Waves generated in the troposphere propagate into the stratosphere and are absorbed, so that stratospheric changes alter where and how they are absorbed, and effects can extend downward into the troposphere through a mechanism called "downward control" (Haynes et al., 1991). The downward propagation of zonal-mean anomalies provides a purely dynamical stratosphere-troposphere link, which may account for the well-documented troposphere-stratosphere anomaly correlations seen in observations (Baldwin et al., 1994; Perlwitz and Graf, 1995). The North Atlantic Oscillation (Section 7.6.4) thus could be coupled with the strength of the wintertime Arctic vortex (Thompson and Wallace, 1998).
The dominant wave-induced forcing in the stratosphere is believed to come from tropospherically generated planetary scale Rossby waves in wintertime. These waves are explicitly resolved in models and data. Thus the meridional mass circulation, although two-celled, is predominantly directed towards the winter pole (e.g., Eluszkiewicz et al., 1996), and leads to a significant warming and weakening of the polar night vortex relative to its radiatively determined state (Andrews et al., 1987). Variability and changes in planetary wave forcing thus lead directly to variability and changes in wintertime polar temperatures, which modulate chemical ozone loss (WMO, 1999).
The principal uncertainties in wave-induced forcing come from gravity waves, which are undetected in analyses, but whose role is inferred from systematic errors in climate models. The most notable such error is the tendency of all atmospheric GCMs to suffer from excessively cold polar temperatures in the winter stratosphere, together with an excessively strong polar night jet, especially in the Southern Hemisphere (Boville, 1995). Enhanced Rayleigh friction improves the results (Manzini and Bengtsson, 1996; Butchart and Austin, 1998), but its physical basis is unclear (Shepherd et al., 1996). The principal forcing of gravity waves arises from unresolved sub-grid scale processes, such as convection, and more physically based gravity wave parametrizations are being developed.
A dominant factor determining the interannual variability of the stratosphere is the quasi-biennial oscillation (QBO). It is driven by wave drag (momentum transport), but it remains unclear exactly which waves are involved (Dunkerton, 1997). Most current atmospheric GCMs do not simulate the QBO and are therefore incomplete in terms of observed phenomena. It appears that QBO-type oscillations are found in models with higher vertical resolution (better than 1 km, Takahashi, 1996; Horinouchi and Yoden, 1998). It is still not clear what aspects of vertical resolution, energy dissipation, and wave spectrum are necessary to generate the QBO in climate models in a self-consistent way.
The meridional mass circulation, known as the Brewer-Dobson circulation, transports chemical species poleward in the stratosphere (Andrews et al., 1987). Air entering the stratosphere in the tropics returns to the troposphere in the extratropics with a time-scale of about five years (Rosenlof, 1995). Current models indicate shorter time-scales (Waugh et al., 1997), but the reasons are not currently well understood. The variation of the height of the troposphere with latitude is also important for meridional transport and troposphere-stratosphere exchange. This is because mid-latitude cross-tropopause mixing is preferentially along isentropic surfaces which are in the troposphere in the tropics but in the stratosphere at higher latitudes. Transport processes in the lowermost stratosphere are important factors affecting tropospheric climate (Pan et al., 1997).
The mean climate and variability of the stratosphere are not well simulated in current models. Because there is increasing evidence of effects of the stratosphere on the troposphere, this increases uncertainty in model results for tropospheric climate change. A key concern is how well mixing on small scales is done in the lowermost stratosphere. While overdue attention is being given to the stratosphere as more resolution in the vertical is added to models, further increases in resolution are desirable.
Other reports in this collection