BEHAVIOUR OF SOUTHERN OCEAN CYCLONES

for

The International Antarctic Weather Forecasting Handbook:

IPY 2007-08 Supplement

Ian Simmonds

School of Earth Sciences

The University of Melbourne

Victoria, 3010

simmonds@unimelb.edu.au

Submitted April 2008

*Contribution relevant to Chapter 2 An Overview of the Meteorology and Climatology of the Antarctic.

Editors’ note: at this time, the contribution has not been adapted to the original Handbook style, especially wrt numbering of figures etc.

Introduction

The forecasting of weather in the Antarctic region is intimately tied up with the behaviour of cyclonic systems in the locality. Given the strong thermal contrast between the Antarctic continent and the oceans to the north (and hence an ample supply of ‘available potential energy’) it is not surprising that the coastal and subAntarctic regions are host to numerous and intense cyclones throughout the year. The appropriate forecasting of these features relies on a broad understanding of their properties and spatial distributions, and indeed on how those characteristics may be changing. In the past obtaining appropriate structural information about these features has been difficult for a number of reasons. Included in these was the fact that the subAntarctic was very poorly observed and also that the analysis techniques that were applied were not always appropriate to the region, given the influence of steep topography, the presence of sea-ice with diverse concentrations, etc. The cyclonic interactions with sea-ice, in particular, present challenges to the understanding and diagnosing of atmospheric structure (e.g., Watkins and Simmonds 1995, 1998, Simmonds et al. 2005, Wassermann et al. 2006).

The first of these problems has essentially been addressed with the continued development of satellite-based monitoring and the advances in sophisticated inversion techniques to obtain estimates of meteorological parameters from raw radiances. Considerable strides were made in the second problem mentioned by using the process of data assimilation within a numerical model. This procedure has now been used for over three decades (see, e.g., Simmonds 1978, Kalnay 2003) and now forms the basis of the analyses of weather services throughout the world.

In obtaining the most reliable climatological picture of cyclonic systems in the Antarctic region it is imperative that the best analyses be used over the period of interest. The operational analyses produced by the data assimilation process referred to above are of immense value but can only be as good as, in part, the numerical model which was used as the assimilating vehicle. According, operational analyses which were generated, say, ten years ago used the model which was current at that time, but which would be clearly inferior to that used today. Hence the quality of the analyses, and hence the cyclone statistics derived from them, could conceivably give a distorted picture on cyclone behaviour. A partial solution to this problem has been to perform ‘re-analyses’. By this is meant making use of a current state-of-the-art global forecast model to assimilate in a time-marching fashion all the data when and where they are available. A number of these re-analysis sets have been constructed and are freely available.

Data sets and techniques

While the reanalysis data sets are of high quality it is important to bear in mind they do exhibit differences, particularly with respect to their cyclone frequency and structure. It is difficult to make a judgment that one particular re-analysis is better than another. The sets have their various strengths and weaknesses (see, e.g., Onogi et al. 2007), and it is important that one particular set should not be seen as the ‘truth’ (see also the comments of Raible et al. (2008)). Comparing cyclone statistics from two (or more) re-analyses allows one to obtain potentially an estimate of the uncertainties in the climatologies so-produced and, accordingly, can give a measure of the level of confidence which can be placed in the various products.

We make use the (global) analyses from the National Centers for Environment Prediction - Department of Energy (NCEP2) (Kanamitsu et al. 2002), and the Japanese Reanalysis (JRA-25) (Onogi et al. 2007). (We include the recent JRA-25 set as it may have a number of advantages over suites previously used for cyclone analysis studies, and may benefit from the fact that it assimilates additional data sources.) Studies by Bromwich et al. (2007) and others have shown that the re-analyses present a reliable description of the polar troposphere, and that they are a powerful tool for climate studies in the polar regions. These data are available on a 2.5^o x 2.5^o latitude-longitude grid every 6 hours. We confine ourselves to displaying results of our survey for the 29 summers (DJF) (from 1 December 1979 – 28 February 1980 to 1 December 2007 – 28 February 2008), and the 29 winters (JJA) (from 1 June – 31 August 1979 to 1 June – 31 August 2007). The starting point of 1979 is chosen as it marks the commencement of the modern ‘satellite era’ (see, e.g., Uppala et al. 2005), and hence the three-decade period explored here represents the best ever global atmospheric coverage. (See also the assessment of Bromwich and Fogt (2004).)

