Impact of aerosols on cloud system-resolving model simulations of deep convection

The indirect impact of atmospheric aerosols (i.e., the impact through cloud processes) is one of the most uncertain aspects of the clouds-in-climate problem. This paper describes simulations of aerosol indirect effects using a two-dimensional cloud system-resolving model with a two-moment liquid and ice microphysics scheme, focusing on deep convection and associated outflow cirrus clouds. Simulations are performed with different aerosol loadings representing either pristine or polluted conditions. Two different quasi-idealized modeling frameworks are utilized. First, the impact of aerosols and cloud microphysics on convective-radiative quasi-equilibrium (CRE) over a surface with fixed characteristics and prescribed solar input, both mimicking the mean conditions on Earth, is described. Next, simulations of a 16-day period of monsoon conditions during the 2006 Tropical Warm Pool - International Cloud Experiment (TWP-ICE) using observed large-scale forcing and initial conditions are discussed. In both setups, aerosols have little impact on tropospheric cooling through modification of clouds; therefore, surface precipitation is insensitive to aerosols given the fixed sea surface temperature. 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 process-oriented approach. 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 (TOA) in nearly all sets of runs. However, for TWP-ICE, the spread of the TOA radiative fluxes among different model realizations (generated from applying very small, random perturbations to lower tropospheric potential temperature) with either pristine or polluted aerosols is large, exceeding 30 W m-2 even when averaged over the entire the 16-day period. This "internal" variability overwhelms any signal from the aerosol indirect effects for a given set of pristine and polluted realizations. Thus, large member ensembles were run to determine statistical significance of the aerosol impacts. Sensitivities of aerosol indirect effects to cloud microphysics parameter settings, model resolution, and domain size using the ensemble approach are also described.