Time scale, power spectra, and bimodality of zonal mean flow variability

This study presents results from probability density function (PDF) and power spectral analyses of the zonally averaged zonal wind from a series of idealized, multi-level, primitive equation model runs. The examination with the PDFs finds that bimodality occurs in the first principal component (PC1) time series if (1) its time scale is very long, (2) it is maintained by a positive zonal mean/eddy feedback, and (3) the background potential vorticity gradient is sharp and large. On the other hand, if all three criteria, which over most of the parameter space are either present or absent, are not satisfied, then bimodality is not found. These results link the occurrence of bimodality to both the properties of the background flow and the type of zonal mean flow variability. For single jet states, the presence of bimodality was found to depend on the height and the number of mountains, whereas for the double jet state no bimodality was observed.

The power spectral analyses of the PC1 time series found that when the zonal index (poleward zonal mean flow anomaly propagation) is dominant the corresponding power spectrum more closely resembles that for an AR(1) (AR(2)) process. These results suggest that the zonal index can be better described by a first order ordinary differential (ODE) and the poleward propagation by a second order ODE.

These model data are further analyzed to gain insight into the physical processes responsible for the characteristics of the power spectra of the observed Southern Hemisphere zonal mean zonal wind.