ANALYSIS OF THE COPOLAR CORRELATION-COEFFICIENT BETWEEN HORIZONTAL AND VERTICAL POLARIZATIONS

Recent research has suggested that the copolar correlation coefficient termed rho(hupsilon)(0) can be used to identify large hail and improve polarization estimates of rainfall. The typical measured values of rho(hupsilon)(0) at S band vary approximately between 0.8 and 1.0. For applications to hail identification, the required accuracy should be within +/-0.01, while for rainfall improvement a higher accuracy is necessary, for example, within +/-0.001. The statistics of the estimator of rho(hupsilon)(0) using the Gaussian spectrum approximation from both an analytical approach and using simulations are discussed. The standard deviation and bias in rho(hupsilon)(0) are computed as a function of number of samples, Doppler spectral width, and mean value of rho(hupsilon)(0). The effect of finite signal-to-noise ratio and phase noise are also studied using simulations. Several other estimators of rho(hupsilon)(0) are evaluated. Time series data collected with the Colorado State University-University of Chicago and Illinois State Water Survey (CSU-CHILL) radar are analyzed and compared with the simulations. Antenna pattern effects as they affect the accuracy of rho(hupsilon)(0) are also discussed.