Estimation of Polarization Errors from Covariance Matrices of CSU-CHILL Radar Data

The construction of covariance matrices from radar data requires that the the data be very well calibrated. Techniques are shown and issues are discussed involving the estimation of both the magnitude and phase of the covariances. Eigenpolarization analysis is applied to the matrices and the results are discussed. The CSU-CHILL radar ideally transmits alternately pure horizontal and vertical polarizations, but in actuality polarization errors exist that can affect the behavior of the covariances. A model is given that predicts what these effects are. Polarization errors are estimated from CSU-CHILL data and are shown to be consistent with the theoretical model. In general the polarization errors can be non orthogonal. The experimental covariance matrices are transformed via a non orthogonal polarization basis transformation thus correcting the data for the polarization errors. Results are shown to agree well with theory.