J. D. Farrara, Y. -J. Kim and C. R. Mechoso, 1997
(Preprints, 7th Conference on Climate Variations, 2-7 February, 1997, Long Beach, CA, Amer. Meteor. Soc., pp 21-26)

ABSTRACT

        This paper discusses aspects of an atmospheric general circulation model (AGCM) whose improvement can have a significant beneficial impact in the simulation of the climatology and interannual variability in the stratosphere. The work is based on the tropospheric-stratospheric version of the UCLA AGCM (top at 1 mb) with both low (5 deg lon. x 4 deg lat.) and high (2.5 deg lon. x 2 deg lat.) horizontal resolution. The model is the most recent version we use in climate studies.
        We start by contrasting decade-long simulations obtained using distributions of ozone mixing ratio that are either prescribed (according to an observed climatology) or predicted (according to a highly simplified photochemistry). Our results show that errors in predicted ozone in the upper troposphere/lower stratosphere can result in large cold biases at these levels and unrealistically strong stratospheric jets. With ozone predicted the simulated zonal mean circulation is much improved.
        Next, we focus on updates in the parameterization of shortwave radiative heating. Specifically, we updated the calculation of the absorption cross-sections for water vapor and ozone. Again, the results of several simulations for the Northern Hemisphere winter show significant improvements in the middle and upper stratosphere.
        Finally, we consider processes that do not directly affect the stratosphere. One of these processes concerns the incorporation of vertical mixing of momentum by dry convective adjustment processes. The other refers to the retention or elimination of small scales in the depth of the planetary boundary layer predicted by the AGCM. The results of several simulations for the northern winter show that these model changes have a relatively minor impact on the troposphere, but a substantial effect on the stratosphere.
        Our results emphasize that the performance of an AGCM in the stratosphere can depend on several model aspects, and that this dependence can be straightforward in some cases but subtle in some other cases. We will describe our new results and the effect that an improved simulation of the seasonal cycle in the stratosphere can result in improvement of the interannual variability in that region.