Geoscience Reference
In-Depth Information
4. Near term enhancements
Currently a significant transformation of the radars is ongoing; it is addition of dual
polarization (Zrnic et al., 2008). By mid 2013 all radars on the network should have this
capability. Although Doppler capability is not a prerequisite for dual polarization, the
coherency of transmit-receive signals within one PRT is for differential phase measurement.
Dual polarization offers ample possibilities for application of spectral analysis to
polarimetric signals and these are being explored (e.g., to discriminate between insects and
birds, Bachman & Zrnic, 2007; to suppress ground clutter, Unal, 2009; or to achieve adaptive
clutter and noise suppression, Moisseev & Chandrasekar, 2009).
Three improvements approved for soon inclusion on the network are pending. These are
staggered PRT, processing of range oversampled signals, and adaptive recognition and
filtering of ground clutter. Brief description follows.
4.1 Staggered PRT
It is planned for mitigating range velocity ambiguities at mid elevation angles with possible
use at the lower elevations. The scheme consists of alternating interval between transmitted
pulses (Fig. 10) and estimating arguments of two autocorrelations at the two lags, arg[ R ( T 1 )]
and arg[ R ( T 2 )]. The velocities estimated from these arguments have a different
unambiguous interval (each inversely proportional to the corresponding separation T i , i=1
or 2) as can be deduced from eq. (8). Therefore the difference of the velocities uniquely tags
the proper unambiguous interval for either PRT so that correct dealiasing can be achieved
(Torres et al., 2004a) up to larger v a than possible with only one of these PRTs . For the
example in Fig. 10, v a = 3 v a2 = 2 v a1 . Consider T 1 =1 ms T 2 =1.5 ms which produces v a = 50 m s -1
(unambiguous interval is -50 to 50 m s -1 ) and unambiguous range of at least 150 km.
Fig. 10. Staggered PRT. The stagger ratio T 1 / T 2 = 2/3. The continuous curve depicts the
return from precipitation extending up to c T 2 /2 but not further (from Torres et al., 2009 and
adapted from Sachidananda & Zrnic, 2003).
Power estimates in range sections I, II, and III (Fig.10) are computed separately for the short
PRT and the long PRT to check if data censoring is needed. Comparison of powers in the
two PRT intervals indicates if there is overlay and how severe it is so that appropriate
censoring can be applied. In Fig. 11 contrasted are two fields of velocities obtained with two
radars (spaced about 20 km apart). The left field comes from the operational WSR-88D in
Oklahoma City and was obtained with the “batch mode” and parameters as indicated. On
the right is the same storm complex but obtained with staggered PRT on the research WSR-
88D radar in Norman OK some 20 km SSW from Oklahoma City. Highlighted in yellow
circles are regions where significant aliasing occurs on the operational radar (exemplified by
abrupt discontinuities in the field, change from red to green) but are absent in the field from
Search WWH ::




Custom Search