Effects of Convective Ice Fall Velocity and Detrainment on Anvil Clouds: Single-Column Model Testing with NCAR CAM6 Evaluated with the ARM TWP-ICE and MC3E Data

Description

Partitioning of convective cloud and precipitating particles between fallout and detrainment in Global Climate Models (GCMs) is of great importance for climate projection because of its strong impacts on convective and stratiform precipitation and large-scale anvil clouds. In this study, we first implement and test two new ice particle fall velocity parameterizations in the convective microphysics scheme of the NCAR Community Atmospheric Model version 6 (CAM6). One parameterization is based on the terminal velocity-diameter (Vt-D) power law relation with the pre-factor and exponential factor varying with temperature, ice water content and pressure derived from in-situ measurements. The other one is physically derived based on the tank experiment considering both the particle and air flow properties. The original convective microphysics scheme in CAM6 only considers the fall velocities of snow and rain and we have further improved it to consider the convective cloud ice crystal sedimentation. Second, in addition to the detrainment of convective cloud liquid droplets and ice crystals, we have also considered the convective snow detrainment in the convective updraft to be fed into the stratiform microphysics scheme, which is not considered in the original CAM6. A series of sensitivity experiments have been performed with the single-column model (SCM). Simulated microphysical properties with the new ice fall speed and detrainment parameterizations are evaluated against observations from the U.S. Department of Energy (DOE) Atmospheric Radiation Measurement (ARM) Program’s Tropical Warm Pool-International Cloud Experiment (TWP-ICE) and Midlatitude Continental Convective Cloud Experiment (MC3E). Our results suggest that the new mass-weighted terminal velocity of snow particles increases by a factor of 2 at lower altitudes compared with the original parameterization. However, the fall speed of snow particles is still not sufficiently large to redistribute the cloud condensate vertically due to the incapability of treating graupel. By taking into account the convective snow detrainment, the large-scale precipitation increases. Moreover, it is found that observed convective updraft vertical velocity favors more frequent and intense updraft velocity over model simulations particularly at higher altitudes, which reveals that the upper bound threshold of 15m/s in the original convective microphysics scheme is problematic.