Testing ice microphysics parameterizations in the NCAR Community Atmospheric Model Version 3 using Tropical Warm Pool–International Cloud Experiment data

Cloud properties have been simulated with a new double‐moment microphysics scheme under the framework of the single‐column version of NCAR Community Atmospheric Model version 3 (CAM3). For comparison, the same simulation was made with the standard single‐moment microphysics scheme of CAM3. Results from both simulations compared favorably with observations during the Tropical Warm Pool–International Cloud Experiment by the U.S. Department of Energy Atmospheric Radiation Measurement Program in terms of the temporal variation and vertical distribution of cloud fraction and cloud condensate. Major differences between the two simulations are in the magnitude and distribution of ice water content within the mixed‐phase cloud during the monsoon period, though the total frozen water (snow plus ice) contents are similar. The ice mass content in the mixed‐phase cloud from the new scheme is larger than that from the standard scheme, and ice water content extends 2 km further downward, which is in better agreement with observations. The dependence of the frozen water mass fraction on temperature from the new scheme is also in better agreement with available observations. Outgoing longwave radiation (OLR) at the top of the atmosphere (TOA) from the simulation with the new scheme is, in general, larger than that with the standard scheme, while the surface downward longwave radiation is similar. Sensitivity tests suggest that different treatments of the ice crystal effective radius contribute significantly to the difference in the calculations of TOA OLR, in addition to cloud water path. Numerical experiments show that cloud properties in the new scheme can respond reasonably to changes in the concentration of aerosols and emphasize the importance of correctly simulating aerosol effects in climate models for aerosol‐cloud interactions. Further evaluation, especially for ice cloud properties based on in situ data, is needed.