The effects of models of aerosol hygroscopicity on the apportionment of extinction

The role that aerosols play in climate forcing and visibility issues has been the subject of research for several decades. Recent research efforts have focused on assessing the contribution of individual species to scattering and absorption under ambient conditions and on how scattering and absorption change as one or more species are removed from the atmosphere. A key concern is the distribution of water among aerosols as a function of mixing assumptions. As an illustrative and relevant example, we examine the roles of sulfates and organics in visibility and climate forcing, and in particular, the implications of assumptions regarding hygroscopic growth behavior upon the assignment of mass-scattering efficiencies to these species. We demonstrate that the total scattering computed for an aerosol sample is relatively insensitive to the choice of internal or external mixture, and can be insensitive to the exact formulation of the hygroscopic growth of the sample. Since the atmospheric aerosol is generally a complex mixture of chemical species, with the precise distribution of species on a particle-by-particle basis not known, the use of semi-empirical models of multicomponent aerosol hygroscopicity is appropriate for the calculation of atmospheric aerosol scattering and/or extinction, particularly since these details appear to be unimportant in most cases. In contrast, the apportionment of percentages of the total scattering to individual chemical species is quite sensitive to the choice of assumption regarding the aerosol microphysical structure. The use of semi-empirical hygroscopic growth models for computing the change in species scattering efficiency can lead to incorrect predictions in the limit of the complete removal of all but one chemical component. We propose a model that invokes the Zdanovskii, Stokes, and Robinson (ZSR) assumptions for the water content of multicomponent mixtures, and demonstrate that this method both approximates the predictions of generally accepted, semi-empirical models and asymptotically approaches the correct single-species scattering contributions upon removal of species. Efficiencies computed with this model can differ significantly from those predicted using the semi-empirical approach.