To obtain an overall view of the behaviour of cyclones using manual techniques is clearly not practicable. Accordingly, we base our investigation on an automatic and objective cyclone identification package. The software we use is The University of Melbourne cyclone scheme (e.g., Simmonds and Keay 2000a, Simmonds et al. 2003). This algorithm has been shown to perform well in a number of independent comparisons of mid- and high-latitude cyclone behaviour (e.g., Leonard et al. 1999, Pinto et al. 2005, Raible et al. 2008). We here confine our attention to mean sea level pressure cyclones (analyses of Antarctic cyclone behaviour at upper levels can be found in, e.g., Keable et al. (2002) and Lim and Simmonds (2007)). The scheme determines many aspects of the morphology of each identified cyclone, including the central pressure, intensity, radius, and depth. In the discussion below lows are considered as ‘weak’ or ‘strong’, depending on whether Ñ²p in the vicinity of the centre of a low assumes values of 0.2-0.7 hPa (deg. lat.)^-2, or greater than 0.7 hPa (deg. lat.)^-2, respectively. The ‘radius’ is taken as the weighted mean distance from the cyclone centre to the points at which Ñ²p is zero around the ‘edge’ of the cyclone, and the ‘depth’ is the difference between the pressure at the edge of a cyclone and that at the center (and as such is very similar to the ‘pressure deficit’ used by synopticians (e.g., Nielsen and Dole, 1992)). It can be shown that the depth for an individual system can be expressed in terms of the product of the Laplacian and the square of the radius. This depth is arguably the most useful cyclone characteristic to quantify strength and it can be shown to be related to the ‘circulation’ around a system. (More details of the meaning and calculation of these parameters can be found in Simmonds et al. (1999) and Simmonds (2000).)

Cyclone tracks and climatology

As mentioned above, the Antarctic region is host to much cyclonic activity. This is evident in Fig. 1 (Cyclones) which shows all the cyclone tracks in July 2006 (blue) and July 2007 (red) in the NCEP2 and the JRA-25 re-analyses. Overall the plots are very similar and many specific cyclones can be identified in both sets. There are also some interesting contrasts which show up in the mean frequency plots. Fig. 2 (a, b) (Cyclones) shows the mean summer distribution of the system density (the mean number of cyclones per analysis found in 10³ (deg. lat.)² in NCEP2 and JRA-25. Both indicate a very high density of cyclonic centres near or south of 60^oS, particularly in the Indian Ocean, and to the north of the Ross, Bellingshausen and Weddell Seas. Similar patterns are apparent in winter (Fig. 2 (c, d) Cyclones)), except that the highest system density hugs the coast in the eastern hemisphere and the numbers tend to be a little smaller in the western hemisphere in winter. The plots allow an appreciation of the subtle differences between the products derived from NCEP2 and JRA-25. These differences are made clearer in the difference maps displayed in Fig. 2 (e, f) (Cyclones). The plots reveal that in the Indian sector around to the Dateline the JRA-25 finds fewer systems in summer but more in winter.

The seasonal dependence of cyclogenic activity in the NCEP2 re-analysis is shown in Fig. 3 (Cyclones). In both seasons the maxima of generation are found poleward of 60^oS, and in general the rate of genesis is greater in winter. A feature of note found in both seasons is the high rate of cyclone development immediately to the east of the northern part of the Antarctic Peninsula.

Figs. 4 and 5 exhibit the summer and winter means (in the NCEP2 re-analysis) of two of our key characteristics of cyclone structure, namely the ‘radius’ and ‘depth’. Not only does the region to the south of 60^oS host the largest number of cyclones but, as Fig. 4 (Cyclones) indicates, on average, the largest systems are also found there. In both seasons the longitudinal distribution indicates that the largest systems lie to the south of Africa, in the Indian Ocean, and to the north of the Amundsen and Bellingshausen Seas. It may seem unexpected at first sight that, in the mean, the summer cyclones assume a larger size than their winter counterparts. The context of this result becomes clear when the seasonal distribution of mean cyclone depth is considered (Fig. 5 (Cyclones)). The mean depth of winter systems is greater than their summer equivalent over virtually the entire subAntarctic region, particularly in the Indian Ocean sector. Broadly speaking these results indicate that subAntarctic cyclones are smaller and more intense (and more influential) in winter.

