have plots to show that?] 3.2 Number of cyclones and bombs The frequency of occurrence for summer cyclones and explosive cyclones from Dec 2004 to Feb 2007 is shown in Table 1. In all, a total of 734 cyclones occurred during these months, and among which 135 explosive cyclones were found, which represent 18.4% of the all cyclones over the Southern Ocean. There were 234 cyclones and 47 explosive cyclones during summer 2004~2005, 247 cyclones and 38 explosive cyclones for 2005~2006, 252 cyclones and 50 explosive cyclones for 2006~2007. January is the month with most frequent explosive cyclones, and most cyclones also develop in January except for year 2006~2007. January 2006 and 2005 are the two special moths when cyclones and explosive cyclones develop with highest frequency respectively. There were about 75 cyclones and 13 explosive cyclones on average with a little difference for December 2006. [Kuo: Your table says 92 total cyclones and 17 explosive cyclones. They are not consistent with your statement.] 3.3 Cyclone and bomb characteristics Because much of the SH landmass (except for Antarctica) is located in the subtropical region, which is different from the NH, cyclones in the SH often display the characteristics of high occurrence frequency, low central pressure, and large horizontal scale. Statistical analysis was used to study the locations of all the cyclones every time they occurred. [Kuo: What does this mean? Do you mean the “genesis”? or the same cyclone would be counted multiple times during its life time?] In order to show the spatial distributions of cyclones clearly, we divided the cyclones into four types according to the central pressure: Type A with the central pressure lower than 970 hPa, Type B between 970 hpa and 980 hpa, Type C between 980 hpa and 990 hpa, Type D higher than 990 hpa. [Kuo: Why would this be a good way to characterize cyclones? Is the FNL analysis reliable enough for such categorization?] a. Locations of entire cyclone Figure 2 shows locations of all the cyclones from their initial stage to their dissipated stage for a specific month. In December 2004 (Fig. 2a), cyclones always occurred within the zone between 50°S and 70°S in latitudes. Between 40°S and 50°S, the distribution of cyclones was not homogeneous, most cyclones appeared around the New Zealand. A small number of cyclones of which the central pressures are about 980 hpa, also existed in the areas between 0°~25°W and 120°~150°W in longitude. The west of Antarctic peninsula between 50°S and 70°S, north of Rose sea and around of 20°E and 90°E between 50°S and 65°S is a favored place of type A. [Kuo: Frankly, I have problem with this type of classification. As you know, in the past, scientists define different types due to different forcing mechanism. Here, we define it purely on central pressure. As a cyclone would go through its life cycle, with varying central pressure, this is not “physically” representative of forcing mechanisms.] The loci of cyclones for Dec 2005 did not differ much from those of Dec 2004. However, the amount, density and the spatial scales of cyclones between 40°S and 50°S were all much higher in Dec 2005 than those of Dec 2004. They occurred not only around the New Zealand, but also expanded to the other areas. Cyclones in Type A still occurred mostly in the Antarctic Peninsula, Rose sea and 20°E, 90°E but the amount and density were also much larger than the previous year. The distributions of cyclones in Dec 2006 is similar to those of Dec 2005, except that more cyclones in Type A clustered around the Antarctic Peninsula Cyclones mainly stayed in the latitude zone between 55°S-75°S in Jan 2005, but it is much higher number than that in Dec 2004 (Fig. 2b). Between 40°S and 50°S, there