Influence of Droplet Sedimentation on Stratocumulus Cloud-Top Entrainment: Impact on Aerosol Indirect Effect in GISS ModelE3

Reduction in cloud droplet sedimentation induced by increasing droplet concentrations has been shown in large-eddy simulations (LES) and direct numerical simulation (DNS) to enhance boundary-layer entrainment, thereby reducing cloud liquid water path and offsetting the Twomey effect for lightly drizzling stratocumulus when the overlying air is sufficiently dry, which is typical of the subtropics. Among recent upgrades to ModelE3, the latest version of the NASA Goddard Institute for Space Studies (GISS) general circulation model (GCM), are a two-moment stratiform cloud microphysics treatment with prognostic precipitation (Gettelman and Morrison 2015) and a moist turbulence scheme that uses an explicit entrainment closure (Bretherton and Park 2009), which provides a means of expressing the effect of cloud droplet sedimentation on entrainment, for which we use the simple parameterization of Bretherton et al. (2007). Under-constrained parameters in the entrainment parameterization are computed by comparing single column model (SCM) simulations with LES results for a stratocumulus case study. Including the cloud droplet sedimentation effect allows liquid water path in the SCM to respond to progressively increasing aerosol concentration comparably to that in the LES. In global simulations the cloud droplet sedimentation effect indeed reduces the dependence of cloud liquid water path on increasing aerosol concentrations. We also assess the global aerosol indirect effect (ERF_aci, the effective radiative forcing from aerosol-cloud interactions) by using offline pre-industrial and present day aerosol fields from a ModelE2.1 simulation with the MATRIX model (Bauer et al. 2008). The cloud droplet sedimentation effect reduces global ERF_aci in ModelE3 by about 30% to roughly 1 W m^-2, depending on assumptions. We compare simulated climatologies of low cloud properties with satellite retrievals to assess biases in our approach.