Observations of ice water content are needed to evaluate the performance of models used for forecasting weather and climate change. Recent papers have demonstrated that dual-wavelength radar technique can reliably determine ice particle concentration, mean volume diameter, and ice water content in ice clouds. But the results still depend widely on the choice of the ice density function. As well as reflectivity ratio the Doppler velocity difference depends on particle size - Mie scattering reduces Z at the higher frequency; it also reduces the observed Doppler velocity, which is weighted by Z. Since, both reflectivity ratio and Doppler velocity difference depend differently on the density function, the relationship between the Doppler velocity difference and reflectivity ratio is dependent on the choice of the density function only. By comparing model output and measurements, we show that a best density function can be inferred.