Improved ENSO simulation from climate system model FGOALS-g1.0 to FGOALS-g2

This study presents an overview of the improvement in the simulation of El Niño–Southern Oscillation (ENSO) in the latest generation of the Institute of Atmospheric Physics’ coupled general circulation model (CGCM), the Flexible Global Ocean–Atmosphere–Land System model Grid-point Version 2 (FGOALS-g2; hereafter referred to as “g2”) from its predecessor FGOALS-g1.0 (referred to as “g1”), including the more realistic amplitude, irregularity, and ENSO cycle. The changes have been analyzed quantitatively based on the Bjerknes stability index, which serves as a measure of ENSO growth rate. The improved simulation of ENSO amplitude is mainly due to the reasonable representation of the thermocline and thermodynamic feedbacks: On the one hand, the deeper mean thermocline results in a weakened thermocline response to the zonal wind stress anomaly, and the looser vertical stratification of mean temperature leads to a weakened response of anomalous subsurface temperature to anomalous thermocline depth, both of which cause the reduced thermocline feedback in g2; on the other hand, the alleviated cold bias of mean sea surface temperature leads to more reasonable thermodynamic feedback in g2. The regular oscillation of ENSO in g1 is associated with its unsuccessful representation of the role of atmospheric noise over the western–central equatorial Pacific (WCEP) in triggering ENSO events, which arises from the weak synoptic–intraseasonal variability of zonal winds over the WCEP in g1. The asymmetric transition of ENSO in g1 is attributed to the asymmetric effect of thermocline feedback, which is due to the annual cycle of mean upwelling in the eastern Pacific. This study highlights the great impact of improving the representation of mean states on the improved simulation of air–sea feedback processes and ultimately more reasonable depiction of ENSO behaviors in CGCMs.