Large scale constraints on stratospheric wave forcing: Implications for Downward Control

Recent studies have shown strong interactions between resolved planetary-scale Rossby wave driving and parameterized small-scale gravity wave drag in atmospheric models. In particular, perturbations in parameterized wave drag are often compensated by an equal and opposite change in the resolved wave driving. Here, we aim to better understand the mechanism behind these interactions, using a conceptual model of the ``stratospheric surf zone'' (McIntyre and Palmer, 1983). The surf zone is a region in which large-amplitude Rossby waves that have propagated upwards from the troposphere break and consequentially mix the potential vorticity (PV). Note that as oppose to Rossby waves which propagate along PV gradients, gravity waves propagate along density gradients associated with stratification, and so can segregate the PV when they break. Suppose, conceptually, that the surf zone is fixed in size, with no PV gradient within, and strong barrier to mixing on either side, as in the PV staircase model. In this limit, the total wave forcing of the stratosphere is essentially fixed, and any perturbations in the gravity wave drag will be compensated. Relaxing this constraint on the surf zone size allows waves to widen/deepen the surf zone, thus increasing the total wave forcing. This framework suggest that Rossby waves will compensate for any tendency for the gravity waves to alter the PV within the surf zone, and that a net response of the meridional circulation to gravity wave forcing only occurs when it tends to extend (or contract) the surf zone. In this case, the Rossby wave forcing may actually amplify the perturbation. This conceptual model is explored and validated in a series of integrations with an idealized GCM.

The subgrid-scale orographic source for the columnwised gravity wave parameterizations in atmospheric models is largely in the mid-latitudes, and so this may explain why the compensation is such a broad phenomenon in model simulations. This is not the case for nonorographic gravity waves, which in contrast have a very strong effect on the total wave driving.