were still a few cyclones, but the density became much lower and most of them belonged to Type D. The latitudinal distributions of Type A are more dispersive than that in Dec 2004, and most cyclones converged towards the higher latitude area. Compared with Jan 2005, many cyclones in Jan 2006 moved a little toward lower latitude areas and there were also a lot of Type C and Type D cyclones existing in the mid-latitude areas. However, in comparison with Dec 2005, most cyclones still accumulated towards higher latitude area in Jan 2006. Different from the former two years, the amount of cyclones became lower and positions were highly dispersed, but the trend of cyclones converging toward the higher latitude as the season progressed didn’t change. Figure 2c shows that the trend that cyclones moved to higher latitude from December 2004 to February 2005 was even more clear. There were some areas where almost no cyclones existing within the latitude zone 40°S~50°S (like 90W~120W and 60E~90E) in February 2005. Latitudinal distributions of type A cyclones became more even and most of them concentrated in the latitude zone 60°S-70°S. Compared with the other months of the same year, positions of cyclones moved to the relatively higher latitude in Feb 2006 and Feb 2007, but the spatial distributions became much dispersed. Many cyclones still appear in the low- and mid-latitude, and Type A and Type B were more evenly centered on the broader latitude zone of 50°S-70°S. As is described above, we can see that most cyclones accumulate towards the higher latitude area from December to February, and the latitude range of 55°S-70°S is the area where most cyclones exist. Type A cyclones are evenly distributed along the latitude range, except for the west to the Antarctic Peninsula where Type A cyclones often have high concentration. Distinct differences among these three summers are found that is the amount of cyclones is larger, the density is higher, and the horizontal distribution is broader in the second and the third year than those in the first year. b. Entire cyclone tracks Figure 3 shows the monthly distributions of cyclone tracks (plotted with 6-h segments over the 24-h of cyclones, the dots denote the locations of formation) compiled for the 6 month of record. [Kuo: Actually, I think identifying the “genesis” and “dying” locations of these cyclones are interesting. We will find “where” will the cyclone form, develop and die. We can make the plots without the cyclone tracks. Another important point is “where” does explosive cyclogenesis take place. To me, the more interesting result is to see “where” does cyclone experience more than 1 bergeron deepening. I am sure that will be different from just the distribution of cyclones.] From these figures, one can find that cyclones form over the east of South America and surround the Antarctic Peninsula at all the nine months. A chain of cyclones are found to extend from these regions to the south of Australia and Tasman Sea. The tracks of the entire cyclone population are very dense near the circumpolar trough. Most of cyclonic systems have a strong southeastward component and some northeast movement of cyclones was observed clearly around 50°S. It is noticed that most of these cyclones move to avoid the sea ice covered regions. [Kuo: It would be useful to present a figure of sea ice for each month. How do they change, and how they affect the cyclone track.] Carleton and Fitch (1993) observed a number of meso-cyclones moving northward in Antarctic and sub-Antarctic latitudes in winter and associated it with the Antarctic sea ice expansion and katabatic outflow. The tracks of summer cyclones in these three years all move towards the high latitude region, which is