An Efficient Representation of Aerosol Mixing State for Atmospheric Models

Poster PDF

Description

Atmospheric aerosol particles influence the Earth’s radiative balance directly by absorbing and scattering the incoming solar radiation and indirectly by serving as cloud condensation nuclei (CCN). The impact of aerosols on the climate system is one of the major contributors to climate prediction uncertainties. Therefore, to reduce the uncertainties of aerosol climate impacts, it is essential to reliably predict CCN concentrations and aerosol optical properties, both of which strongly depend on particle size, chemical composition, and mixing state. Light absorption by black carbon (BC) particles emitted from fossil fuel combustion depends on their size and how thickly they are coated with non-refractory species such as ammonium, sulfate, nitrate, organics, and water. The cloud condensation nuclei (CCN) activation behavior of a particle depends on its dry size and the hygroscopicities of all the individual species mixed together. It is therefore necessary to represent both size and mixing state of aerosols to reliably predict their climate-relevant properties in atmospheric models. Here we describe and evaluate a novel sectional framework in a box-model version of the Model for Simulating Aerosol Interactions and Chemistry, referred to as MOSAIC-mix, that represents the mixing state by resolving aerosol dry size (Ddry), BC dry mass fraction (wBC), and overall particle hygroscopicity (k). Using 10 idealized urban plume scenarios in which different types of aerosols evolve over 24 h under a range of atmospherically relevant conditions, we examine errors in CCN concentrations and optical properties with respect to the level of detail of the aerosol mixing state representation. We find that a small number of wBC and bins can achieve significant reductions in the errors and propose a configuration with 24 Ddry bins, 2 wBC bins, and 2 k bins that give average errors of about 5% or less in CCN concentrations and optical properties, which are 3–4 times lower than those from size-only resolved (i.e., internally mixed) simulations. We further evaluate the performance of MOSAIC-mix along with the particle-resolved model PartMC-MOSAIC in Lagrangian box-model frameworks against single-particle data obtained during the 2010 Carbonaceous Aerosols and Radiative Effects Study (CARES). These results suggest that MOSAIC-mix is suitable for use in large-scale models to examine the effects of mixing state on aerosol-radiation-cloud feedbacks.