Temporal growth and vertical propagation of perturbations in the winter atmosphere

We present a general-circulation model study of the temporal growth and vertical propagation of perturbations following transient, vertically confined forcings. Five ensembles of perturbation experiments are performed, each with eight different initial conditions and with the perturbations enforced at ten different vertical levels from the lower troposphere to the upper stratosphere. The motivation for the study is the recent recognition of downward propagation of anomalies from the stratosphere to the troposphere and its implications both for medium-range forecasts and for a possible physical mechanism for stratospheric impacts on weather and climate. The perturbations grow in an initial period after which they saturate to a level comparable to the natural variability. In the initial period the growth of the perturbations is quadratic in time. This power-law growth suggests a predictability time which converges towards a finite value in the limit of small perturbations. In the troposphere the predictability time, defined as when perturbations reach a size of 50% of the natural variability, is 20-30 days. In the stratosphere the response is delayed and a predictability time of 30-40 days is observed. In the initial period of growth, the vertical spread of perturbations seems decoupled from the background vacillations in the sense that no vertical propagation of perturbations with the time-scale of the vacillations is observed. Forcings confined to the stratosphere result instantaneously in perturbations of the troposphere and vice versa. Thus, stratospheric vacillations seem to have no active role in the propagation of perturbations from the stratosphere to the troposphere: stratospheric vacillations are not a vehicle for transient perturbations in the stratosphere to be transported to the troposphere. The disconnected nature of the stratospheric vacillations and the growth of perturbations is further substantiated in experiments with a time-independent troposphere. In these experiments the periodic stratospheric vacillations are shown not to support the growth of perturbations.