Dust aerosols play a crucial role in the global climate system and radiation budget. However, current aerosol radiative transfer models have two major limitations in microphysical characterization. First, the mass loading of coarse-mode large dust particles is severely underestimated. These models typically truncate the maximum particle size at 10 μm and omit the super-coarse particles that are abundant in actual observations. Second, particle shape is oversimplified by the widespread use of highly symmetric spherical or spheroidal assumptions, which fail to capture their true irregular morphology. These limitations cause large biases in radiative forcing calculations and systematic errors in polarized remote sensing retrievals. To solve this problem, this study quantifies the respective effects of dust particle shape and size range on shortwave radiation and polarization characteristics at the top and bottom of the atmosphere.
Two typical shortwave bands, namely 1.55 μm and 0.532 μm, are selected for sensitivity analysis. Regarding size distribution, this study uses airborne in-situ data from the Sahara Desert for comparison. It contrasts the traditional distribution with a 10 μm upper cutoff and the extended distribution with a 50 μm upper cutoff that covers coarse-mode particles. For shape characterization, comparisons are conducted among spherical, spheroidal, and tri-axial ellipsoid models based on globally measured statistical distributions. Furthermore, the proposed model couples the desert surface bidirectional reflectance function and the polarization distribution function. The model also uses the vector successive order of scattering method to compute the full-vector radiation field in the atmosphere-surface system, with multiple scattering included in calculations.
The results show that particle morphology and size exert an obvious joint regulatory effect on radiative transfer. The traditional spherical model has high geometric symmetry. It induces coherent interference of internal light waves and intense Mie scattering resonance. It also greatly overestimates polarization in the forward scattering region. In contrast, the tri-axial ellipsoid model breaks central symmetry. It takes advantage of the phase smoothing effect from random particle orientation. This model completely eliminates spurious resonance. Its polarization features are smooth and agree well with real atmospheric conditions. Neglecting coarse-mode particles will overestimate the single scattering albedo systematically. It hides the actual absorption of shortwave radiation by dust aerosols. In addition, this study carries out extreme sensitivity tests on the morphological parameters of the tri-axial ellipsoid. Extreme parameter perturbations only cause minor changes in radiation and polarization. The maximum relative deviation is less than 3.05%. Such deviations are much smaller than the errors caused by traditional simplified shape assumptions. This proves that the proposed baseline model has strong physical robustness.
In conclusion, particle non-sphericity and coarse-mode contributions to size distribution have significant effects on radiation simulations. Traditional shape assumptions or neglect of coarse-mode particles will introduce systematic biases in polarized remote sensing retrievals. Therefore, to improve dust aerosol climate assessment and satellite retrieval accuracy, radiative transfer models should use observation-based particle models and size distributions to accurately simulate radiative transfer processes in the real atmosphere. High-precision physical characterization brings a large computational overhead. Under the same computing conditions, the spherical model completes calculations within a few hours, while the tri-axial ellipsoid model with full-range optical properties takes nearly one month. Accordingly, future work shall integrate high-precision non-spherical scattering databases with lightweight surrogate models such as machine learning and high-dimensional look-up tables. This can overcome computational constraints and enable online operational application of aerosol high-precision multispectral radiative effects.