Impact of Deep Convective Parameterization on Low Cloud Feedback and Climate Sensitivity
Inter-model differences in the shortwave cloud feedback associated with low cloud change contribute significantly to the spread of climate sensitivity. The perturbed physical experiments (PPE) with modified deep convective parameters such as the convective entrainment rate produce similar spread in low cloud response to surface warming to CMIP5 and CMIP6 multi-model ensembles. In this study, we show that different deep convective parameterizations produce different characteristics in convective initiation and frequency, leading to differences in tropical overturning circulation and high/low cloud sensitivity to surface warming. A number of physical pathways are identified that link the model diversity in deep convection to that in low cloud feedback and climate sensitivity. Our study highlights the importance of improving deep convective parameterization in climate models to reducing the uncertainty of climate sensitivity estimate.