Improved observations of ice hydrometeors can lead to better weather predictions and understanding of the hydrological cycle. Global coverage is best achieved by satellites, using active and passive microwave remote sensing due to the inherent penetration capability and sensitivity to the snow and ice cloud particles which scatter and absorb the radiation. Snow and cirrus cloud particles are to a large degree composed of aggregates. While it is known that the shape of these aggregates influence microwave measurements to various degrees, it is currently not fully clear to what extent certain particle features are of importance. For example, of what importance is the shape of individual crystals in the aggregates? This work is an attempt at improving our knowledge on the impact of ice aggregate microphysics on microwave and sub-millimetre scattering properties from a modelling point of view. A large amount of aggregates (roughly 4000) where modelled through several semi-physical stochastic simulations. The aggregates are composed of hexagonal ice crystals of varying axis ratio, ranging from 1/15 (plates) to 15 (columns), and assumed to be oriented in the horizontal plane. Single scattering properties of over 1000 aggregates were then assessed for zenith/nadir observations, using the discrete dipole approximation (DDA) at three typical radar bands (13.4, 35.6, 94.1 GHz) and three passive microwave frequencies (183.3, 325.15 and 664 GHz). An analysis on the sensitivity of these scattering data to various aggregate parameters is presented. In general, extinction was found to be less sensitive to shape than back-scattering at investigated at frequencies. Extinction at 664 GHz in particular was found to be shape insensitive; promising for the sake of the Ice Cloud Imager (ICI) on the upcoming Metop-SG satellite. In contrast, evaluation of triple frequency signatures showed relatively high shape sensitivity; of relevance to future multi-frequency radars.