Search

A 3D growth model is used to simulate pristine ice crystals, which are aggregated using a collection algorithm to create larger, multicrystal particles. The simulated crystals and aggregates have mass-versus-size and fractal properties that are consistent with field observations. The growth/collection model is used to generate a large database of snow particles, and the single-scattering properties of each particle are computed using the discrete dipole approximation to account for the nonspherical geometries of the particles. At 13.6 and 35.5 GHz, the bulk radar reflectivities of nonspherical snow particle polydispersions differ from those of more approximate spherical, homogeneous, ice-air particle polydispersions that have the same particle size distributions, although the reflectivities of the nonspherical particles are roughly approximated by polydispersions of spheres of 0.1-0.2 g cm−3 density. At higher microwave frequencies, such as 165.5 GHz, the bulk extinction (and scattering) coefficients of the nonspherical snow polydispersions are comparable to those of low-density spheres, but the asymmetry parameters of the nonspherical particles are substantially less than those of spheres for a broad range of assumed spherical particle densities. Because of differences in the asymmetry of scatter, simulated microwave-scattering depressions using nonspherical particles may well exceed those of spheres for snow layers with the same vertical water path. It may be concluded that, in precipitation remote sensing applications that draw upon input from radar and/or radiometer observations spanning a range of microwave frequencies, nonspherical snow particle models should be used to properly interpret the observations. [ABSTRACT FROM AUTHOR]