Abstract
Bistatic dual-polarization radar parameters at S- and C-band frequencies are simulated for rain and hail. The goal is to determine their potential for discriminating the two precipitation types and for estimating the parameters of an exponential size distribution for hail. Raindrops and hailstones are modeled as oblate spheroids with canting distributions representing their fall behavior. Three hailstone composition models are used to illustrate the effects of melting. Most of the bistatic radar parameters are significantly affected by the amount of liquid water in the hailstones, which may prove useful in determining the melting level from the vertical profiles of these parameters. For single-polarized transmission, such as vertical (v) or horizontal (h) polarization, the four bistatic radar parameters of interest are effective reflectivity factor (Zv or Zh), bistatic-to-backscattering reflectivity ratio (BBRv or BBRh), linear depolarization ratio (LDRv or LDRh), and magnitude of the correlation coefficient between the co- and cross-polarized signals (ρv or ρh). If the transmission is dual polarized, then in addition to these two sets of parameters, the bistatic differential reflectivity (ZDR) and the magnitude of the copolarized correlation coefficient (ρhv) will be available. For low elevation angles of the transmitter and receiver the parameters resulting from h-polarized transmission may be difficult to measure near the bistatic azimuth angle of 90° due to very low signal levels. This may not be an issue for precipitation involving large hailstones.
When parameter pairs such as (LDRv, ρv) and (BBRv, Zv) are plotted, it is observed that rain and hail tend to cluster in different regions on these planes. This indicates a potential for using bistatic radar parameters for differentiating rain from hail. Similar pairs are possible for h-polarization. Various other combinations of these parameters lead to similar results. The use of more than one pair of parameters and/or several bistatic receiver locations should enhance the level of confidence in the discrimination process. It should also be noted that in some cases there are regions on these planes where rain and hail overlap and discrimination may not always be possible.
Other than Zv and Zh, all of the bistatic radar parameters mentioned above are in the form of ratios. As a result, given an exponential size distribution, N0 exp(−3.67D/D0), they depend only on the median volume diameter D0 and not on N0. Assuming that the amount of liquid water and ice in the composition of the hailstones are known, the ratio parameters may be used for estimating D0. However, among these parameters only BBRv and BBRh are negligibly affected by variations in the axial ratio and the mean orientation of hailstones, making them preferable for D0 estimation. Once D0 is obtained, N0 may be estimated using Zv or Zh.
Corresponding author address: Dr. Kultegin Aydin, Department of Electrical Engineering, The Pennsylvania State University, 314 Electrical Engineering East, University Park, PA 16802.
Email: k-aydin@psu.edu
