Atmospheric nonlinearity controls ENSO asymmetry in a hybrid statistical-dynamical model

The El Niño Southern Oscillation (ENSO) phases are asymmetrical, with El Niño exhibiting greater sea surface temperature (SST) anomalies compared to La Niña, though La Niña can persist for longer. This asymmetry results in a positive skewness of SST variations in the eastern equatorial Pacific—a feature that climate models struggle to replicate. To understand the roles of oceanic and atmospheric nonlinearities in this asymmetry, we use a hybrid statistical-dynamical model, based on the Community Earth System Model version 2 (CESM2), that couples the ocean component of CESM2 to a partially statistical atmospheric component. We find that without the nonlinear atmospheric wind-stress response to SST anomalies, ENSO exhibits stronger La Niñas than El Niños and negative skewness, contrary to observations, which is caused by oceanic nonlinearities. Only by incorporating the observed wind-stress nonlinearities can the model reproduce the observed ENSO asymmetry, highlighting the atmosphere's critical role within this framework.