Evaluation of CMIP6 Model Performance in Simulating the Stratospheric Semiannual Oscillation

Abstract The semiannual oscillation (SAO) is the dominant mode of zonal mean wind in the tropical upper stratosphere to the middle mesosphere, significantly influencing the distribution of stratospheric trace gases and circulation. This study utilizes SABER satellite observations and ERA5 reanalysis data to evaluate the performance of 17 CMIP6 high-top models in simulating the SAO in the tropical upper stratosphere, filling the gap in the evaluation of the SAO simulation performance. We find that the phase transition times of the SAO, as simulated by the majority of CMIP6 models, align closely with both SABER observations and ERA5 reanalysis data. Furthermore, all 17 CMIP6 models accurately reproduce the SAO’s half-year periodicity, with MIROC6 exhibiting the most similar seasonal cycle to that observed in both SABER data and ERA5 reanalysis. Regarding the simulation of the SAO amplitude, most of the 17 selected CMIP6 models effectively capture the shift of amplitude peaks toward the Southern Hemisphere. The peaks of these models at 1 hPa are close to the peak of ERA5 and slightly stronger than those observed by the SABER. In addition, we employed the transformed Eulerian mean momentum equation to examine the various forcing terms driving the SAO. Our analysis highlights the significant role of both parameterized and resolved gravity waves in simulating the SAO. To improve the representation of the SAO in the upper stratosphere, CMIP6 models may need to enhance the realism of their simulations of both gravity waves and resolved wave processes.