## Abstract

Recent advances in our understanding of the dynamics of the stratospheric circulation have led to the concepts of “downward control” and the “extratropical pump.” However, under the assumptions on which these concepts are based, midlatitude wave driving cannot explain the fact that mean stratospheric upwelling is located in the Tropics. Nevertheless, using a nonlinear two-dimensional model it is shown here that a steady and (in the lower stratosphere) linear circulation with a qualitatively reasonable upwelling can be produced, provided the wave drag extends to within about 20° of the equator. In a linear analysis of the problem, it is shown that the effects of weak model viscosity (some 50 times weaker than thermal relaxation) are crucial in permitting flow across angular momentum contours within a tropical boundary layer whose width is of order *L*_{R}*P*^{1/4}, where *L*_{R} is the equatorial Rossby radius and *P* a Prandtl number (the ratio of radiative to viscous relaxation times). Provided the wave drag extends into this boundary layer, upwelling is distributed across the Tropics. These considerations put limits on the generality of the concepts of the extratropical pump and downward control and, inter alia, open the possibility that diabatic heating alone can drive a meridional circulation within the Tropics. On the basis of simple representations of wave drag and diabatic heating in a nonlinear, zonally symmetric model, it is found that, although driving by wave drag is the dominant mechanism, stratospheric (and perhaps tropospheric) heating may make a significant contribution to the net upwelling and may help explain its structure. Just what, in reality, might play a role analogous to that of viscosity in the model is an open question.

*Corresponding author address:* Dr. Alan Plumb, Dept. of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, Room 54-1726, Cambridge, MA 02139.

Email: rap@rossby.mit.edu