Percentage frequency

Fig. 2.9. Mean distribution of Streten vortex types as a function of latitude, based on satellite photographs over 5 winters 1973-77. Frequencies in each 5-degree latitude band are presented as the percentage of the total for that type over all latitudes. Values in brackets give the total numbers of cyclones for the five years studied (from Carleton, 1979).

The overall picture then is one of extra-tropical cyclones originating in the middle latitude westerlies and moving gradually southward as they track around the hemisphere, eventually to decay in higher latitudes. Although the cloud climatologies of Carleton and others demonstrate the considerable longitudinal uniformity that we have come to expect in Southern Ocean features, some regional differences have been discovered. A plan view of the frequency of developing vortices in summer (December to March inclusive) is given in Fig. 2.10, taken from the climatology of Streten and Troup (1973). We see that there is a maximum frequency of developing or intensifying cyclones in a concentric band about the Pole near 50°S, but there is also clear evidence of enhanced lower latitude development in the South Pacific and South Atlantic. However, in the Atlantic, the cyclogenetic region is continuous over a broad latitude range, while in the Pacific cyclogenesis seems to occur only within two well-defined bands, with a wedge-shaped region in between that shows significantly weaker activity.

Fig. 2.10. Areal frequency of developing vortices (types W, A and B) in Southern Hemisphere summer, as observed by satellite over the period 1972-77. Full line is the position of the atmospheric polar front in summer (after Taljaard, 1968). Highest frequency regions are shaded (from Streten and Troup, 1973).

Fig. 2.10. Areal frequency of developing vortices (types W, A and B) in Southern Hemisphere summer, as observed by satellite over the period 1972-77. Full line is the position of the atmospheric polar front in summer (after Taljaard, 1968). Highest frequency regions are shaded (from Streten and Troup, 1973).

Carleton's (1981a) analysis for the winter season shows the same pattern so that the double-band structure is evidently a semi-permanent feature of the South Pacific circulation throughout the year.

The symmetric 50°S band of maximum cyclone frequency is closely related to, and occurs on the equatorward side of, the position of the oceanic Polar Front. This oceanic feature is a zone of steep latitudinal gradient in sea surface temperature, marking the axis of the Antarctic Circumpolar Current (see Patterson and Whitworth, this volume). The lower latitude extension of maximum cyclone frequency in the Pacific occurs in a region where there is a secondary maximum in the north-south atmospheric temperature gradient (Fig. 2.3), and is associated with a feature known as the South Pacific Convergence Zone (see below).

Seasonal variation in cyclone activity has been noted in a number of studies, and some attempts have been made to relate these variations to changes in north-south temperature gradient, location of the Antarctic Circumpolar Trough, and position of the sea-ice boundary. Carleton (1981a) showed in his study of the 1973-77 period that cyclonic vortices of all types were about two and a half times more frequent in the winter months than in summer. August was the month with the highest total frequency of cyclonic vortices, but a latitudinal shift in the occurrence frequencies over the winter months was also apparent. Dividing the hemisphere up into latitude zones (30-39, 40-49, 50-59, and 60-75°S), Carleton found that in July cyclonic activity was highest in the 40-49°S zone and decreased southwards whereas in September the maximum activity was observed in the southernmost zone of 60-75°S.

We have taken Carleton's (1981a) figures for vortex frequencies in various latitude bands, and calculated the latitude of peak frequency, which is plotted in Fig. 2.11. Also shown is the monthly location of the latitude of the Circumpolar Trough and the northern boundary of the Antarctic sea ice, as determined by Streten and Pike (1980) from data over the 5-year period 1972-77. Carleton claimed that the northward shift of maximum cyclonic activity in July and the southward return in September was evidence of the controlling influence of the Circumpolar Trough and, taken in isolation, this result certainly looks reasonable. Also plotted in the figure, however, are the latitudes of peak frequencies of cyclonic vortices calculated from the results of Schwerdtfeger and Kachelhoffer (1973) for the period 1967-70. These authors counted the total number of vortices over two September and three October months (constituting spring), and over three February and March months (autumn). Schwerdtfeger and Kachelhoffer concluded from their limited study that there was a highly significant relationship between the pack-ice border and the latitudinal band of maximum frequency of cyclone occurrence which, of course, is completely at variance with Carleton's (1981a) result. If nothing else, Fig. 2.11 illustrates the difficulty of drawing conclusions about feedbacks between the different aspects of Southern Ocean climate from a limited data base.

Fig. 2.11. Seasonal cycle of zonally-averaged latitude of Antarctic sea ice margin (solid line) and latitude of axis of the Circumpolar Trough (dotted line) (after Streten and Pike, 1980). Latitudes of maximum cyclonic activity at various times of the year are indicated by crosses (determined from Table 1 in Carleton, 1981a) and open circles (determined from table in Schwerdtfeger and Kachelhoffer, 1973).

Fig. 2.11. Seasonal cycle of zonally-averaged latitude of Antarctic sea ice margin (solid line) and latitude of axis of the Circumpolar Trough (dotted line) (after Streten and Pike, 1980). Latitudes of maximum cyclonic activity at various times of the year are indicated by crosses (determined from Table 1 in Carleton, 1981a) and open circles (determined from table in Schwerdtfeger and Kachelhoffer, 1973).

