Cloud Response Is Significantly Biased by CMIP6 Over the Tibetan Plateau

Abstract To reduce the uncertainty of cloud feedback, models should prioritize the evaluation of cloud response to surface warming. Using International Satellite Cloud Climatology Project‐H data sets, we quantified the cloud fraction response of different cloud types to surface warming in the Tibetan Plateau (TP) and examined the performance of Coupled Model Intercomparison Project Phase 6 models in reproducing cloud response. Results showed multimodel ensemble mean (MEM) exhibited a similar pattern of total cloud fraction response with the observation but severely underestimated it over the western TP. MEM also obviously underestimated cloud fraction of optically thin cirrus (bias >15%) and had a wrong trend of −0.02% 10a−1 against observed +2.44% 10a−1. Moreover, models failed to reproduce extensive negative response of optically thick clouds and exhibited significant underestimation for the negative response of optically thick but lower‐top clouds. Especially, models even simulated an opposite response with International Satellite Cloud Climatology Project for optically thin but lower‐top cloud (bias hit −0.4% K−1).