Statistical-Dynamic Diagnosis of Midsummer Stratospheric Wind Anomalies and the Evaluation of Quasi-Biennial Oscillation Simulation in CAS-ESM 1.0 Model

To evaluate the performance of the Chinese Academy of Sciences Earth System Model version 1.0 in simulating the stratospheric quasi-biennial oscillation (QBO), this study uses 84 years of simulated July wind field data over the equatorial Pacific stratosphere. A statistical–dynamical diagnosis, namely, Complex Empirical Orthogonal Function（CEOF）analysis, is applied to the wind field anomalies, and the results were compared with observations from the NCAR/NCEP reanalysis. The findings reveal that the first and second modes account for 21.5% and 15.5% of the variance, respectively, for a cumulative contribution of 37.0%. The principal component of the first mode exhibits notable interannual variation periods of 2.4, 2.9, 3.5, 4.1, and 4.9 years. The first two periods correspond closely to the QBO cycle, whereas the latter three align with the El Niño-Southern Oscillation(ENSO)period. In addition, a pronounced decadal variation with a period of 11.6 years is observed. In the mid, upper-mid, and high levels of the first-mode spatial field in the stratosphere, particularly over the 10°S–15°S Pacific region, pronounced easterly wind anomalies are observed, strengthening upward and generally exhibiting characteristics of an equivalent-barotropic Kelvin wave. The principal component of the second mode shows notable interannual variation periods of 3.5, 4.1, and 4.9 years, with the 4.1-year period being the most prominent. These periods also lie within the interannual variability range of ENSO. Similar to the first mode, an 11.6-year decadal variation is present. In the spatial pattern of the second mode, notable meridional wind and cross-equatorial flow anomalies are observed over the equatorial Pacific stratosphere. However, these features differ from the observations and do not display the characteristics of a baroclinic Kelvin wave. This mode primarily reflects the stratospheric signature of ENSO. The evaluation indicates that the first mode of the model can represent QBO and generally captures the barotropic Kelvin wave in the stratosphere, whereas the second mode reflects the influence of ENSO. These discrepancies may result from the insufficient vertical resolution of the model, imperfect representation of physical processes, and overemphasis on tuning for ENSO simulation. Future improvements in these areas are necessary to optimize the model’s performance.