Parametric Sensitivity Analysis of <scp>ENSO</scp> ‐Related Shortwave Feedback Simulations by <scp>PPE</scp> Experiments in <scp>CAM6</scp>

ABSTRACT Accurately representing the shortwave feedback during the El Niño‐Southern Oscillation (ENSO) cycle (α SW ) remains a major challenge for climate models. This study examined the key parameters within the deep convection scheme and cloud physics scheme of Community Atmosphere Model version 6 (CAM6) to ascertain their impacts on simulating α SW . A suite of 128‐member Perturbed Parameter Ensemble (PPE) experiments was first conducted by simultaneously perturbing 12 selected parameters. Empirical Orthogonal Function (EOF) analysis identifies two dominant modes of α SW simulation biases. The first mode is primarily controlled by cloud ice‐related parameters ( dcs , eii , ai, and ai*dcs ), which exert their influence on α SW through modulating the ice water path (IWP) feedback; while the second mode is predominantly influenced by convection‐related parameters ( dmpdz , dmpdz*dmpdz , and tau ), which act through the liquid water path (LWP) feedback. Our further physical interpretation reveals how these identified key parameters, as well as the nonlinear term ai*dcs (i.e., the interplay between ai and dcs ), modulate cloud radiative processes and hence shape the simulated α SW in CAM6. Our findings regarding the identified key parameters and their underlying mechanisms are not only useful for the development of the widely‐used CESM model, but also provide valuable insights for the advancement of other AGCMs.