Since the SAR there has been more emphasis on analysis of decadal variability over the Southern Hemisphere. The Southern Hemisphere gridded SLP data for the period 1950 to 1994 show two dominant modes in annual average values, similar to those identified by Karoly et al. (1996) using station data. The first mode unambiguously represents the Southern Oscillation and reflects the tendency towards more frequent and intense negative phases over the past several decades. The second mode represents anomalies throughout the mid-latitude regions across the Indian Ocean and western Pacific, which contrast with anomalies elsewhere.

Figure 2.31: The High Latitude Mode (Kidson, 1988) or Antarctic Oscillation (AAO), defined as the first orthogonal pattern (covariance eigenvector of the Southern Hemisphere monthly surface pressure, January 1958 to December 1997) (Gong and Wang, 1999c; Kiladis and Mo, 1999). Data from NCAR/NCEP Reanalysis (Kalnay et al., 1996). Note that Thompson and Wallace (2000) use 850 hPa height to define their AAO.

The Trans-Polar Index (TPI) is the only large-scale station pressure-based extra-tropical Southern Hemisphere circulation index in regular use. It is based on the normalised pressure difference between New Zealand and South America and has been recalculated and extended by Jones et al. (1999b). On decadal and longer time-scales the TPI reflects movement in the phase of wave number one around the Southern Hemisphere. Troughing (low pressure) was more frequent in the New Zealand region in the 1920s, and at a maximum in the 1940s. Anticyclonicity was favoured from the late 1950s to 1976, with troughing in the South American sector. Troughing was again apparent in the New Zealand sector in the 1990s (Salinger et al., 1996).

A leading mode of variability in the extra-tropical Southern Hemisphere circulation on interannual to multi-decadal time-scales is a zonally elongated north-south dipole structure over the Pacific, stretching from the sub-tropics to the Antarctic coast (Mo and Higgins, 1998; Kidson, 1999; Kiladis and Mo, 1999). It is strongly related to ENSO variability. The lower-frequency dipole structure contributes to variability in blocking frequency across the far south Pacific (Renwick, 1998; Renwick and Revell, 1999).

ENSO variability is also implicated in the modulation of a �High Latitude Mode� (HLM) (Kidson, 1988; Karoly, 1990), especially over the austral summer. The HLM is now also called the �Antarctic Oscillation� (AAO); they appear to be the same phenomenon with the same structure (Thompson and Wallace, 2000). The AAO is a zonal pressure fluctuation between mid- and high latitudes of the Southern Hemisphere, centred on 55 to 60°S. It has recently been further studied (Gong and Wang, 1999c; Kidson, 1999; Thompson and Wallace, 2001; Figure 2.31) and shown to extend into the lower stratosphere between the Antarctic and the sub-tropical latitudes of the Southern Hemisphere. The AAO appears to persist all year but may be most active from mid-October to mid-December when it extends into the stratosphere (Thompson and Wallace, 2001). In its high index phase, it consists of low pressure or heights above the Antarctic and the near Southern Ocean with high heights north of about 50oS. Although the data are sparse, there is evidence that, like the NAO, the AAO has tended to move more towards a positive index phase, despite lower pressures being observed over the New Zealand region during the 1990s. This change is also associated with with increasing westerly winds in mid-latitudes. Thompson and Wallace (2001) show that most of Antarctica is rather cold in this phase, except for the Antarctic Peninsula which is warm due to additional advection of relatively warm air from seas to the west. This may explain some of the behaviour of Antarctic temperatures in the last two decades (Figure 2.10; Comiso, 2000).

Other work has identified the likely existence of an Antarctic Circumpolar Wave (ACW) (Jacobs and Mitchell, 1996; White and Peterson, 1996), a multi-annual climate signal in the Southern Ocean, with co-varying and perhaps coupled SST and SLP anomalies that move around the Southern Ocean. Its long-term variability is not yet known.

The interannual variability of ENSO has fluctuated substantially over the last century, with notably reduced variability during the period 1920 to 1960, compared with adjacent periods. It remains unclear whether global warming has influenced the shift towards less frequent La Niña episodes from 1976 to 1998, including the abnormally protracted ENSO 1990 to 1995 event and the exceptionally strong 1982/83 and 1997/98 events. Analysis of SST patterns indicates that a global warming pattern may have increased the background temperature in the region most affected by ENSO, but there is some ambiguity in the details of this pattern.

Since the SAR, ‘ENSO-like� features operating on decadal to multi-decadal time-scales have been identified, such as the PDO and IPO. They appear to be part of a continuous spectrum of ENSO variability that has subtly changing SST patterns as time-scales increase and which may have distinctive effects on regional climate around the Pacific basin. For the period since 1900, El Niño (La Niña) events are more prevalent during positive (negative) phases of the IPO.

In the Northern Hemisphere, pronounced changes in winter atmospheric and oceanic circulations over the North Pacific in the 1970s (the North Pacific Oscillation) have been paralleled by wintertime circulation changes over the North Atlantic, recorded by the NAO. In the North Pacific, spatially coherent changes have occurred in surface temperature across the North Pacific and western North America, while the enhanced westerly phase of the NAO has caused considerable winter half-year temperature and precipitation changes over a vast area of extra-tropical Eurasia. In the Southern Hemisphere, a feature quite like the NAO, the HLM or the AAO, also appears to have moved into an enhanced westerly phase in middle latitudes.

IPCC Homepage