Effects of gravity wave drag induced by cumulus convection on the atmospheric general circulation

A parameterization scheme of gravity wave drag induced by cumulus convection (GWDC) is implemented in the Yonsei University atmospheric general circulation model (GCM) and the effects of GWDC on the zonal-mean flow and planetary waves are investigated through perpetual July simulations. The GWDC parameterization scheme used in this study includes a momentum gain in the cloud region to conserve the momentum. The gravity wave stress at the cloud top is concentrated in the intertropical convergence zone (ITCZ) with its maximum value of 0.14 N m^-2 near the tropopause due to deep cumulus clouds. The wave breaking occurs mainly in the upper troposphere and lower stratosphere. The maximum westerly acceleration in the ITCZ is 0.6 m s^-1 day^-1, which is close to that observed. It is surprising to observe that the zonal wind difference between the simulations with and without the GWDC parameterization is largest in the Southern Hemisphere (SH) midlatitude stratosphere, where a westerly jet exists, rather than in the major drag forcing region and that there is an associated warming in the SH polar stratosphere. The excessive westerly jet in the SH that appears in the simulation without the GWDC parameterization is alleviated significantly (7 m s^-1) by its inclusion. This result implies that the nonlinear process through planetary waves rather than by direct drag forcing might play an important role in changing the zonal-mean flow. The analysis of the geopotential height perturbation reveals that the amplification of the waves of zonal wavenumbers 1 and 2 in the SH stratosphere is responsible for the change in the zonal-mean flow there. In particular, the wave amplitude of zonal wavenumber 2 significantly increases (more than three times) by the GWDC process in the SH midlatitude upper stratosphere. It is suggested that understanding interactions between the gravity wave drag, zonal-mean flow, and planetary waves is necessary to better parameterize the gravity wave drag. This study is particularly encouraging in that including the GWDC parameterization can alleviate the excessive westerly bias in the SH midlatitude and its associated cold temperature bias in the SH polar region reported for many GCMs.