双偏振雷达差分传播相移的五种滤波方法对比分析

Comparative Analysis of Five Filtering Methods of Differential Propagation Phase Shift for Polarization Radar Subtitle as Needed

为了对双线偏振雷达的差分传播相移ΦDP进行滤波降噪处理,提高双线偏振雷达识别粒子、估测降水的能力。通过对C波段双线偏振雷达差分传播相移分别采用滑动平均、中值滤波、FIR滤波、卡尔曼滤波和小波分析5种不同方法地介绍,利用中国气象科学研究院灾害天气国家重点实验室移动式C波段双线偏振雷达(POLC)2013年外场试验观测资料,对比分析了经5种方法处理后的差分传播相移ΦDP径向数据,及其分别在降水估测中应用的效果。结果表明:采用FIR滤波方法、卡尔曼滤波方法和小波分析方法对ΦDP径向数据进行处理的效果比传统的滑动平均和中值滤波方法好;其中小波分析处理的ΦDP径向数据波动最小,再经过变距离法拟合的KDP数据的负值能够得到最有效地抑制,最后应用到R(KDP)降水估测中的精度最高。

In this paper,five different filtering methods of differential propagation phase shift for C-band Polarization Radar are introduced. Based on the observation in field experiments by a C-band Polarization Doppler Radar on wheel which was built in the State Key Laboratory of Severe Weather,Chinese Academy of Meteorological Sciences（ CAMS）,running average method,median average method,finite-impulse response filter（ FIR）,Kalman filter,and wavelet analysis are respectively used to filter radial data of differential propagation phase shift ΦDP. In addition,the biases are compared between the specific differential propagation phase shift KDP-based rainfall estimation which fitted from the ΦDPradial profile after filtered with these methods. The results show that the complicate FIR,Kalman,wavelet methods have better filtering effect than traditional running filer and median filter. In terms of wavelet analysis method,the data fluctuation of ΦDPradial profile is the minimum; the negative value of KDPfitted by ΦDPis the least; the accuracy of KDP-based rainfall estimation is the highest.