Uncertainty of Zdr Calibration Techniques

Dual polarized radars promise to increase the accuracy of radar rainfall measurements. The copolar differential power measurement, Zdr, as well as the specific differential phase, phi_dp, contain additional information about the scattering medium that can be used to increase the accuracy of precipitation measurements that are based solely on radar reflectivity. However, to realize this benefit, Zdr must be calibrated to about one tenth of a dB.

One widely accepted way to calibrate Zdr is to point the radar dish vertically in light rain and measure Zdr while turning the dish 360 degrees. Since raindrops have no preferred orientation (i.e., distributed uniform randomly in the plane of polarization) intrinsic 360 degree integrated Zdr is zero dB. Thus, any measured non-zero dB Zdr yields the radar system Zdr offset. This technique works particularly well with radar systems that employ a single copolar receiver so that any receiver drift will not affect the calibration. However if separate horizontal (H) and vertical (V) receivers are employed, the temporal drift of the receivers likely needs to be monitored. There are additional problems with the vertical pointing method: 1) the radar can only be calibrated when there is precipitation over the radar, 2) the precipitation needs to be deep enough so that measurements can be made in the far field and so that measurements are not affected by transmitter transients.

NCAR has been tasked by National Weather Service (NWS) to evaluate Zdr calibration techniques for the NWS's WSR-88D network which will be converted to dual polarization within the next few years. The WSR-88D will use simultaneous horizontal and vertical polarization transmit mode with H and V channel receivers. Thus, differential receiver drift is likely to be an issue for Zdr calibration. Additionally, the WSR-88Ds do not point vertically (60 degree elevation angle maximum). Other techniques for Zdr calibration for the WSR-88Ds must then be used.

This paper, then, investigates two other methods for Zdr calibration:1) an “engineering” calibration technique and 2) a crosspolar power technique. The engineering calibration relies on passive power measurements of sun power and on measurement of injected test signals to establish the differential gains and losses of the passive sections of the signal path (wave guides, etc.). The crosspolar power method is a non-obtrusive technique (no injection of test signals or power meter measurements) that relies on radar reciprocity: crosspolar scattering amplitudes are equal (Svh=Shv). Both techniques also rely on injected test pulses to monitor any temporal differential drift of the H and V receivers. The crosspolar power technique has been used successfully on research radars such as S-Pol (NCAR's S-band polarimetric radar) and CSU-CHILL (Colorado State University's NSF supported S-band radar) both of which use fast alternating H and V transmit pulses to achieve dual polarization. The crosspolar power technique has not been tested on a simultaneous H and V transmit radar.

This paper investigates these two Zdr calibration techniques and describes the assumptions and the limitations for their practical implementation. Measurement uncertainty analysis is performed. Concepts are illustrated with data from S-Pol.