3.5.5 The Stratosphere

The dynamically stable stratospheric circulation is dominated in mid-latitudes by westerlies in the winter hemisphere and easterlies in the summer hemisphere, and the associated meridional overturning ‘Brewer-Dobson’ circulation. In the tropics, zonal winds reverse direction approximately every two years, in the downward-propagating QBO (Andrews et al., 1987). Ozone is formed predominantly in the tropics and transported to higher latitudes by the Brewer-Dobson circulation. Climatological stratospheric zonal-mean zonal winds (i.e., the westerly wind averaged over latitude circles) from different data sets show overall good agreement in the extratropics, whereas relatively large differences occur in the tropics (Randel et al., 2004b).

The breaking of vertically propagating waves, originating from the troposphere, decelerates the stratospheric westerlies (see Box 3.3). This sometimes triggers ‘sudden warmings’ when the westerly polar vortex breaks down with an accompanying warming of the polar stratosphere, which can quickly reverse the latitudinal temperature gradient (Kodera et al., 2000). While no major warming occurred in the NH in nine consecutive winters during 1990 to 1998, seven major warmings occurred during 1999 to 2004 (Manney et al., 2005). As noted by Naujokat et al. (2002), many of the recent stratospheric warmings after 2000 have been atypically early and the cold vortex recovered in March. In September 2002, a major warming was observed for the first time in the SH (e.g., Krüger et al., 2005; Simmons et al., 2005). This major warming followed a relatively weak winter polar vortex (Newman and Nash, 2005).

The analysis of past stratospheric changes relies on a combination of radiosonde information (available since the 1950s), satellite information (available from the 1970s) and global reanalyses. During the mid-1990s, the NH exhibited a number of years when the Arctic winter vortex was colder, stronger (Kodera and Koide, 1997; Pawson and Naujokat, 1999) and more persistent (Waugh et al., 1999; Zhou et al., 2000). Some analyses show a downward trend in the NH wave forcing in the period 1979 to 2000, particularly in January and February (Newman and Nash, 2000; Randel et al., 2002). Trend calculations are, however, very sensitive to the month and period of calculation, so the detection of long-term change from a relatively short stratospheric data series is still problematic (Labitzke and Kunze, 2005).

In the SH, using radiosonde data, Thompson and Solomon (2002) reported a significant decrease of the lower-stratospheric geopotential height averaged over the SH polar cap in October to March and May between 1969 and 1998. The ERA-40 and NRA stratospheric height reanalyses indicate a trend towards a strengthening antarctic vortex since 1980 during summer (DJF; Renwick, 2004; Section 3.5.2), largely related to ozone depletion (Ramaswamy et al., 2001; Gillett and Thompson, 2003). The ozone hole has led to a cooling of the stratospheric polar vortex in late spring (October–November; Randel and Wu, 1999), and to a two- to three-week delay in vortex breakdown (Waugh et al., 1999).