Improving Polarimetric radar parameter estimates and Target Identification: a comparison of different approaches

Elimination of clutter contamination from weather radar signals is the objective of many techniques. Polarimetric radar data is extremely useful in hydrometeor classification as well as in the identification of clutter and biological targets . However at low SNR levels, polarimetric parameters are very noisy and target identification based on these parameters is no longer reliable. The merits of different approaches for clutter mitigation, target identification and improvement of the polarimetric radar parameter estimation at low SNR levels are explored. McGill University's S band weather radar data is used in this study. The approaches we have considered in this study are spectral analysis based target identification and clutter mitigation, Multilag correlation estimators for improvement of polarimetric parameters at low SNR levels and Autocorrelation Spectral Density based estimator for clutter mitigation. The question " could the information from these various approaches be combined to generate better estimates for polarimetric radar parameters and target identification " is addressed. New formulas had to be developed to properly correct for biases in RhoCO in low SNR as existing ones are biased. Multilag estimators appear to be more noisy than spectral analysis based estimators. The jury is still out as for the relative performances of ASD and spectral based techniques for clutter mitigation.