Improved cirrus simulations in a general circulation model using CARMA sectional microphysics

We have developed a new cirrus model incorporating sectional ice microphysics from the Community Aerosol and Radiation Model for Atmospheres (CARMA) in the latest version of NCAR's Community Atmosphere Model (CAM5). Comparisons with DARDAR and 2C-ICE show that CAM5/CARMA improves cloud fraction, ice water content, and ice water path compared to the standard CAM5. Prognostic snow in CAM5/CARMA increases overall ice mass and results in a melting layer at ~4 km in the tropics that is largely absent in CAM5. Subgrid scale supersaturation following Wilson and Ballard (1999) improves ice mass and relative humidity. Increased middle and upper tropospheric condensate in CAM5/CARMA requires a reduction in low-level cloud for energy balance, resulting in a 3.1 W m−2 improvement in shortwave cloud forcing and a 3.8 W m−2 improvement in downwelling shortwave flux at the surface compared to CAM5 and Clouds and Earth's Radiant Energy Systems (CERES). Total and clear-sky longwave upwelling flux at the top are improved in CAM5/CARMA by 1.0 and 2.6 W m−2, respectively. CAM has a 2–3 K cold bias at the tropical tropopause mostly from the prescribed ozone file. Correction of the prescribed ozone or nudging the CAM5/CARMA model to GEOS5-DAS meteorology yields tropical tropopause temperatures and water vapor that agree with the Constellation Observing System for Meteorology, Ionosphere, and Climate (COSMIC) and the Microwave Limb Sounder (MLS). CAM5 relative humidity appears to be too large resulting in a +1.5 ppmv water vapor bias at the tropical tropopause when using GEOS5-DAS meteorology. In CAM5/CARMA, 75% of the cloud ice mass originates from ice particles detrained from convection compared to 25% from in situ nucleation.