SEARCH

SEARCH BY CITATION

In this paper we study the behavior of backscattering quantities that are of primary importance for the interpretation of millimeter-wavelength multiparameter radar observations (i.e., the backscattering cross sections at horizontal and vertical polarization, the differential reflectivity, and the linear and circular depolarization ratios) for rotationally symmetric nonspherical particles having a conical-elongated form that can be continuously varied. The scattering calculations at 35.8 GHz have been carried out using a computer code based on the Extended Boundary Condition Method for liquid water and soft-ice hydrometeors in the equal-volume sphere size ranges 0.05 ≤ rev ≤ 0.4 cm and 0.1 ≤ rev ≤ 1.0 cm, respectively. The calculated backscattering quantities are examined for the effects of four factors: particle size, conicity, elongation, and orientation with respect to the incident field (i.e., radar elevation and particle canting). Results of this sensitivity study show that the effects of nonsphericity tend to increase as size increases; for hydrometeors smaller than rev ≈ 0.4–0.5 cm, however, oscillations dominate the behavior of the considered backscattering quantities with size. Below such sizes, particle elongation and radar elevation are by far the main factors affecting the backscattering. As size increases, however, particle conicity becomes more and more important; at the largest sizes it is comparable to particle elongation. Our results also show that depending on particle shape, even limited variations in radar elevation or in particle orientation due to hydrometeor canting may significantly affect millimeter radar measurements of precipitating clouds. In any case, generally there is a considerable (often, large) sensitivity to particle parameters; in our view this augurs well for the existence of backscattering “signatures” that could be used to characterize the observed hydrometeors. On the other hand, it also implies that interpretation of millimeter radar measurements should be based on a statistical approach, rather than on a deterministic one, due to the variability of hydrometeor shape and orientation in nature and to their imperfect a priori information.