July 22, 2021 - 2:00 pm
July 22, 2021 - 3:00 pm
CategoriesSchool of Meteorology (Defense)
School of Meteorology MS Thesis Defense
Quantifying and Mitigating Debris-Induced Bias in Radar Measurements of Tornadic Winds
Thursday, July 22nd
Join Google Meet:
Or dial: (US) +1 857-309-5338 PIN: 716 530 781#
The centrifuging of lofted tornadic debris is known to cause bias in Doppler radar measurements of tornado wind speeds. Debris presence in a radar volume is associated with anomalous radial divergence, underestimation of azimuthal wind speeds, and negative bias in vertical velocities, potentially resulting in erroneous interpretations of tornado structure. Using a simulation-based framework to study these errors, a variety of polarimetric radar time-series simulations from SimRadar are analyzed and compared in order to establish the relationships between debris field characteristics—such as debris size and number concentration—and the magnitude of bias in Doppler velocity and retrieved wind fields. Since debris characteristics also influence polarimetric measurements, we additionally seek to assess the relationships between velocity bias magnitude and relevant polarimetric variables. Establishing such relationships could support the development of a new moment-based approach to Doppler velocity bias correction for mobile research radars.
The latter half of this work introduces an alternative method for Doppler velocity bias mitigation utilizing novel spectral filtering techniques. Since debris is associated with unique polarimetric signatures as well as substantial velocity bias, this method incorporates dual-polarization spectral density (DPSD) estimation and fuzzy logic scatterer classification to identify debris-dominated signal contributions in a Doppler spectrum based on the velocity distribution of polarimetric characteristics. Outputs from the scatterer classification algorithm are used to suppress and filter the identified debris contributions within the original Doppler spectrum. New Doppler velocity estimates are recalculated from the filtered signals, and comparisons are made against both the original velocity estimate and the true Doppler velocity to evaluate the performance of these spectral filtering methods in reducing debris-related bias. In the future, these algorithms will be applied to observational data sets from mobile research radars.