Influence of Long Waves on Patterns of Cyclogenesis

We saw in Fig. 2.10 that the South Pacific was unusual in that it had two distinct regions where cyclone activity was high: a zonal belt at approximately 50°S and a second band oriented northwest-southwest in the central South Pacific. This lower-middle latitude zone of frequent cyclogenesis and cloudiness is known as the South Pacific Convergence Zone (SPCZ), and is a semi-permanent feature of the region, although its exact position may vary from month to month. A satellite composite cloud picture over the Southern Hemisphere (Fig. 2.12) shows a well-developed cloud band to the north-east of New Zealand, highlighting the location of the SPCZ at that time. (The reader may also recognize some other cloud vortex patterns with reference to Fig. 2.8.)

A useful start to understanding variations in the behaviour of weather systems at different longitudes is gained by studying the pattern of very long wavelength stationary waves* in the atmosphere. Because of vertical motions associated with these "long waves", weather systems within and eastwards of a long-wave trough

Fig. 2.12. Hemispheric cloud mosaic (visible) for 8 January 1972, photographed by ESSA-9 satellite (from NOAA catalogue, 1974).

* These waves are identified by a wave number (1, 2, 3, etc.) that describes how many complete wavelengths fit around the hemisphere. For example, wave 1 has one trough and one ridge around a latitude circle; wave 3 has three troughs and ridges so that consecutive troughs are separated by 120° of longitude; etc.

are in a more favourable environment for development of cloud systems and cyclonic vortices. Conversely, a cyclone moving into a long-wave ridge position is likely to weaken. The mean locations of the trough and ridge axes of the longest waves at 500 hPa are shown schematically in Fig. 2.13. In terms of amplitude, wave 1 is the dominant wave north of about 40°S, wave 3 is strongest between 40° and 50°S, and waves 1 and 2 are of approximately equal strength in higher latitudes. Trenberth (1980) discussed these details at length, but our interest here is mainly in the overall pattern produced. We see from Fig. 2.13 that a clustering of trough axes occurs in the eastern Indian Ocean, whereas the Tasman Sea and region southeast of New Zealand shows persistent ridging. The northwest-southeast band of trough activity in the central Pacific also extends into lower latitudes, and is associated with the aforementioned South Pacific Convergence Zone.

The long-wave trough-ridge pattern between the Indian and Pacific Oceans influences the movement of cyclones travelling through this region of the hemisphere. Depressions coming out of the Indian Ocean frequently move quickly

Fig. 2.13. Positions of mean long-wave trough and ridge axes at 35°, 45° and 55°S, at the 500 hPa level. Ridges are shaded and troughs hatched, with central value indicating the wave number. Based on 7 years of daily 500 hPa height data 1976-83. Over this period, wave 2 at 35°S showed no preferred geographic location (adapted from Mullan, 1985).

Fig. 2.13. Positions of mean long-wave trough and ridge axes at 35°, 45° and 55°S, at the 500 hPa level. Ridges are shaded and troughs hatched, with central value indicating the wave number. Based on 7 years of daily 500 hPa height data 1976-83. Over this period, wave 2 at 35°S showed no preferred geographic location (adapted from Mullan, 1985).

into higher latitudes as they pass south of Australia and New Zealand (this may be a partial explanation for the persistent pressure minimum in the Ross Sea Sector of the Circumpolar Trough). To the east of New Zealand, stationary longwave troughs at 35°S coincide with ridges further south at the same longitude. This trough-ridge juxtaposition weakens the north-south pressure gradient and hence the westerlies in the southwestern Pacific. The weaker westerlies mean that the cyclones travel eastward more slowly in this part of the hemisphere, as is shown in Fig. 2.14 from Mullan (1985). In this study, daily 500 hPa height data from 1976-83 were subjected to Fourier analysis and separated into wave components, and the contributions from wave numbers 4 to 9 then recombined to produce what was termed the "medium-wave" field (the medium-wave troughs and ridges can be identified with the travelling cyclones). By tracking the movement of the medium-wave troughs from day to day, the mean speed as a function of longitude at 45°S was calculated. Fig. 2.14 also shows the trough amplitude, and it is apparent that cyclones not only tend to travel faster in the Indian Ocean than in the South Pacific but also are more intense in the Indian Ocean.

The existence of stationary long-wave ridges south of New Zealand, and the reduced speed of weather systems through this region, are key factors in the high incidence of "blocking" in these longitudes. Blocking is associated with a slow-moving long-lived surface anticyclone in high latitudes, commonly with a cut-off low on its equatorward side, that together "block" the progress of the westerlies.

Fig. 2.14. Amplitude (in geopotential metres) and daily movement (in m.s-1) of 500 hPa medium wave trough axes at 45°S. Medium waves defined as the composite over wave numbers 4 to 9 (adapted from Mullan, 1985).

Persistent anomalies created by blocking patterns are of major importance to extended-range forecasting. Although occasional blocks occur southeast of South America and South Africa, the Australia-New Zealand region shows the primary maximum, with more frequent and longer-lasting eposides (Lejenas, 1984; Tren-berth and Mo, 1985).

0 0

Post a comment