atmospheric-modeling studies by James and James (1989) and Robinson (1992) have shown that it is also the dominant mode of low-frequency variations in extended simulations with simplified general-circulation models with zonally uniform boundary conditions. Hence, this zonally symmetric high-latitude mode appears to be a ubiquitous mode of low-frequency variation of the SH, associated with interaction between the mean flow and the transient eddies in the SH storm track.

Little study has been made of the zonal wave-train modes, but it is likely that the quasi-stationary interannual wave-number-three modes are associated with the three SH land masses and continent-ocean thermal contrasts. In addition, the zonal wave-number-three Rossby wave is almost stationary in the mean SH zonal wind in mid-latitudes.


Since upper-air observations are available for the SH for about three decades only, it is not possible to use such data to describe the interdecadal variations of the SH circulation. The only analyses of longer-term variations of the SH troposphere using upper-air station data have concentrated on temperature trends (Angell, 1986; Karoly, 1987). These have shown a small warming trend in the SH lower troposphere over the last three decades.

Some evidence of longer-term changes in the SH circulation is found when the mean sea-level pressure analyses prepared by the South African Weather Bureau for 19511958 (Taljaard et al., 1961) are compared with analyses from the Australian Bureau of Meteorology from 1972. Although some differences are due to different analysis methods, some pressure changes appear to be real (Mo and van Loon, 1984; J91).

The equivalent-barotropic vertical structure of low-frequency variations in the SH (Szeredi and Karoly, 1987a) means that longer records of surface observations may be used to describe interdecadal variations. Surface-pressure observations are available for some SH stations for periods extending back to the last century, and sea surface pressure observations over limited SH ship tracks are available for a similar period. It should be noted that these records are often incomplete and are representative of only limited SH regions.

J91 has used station data to reconstruct gridded SH mean sea level pressure (MSLP) fields for most of this century. The quality of this MSLP data set has been assessed by J91 and Barnett and Jones (1992). It is considered to be of reasonable quality over the region 15°S to 60°S from 1951 on, but the reconstructions back to 1911 are useful only near the continents and over the southwestern Pacific Ocean. Jones has used these data to assess changes since 1951 in the strength of the three subtropical anticyclones and the southern high-latitude westerlies; time series of indices of the strength of the South Pacific anticyclone and the high-latitude westerlies from J91 are shown in Figure 4. They


Seasonal and annual time series of two indices of the SH circulation: upper panel, South Pacific anticyclone high (average pressure in the region 25°S to 35°S, 80°W to 110°W); lower panel, strength of the high-latitude westerlies (zonal-mean pressure difference 40°S to 60°S). The filtered curve is from a 13-term Gaussian filter designed to suppress variations on time scales less than 10 years. (From Jones, 1991; reprinted with permission of the Royal Meteorological Society.)

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement