A Multi-Model Analysis of the Cloud Phase Transition in 16 GCMs Using Satellite Observations (CALIPSO/GPCP) and Reanalysis Data (ECMWF/MERRA).

Thursday, 18 December 2014
Grégory Cesana1, Duane Edward Waliser1, Xianan Jiang2 and Jui-Lin F Li1, (1)NASA Jet Propulsion Laboratory, Pasadena, CA, United States, (2)JIFRESSE/UCLA, Pasadena, CA, United States
The ubiquitous presence of clouds within the troposphere contributes to modulate the radiative balance of the earth-atmosphere system. Depending on their cloud phase, clouds may have different microphysical and macrophysical properties, and hence, different radiative effects.

In this study, we took advantage of climate runs from the GASS-YoTC and AMIP multi-model experiments to document the differences associated to the cloud phase parameterizations of 16 GCMs. A particular emphasize has been put on the vertical structure of the transition between liquid and ice in clouds. A way to intercompare the models regardless of their cloud fraction is to study the ratio of the ice mass to the total mass of the condensed water. To address the challenge of evaluating the modeled cloud phase, we profited from the cloud phase climatology so called CALIPSO-GOCCP, which separates liquid clouds from ice clouds at global scale, with a high vertical resolution (480m), above all surfaces.

We also used reanalysis data and GPCP satellite observations to investigate the influence of the temperature, the relative humidity, the vertical wind speed and the precipitations on the cloud phase transition. In 12 (of 16) models, there are too few super cooled liquid in clouds compared to observations, mostly in the high troposphere. We exhibited evidences of the link between the cloud phase transition and the humidity, the vertical wind speed as well as the precipitations. Some cloud phase schemes are more affected by the humidity and the vertical velocity and some other by the precipitations.

Although a few models can reproduce the observe relation between cloud phase and temperature, humidity, vertical velocity or precipitations, none of them perform well for all the parameters.

An important result of this study is that the T-dependent phase parameterizations do not allow simulating the complexity of the observed cloud phase transition. Unfortunately, more complex microphysics schemes do not succeed to reproduce all the processes neither. Finally, thanks to the combined use of CALIPSO-GOCCP and ECMWF water vapor pressure, we showed an updated version of the Clausius-Clapeyron water vapor phase diagram. This diagram represents a new tool to improve the simulation of the cloud phase transition in climate models.