In this study, a definition of the COWL pattern is proposed using concepts derived from the theory of thermal equilibration of planetary waves. Alternation of planetary waves in which the tropospheric temperature pattern is either in-phase or out-of-phase with the surface-temperature eddy field is predicted by such a theory, providing a likely dynamical explanation of COWL internal variability.
Using a centennial AGCM experiment with perpetual-January surface temperature and radiative forcing, it is shown that regressions against measures of integrated surface energy fluxes produce a COWL-like pattern which agrees with predictions from thermal equilibration theories, and is close to the leading EOF of monthly-mean fields.
A second experiment, in which the zonal gradient of surface temperature is reduced to spring-like values while retaining January zonal means, is used to assess the impact of the reduction in land-sea thermal contrast on low-frequency variability. The hemispheric pattern for which the largest reduction in monthly-mean variance is found (with respect to the real-January experiment) is close to the flux-based definition of the COWL, confirming the dynamical relevance of thermal equilibration.
Finally, the possible role of ocean-atmosphere interactions and stratospheric changes in forcing and/or amplyfing interdecadal COWL variations is discussed.