A comparison of the microphysical properties of ice clouds, using lidar and radar data, is made for three sites: Cabauw (Netherlands), Atmospheric Radiation Measurement Program Southern Great Plains (ARM-SGP) site (United States), and Chilbolton (United Kingdom). The effective particle size (Reff), extinction, and ice water content (IWC) are derived and correlated to each other, temperature, radar reflectivity, and depth into the cloud from cloud top (ΔZt). There is no indication for large seasonal differences of the ice microphysical properties; however, the Reff differences observed at the ARM-SGP site are of the same magnitude as the error. The Chilbolton and Cabauw sites exhibit similar behavior in all cases while the ARM site shows large differences for some relationships, e.g., Reff(T, IWC). Within the sensitivity studies performed, it is not possible to construct a single Reff(T, IWC) parameterization valid at all three sites, and therefore it is not applicable in global models. It is possible to construct a single parameterization of ice water content related to temperature or to radar reflectivity. In all cases, an ice habit and particle size distribution assumption has to be made, resulting in different fits for different habits. When Reff is correlated to ΔZt for different classes of total cloud thicknesses (H), one can define a single parameterization, using parabolic descriptions, valid at the three sites and possibly on a global scale.