The Brewer-Dobson circulation is often quantified by the integrated transport measure age of air (AoA). AoA is affected by all transport processes, including transport along the residual mean mass circulation and two-way mixing. In global models, two-way mixing results from stirring by resolved winds and from unresolved parametrized and/or numerical diffusion.

We present a method to disentangle the effects of residual circulation transport, resolved two-way mixing and unresolved diffusion on AoA. The method is applied to global model data, and it is shown that the large spread in the simulation of AoA between models can in large parts be attributed to differences in the relative strength of mixing (i.e. the ratio of mixing mass flux to net mass flux between tropics and extratropics, which we define as the “mixing efficiency”).

Furthermore, we explore whether (resolved) two-way mixing and the residual circulation are coupled. This would be expected in a steady-state climate as the residual meridional velocity is then proportional to PV eddy mixing (in the absence of unresolved wave driving). A constant mixing efficiency is found for different climate states in equilibrium simulations with the ECHAM model. We use an idealized version of the ECHAM5/MESSy model system to further explore the relationship between residual circulation transport and mixing under different scenarios (e.g. different climate states, with or without additional parametrized wave forcing, steady versus transient forcing).

A better understanding of the processes that control AoA will help to reconcile current discrepancies between simulated and observed AoA and its long-term changes.