Bootstrap Dual-Polarimetric Spectral Density Estimator

Weather radar variables provide useful information about the characteristics and motion of hydrometeors. However, the bulk information may be masked, when the meteorological signal of interest is contaminated by clutter. The dual-polarimetric spectral densities (DPSDs) may unveil additional information about the polarimetric characteristics of the groups of scatterers moving at different Doppler velocities in a given radar resolution volume. Previous DPSD estimation methods required averaging a large number of spectra (obtained from different spatial locations or times), or averaging in frequency to get accurate estimates; though by doing so, the resolution is degraded, and the important features of the meteorological phenomenon may be masked. In an attempt to overcome these limitations, the Bootstrap DPSD estimator is proposed, which allows the estimation of DPSDs from a single dwell with minimal spatial, temporal, or spectral resolution loss. The performance and the limitations of the Bootstrap and conventional DPSD estimators are assessed when identifying signals with different polarimetric signatures of scatterers moving at different radial velocities in the radar volume. The advantages of the Bootstrap DPSD estimator as a tool for the polarimetric spectral analysis are demonstrated with a few examples of polarimetric spectral signatures in data from tornado cases. It is expected that, with the Bootstrap DPSD and the polarimetric spectral analysis, it will be possible to better understand tornado dynamics and their connection to weather radar measurements, as well as to elucidate important scientific questions that motivated this paper.