similar to the distributions of their positions. However, the important difference among these three years is that activities of the cyclones in the first year always exist in the area closer to the Antarctica than the second and the third year. [Kuo: Are there significant differences in the sea ice extend between these three years? We should find out.] c. Distributions and tracks of bomb Figure 4 shows the tracks of bombs, locations of formation, maximum deepening, and minimum center pressure. The distribution of bomb tracks is much like that of the entire cyclone population and both of them show spiral bands stretching from South America and Australia (Jones and Simmonds 1993). From these figures, one can see that most explosive cyclones form on the mid-latitude. Southwest of Australia, northeast of Antarctic Peninsula, and the vicinity of 20°E and 60°E are still the favored places for the bombs formation for all nine months. [Kuo: Are these areas different from the regular cyclones?] The maximum deepening almost always occurred over the ocean at latitudes zone 45°S-60°S, and the locations of minimum center pressure are centered on the high-latitude zone 60°S-70°S. Most of the tracks move east-southeastward and some specials are mostly found south of the South America and around of the Antarctic Peninsula. d. Intensity of entire cyclones The central pressure of cyclones in the SH is much lower than in the NH, of which the lowest can reach 940hPa and the explosive cyclone can reach 946hPa in this study. [Kuo: It is strange to have the regular cyclone be deeper than an explosive cyclone. Is there something wrong?] Table 2 shows the statistics of each type of cyclones based on 6-h FNL analysis. About 67% of the SH cyclones are with central pressures lower than 980 hPa. The number of type A and type B increases gradually from summer to autumn with decreasing number of type C and type D, indicating the increasing intensity of SH cyclones during this seasonal transition. [Kuo: I suggest that we don’t use “type” to indicate cyclone of different intensity. May be “category” would be better than “type”. When people use “type” they usually mean significant different forcing mechanisms for cyclones.] e. The life-time and horizontal scales The life-time of each cyclone can be estimated by the analysis dataset. Results show that the average life-time of a cyclone is about 2-6 days and the maximum life-time can reach 10.5 days. The life time of explosive cyclones is about one week. The diameter range of a SH cyclone is about 1000Km, and the maximum can reach 5000-6000Km. Most explosive cyclones are extratropical cyclones which is always associated with a front in the structures. [Kuo: Are you saying regular cyclones are not associated with a front? It seems cyclones are always associated with fronts. I would be interested in seeing cyclones that are not associated with fronts.] The average diameter range of a SH explosive cyclone is about 3000 km. f. Satellite images Figure 5 shows the GOES-9 infrared satellite image over the southern oceans at 18 UTC 31 January 2005. Symbols A, B and C indicate ordinary cyclones, while D and E represent the explosive cyclones. The explosive cyclone E whose deepening rate is 1.85 Bergeron and central pressure is 954 hPa, and is reaching its maximum intensity. The deepening rate fits the “strong intensification category” (Table 3). The explosive cyclone D whose deepening rate is 1.01 Bergeron and central pressure is 992 hPa, is in its initial development stage. The deepening rate fits the weak intensification category (Table 3). Cyclones A, B, and C are all mature ordinary cyclones, with central pressures of 971 hPa, 970hPa, and 960 hPa, respectively. SH cyclones rotate clockwise which