Dried sea salt aerosol is commonly represented in aerosol-optics models
by ideal cubical particles, while samples reveal that marine aerosol
particles frequently display distorted cubical shapes, and they can have
more or less rounded edges. In this study three types of non-ideal
cuboidal model geometries are investigated, namely, convex polyhedra,
Gaussian random cubes, and superellipsoids. Optical calculations were
performed at a wavelength of 532 nm using the discrete dipole
approximation and the T-matrix method. The main focus is on optical
properties relevant to lidar remote sensing, namely, the linear
depolarization ratio in the backscattering direction, and the
extinction-to-backscatter or lidar ratio. Gaussian random distortions
tend to increase the depolarization ratio in relation to that of perfect
cubes, while superellipsoids mimicking cubes with rounded edges
generally decrease the depolarization ratio. Convex polyhedra can
describe randomly distorted cubes. Their computed depolarisation ratios
display random fluctuations about those computed for ideal cubes. The
results suggest that Gaussian random cubes and superellipsoids are most
consistent with the observations if the geometries deviate only mildly
from that of an ideal cube. Gaussian random cubes that strongly diverge
from cubical shape pose a risk of overestimating both depolarization and
extinction-to-backscatter ratio. Superellipsoids that approach
octahedral shape yield unrealistically high depolarization ratios.
Investigation of size-averaged optical properties of superellipsoids
demonstrate that the presence of absorbing material in marine aerosols
can have a dramatic effect on the lidar ratio, and its effect on the
depolarization ratio can be of comparable magnitude as that caused by
rounding of edges.