Impact of nucleation schemes on cirrus cloud formation in a GCM with sectional microphysics

We have implemented a sectional microphysics scheme for ice clouds based upon the Community Aerosol and Radiation Model for Atmospheres (CARMA) in the Community Atmosphere Model version 5 (CAM5), which allows for a size resolved treatment of ice particle nucleation, condensational growth, coagulation, sedimentation and detrainment. Detrained and in situ formed ice particles are tracked separately in the model allowing for different microphysical assumptions and separate analysis. Cloud ice from CAM5/CARMA simulations compare better with satellite observations than those with the standard CAM5 two-moment microphysics. CAM5/CARMA has a prognostic treatment for snow, which results in improved ice mass and representation of a melting layer that is absent in CAM5. Here we explore the sensitivity of the simulations to different nucleation schemes including: homogeneous freezing based on Koop et al. (2000), homogeneous freezing based upon Aerosols Interaction and Dynamics in the Atmosphere (AIDA) chamber measurement (Möhler et al., 2010), heterogeneous nucleation with dust aerosols, and heterogeous nucleation with glassy aerosols (Murray et al. 2010). The initial size for detrained ice particles in CAM5/CARMA is temperature dependent based upon a fits to observations from Heymsfield et al. (2010). We explore the sensitivity of the model to different choices for these fits. Results from these simulations are compared to retrievals of water vapor from the Microwave Limb Sounder (MLS) and the Atmospheric Infrared Sounder (AIRS), ice cloud properties from CloudSat-CALIPSO observations (Delanoë and Hogan, 2010; Deng et al. 2010) and to aircraft observations from several field campaigns including: the Costa Rica Aura Validation Experiment (CR-AVE), the Tropical Composition, Cloud and Climate Coupling (TC4), the Mid-latitude Airborne Cirrus Properties Experiment (MACPEX) and the Airborne Tropical Tropopause Experiment (ATTREX).