Revisiting the modelled cloud radiative feedback on ENSO — the source of model uncertainty

The El Niño-Southern Oscillation (ENSO) is a dominant source of global climate variability. However, climate models exhibit persistent and substantial spread in simulating the ENSO-related cloud radiative feedback (CRFB), which directly limits the fidelity of ENSO amplitude and period projections. This study re-evaluates the representation of ENSO-related CRFB in CMIP6 models, which generally exhibit a weak amplitude and a westward shift of the negative feedback center in the tropical Pacific. To identify the sources of model uncertainty in ENSO CRFB, we analyzed experiments conducted with the atmospheric models Community Earth System Model (CESM) and a modified version of the Max Planck Institute for the Meteorology Earth System Model at low resolution (MPI-ESM-LR). Results show that compared to CESM, MPI-ESM-LR fails to accurately simulate mid-level cloud properties, which largely govern the cloud radiative effect. In contrast, biases in mean sea surface temperature (SST) and ENSO amplitude also considerably influence the simulation of CRFB. The CRFB bias in CMIP6 is strongly linked with that in the corresponding models from the Atmospheric Model Intercomparison Project (AMIP), further indicating the important role of atmosphere model (especially the cloud and convective parameterization) in simulating the CRFB on ENSO.