The Study Of Quantitative Precipitation Estimation From X-band Polarimetric Doppler Radar

Polarimetric Doppler Radar is a new generation radar, which can transmit and receive the horizontal and vertical polarized waves. Compared with the other ordinary radars, this kind of radar can observe more parameters from cloud and rain particles. And it has a bright application prospect.High time resolution X-Band Polarimetric Doppler Radar data from the measurements of Jilin Province convective precipitation were analyzed to study the algorithms for retrieving rain rate and the potential for quantitative precipitation estimation of this kind of radar. And enhance the universality of the empirical formulas of rain rate measurements.First of all, we introduce the basic composition and main performance parameters of this kind of Polarimetric Doppler Radar. The definitions and physical meanings of the various parameters are also presented. These introduces provide the theoretical basis of quantitative precipitation estimation in order to understand the characteristics of all the polarimetric parameters.Second, on the basis of Mueller algorithms, this paper gives a preliminary explains of the formation mechanism of differential phase shiftΦDP. And also discusses where the noises ofΦDP come from. These are the theoretical foundations of the radar data preprocessing. Introducing from the physical properties of the polarized waves, it is proved that differential phase shift ODP is composed by differential propagation phase shiftΦDP and differential backscatter phase shiftδ. Because of the significantδeffect during the observation,δshould be eliminated as the noise in theΦDP signals.Then, we analyze the case of convective precipitation and process the radar data profoundly. In this paper, we apply the Kalman filtering method to eliminate the random noises ofΦDP signals. After filtering, the signals are more smoothing, and keep the original features as well. As a conclusion, Kalman filtering method can eliminate theδinΦDP signals efficiently. Meanwhile, this paper makes use of the filteredΦDP to correct ZH and ZDR. After corrected, ZH and ZDR enhance obviously.According to the multivariate linear regression equation,5 empirical formulas are fitted. With 3 different gauges data of different range from the radar, the effect of precipitation estimation of the 5 methods is tested and the errors are analyzed. The results show that, R(KDP,Zh) is much better than any other four methods. The average SD is better than 14%.In the end, this paper gives comprehensive conclusions of the study. And the future study direction is also discussed.