This study examines the prediction of future changes in clouds under the effects of anthropogenic global warming, using 32 coupled models that participated in phase five of the Coupled Model Intercomparison Project (CMIP5), by comparing two runs: the historical run for 1850–2005 and the Representative Concentration Pathway (RCP) 4.5 run for 2006–2100. Metrics for evaluating the models’ performance on cloud variation are designed to document performance for the period 1980–2005. The metrics include (1) the annual mean cloud; (2) the annual variation in cloud; and (3) the interannual variation in cloud. The evaluation metrics mostly consist of global and regional coefficients of spatial pattern correlation and root-mean-squared errors between the observed and the simulated results. In addition, the models’ performances regarding the interannual variation in low cloud are considered through their ability to simulate the cloud meteorology correlation proposed by Clement et al . (2009). Based on the models’ performances in simulation of the cloud variation metrics, we selected some best models for further assessment of future. Using a multi-model ensemble with the three best-performing models (B3MME) the following changes are projected in the twenty-first century under the RCP4.5 scenario: (1) The annual mean and range of cloud and cloud radiative cooling effects will decrease a significant amount, especially over the eastern north subtropical ocean, suggesting the role of positive feedback in the climate change. (2) Changes in cloud exhibit huge differences between the northern hemisphere (NH) and the Southern Hemisphere (SH). (3) There will be a more prominent temperature asymmetry between the northern and southern hemispheres due to the more evident reduction of cloud in the NH than in the SH. (4) The model spread is substantially reduced by carefully applying the evaluation metrics and selecting most reliable models. Remaining models show a good agreement in the sign and the magnitude of the cloud radiation feedback and the geographical pattern of cloud reduction in the future climate. The results of this study address the degree to which the uncertainty in global climate temperature projections can be reduced through the increased confidence in cloud feedback.

1. INTRODUCTION

2. MODEL AND DATA

2.1 Models and experiment

2.2 Observation

3. Evaluation and selection of CMIP5 models

3.1 Annual mean cloud

3.2 Seasonal variation in cloud amount

3.3 Interannual variation in cloud amount

3.4 Best model selection

4. Future changes in cloud radiative forcing by global warming

5. Transient trends

5.1 Cloud cover and CRF

6. Summary