Trends and variability in cyclone behaviour

It is well appreciated that the weather and climate systems are inherently variable. In particular the atmosphere over the Southern Ocean is influenced by a myriad of modes of variability and instability (see, e.g., Simmonds 2003, Simmonds and King 2004, White and Simmonds 2006, Pezza et al. 2007). It follows that the mean structures we have just examined need to be seen against this background of variability.

To set the scene for this examination we show in Fig. 6 (Cyclones) the linear trend (least squares best fit) in the summer and winter cyclone system density (NCEP2) over the 29 seasons or our records. The local statistical significance (95% confidence level indicated by stippling) of these trends is quite modest (although it is more marked in winter). However, it is apparent that much of the domain of interest here is host to decreases in cyclone numbers. (This result is consistent with the earlier findings of Simmonds and Keay (2000b) There are a few regions in the vicinity of the Antarctic coast which exhibit increases in cyclone numbers. These include the Drake Passage and the Weddell Sea which have been host to increasing numbers of cyclones in both seasons.

Figure 7 (Cyclones) displays the time series of the mean number of cyclones per analysis between 50 and 70^oS in the NCEP2 and JRA-25 re-analyses for summer and winter. These series have also been drawn separately for the counts of ‘strong’ and ‘weak’ systems. The plots reflect considerable interannual variability. It is comforting to note that the nature of the variability is rather similar in the NCEP2 and JRA-25 compilations, although the total number of summer systems in the Japanese re-analysis is smaller than in NCEP2. Part (a) of the Figure suggest a downward trend in the number of weak summer cyclones, and it is interesting to note that such weak cyclones make up about one third of the cyclone population in that season. By contrast, in winter strong cyclones make up a greater proportion of the total population.

The radius and depth of cyclonic systems are related to the structure of the diagnosed cyclones, and are hence more subtle and perhaps more informative than simply counting cyclone occurrence. Given this, it is not surprising that their time series exhibit much variability and that the time series derived from the NCEP2 and JRA-25 sets show noticeable differences. Fig. 8 (Cyclones) shows the time series of the mean radius of cyclones which lie in the 50-70^oS latitude band. Both data sets indicate that the mean radius of summer cyclones (part (a)) exhibited an increase up to the mid 1990s (an increase in size of about 5% from the start of our record) with the suggestion of systems becoming smaller since that time. In contrast, in winter (Fig. 8 (b)) while NCEP2 indicated a trend to systems being larger on average this is not nearly so apparent in the JRA-25 set. Cyclones diagnosed from the JRA-25 are smaller in every year than those in NCEP2, a difference which is particularly marked after the late 1980s.

The time series of the mean depth of cyclones in our subAntarctic latitude belt (Fig. 9 (Cylones)) show trends to deeper systems in both seasons and in both re-analysis sets. While the time series are fairly noisy they reflect an increase of 10-15% in cyclone depth over our 29 years of record. We mentioned above that our cyclone depth parameter is perhaps the single most informative parameter on net cyclone influence so these changes can be seen as quite dramatic.

References cited by Simmonds

Bromwich, D. H., and R. L. Fogt, 2004: Strong trends in the skill of the ERA-40 and NCEP-NCAR reanalyses in the high and midlatitudes of the Southern Hemisphere, 1958-2001. Journal of Climate, 17, 4603-4619.

Bromwich, D. H., R. L. Fogt, K. I. Hodges and J. E. Walsh, 2007: A tropospheric assessment of the ERA-40, NCEP, and JRA-25 global reanalyses in the polar regions. Journal of Geophysical Research, 112, D10111, doi:10.1029/2006JD007859.

Kalnay, E., 2003: Atmospheric Modeling, Data Assimilation and Predictability. Cambridge University Press, 341 pp.

