State dependence of cloud feedback and its implications for climate sensitivity

The time-dependence of climate feedbacks is critical for understanding and constraining climate sensitivity. Although time-varying feedbacks are typically attributed to the geographic pattern of surface warming, it can also arise from a state-dependence of feedbacks on global temperature. Utilizing a coordinated set of CMIP6 model simulations, we demonstrate a systematic increase in climate sensitivity as the base state warms. This state-dependence is further explored in a series of GFDL model simulations that span a larger range of base states. Both sets of simulations exhibit a monotonic increase in climate sensitivity due to a strengthening of cloud feedback in warmer climates. This occurs primarily over middle and high latitudes, where a weakening in the negative optical depth feedback consistently overwhelms changes or shifts in cloud amount. These changes are not attributable to differences in mixed-phase cloud microphysics or meridional shifts of mid-latitude storm tracks as have been previously suggested. Instead, the warmer moist adiabat—characterized by an increase in adiabatic cloud water content with surface warming, driven by a modest reduction in lapse rate and a significant rise in cloud-base saturation vapor pressure—and the weakened adiabaticity dominate the state-dependence of cloud feedback, alongside the intrinsic logarithmic dependence of cloud albedo on optical depth. This represents a fundamental state-dependence of cloud feedback, implying a continuing increase in climate sensitivity but reaching a maximum value earlier than previous studies and assessments suggested. Main Text