Quantifying the Barotropic and Baroclinic Mechanisms in the Persistence of Annular Modes

The persistence of the annular modes, the leading mode of variability in the extratropics, is important for weather prediction on the subseasonal timescale and for the circulation change in a warming climate. The annular modes in global climate models (GCMs) are often overly persistent as compared with the reanalysis, particularly in the Southern Hemisphere summer season where greenhouse gas warming and Antarctic ozone depletion are both important for future climate change. The biases in the annular mode persistence have been related to the biases in the mean jet location or the planetary wave feedback to the zonal flow variability. Feedback mechanisms have been proposed to elucidate this persistence including a barotropic mechanism of upper-tropospheric wave propagation and wave breaking and a baroclinic mechanism of low-tropospheric baroclinic wave generation.

Here we introduce a method to quantify the strength of barotropic and baroclinic feedback mechanisms through the use of finite-amplitude wave activity formalism. Particularly, the barotropic mechanism is quantified by the variability of wave activity and the eddy mixing of potential vorticity in the upper troposphere and the baroclinic mechanism by the upward wave activity flux from the lower troposphere to the upper troposphere. It is shown that while the wave activity provides a negative feedback to the annular mode variability, eddy mixing and upward wave activity flux provide a positive feedback. These relationships are further investigated in an idealized dry atmospheric model in which varying temperature profiles are used to shift the climatological jet to various latitudes. As the latitude of the jet shifts poleward with tropical warming, there is a substantial reduction in the strength of the baroclinic feedback, while a modest decrease in the strength of the barotropic feedback occurs. This suggests that a combination of both mechanisms is responsible for the decrease in annular mode timescales with increasing latitude in this model, which may have implications for a similar relationship in the CMIP models.