is opposite to NH cyclones, and they are often connected with long horizontal fronts (E is over 3000 km, and C can reach 5000 km) and with no obvious eyes found in some NH explosive cyclones (Kuo et al. 1992). g. Intensity of bombs Followed by Sanders (1986), cyclones identified as explosive developing cyclones can further be classified into three categories of intensification: strong intensification (ST) cyclone (deepening rate larger than 1.8 Bergeron), moderate intensification (MO) cyclone (deepening rate between 1.3 and 1.8 Bergeron), and weak intensification (WE) cyclone (deepening rate is less than 1.3 Bergeron). The occurrence frequency of each type is summarized in Table 3. During the three summers from Dec 2004 to Feb 2007, WE cyclones occurred most frequently (79 cases), followed by MO cyclones (44 cases), and ST cyclones (12 cases). The average deepening rate of these cyclones is 1.32 Bergeron, and the strongest deepening rate is about 2.35 Bergeron. Cyclones occurred in the first summer are the weakest of the three, and the average deepening rate is within 1.2-1.25 Bergeron. The last summer cyclones are middle, and the second summer cyclones are the strongest. So, there are also noticeable interannual variability. 4. Conclusions To better understand the characteristics of cyclones and explosively developing cyclones (or “bombs”) over the Southern Oceans in the summer (December, January, February), statistical analysis are used in this study by analyzing the Final Analysis (FNL) data provided by the National Centers for Environmental Prediction (NCEP) from 2004 to 2007. Our results reveal that cyclones and explosively developing cyclones frequently occurred in January. During the seasonal transition from summer to autumn, with the increasing land-ocean temperature differences, cyclones gradually move to the Antarctic Continent, and most cyclones accumulate within the latitude zone 55°S-70°S. [Kuo: I think we need to look at the sea-ice and cyclone location connection.] Type A cyclones are evenly distributed along the latitude, except for the west to the Antarctic Peninsula where type A cyclones occur a lot. Cyclones often form over the east of South America and around the Antarctic Peninsula area, move across the Atlantic Ocean and Indian Ocean, result in a formation of a long chain of cyclones between the south of Australia and South America. [Kuo: Are these cyclone triggered by mountains of Antarctic Peninsula? Are they orographic cyclogenesis?] In general, SH cyclones move east-southeastward, with some exceptions moving northeast mostly found near 50°S and around the Antarctic Continent. Distributions of cyclone tracks are often found around the persistent circumpolar trough. Most explosive cyclones form on the mid-latitude, specifically the west south of Australia, east north of Antarctic Peninsula, the vicinity of 20°E and 60°E are the favored places for the bombs formation. The maximum deepening of explosive cyclones always occur within latitude range 45°S-60°S [Kuo: Are these some special “longitude zones” that favor explosive cyclogenesis? If so, why? Can we speculate as to why explosive cyclones have their deepening over these regions? Actually, I am thinking that one can plot the distribution of the “deepening rate” horizontally for each month, and see if we can identify some clear pattern.], while the locations of minimum center pressure are found within the high-latitude zone 60°S-70°S. Most of the summer explosive cyclones move east-southeast, while some moving northeast are mostly found south of the South America and around of the Antarctic Peninsula. The lifetimes of cyclone over the southern oceans are around 2-6 days, and their horizontal scales are around 1000 km. For explosive cyclones, the lifetimes are about 1 week and horizontal scales are nearly 3000 km. Overall comments: This is a very interesting study. However, a review may ask the following questions: 1. What are the same and different from previous studies of Antarctic cyclones? Previous work cover the entire year, here you only cover the summer. 