Quasi-periodic poleward propagation of zonal mean anomalies

It has been shown in observations that at times there are tropospheric zonal mean flow anomalies that undergo quasi-periodic poleward propagation. The parameter study of Son and Lee (2005) shows that this poleward propagation is the most dominant form of zonal mean flow variability in model climates where the potential vorticity (PV) gradient associated with the subtropical jet is weak. One of these model runs is examined in detail to investigate the physical mechanism behind the poleward propagation.

It is found that the initiation of the poleward propagation is marked by the formation of negative zonal wind anomalies in the tropics. These negative anomalies arise from meridional overturning/breaking of waves that originate in midlatitudes. This wave overturning/breaking homogenizes potential vorticity (PV) within the region of the negative zonal wind anomalies. Subsequent equatorward radiation of midlatitude waves is halted, resulting in wave breaking at the poleward end of the homogenized PV region. This in turn generates new negative zonal wind anomalies, which enables a continuation of the poleward propagation. The shielding of the homogenized PV region from equatorward wave propagation allows the model's radiative relaxation to re-establish undisturbed westerlies on its equatorward side. This forms a band of positive zonal wind anomalies in that region. As radiative relaxation causes this positive band to propagate poleward, subsequent equatorward wave propagation into the tropics resumes. This process initiates the next cycle of poleward propagation.

The above process indicates that the poleward propagation is caused by an orchestrated combination of linear Rossby wave propagation, nonlinear wave breaking, and radiative relaxation. The importance of the meridional wave propagation and breaking is consistent with the fact that the poleward propagation takes place only in the parameter space where the PV gradient is of moderate strength.