Evaluating Observational Constraints on Intermodel Spread in Cloud, Temperature, and Humidity Feedbacks.

Uncertainty in climate feedbacks is the primary source of the spread in projected surface temperature responses to anthropogenic forcing. Cloud feedback persistently appears as the main source of disagreement in future projections while the combined lapse‐rate plus water vapor (LR + WV) feedback is a smaller (30%), but non‐trivial source of uncertainty in climate sensitivity. Here we attempt to observationally constrain the feedbacks in an effort to reduce their intermodel uncertainties. The observed interannual variation provides a useful constraint on the long‐term cloud feedback, as evidenced by the consistency of global‐mean values and regional contributions to the intermodel spread on both interannual and long‐term timescales. However, interannual variability does not serve to constrain the long‐term LR + WV feedback spread, which we find is dominated by the varying tropical relative humidity (RH) response to interhemispheric warming differences under clear‐sky conditions and the RH‐fixed LR feedback under all‐sky conditions. Plain Language Summary: How much the Earth warms in response to greenhouse gas increases depends on the Earth's efficiency in restoring radiative equilibrium. This efficiency differs significantly among climate models due to differences in feedback processes, particularly the responses of clouds, temperature and water vapor to the initial perturbation. One approach to narrowing the intermodel spread of feedbacks is to only consider models whose observable variability is consistent with available measurements. The magnitude of cloud feedbacks on interannual and long‐term timescales are closely related, which allows us to employ this approach with observational estimates of the interannual cloud feedback to constrain the long‐term cloud feedback. However, this approach does not work for the feedback resulting from changes in the vertical distribution of temperature and water vapor (the combined lapse‐rate plus water vapor feedback). Key Points: Observed interannual variation provides a useful constraint to narrow the uncertainty in long‐term cloud feedbackIt is not possible to constrain the long‐term LR + WV feedback uncertainty with available observations of interannual variabilityDisagreements in the response of tropical relative humidity are responsible for the long‐term clear‐sky LR + WV feedback spread [ABSTRACT FROM AUTHOR]