The fact that the difference in the two-way phase change between horizontally and vertically polarized waves (ΦDP) increases with distance through rain is another basis for polarimetric-based rainfall estimation (Seliga and Bringi 1978; Mueller 1984; Sachidananda and Zrnić 1987). As discussed by Zrnić and Ryzhkov (1996), use of specific differential phase shift (KDP) has many practical advantages. It is less susceptible to variations in DSD, does not depend on absolute radar calibration, is not affected by attenuation, is immune to beam blockage effects, and can be used to detect anomalous propagation and mixed phase conditions. However, as pointed out by Chandrasekar et al. (1990) and Keenan et al. (1997), it is susceptible to the assumed form of the relation between the raindrop axial ratio and the equivalent diameter. Bringi et al. (1978) and Scarchilli et al. (1993) also propose that ΦDP be employed for attenuation correction in rainfall estimation studies.
microwaves and radar by giridhar 16
Download File: https://miimms.com/2vFgM8
Not suprisingly, with the advent of polarimetric radars that can obtain KDP estimates this latter approach to rainfall estimation has been actively pursued in recent years. Goddard and Cherry (1984), Blackman and Illingworth (1995), Ryzhkov and Zrnić (1995a), Ryzhkov and Zrnić (1996), and others have presented rainfall estimation results based on the use of KDP that are extremely promising.
The purpose of this paper is to further explore the use of polarimetric and conventional radar rainfall estimation techniques in a high rainfall tropical regime using the C-band (5-cm wavelength) polarimetric (C-POL) radar described by Keenan et al. (1998). To date, with the exception of European-based work (e.g., Meischner et al. 1991), most polarimetric radars have undertaken rainfall measurements at 10-cm wavelengths (S band). Compared to S-band platforms, C-band radars have some advantages. A narrow beam is achieved with a smaller dish, which in turn means less wind loading and hence less demands on the system design (i.e., overall the system is less expensive and is more easily made portable). Furthermore, the propagation differential phase ΦDP is approximately inversely proportional to wavelength, so that random statistical errors of KDP are less for a given rain rate and hence more accurate measurements can be made at lower rain rates. However the cost is the increase in attenuation and the potential that resonance-scattering effects could distort the measurements, especially in tropical regimes where large raindrops are known to occur (Keenan et al. 1998).
Polarimetric rainfall estimation techniques based solely on KDP are essentially independent of attenuation effects. However, estimation of KDP is affected by resonance-scattering effects, specifically through the introduction of the differential backscatter phase shift (δ), which must be removed. The technical aspects of the derivation of the KDP estimate are therefore important, particularly at higher radar frequencies.
C-POL is a 1 beamwidth C-band polarimetric/Doppler radar capable of transmitting and receiving linear horizontal (H) and vertical (V) polarizations (Keenan et al. 1998). Using the algorithms of Zahrai and Zrnić (1993), the available polarimetric variables include ZDR and ΦDP, the correlation between H and V at zero lag [ρHV(0)], as well as the mean Doppler velocity and spectral width. The number of samples for extracting the polarimetric variables was 128 (64H, 64V interleaved) at a range spacing of 300 m with an azimuthal sampling of 1.0. The pulse width is 1 μs and the pulse repetition frequency is 1000 Hz. C-POL does not employ a radome.
To gain a first-order estimate of the degree to which beam blockage was affecting the radar measurements of ZH, a composite of clear-air radar signatures was undertaken. It is assumed here that without blockage the average clear-air signature should be relatively uniform as a function of azimuth over the islands. There is a small decrease associated with loss in sensitivity with range. This is because we are dealing with a small azimuthal span so that variations in the orientation of insects with wind direction should be small. This is supported by the relatively small variation of reflectivity as a function of azimuth at the third elevation. The resulting composites of the clear-air ZH signature have been used in assessing the azimuthal signal variation and the implied blockage effects are summarized in Fig. 2. Clutter associated with the hills to the north, northeast, and directly west of the radar is clearly evident. The lowest two tilt angles show three significant blockages over the gauge network. At the 1.0 (1.9) tilt the maximum signal loss is up to 5 (4) dB directly north of the radar. For the 3.4 tilt angle the maximum loss is about 2 dB. The maximum signal at the upper tilt is reduced because of the vertical gradient of the (clear air) reflectivity. Note that the elevations used for the blockage calculation are not exactly those used for the storm conditions as the tilts were modified to sample the whole storm in a timely manner.
For this study the radar-derived rain rates were made by first constructing an average of the KDP and ZH over five consecutive range gates (300 m apart) and five consecutive azimuth rays (a total of 25 values) centered over each rain gauge. The resulting average is a radial average of 2.3 km for the KDP estimates and 1.5 km for the reflectivity-based estimates. At the range of the gauges the azimuth average represents a spatial scale of about 2.2 km. As each KDP estimate is an average over 1.8 km, the five sets of radial KDP information are not independent. The areally averaged radar rainfall estimate is almost matched with the gauge network (one gauge per 5 km2).
This paper has examined the use of a polarimetric C-band (5-cm wavelength) radar for rainfall measurements with the primary focus on KDP- and Z-based estimators. The use of polarimetric measurements have considerable potential advantages compared with conventional radar observations. In particular, the phase-based estimators are immune to beam blockage and attenuation effects (Ryzhkov and Zrnić 1996) and simulations suggest that the polarimetric variables are less sensitive to drop size distribution (DSD) effects (Keenan et al. 1997). C-band systems offer some advantages over polarimetric S-band radars as they have smaller antennas for a given angular resolution and the differential propagation phase is approximately twice that observed at S band. However, there are significantly more questions resulting from important resonance-scattering effects (Keenan et al. 1997). A new consensus-based KDP estimator has been used in these analyses since the sensitivity to resonance scatter effects occasionally produce significant backscatter phase discontinuities.
The R(Z) and R(Z, B) estimates were all heavily biased. However, when a polarimetric-based attenuation correction was applied, good agreement was obtained at the upper two elevations (2.2 and 3.9 for two cases;1.5 and 2.7 for the other two cases). Beam blockage effects in the lowest tilt had overwhelming consequences, although there may also have been some microphysical influence on the chronic underestimation of rain rate. The rms differences between radar and gauges for the R(Z, A, B) are larger than those for the R(KDP) estimates by about 50%; a result that is also reflected in lower cross correlations. Variations from case to case associated with variations in mean DSD characteristics were more pronounced in the Z-based estimates.
Schematic of the C-POL radar system. The transmitted polarization is achieved through use of the ferrite switch S1. Circulators CR1 and CR2 switch energy through either the transmit or receive channels. Switch S2 selects the polarization mode processed by the receiver.
The development of the first Australian C-band polarimetric/Doppler meteorological radar system (C-POL) is described. Motivated by the need to obtain improved rainfall estimation and the vertical profile of hydrometeors, C-POL was developed jointly by the Bureau of Meteorology (BOM), the Commonwealth Scientific and Industrial Research Organisation of Australia, and the National Center for Atmospheric Research. C-POL is based on a standard operational C-band radar employed by the BOM but modified to be capable of transmitting linear horizontal and vertical polarizations and receiving the co- and cross polarizations on a pulse-to-pulse basis. Standard variables extracted include horizontal reflectivity (ZHH), radial velocity (Vr), spectral width (συ), differential reflectivity (ZDR), differential phase shift (ΦDP), and zero lag correlation coefficient [ρHV(0)]. With the addition of a second receiver chain, the linear depolarization ratio will soon be available. 2ff7e9595c
Comments