Simulation and Evaluation of the CAS-ESM 1.0 Model Used for Studying Upper Layer Flow-Field Abnormal Modes in the Tropical Pacific Ocean in July

Using simulated data from 1922 to 2005 for the month of July obtained from the CAS-ESM 1.0 model, we performed a complex empirical orthogonal function analysis of flow-field anomalies. The results were compared with the corresponding simulated data for January, actual observations (SODA data), and theoretical results. The main conclusions were as follows: In the surface, near-surface, subsurface, and upper bottom layers significant flow-field anomalies were distributed between 10°S and 10°N in the first and second modes, centered around the equator. Overall, the flow-field anomalies were mainly characterized by pure latitudinal flows. They all corresponded well with the anomalies of equatorial flow systems and fluctuations. The flow-field anomalies from the surface to upper bottom layers significantly weakened with increasing ocean depth. Both the first and second modes showed 3–5-year interannual variation, with the most obvious changes observed over a four-year period, and all variations fell within the range of ENSO. There was no significant interdecadal variation in either mode. The SSTA of the first mode in the subsurface layer was characterized by negative anomalies in the western Pacific warm pool and near the equator in the western and central Pacific, as well as positive anomalies in the equatorial region of the eastern Pacific and along the coast of South America. Further, both anomalies exhibited a seesaw pattern over time. The SSTA of the second mode showed negative anomalies dominating the central and western Pacific, while positive anomalies prevailed in the central and eastern Pacific, which also showed a seesaw pattern. This change encompassed the seesaw pattern observed in the first mode. The characteristics of the first and second modes described in this paper correspond to the primary and secondary modes of ENSO, respectively. The absence of significant interdecadal variation in the time coefficient is attributed to the model’s inability to accurately represent the Indonesian Throughflow. The width of the pronounced flow-field anomalies on both sides of the equator, as simulated by the model, is twice that of the actual case. This indicates that the model resolution near the equator is insufficient and needs further improvement. Additionally, the physical processes and coupled air–sea interactions within this model require further refinement and optimization.