The three-dimensional structure of thermal and eddy driven eddy-mean flow interactions

Previous work (Li and Wettstein, 2012) has demonstrated that indices of two fundamental processes in the atmospheric general circulation (iT: thermal driving in the tropics and iE: momentum flux convergence in the midlatitudes) are associated both with simple reorganizations of storm track / jet covariability and with familiar leading patterns of climate variability in the Northern Hemisphere. North Atlantic zonal wind variability is mainly associated with eddy momentum flux convergence (NAO-like variability), whereas zonal wind variability in the Pacific is associated with both driving processes, providing evidence the Pacific jet is both thermally-driven (PNA-like variability) and eddy-driven (WP-like variability). The present study expands on that previous work by illustrating these driving processes are associated with coherent three-dimensional zonal wind variability in both hemispheres, particularly in the vertical. Results using both reanalysis and control model simulations are presented and related. Zonal wind variability is analyzed on pressure surfaces chosen to emphasize one or the other of the thermal or eddy-driven processes. For example, the leading Northern Hemisphere pattern of extratropical zonal wind variability in the lower troposphere (850 hPa) is an NAO-like pattern of zonal wind variability that resides almost exclusively in the (eddy-driven) North Atlantic. Conversely, the leading Southern Hemisphere pattern of extratropical zonal wind variability near the tropopause (200 hPa) is a Pacific South America-like pattern of zonal wind variability that resides almost exclusively in the (thermally-driven) South Pacific. An eddy-driven Southern Annular Mode-like pattern of zonal wind variability concentrated in the South Indian Ocean is the second leading pattern of 200 hPa zonal wind variability in the Southern Hemisphere. Choosing different variables (e.g., geopotential height) and pressure levels for analysis will emphasize or convolve relatively clear patterns of variability associated with iT and iE. The three-dimensional (and especially vertical) structure of zonal wind covariance associated with the leading patterns of climate variability corroborate these results. Taken together, the results in Li and Wettstein (2012) and those in the present study suggest that the leading patterns of climate variability are predominantly tropospheric and sectoral in both the Northern and Southern Hemispheres. Furthermore, the relative position and intensity of the subtropical jet with respect to the eddy-driven jet seems to determine the leading pattern of variability, in agreement with many previous idealized modeling studies.