Observational Constraints on Mixed-Phase Clouds Imply Higher Climate Sensitivity
Given the known tendency for models to simulate clouds with too little liquid below freezing, the authors explored the implications of changing the supercooled liquid fraction (SLF) in a single model for its cloud feedback and climate sensitivity. When the model was modified to create SLFs similar to clouds observed from space, the stabilizing cloud phase feedback was weakened and the model's climate sensitivity increased by 1-1.3 degrees C.
A common feature of climate model simulations of global warming is an increase in the reflectivity of cold clouds at middle and high latitudes. This increase is primarily due to the increase in cloud liquid water content, part of which arises from phase transitions from ice to liquid. Because liquid clouds tend to be more reflective than ice clouds, this phase transition acts as a stabilizing feedback, limiting the amount of warming in response to increased carbon dioxide. The icier the clouds to begin with, the more liquid is gained as the planet warms, so this stabilizing feedback is stronger in models containing less liquid relative to ice at sub-freezing temperatures. Because most models’ clouds contain too much ice that is susceptible to becoming liquid with warming, their stabilizing cloud phase feedback may be unrealistically strong. It is important to assess the implications of contraining this feedback with the best available observations.
To explore the implications of changing the ice-to-liquid ratio for cloud feedback and climate sensitivity, the researchers modified parameters in CESM version 1.0.6. They saw a systematic weakening of the cloud phase feedback and increase in climate sensitivity as they transitioned from model versions that readily convert liquid to ice below freezing to model versions that can maintain liquid down to colder temperatures. In the two model versions in which the ice-to-liquid ratio were in close agreement with clouds observed in nature by the CALIOP instrument, the model’s climate sensitivity was 1-1.3 degrees C larger than in the default model. Given that most climate models too readily convert liquid to ice below freezing, the results imply that they may also exaggerate the increase in cloud reflectivity as the atmosphere warms and hence underestimate climate sensitivity. Independent confirmation with other, less sensitive, models will be important for establishing the robustness of the result.