Toward a better representation of iron oxide effects upon dust shortwave radiative effect in the Community Atmospheric Model

Abstract:

The complex refractive index (CRI) of dust aerosols that results from mixtures of common dust minerals remains highly uncertain. Mixing states of different dust minerals and the method to compute the bulk-dust optical properties are among factors determining the dust CRI and thus the accuracy of the direct radiative effect (DRE) estimate. The recently developed Community Atmospheric Model of version 6 (CAM6) simulates regional variations in dust mineral composition using the four-modes Mode Aerosol Model (MAM4). The MAM4 in CAM6 calculates the bulk-aerosol CRI using a volume-averaging method that assumes all dust minerals and other aerosol species are completely internally mixed within each mode. Applying the volume-averaging method to the mixture of dust minerals to calculate the effective CRI, however, could lead to a low single scattering albedo compared to AErosol Robotic NET-work (AERONET) observations and hence a considerable bias toward warming in the dust DRE estimate. Here, the shortwave DRE by dust at the top of the atmosphere is reevaluated with CAM6. As an extension, a more accurate method of deriving the bulk-dust CRI from insoluble mineral mixtures, the Maxwell-Garnett Approximation (MGA), is introduced. In MGA, dust is assumed as a spatially uniform matrix medium taking iron oxide as an isotropic inclusion on average. Contrasting with the random distribution of iron oxide within the dust particles assumed by MGA, dust DRE by an external mixture of iron oxide with other species is also calculated. In addition, we consider how certain components, which are not routinely included in the DRE estimate (e.g., dark rock particles like particles derived from many kinds of igneous and metamorphic rocks commonly containing ferromagnesian and iron-titanium oxide minerals but are omitted from soil mineral datasets used to compute dust mineralogy), might influence the dust DRE under different mixing-state assumptions. Our preliminary results highlight the sensitivity of dust DRE to (1) the mixing state of iron oxide with other minerals and (2) the method used to compute the bulk-dust optical properties for the dust-DRE estimate.