Kanamitsu, M., W. Ebisuzaki, J. Woollen, S.-K. Yang, J. J. Hnilo, M. Fiorino and G. L. Potter, 2002: NCEP-DOE AMIP-II Reanalysis (R-2). Bulletin of the American Meteorological Society, 83, 1631-1643.

Keable, M., I. Simmonds and K. Keay, 2002: Distribution and temporal variability of 500 hPa cyclone characteristics in the Southern Hemisphere. International Journal of Climatology, 22, 131-150.

Leonard, S. R., J. Turner and A. Van Der Wal, 1999: An assessment of three automatic depression tracking schemes. Meteorological Applications, 6, 173-183.

Lim, E.-P., and I. Simmonds, 2007: Southern Hemisphere winter extratropical cyclone characteristics and vertical organization observed with the ERA-40 reanalysis data in 1979-2001. Journal of Climate, 20, 2675-2690.

Nielsen, J. W., and R. M. Dole, 1992: A survey of extratropical cyclone characteristics during GALE. Monthly Weather Review, 120, 1156-1167.

Onogi, K., J. Tslttsui, H. Koide, M. Sakamoto, S. Kobayashi, H. Hatsushika, T. Matsumoto, N. Yamazaki, H. Kaalhori, K. Takahashi, S. Kadokura, K. Wada, K. Kato, R. Oyama, T. Ose, N. Mannoji and R. Taira, 2007: The JRA-25 reanalysis. Journal of the Meteorological Society of Japan, 85, 369-432.

Pezza, A. B., I. Simmonds and J. A. Renwick, 2007: Southern Hemisphere cyclones and anticyclones: Recent trends and links with decadal variability in the Pacific Ocean. International Journal of Climatology, 27, 1403-1419.

Pinto, J. G., T. Spangehl, U. Ulbrich and P. Speth, 2005: Sensitivities of a cyclone detection and tracking algorithm: Individual tracks and climatology. Meteorologische Zeitschrift, 14, 823-838.

Raible, C. C., P. M. Della-Marta, C. Schwierz, H. Wernli and R. Blender, 2008: Northern Hemisphere extratropical cyclones: A comparison of detection and tracking methods and different reanalyses. Monthly Weather Review, 136, (in press).

Simmonds, I., 1978: The application of a multi-level spectral model to data assimilation. Journal of the Atmospheric Sciences, 35, 1321-1339.

Simmonds, I., 2000: Size changes over the life of sea level cyclones in the NCEP reanalysis. Monthly Weather Review, 128, 4118-4125.

Simmonds, I., 2003: Modes of atmospheric variability over the Southern Ocean. Journal of Geophysical Research, 108, 8078, doi: 10.1029/2000JC000542.

Simmonds, I., and K. Keay, 2000a: Mean Southern Hemisphere extratropical cyclone behavior in the 40-year NCEP-NCAR reanalysis. Journal of Climate, 13, 873-885.

Simmonds, I., and K. Keay, 2000b: Variability of Southern Hemisphere extratropical cyclone behavior 1958-97. Journal of Climate, 13, 550-561.

Simmonds, I., K. Keay and E.-P. Lim, 2003: Synoptic activity in the seas around Antarctica. Monthly Weather Review, 131, 272-288.

Simmonds, I., and J. C. King, 2004: Global and hemispheric climate variations affecting the Southern Ocean. Antarctic Science, 16, 401-413.

Simmonds, I., R. J. Murray and R. M. Leighton, 1999: A refinement of cyclone tracking methods with data from FROST. Australian Meteorological Magazine, Special Edition, 35-49.

Simmonds, I., A. Rafter, T. Cowan, A. B. Watkins and K. Keay, 2005: Large-scale vertical momentum, kinetic energy and moisture fluxes in the Antarctic sea-ice region. Boundary-Layer Meteorology, 117, 149-177.

