Radar parameters simulation for populations of spherical and non-spherical hydrometeors: dependence on size distributions, shapes and composition

Two codes based on the T-matrix approach have been modified. In addition to the single-scattering evaluations of C-ext, C-sca, C-abs, w (albedo), [cosq] (asymmetry parameter) and the non-zero elements of Mueller matrix at different scattering angles, a set of parameters which are even more useful to radar applications can now be estimated: backscattering cross-section, attenuation coefficient, effective reflectivity Z(e) (expressed in dBZ), IW C and LW C, and the depolarisation ratios (LDR and CDR) either for ice or water clouds. Hydrometeor model shapes can be spheres, prolate and oblate spheroids, conical-elongated particles and Chebyshev particles of even order. Computations are made at a frequency of 35 and 94 GHz. In the first part of this work scattering by cloud populations of randomly oriented polydisperse oblate and prolate spheroids is computed using the size distribution of Hansen and Travis, and monodisperse conical-elongated graupels, dry, soft with air inclusions and wet with water inclusions, at a frequency of 35 and 94 GHz. Using the above cloud models, some realistic scenarios representing vertical hydrometeors profiles have been built. In the second part three different theoretical size distributions have been selected: power law Hansen and Travis, Gamma modified and Log-normal. Reflectivity and attenuation have been computed for fixed water content wc, concentration n(0) and median radius R-m and the results for the different size distributions have been compared. Computations have been made for different types of cloud models (some of which with air or water inclusion in ice), assumed as statistically uniform ensembles for populations of randomly oriented polydisperse oblate and prolate spheroids, spheres and Chebyshev particles of even order, focalising at a frequency of 94 GHz. The particular study of output sensitivity on size distribution parameters has shown that effective reflectivity Z(e) is quite influenced by the type of size distributions, while shapes are less important. (C) 1999 Elsevier Science Ltd. All rights reserved.