Inversion algorithm for non-spherical dust particle size distributions

Abstract

Dust particles are the main aerosol component of the atmosphere, and can influence human, environmental, and ecological health. Particle size distribution is an important aerosol micro-physical parameter that denotes the concentration distribution of particles of different radii and can determine the extinction characteristics of these particles. In traditional inversion algorithms, the aerosol is generally assumed to be spherical according to Mie theory, and the relationship between aerosol optical thickness and particle size distribution is described by the Fredholm integral equation of the first kind. For non-spherical dust particles, this spherical assumption is obviously unreasonable and yields unreliable results. Therefore, we developed an algorithm assuming non-spherical particles for inversion of dust particle size distributions. In the case of non-spherical particles, the extinction efficiency factor kernel functions of the ellipsoid were calculated using the anomalous diffraction approximation method, and the kernel function of Mie scattering theory was substituted with these new kernel functions. Moreover, the Phillips–Twomey method was employed to solve the Fredholm integral equation of the first kind using aerosol optical thickness data from a CE-318 sun photometer. To verify the feasibility of the anomalous diffraction approximation method, experiments were carried out under sunny, dusty, windy and hazy weather conditions. These experiments showed that the extinction kernel function for non-spherical particles obtained using the anomalous diffraction approximation method is suitable for inversion of non-spherical dust particle size distributions under different weather conditions in the Yinchuan area.