Observational Constraint on a Feedback from Supercooled Clouds Reduces Projected Warming Uncertainty
Supercooled clouds become less precipitation-efficient warm clouds in response to warming, which increases overall liquid cloud amount and optical thickness, and therefore their reflection, constituting a negative feedback.
Greg Cesana
Scientists at Lawrence Livermore National Laboratory collaborated with colleagues at Columbia University and elsewhere in investigating the radiative feedback arising from changes in cloud optical depth with warming and their dependence on mean-state cloud properties. Previous work suggested that models with too much mean-state cloud ice experienced a too-strong negative “phase feedback” from the transition of ice clouds to liquid clouds. In this work, instrument simulators were used to rigorously inter-compare model cloud phase distributions to each other and to satellite observations. The relationship between the mean-state cloud phase and cloud feedback was then elucidated using single-model experimentation and multi-model analysis.
The team made two major findings in this work: First they showed that the most recent generation of models overestimates the relative amount of supercooled cloud liquid water at sub-freezing temperatures, contrary to previous studies that suggested the opposite. Secondly, they showed that models with a greater relative amount of supercooled liquid tend to have stronger negative cloud feedbacks, contrary to previous studies based on single model analysis. This stronger negative feedback is tied to the shift from supercooled clouds to warm clouds that precipitate less readily, leading to longer-lived and thicker clouds that reflect more sunlight back to space. Because of the linkage between mean-state cloud phase and the cloud feedback, observationally constraining the mean-state phase helps constrain climate sensitivity.
The increase of carbon-dioxide-doubling-induced warming (climate sensitivity) in the latest climate models is primarily attributed to a larger extratropical cloud feedback. This is thought to be partly driven by a greater ratio of supercooled liquid-phase clouds to all clouds, termed liquid phase ratio. We use an instrument simulator approach to show that this ratio has increased in the latest climate models and is overestimated rather than underestimated as previously thought. In our analysis of multiple models, a greater ratio corresponds to stronger negative cloud feedback, in contradiction with single model-based studies. We trace this unexpected result to a cloud feedback involving a shift from supercooled to warm clouds as climate warms, which corresponds to greater cloud amount and optical depth and weakens the extratropical cloud feedback. Better constraining this ratio in climate models – and thus this supercooled cloud feedback – impacts their climate sensitivities by up to 1 °C and reduces inter-model spread.