Responses to Lower‐Tropospheric Stability Dominate Intermodel Differences in the Historical Pattern Effect

Abstract The sea surface temperature (SST) pattern modulates the Earth's radiative feedbacks at decadal timescales. However, decadal feedback variations over the historical period depend on the choice of climate model, even when forced with the same SST data set. We investigate these differences through analyzing amip‐piForcing output, focusing on low cloud feedback. Time‐variation in low cloud feedback results primarily from the SST pattern, mediated by changes in near‐global lower tropospheric stability (LTS). We investigate the meteorological factors contributing to low cloud feedback variation through the cloud controlling factor analysis method, diagnosing cloud radiative sensitivities from 16 CMIP5 and CMIP6 models, and an observational constraint from MODIS. We find the radiative sensitivity of models to LTS in the stratocumulus regions dominates the magnitude of cloud feedback time variation due to the pattern effect. Additionally, models underestimate low cloud radiative sensitivity to changes in LTS and SST and thus the time‐variation in feedback.