The uncertainties in the laboratory-measured short-wave refractive indices of mineral dust aerosols and the derived optical properties: A theoretical assessment

Abstract. Mineral dust particles are typically nonspherical and inhomogeneous; however, they are often simplified as homogeneous spherical particles for retrieving the refractive indices from laboratory measurements of scattering and absorption coefficients. This study theoretically investigated uncertainties in refractive indices and corresponding optical properties resulting from this simplification at various sizes within the wavelength range of 355 to 1064 nm. Different numerical experiments were conducted under both ideal and realistic scenarios, taking into account instrumental bias in the realistic scenarios. In the numerical experiments, the inhomogeneous super-spheroid models were considered as the dust samples, while the homogeneous super-spheroid models and sphere models were used to retrieve the refractive indices. Under the ideal scenario, the look-up tables for the homogeneous super-spheroid models satisfactorily covered the measurements at any size and wavelength, while those for the sphere models failed when considering large sizes. Under the realistic scenario, both the homogeneous super-spheroid models and sphere models were ineffective for large sizes due to discrepancies in size distribution resulting from the measurements using an optical particle counter. Nevertheless, it was possible to retrieve the imaginary parts of the refractive indices based solely on the absorption coefficients. The imaginary parts obtained from the sphere models were generally consistent with those from the super-spheroid models under ideal conditions, while the former was significantly smaller than the latter under the realistic conditions. In addition, the retrieved imaginary parts were found to be size-dependent, which could be attributed to the inherent limitations of homogeneous models in characterizing inhomogeneous particles. Results showed that the uncertainties in the imaginary part and single scattering albedo should be smaller than 0.002 (0.0007) and 0.03 (0.01), respectively, under conditions of high (low) absorption. The sphere models tended to overestimate the asymmetry factor. The uncertainty in the asymmetry factor exhibited a significant variation, reaching up to 0.04 or even larger. Nonetheless, the uncertainties in the phase matrices resulting from the uncertainties in refractive indices were generally acceptable within a specific model.