Methods of Evaluating the Interannual Cloud Feedback and Its Uncertainties

The cloud feedback is the largest source of uncertainty among climate feedbacks. It is defined as the change of cloud-induced global top-of-atmosphere (TOA) flux per unit change in global mean surface temperature. However, this definition is impractical since the radiative flux is also sensitive to the spatial structure of surface temperature, which results in large spread among models. Here we test an alternative framework in which surface temperature is replaced by tropical 500-hPa temperature and we show that it can reduce the uncertainty in cloud feedback. We examine interannual cloud feedback using both frameworks in observations/reanalysis and in global climate models. The major discovery is that calculating feedbacks using 500-hPa temperature yields significant reductions in both the uncertainty of individual estimates and in the spread of the cloud feedback between model runs. Furthermore, the spatial and temporal variations of cloud feedback are also reduced. The diminution is mainly contributed by the shortwave component of cloud feedback. This research implies a simple yet powerful approach to constrain cloud feedback and thus the climate sensitivity.