Uppala, S. M., P. W. Kallberg, A. J. Simmons, U. Andrae, V. D. Bechtold, M. Fiorino, J. K. Gibson, J. Haseler, A. Hernandez, G. A. Kelly, X. Li, K. Onogi, S. Saarinen, N. Sokka, R. P. Allan, E. Andersson, K. Arpe, M. A. Balmaseda, A. C. M. Beljaars, L. Van De Berg, J. Bidlot, N. Bormann, S. Caires, F. Chevallier, A. Dethof, M. Dragosavac, M. Fisher, M. Fuentes, S. Hagemann, E. Holm, B. J. Hoskins, L. Isaksen, P. Janssen, R. Jenne, A. P. McNally, J. F. Mahfouf, J. J. Morcrette, N. A. Rayner, R. W. Saunders, P. Simon, A. Sterl, K. E. Trenberth, A. Untch, D. Vasiljevic, P. Viterbo and J. Woollen, 2005: The ERA-40 re-analysis. Quarterly Journal of the Royal Meteorological Society, 131, 2961-3012.

Wassermann, S., C. Schmitt, C. Kottmeier and I. Simmonds, 2006: Coincident vortices in Antarctic wind fields and sea ice motion. Geophysical Research Letters, 33, L15810, doi:10.1029/2006GL026005.

Watkins, A. B., and I. Simmonds, 1995: Sensitivity of numerical prognoses to Antarctic sea ice distribution. Journal of Geophysical Research, 100, 22,681-22,696.

Watkins, A. B., and I. Simmonds, 1998: Relationships between Antarctic sea-ice concentration, wind stress and temperature temporal variability, and their changes with distance from the coast. Annals of Glaciology, 27, 409-412.

White, W. B., and I. Simmonds, 2006: Sea surface temperature-induced cyclogenesis in the Antarctic circumpolar wave. Journal of Geophysical Research, 111, C08011, doi:10.1029/2004JC002395.

(a)

(b)

Figure 1 (Cyclones): July cyclone tracks in 2006 (blue) and 2007 (red) in the (a) NCEP2 and (b) JRA-25 re-analyses.
(a)

(b)

Figure 2 ((a)-(b)) (Cyclones): Mean (1980 -2008) summer cyclone system density (number of systems per 10³ (deg. lat.)² area) in the (a) NCEP2 and (b) JRA-25 re-analyses.
(c)

(d)

Figure 2 ((c)-(d) (Cyclones)): Mean (1979 -2007) winter cyclone system density (number of systems per 10³ (deg. lat.)² area) in the (a) NCEP2 and (b) JRA-25 re-analyses.
(e)

(f)

Figure 2 ((e)-(f) (Cyclones)): Difference (JRA-25 minus NCEP2 re-analyses) between the mean cyclone system density in (e) summer and (f) winter.
(a)

(b)

Figure 3 (Cyclones): Mean cyclogenesis density (number cyclogenesis events found per 10³ (deg. lat.)² area per day) in (a) summer and (b) winter in the NCEP2 re-analyses.
(a)

(b)

Figure 4 (Cyclones): Mean cyclone radius (in units of deg. lat.) in (a) summer and (b) winter in the NCEP2 re-analysis.
(a)

(b)

Figure 5 (Cyclones): Mean cyclone depth (in units of hPa) in (a) summer and (b) winter in the NCEP2 re-analysis.
(a)

(b)

Figure 6 (Cyclones): Trend in cyclone system density (in units of 0.1 systems per 10³ (deg. lat.)² area per decade) in (a) summer and (b) winter in the NCEP2 re-analysis. Stippling denotes areas over which the trends are significantly different from zero at the 95% confidence level.

Figure 7 (Cyclones): Time series of the mean number cyclones per analysis between 50 and 70^oS in the NCEP2 and JRA-25 re-analyses for (a, top) summer and (b, bottom) winter. ‘S’ and ‘W’ denote ‘strong’ and ‘weak’ systems (see text), while ‘SW’ denotes all systems together.

Figure 8 (Cyclones): Time series of the mean radius (in deg. lat.) of cyclones between 50 and 70^oS in the NCEP2 and JRA-25 re-analyses for (a, top) summer and (b, bottom) winter.

Figure 9 (Cyclones): Time series of the mean depth (in hPa) of cyclones between 50 and 70^oS in the NCEP2 and JRA-25 re-analyses for (a, top) summer and (b, bottom) winter.