2. Did the previous study also identify the migration of cyclones toward the Antarctic continent as the season progresses? 3. Are there preferred locations for “explosive development”? Why? 4. Are explosive cyclones over the SH distinctly different from the regular cyclones? Or, are they just a little bid stronger? Remember that in the Sanders and Gyakum paper, the fundamental question is: “Are explosive cyclones fundamentally different types of cyclones from the non-explosive cyclones?” 5. It seems that if you want to pursue numerical simulations, we should look at one “ordinary cyclone” and one “explosive cyclone”, and study their differences. 6. Are there ways we can examine the precipitation (and latent heat releases) associated with these explosive cyclones? Are they stronger because of stronger “latent heating?” 7. Are the distribution of cyclones affected by the extend of sea ice? Can you quantify the relationships? Acknowledge References: Carleton, A.m., and M. Fitch, 1993: Synoptic aspects of Antarctic mesocyclones. J. Geohpys. Res., 98, 12 997-13 018. Chen, J.-N., K.-T. Le, C.-M. Jia, and Y. Peng, 2000: The change of the frequency of cyclone occurrence in the southern hemisphere influence on SST in the East Pacific and Southern Oscillation. Acta. Oceanologica. Sinica., 22(3), 86-93. Chen, S.-J., Y.-H. Kuo, P.-Z. Zhang, and Q.-F. Bai, 1991: Synoptic climatology of cyclogenesis over east Asia, 1958-1987. Mon. Wea. Rev., 119, 1407-1418. Kuo, Y.-H., R. J. Reed and S. Low-Nam, 1992: Thermal structure and airflow in a model simulation of an occluded marine cyclone. Monthly Weather Review, 120, 2280-2297 Lim, E.-P., and I. Simmonds, 2002: Explosive cyclone development in the Southern Hemisphere and a comparison with Northern Hemisphere events. Mon. Wea. Rev., 130, 2188-2209. Qu, W.-Z., X.-R. Chen, R.-S. Shen, and H.-G. Wang, 2001: Some Characteristics of General Circulation at Sea Level of the Southern Hemisphere. Journal Of Oceanography Of Hanghai & Bohai Seas, 19(1), 9-16. Sanders, F., and J. R. Gyakum, 1980: Synoptic-dynamic climatology of the ‘bomb’. Mon. Wea. Rev., 108, 1589-1606. Simmonds, I., K. Keay, 2000: Variability of Southern Hemisphere extratropical cyclones behavior, 1958-97. J. Climate, 13, 550-561. Simmonds, I., K. Keay, 2000:Mean Southern Hemisphere extratropical cyclones behavior in the 40-year NCEP-NCAR reanalysis. J. Climate, 13, 873-885. Simmonds, I., K. Keay, and E.-P. Lim, 2003: Synoptic activity in the seas around Antarctica. Mon. Wea. Rev., 131, 272-288. Sinclair, M. R., 1995: A Climatology of Cyclogenesis for the Southern Hemisphere. Mon. Wea. Rev., 123, 1601-1919. Yoshida, A., and Y. Asuma, 2003: Structures and environment of explosively developing extratropical cyclones in the Northwestern Pacific region. Mon. Wea. Rev., 132, 1121-1142. Pacific South America Antarctica peninsula New Zealand Ross Sea Tasmania Antarctica Woddell Sea O Indian 80 S 70O S Atlantic 60O S 50O S Figure 1: Geographic map of the southern hemisphere (latitude ranging from 90oS-40oS) o 60 E 2006-01 0o 0o oE 120 0o 8 o 0S 8 o 0S 7 o 0S 7 o 0S 6 o 0S e 6 o 0S 5 o 0S 5 o 0S 18 0oW 0oW 18 120 oW o 80 o S 60 o S 2004-12 6 0W 70 o S o 60 E 0o 0o 0o 8 o 0S 8 o 0S 80 o S 7 o 0S 7 o 0S 70 o S 60 o S 6 o 0S 6 o 0S o a 6 0W o 6 0W oE 120 2005-12 d oE 120 oW 60 oE 120 6 0 oE o 6 0 oE 6 0W oE 120 50 o S oW 60 oE 120 18 0o W 5 o 0S 5 o 0S 50 o S 18 18 0oW 0oW 18 0o W 120 oW o 120 W o 120 W b o 60 E 2005-01 o 120 W c 2005-02 o 120 W o 60 E f 2006-02 o 120 W o 120 W 0oW 6 o 0S 7 o 0S 8 o 0S 80 o S 8 o 0S 70 o S 7 o 0S g o 60 E 2007-01 0o 0o 0o oE 120 oE 120 2006-12 o 6 o 0S o oW oE 120 60 oE 6 0W 6 0W 60 60 o S 5 o 0S 50 o S 5 o 0S 18 18 18 0o W 0oW 120 oW h i o 60 E 2007-02 Figure 2: Location distributions of cyclones (symbol ●denotes type A,▼is type B,□is type C,╋is type D)。 