Andrew Gregory, John Thuburn and Alan O'Neill
Department of Meteorology, University of Reading.
A new stratosphere-mesosphere model has been designed with many dynamical and numerical advantages, notably the use of:
Potential vorticity (PV) as a conserved prognostic variable.
Isentropic vertical coordinates.
A hexagonal-icosahedral horizontal grid.
It has often been assumed that stratospheric sudden warmings are initiated by the development of wave amplitudes in the upper troposphere. The model has been used to test the following hypothesis:
Warmings can occur without any tropospheric transience as a result of the stratosphere's non-linear response to steady tropospheric forcing.
Many model integrations have been performed forcing the model with the annual solar cycle and constant amplitudes of zonal wavenumbers 1 and 2 at its artificial boundary near the tropopause. These long waves have the greatest potential to influence the stratosphere.
Despite the constant forcing, for large enough amplitudes sudden warmings occur (fig. 2). With 200m forcing there is a spring warming as is observed in the southern hemisphere. With 300m there are winter warmings as often occurs in the northern hemisphere. Vacillations with periods of 1-2 weeks are present in both integrations.
The vacillations are associated with the generation of successive travelling anticyclones in the upper stratosphere (fig. 3). These help maintain a quasi-stationary anticyclone by the polewards advection of low PV. The stationary anticyclone acts to reduce the strength of the polar cyclone during the warming period. The behaviour is very similar to that observed during real stratospheric warmings.
Again sudden warmings occur for stronger forcing (fig. 4). For 150m there is a final warming, while for 300m there are two weak winter warmings. By 350m the winter warmings are very strong.
The PV maps of the wave 2 experiments show a different evolution to the wave 1 results. There are no successive travelling anticyclones in the wave 2 case. Two anticyclones form and split the polar cyclone into two as they merge over the pole (fig. 5). Warmings of this type are also observed in the northern hemisphere.
The hypothesis has been proved correct - sudden warmings can occur without tropospheric transience provided wave amplitudes are large enough. The synoptic events are similar to real warmings.
At lower amplitude forcing (e.g. 200m) there are spring warmings similar to the observed behaviour in the southern hemisphere, while at larger amplitudes (e.g. 300-350m) there are strong winter warmings as observed in some northern hemisphere winters.
The warmings forced by wave 1 and wave 2 are different in character. The wave 1 forced warmings are associated with the generation of a series of travelling anticyclones which continually strengthen a quasi-stationary anticyclone. Only two anticyclones form in the warmings forced by wave 2.