A large data set of volume element models of aggregate snowflakes was created, building the snowflakes from various models of ice crystals found in the atmosphere: dendrites, needles, plates, and bullet rosettes, as well as spheroidal crystals for comparison. Several different sizes for the constituent crystals were also used. The radar backscattering cross sections of the snowflakes were computed from the models using the discrete dipole approximation (DDA) at 13.6 GHz (K_u band), 35.6 GHz (K_a band) and 94.0 GHz (W band), and the effects of the choice of crystal model and size on the K_u/K_a band and K_a/W band dual‐wavelength ratios (DWR) was investigated. It was found that the aggregate DWRs were very similar for all naturally occurring ice crystal types investigated in this study.

The discrete dipole approximation was then used to compute the radar backscattering properties of the snowflakes at frequencies of 9.7, 13.6, 35.6, and 94 GHz. In two of the three growth scenarios, the rimed snowflakes exhibit large differences between the backscattering cross sections of the detailed three‐dimensional models and the equivalent homogeneous spheroidal models, similarly to earlier results for unrimed snowflakes. When three frequencies are used simultaneously, riming appears to be detectable in a robust manner across all three scenarios. In spite of the differences in backscattering cross sections, the triple‐frequency signatures of heavily rimed particles resemble those of the homogeneous spheroids, thus explaining earlier observational results that were compatible with such spheroids.