Cloud-resolving model simulations of aerosol indirect effects in tropical deep convection: Sensitivity to microphysics parameterization and model resolution

A two-dimensional cloud model is coupled with a new two-moment bulk ice microphysics scheme (Morrison and Grabowski, 2008a, JAS) and liquid microphysics scheme (Morrison and Grabowski, 2007, 2008b, JAS). The new ice microphysics scheme does not separate snow and graupel into different species as in most existing schemes, and instead separately predicts ice mixing ratios grown by vapor diffusion and riming. Rime mass fraction is locally derived from the vapor diffusion and riming mixing ratios, with various characteristics (mass-size, projected area-size, and terminal velocity-size relationships) derived as a function of the particle size and rime mass fraction. This approach allows for smooth transitions between pristine ice, snow, and graupel, and avoids arbitrary or tuned conversion parameters and threshold behavior.

The model is applied to a 16-day period of the Jan.-Feb. 2006 Tropical Warm Pool - International Cloud Experiment (TWP-ICE) and compared against available observations and microphysical retrievals. Sets of simulations with varying microphysics parameters and model grid spacing are run with either pristine aerosol concentrations as observed during TWP-ICE, or relatively polluted conditions as observed in the region during the Nov. 2005 Aerosol and Chemical Transport in Tropical Convection (ACTIVE) experiment. Precipitation is mostly determined by the specified large-scale forcing and hence is not sensitive to aerosol. This lack of sensitivity illustrates differences between aerosol effects on precipitation using a system-dynamics approach for an ensemble of clouds versus the effects on a single cloud or cloud system using a more process-level approach. While modification of cloud characteristics as a result of aerosol loading increases outgoing shortwave radiative flux and decreases outgoing longwave flux at the top-of-atmosphere in all sets of runs, the nature and magnitude of this effect is sensitive to microphysical parameter settings, including density of heavily-rimed ice (graupel).