Traditional Mie theory, assuming spherical particles, is inadequate for characterizing the scattering of atmospheric non-spherical ice crystals. Existing studies are largely limited to single frequency (e.g., 94 GHz), lacking systematic quantification of key dual-polarization parameters across the millimeter/submillimeter wave spectrum, which constrains the accuracy of the polarimetric radar for meteorological target detection and classification.

This study aims to systematically investigate the dual-polarization scattering properties of six typical non-spherical ice crystals—hexagonal columns, plates, hollow columns, bullet rosettes, aggregates, and supercooled water droplets—across 35, 94, 140, and 220 GHz bands. It quantifies the responses of differential reflectivity (ZDR) and linear depolarization ratio (LDR) to particle shape and orientation, providing crucial theoretical support for wideband polarimetric radar meteorology.

Scattering models were developed using the Discrete Dipole Approximation (DDA) and Finite-Difference Time-Domain (FDTD) methods, cross-validated with commercial software (XFDTD, HFSS). Backscattering cross-sections, ZDR, and LDR were computed for different ice crystals across the frequency bands, analyzing the influence of particle size, geometry, and frequency.

1）The reliability of DDA was systematically validated across the 35–220 GHz range. Calculation errors for backscattering cross-sections were ≤1.5 dB for all particles except highly random aggregates. 2) Radar reflectivity factor showed a coupled wavelength dependence: small particles (equivalent radius <100 μm) were wavelength-insensitive (<1 dB difference), while large particles (>100 μm) exhibited significant shape-dependent resonance. The equivalent radius corresponding to resonance extrema increased with wavelength. 3) Characteristic ranges of ZDR and LDR for the six ice crystal types were quantified. Hexagonal plates showed the widest ZDR range (9 dB to –9 dB), while axisymmetric particles exhibited stable LDR values (–40 dB to –50 dB).

This wideband, multi-particle study addresses prior limitations in frequency coverage and parameter quantification. It demonstrates that the shape-sensitive ZDR and LDR parameters can reduce dependence on particle size distribution and significantly improve ice crystal identification accuracy, providing a key theoretical basis for millimeter/submillimeter wave polarimetric radar applications in cloud microphysics and meteorological target classification.