Search

Among the many ways investigated for radar estimation of rainfall, polarimetric methods are the most promising. However most polarimetric algorithms are degraded by attenuation by precipitation and clouds and by calibration error. A new method was recently proposed in which the differential polarimetric attenuation is used to perform an accurate rain rate measurement. The method is independent of the radar calibration and of the attenuation by undetected clouds. This algorithm is also usable as a qualitative hail detector, as well as a detector of anomalous propagation. The goal of the paper is to describe the results of the first experimental implementation of this method using the 35 GHz RABELAIS radar, as attenuated radar, and the 3 GHz CAMRa radar as a reference. We show that the proposed algorithm is stable and enables us to retrieve the actual rain rate even from an observed signal attenuated by more than 30 dB. The results are insensitive to the value used for the power coefficient of the Z(R) relation.

The Olympus program was designed to develop new telecommunications technology in space and develop new market applications. This study analyzes attenuation data from two of the Olympus satellite beacons, those operating at 19.77 and 12.502 GHz. It is shown that the difference in the two attenuations provides estimates of rainfall rate that are in close agreement with rain gauge and distrometer measurements at the ground receiver site and along the propagation path. Modeling demonstrates that the attenuation difference using 19.77 and 12.502 GHz is a useful estimator of rainfall because it is largely independent of drop shape and size distribution, making it linearly related to rainfall rate. Attenuation along the propagation path due to the melting layer or ice introduces inaccuracies into estimates of rainfall. However, this difficulty would be eliminated if the method were to be used on terrestrial line-of-site links. Thus attenuation difference provides a means of improving the quality of rainfall estimates when used together with gauges and radar observations.

Whereas different hydrometeor types cannot be readily distinguished using radar reflectivity alone, this paper shows how the differential reflectivity available from dual-polarization radars can be used as an indicator of hydrometeor type (when supported by reflectivity factor also), and that the spatial variability of differential reflectivity is a useful addition. Various hydrometeor types, ground targets, and burnt chaff from straw fires may be identified, whereas conventional radar reflectivity measurement alone would require considerable pattern recognition to make the same distinctions.

Dual-polarization radar measurements of rain (using horizontal and vertical polarizations with copolar transmit and receive) are compared with simultaneous measurements using a Joss-Waldvogel distrometer and a rapid response rain gauge. The radar pulse volume was situated 200 m above the ground-based instruments, at a range of 8 km. The parameters compared are absolute and differential radar reflectivity factors and rainfall rate. A correlation of 0.95 was found between the radar and distrometer measurements of absolute and differential reflectivity factor, with the radar estimate of absolute reflectivity exceeding the distrometer estimate by 1.6 dB on average, and the distrometer estimate of differential reflectivity exceeding the radar estimate (assuming Pruppacher and Pitter drop shapes) by 0.1 dB on average. The latter difference is small but significant and is possibly attributable to drop canting or oscillation. A new relationship between drop shape and drop size is proposed, reducing the mean difference in differential reflectivity to ±0.02 dB. Radar estimates of rainfall rate, assuming an exponential drop-size distribution, tend to produce significant overestimates compared with the direct measurements of the two ground-based devices. This is shown to be largely attributable to departure from a strictly exponential distribution of drop sizes, and a correction factor is derived to compensate for this.

An experiment is described where rain at heights well above ground level was observed simultaneously by an airborne distrometer and a dual-polarization radar. The parameters of the distribution of drop sizes and the rainfall rate were computed from both sets of observations assuming the distribution to have a negative exponential form. The parameters derived from the two sources of data are compared, and differences discussed.