Y.-J. Kim, J. D. Farrara and C. R. Mechoso, 1998
(JMSJ, 76, 695-709)
(text.ps& figs.ps.tar.Z)

ABSTRACT

    The sensitivity of the stratospheric circulation simulated by an atmospheric general circulation model (UCLA AGCM) to modifications in the parameterization of several key physical processes is investigated. The focus is on the northern hemisphere winter circulation simulated by the troposphere-stratosphere version of the model.
    The parameterization of ozone mixing ratio is first investigated by comparing multi-year simulations with distributions of ozone mixing ratio that are either prescribed according to an observed climatology or predicted using a photochemistry formulation. The prescribed ozone mixing ratios produce a considerably more realistic circulation although the results are degraded in some regions of the stratosphere. This lead to closer look into the radiation parameterization. Several simulations for the northern winter season are performed with adjustments leading to obtain more realistic radiative cooling rates mainly in the upper stratosphere. It is found that the simulation of the stratospheric polar night jet is very sensitive to these adjustments. Additional experiments performed by adding Rayleigh friction and Newtonian cooling reveal a similar sensitivity. The results of these experiments emphasize a need for caution when ad-hoc adjustments are applied in the model. They also confirm that the performance of an AGCM in the stratosphere can depend on several different model aspects, and that this dependence is not always straightforward.
    The results of this study also raise questions in contemporary parameterizations of orographic gravity-wave drag (GWD), a phenomenon that plays an important role in the stratospheric circulation. It is argued that the results of this study are consistent with the notion that GWD parameterizations that neglect or underestimate the drag at low levels may also overestimate the drag at upper levels.