Oversampling in azimuth to improve detection of vortices

WSR-88D radar data are collected within 1-degree azimuthal sectors in steps of 1 or 0.5 degree (so called standard and super resolution (SupRes)). The 1o sector equals the WSR-88D's beamwidth when the antenna is stationary. Because the lateral dimension of the radar resolution volume increases with range, the detectability of vortex signatures degrades and therefore determining vortex diameter and maximum rotational speed is impaired. The mesocyclone's and tornadic vortex signature's rotation speeds are estimated from the measurements of the maximal Doppler velocity. The value of this velocity depends on the relative position of the mesocyclones/TVS with respect to the radar beam center as well as the ratio of its diameter to the lateral dimension of the resolution volume. If the vortex radius to lateral dimension of resolution volume is larger than one, the maximum Doppler velocity is measured if beam center is pointed at the location of maximum tangential speed. Otherwise the position where the maximum Doppler is registered shifts further from the circulation center. Note that the estimated Doppler velocity is the key parameter forecasters use to quantify the strength of mesocyclones and TVSs. But either sampling at 1 or 0.5 deg in azimuth (as on the WSR-88D) would most likely miss the peak Doppler velocity. We propose a running average processing and presentation to capture the maximum speeds and other fine-detail weather features as follows. Consider radar that scans at α deg/s, has a PRT of Ts and obtains data from azimuthal swaths of αMTs which for the WSR-88D would equal 1 deg. In the SupRes mode the spectral moments are obtained from the same swath widths but swaths are overlapped 50% so that the data are presented at 0.5 deg. The closest spacing at which moments can be presented is αTs which corresponds to the shift in beam position that occurs during one PRT. Such presentation while processing the closest-spaced overlapped samples we call Finest Over Sampling (FOS). We demonstrate enhanced features of mesocyclones and TVSs in the FOS made on time series data obtained with a WSR-88D. Fields of radar moments obtained with SuperRes and FOS in tornadic storms are compared. It is shown that the gradients of all radar variables in mesocyclones and TVSs (for storms close to the radar) can be obtained more accurately in the FOS than in SuperRes. This makes the boundaries of vortices more pronounced. Moreover, the correlation coefficient between the returns at horizontal and vertical polarization exhibits a deeper minimum at the location of tornadoes in the case of the FOS processing. Advantages of FOSs in detection of distant tornadoes are also discussed.