8 o 0S 8 o 0S 7 o 0S 7 o 0S 6 o 0S o 60 E e 2006-01 6 o 0S 5 o 0S 5 o 0S 18 18 0oW 0oW W 0o 0o oE 120 0o 120 o o 70 o S 2004-12 6 0W 80 o S 50 o S o 60 E 0o 0o 0o 80 o S 8 o 0S 80 o S 70 o S 7 o 0S 70 o S 60 o S 6 o 0S 60 o S oW a 60 oE 120 o 2005-12 d 6 0W oE 120 oE 120 6 0 oE o 6 0 oE 6 0W oE 120 oW 60 60 o S oW 60 oE 120 18 0o W 50 o S 5 o 0S 50 o S 0oW 18 0o W 18 18 0o W 120 oW o 120 W 120 oW b 6 0 oE 2005-01 o 120 W c 2005-02 o 120 W o 60 E f 2006-02 o 120 W o 120 W 5 o 0S 6 o 0S 60 o S 8 o 0S 8 o 0S 7 o 0S 7 o 0S 70 o S 80 o S o o oW oE 120 0o 0o 0o oE 120 2006-12 6 0W 6 0W 60 oE 120 60 oE g h o 60 E 2007-01 Figure 3: Tracks of cyclones over the southern oceans (black dot track). 6 o 0S 50 o S 5 o 0S 18 18 18 0o W 0oW 0oW 120 oW i o 60 E 2007-02 ● indicates the initial position of cyclone, solid line represents the moving o 60 E 2006-01 8 o 0S 8 o 0S 7 o 0S 7 o 0S 6 o 0S e 6 o 0S 5 o 0S 5 o 0S 18 18 0oW 0oW W 0o 0o oE 120 0o 120 o o 70 o S 2004-12 6 0W 80 o S 50 o S o 60 E 2005-01 0o 0o 0o 8 o 0S 8 o 0S 80 o S 7 o 0S 7 o 0S 70 o S 60 o S 6 o 0S 6 o 0S o a 6 0W oE 120 o 2005-12 d 6 0W oE 120 oE 120 6 0 oE o 6 0 oE 6 0W oE 120 oW 60 60 o S oW 60 oE 120 18 0o W 5 o 0S 5 o 0S 50 o S 18 18 0oW 0oW 18 0o W 120 oW o 120 W o 120 W b o 60 E o 120 W c 2005-02 o 120 W o 60 E 2006-02 f o 120 W o 120 W 0oW 18 5 o 0S 6 o 0S o 6 o 0S 7 o 0S 7 o 0S 8 o 0S 8 o 0S 70 o S 80 o S g o 60 E 2007-01 h 0o 0o 0o oE 120 oE 120 2006-12 o oW oE 120 6 0 oE 6 0W 60 60 o S 6 0W 50 o S 5 o 0S 18 0o W 18 0oW 120 oW o 60 E i 2007-02 Figure 4:Tracks of explosive cyclones over the southern oceans (black diamond ◆ indicates the initial position of cyclone, star ★ denotes the position of maximum surface pressure depressing, and triangle ▼represents the location of lowest central pressure). B A E C D Figure 5: GOES-9 infrared satellite imagery over the southern oceans at 18 UTC 31 January 2005. Symbol A, B and C indicate ordinary cyclones, while D and E represent the explosive cyclones. Table 1 Monthly numbers of cyclones and explosive cyclones in the summer over the Southern Ocean from 2004 to 2006 2004~2005 Dec Jan Feb Cyclone number 75 83 76 Explosive cyclone number 12 20 The proportion of explosive cyclone 16 24.1 2005~2006 total 2006~2007 Total Dec Jan Feb total Dec Jan Feb total 234 76 95 75 247 92 82 79 253 734 15 47 13 13 12 38 17 19 14 50 135 19.7 20 17.1 13.7 16 15.4 18.5 23.1 17.7 19.8 18.4 Table. 2 Monthly numbers of various types of cyclones Month Number of type A(<970 hPa) 2004-12 2005-01 2005-02 367 337 481 Number of type B(970-980 hPa) 383 417 373 2005-12 2006-01 2006-02 338 352 370 382 352 420 2006-12 427 2007-01 Number of type C(980-99 0 hPa) 418 300 200 Number of type D(>990 hPa) Total 128 149 48 1296 1203 1102 346 250 191 166 174 95 1232 1128 1076 386 241 149 1203 462 345 182 105 1094 2007-02 342 346 231 90 1009 Total 3476 3404 2359 1104 10343 Table 3. Monthly number of explosive cyclones over the southern oceans Month Strong intensification cyclone (>1.8Bergeron) Moderate intensification cyclone(1.3-1.8 Bergeron) Weak intensification cyclone(<1.3 Bergeron) Total Average cyclone deepening rate 2004-12 2005-01 2005-02 0 1 0 3 5 3 9 14 12 12 20 15 1.23 1.24 1.20 2005-12 2006-01 2006-02 2 1 3 6 5 2 5 7 7 13 13 12 1.39 1.34 1.43 2006-12 1 6 10 17 1.29 2007-01 2 11 6 19 1.46 2007-02 2 3 9 14 1.31 Total 12 44 79 135 1.32