The predictive quality of an ensemble model of cirrus ice crystals to model passive and active measurements of ice cloud, from the ultraviolet (UV) to the microwave, is tested. The ensemble model predicts m ∝ D², where D is the maximum dimension of the ice crystal, and m is its mass. This predicted m‐D relationship is applied to a moment estimation parametrization of the particle size distribution (PSD), to estimate the PSD shape, given ice water content (IWC) and in‐cloud temperature. The same microphysics is applied across the electromagnetic spectrum to model UV, infrared, microwave and radar observations. The short‐wave measurements consist of airborne UV backscatter lidar (light detection and ranging) estimates of the volume extinction coefficient, total solar optical depth, and space‐based multi‐directional spherical albedo retrievals, at 0.865 µm, between the scattering angles 85° and 125°. The airborne long‐wave measurements consist of high‐resolution interferometer upwelling brightness temperatures, obtained between the wavelengths of about 3.45 µm and 4.1 µm, and 8.0 µm to 12.0 µm. The low‐frequency measurements consist of ground‐based Chilbolton 35 GHz radar reflectivity measurements and space‐based upwelling 190 GHz brightness temperature measurements. The predictive quality of the ensemble model is demonstrated to be generally within the experimental uncertainty of the lidar backscatter estimates of the volume extinction coefficient and total solar optical depth. The ensemble model prediction of the high‐resolution brightness temperature measurements is generally within ±2 K and ±1 K at solar and infrared wavelengths, respectively. The 35 GHz radar reflectivity and 190 GHz brightness temperatures are generally simulated to within ±2 dBZ_e, and ±2 K, respectively. The directional spherical albedo observations suggest that the scattering phase function of the most randomized ensemble model gives the best fit to the measurements (generally within ±3%). This article demonstrates that the ensemble model, assuming the same microphysics, is physically consistent across the electromagnetic spectrum.

We calculated the scattering and absorption properties of randomly oriented hexagonal ice columns using T-matrix theory, employing analytic orientation averaging, and the finite-difference time-domain method, which uses a numerical procedure to simulate random orientation. The total optical properties calculated are the extinction efficiency, absorption efficiency, single-scattering albedo, and the asymmetry parameter. The optical properties are calculated at the wavelengths of 0.66, 8.5, and 12 µm, up to a size parameter of 20 at 0.66 µm and 15 at the two other wavelengths. The phase-matrix elements P11, P12, and P22 are also calculated and compared, up to a size parameter of 20 at 0.66 µm and 15 at 12.0 µm. The scattering and absorption solutions obtained from the two independent electromagnetic methods are compared and contrasted, as well as the central processing unit time and memory load for each size parameter. It is found that the total optical properties calculated by the two methods are well within 3% of each other for all three wavelengths and size parameters. In terms of the phase-matrix elements it is found that there are some differences between the T-matrix and the finite-difference time-domain methods appearing in all three elements. Differences between the two methods for the P11 element are seen particularly at scattering angles from approximately 120° to 180°; and at the scattering angle of 180°, relative differences are less than 16%. At scattering angles less than 100°, agreement is generally within a few percent. Similar results are also found for the P12 and P22 elements of the phase matrix. The validity of approximating randomly oriented hexagonal ice columns by randomly oriented equal surface area circular cylinders is also investigated in terms of the linear depolarization ratio.