Introducing random mixtures of irregular hexahedrals and spheroids to reproduce polarimetric lidar observations of desert dust

In this work, we present the use of random shape mixtures of irregular hexahedrals and spheroids to simulate the spectral dependence of lidar-derived dust depolarization ratio and lidar ratio in 3 wavelengths traditionally used in aerosol research: 355, 532 and 1064nm.

Vertically-resolved polarimetric remote sensing provides a comprehensive understanding of atmospheric dust properties and its’ effects on radiation, weather and climate. Optical property profiles derived from polarimetric lidar observations such as the lidar ratio (Sp) and the depolarization ratio (dp), are sensitive to the particles’ morphology. However, in order to extract particle microphysical properties from these observations, accurate modelling of dust scattering properties is required. Dust particles appear to be highly-irregular and while complex shape models have been developed to describe them (e.g. Gasteiger et al. 2011), the scattering calculations are expensive in terms of computational power, which limits their applicability.

Thus, in most cases dust particle shapes are modelled using simplified representations such as spheroids. Spheroid shape mixtures have been demonstrated to successfully reproduce the angular dependence of light scattering from dust aerosols, nevertheless deviations are observed, particularly close to backscattering angles (Dubovik et al., 2006). Recent developments show that irregular hexahedral ensembles can better reproduce the measured dust lidar-relevant properties (Saito & Yang, 2021; Saito et al., 2021), however it is still challenging to reproduce their spectral dependence.

In all cases, additional assumptions are made with respect to the distribution of the different particle shapes considered in the ensemble (i.e. a shape distribution).

Herein we explore a different pathway, using random shape mixtures of irregular hexahedrals and spheroids to simulate the spectral dependence of lidar-derived dust dp and Sp. Since the considered particle shapes are not realistic, we do not constrain the simulations with measured shape distributions, but rather allow the different particle shapes to vary randomly in the mixtures. For the simulations we utilize the MOPSMAP (Gasteiger & Wiegner, 2018) and the TAMUdust2020 (Saito & Yang, 2021; Saito et al., 2021) scattering databases.

The size distributions and complex refractive indices considered for the calculations, are provided by AERONET retrievals collocated with the lidar observations, and height-resolved airborne in-situ data, acquired during the ASKOS-ESA campaign, implemented in Mindelo, Cabo Verde (Marinou et al., 2023). The simulated d_p and S_p for dust particles are evaluated against multi-wavelength polarization lidar data from ASKOS and lidar-derived climatological values.

As an independent consistency check of the simulation results, the derived random spheroids/hexahedral mixtures are utilized in radiative transfer calculations to simulate multi-wavelength sky-radiances from AERONET almucantar sequences. More specifically, the sun-photometer observational geometry is considered for two cases: i) assuming the spheroid shape distribution used in AERONET (Dubovik et all, 2006) and ii) assuming the random spheroids/hexahedral mixture found to better reproduce the lidar data. The modelled sky radiances are then compared to the co-located sun-photometer measurements.  

First results show that the random hexahedral/spheroid shape mixtures can accurately reproduce the spectral dependence of d_p and S_p for a selected dust case study of ASKOS, while the results are also within the uncertainty of the corresponding climatological lidar data.