The performance of the satellite communication systems employing polarization diversity or frequency-reuse
schemes to improve the spectral efficiency is degraded due to the depolarization-induced interference originated by
raindrops and ice particles present along the Earth-Space propagation path. Two models account for both rain and ice
contributions. One predicts the long-term Cumulative Distribution Function (CDF) of cross-polarization discrimination (XPD).
The other predicts the relationship between XPD and co-polar attenuation (CPA) and it was derived considering exclusively
data at V-band. In this paper, the former model is improved by considering the individual ice and rain contributions and
their combined effects, while the latter is validated against new measurements at Ka band. New models are then proposed
for the XPD-CPA relationships at Ka-band taking into account both rain and ice contributions and also their combined
effects. Finally, the predictions provided by the first model are usually converted to the corresponding XPD-CPA
relationship using the long-term first-order statistics of rain attenuation, (incorrectly) considering that the equiprobability
hypothesis applies. A new approach for the conversion of the XPD CDF into the corresponding XPD-CPA relationship is
presented.