Examination of the New Mesocyclone Detection Algorithm Proposed for the WSR-88D Radar Network (99)

Brandon Smith, Cooperative Institute for Mesoscale Meteorological Studies / National Severe Storms Laboratory, Norman, OK

Matthew Mahalik, Cooperative Institute for Mesoscale Meteorological Studies / National Severe Storms Laboratory



The NSSL/ROC Tech Transfer MOU—Severe Storms Section, allows staff at the ROC and NSSL/CIMMS to work together to develop new thunderstorm-related algorithms and improve existing algorithms for the WSR-88D network. One of the current MOU projects is the development of a New Mesocyclone Detection Algorithm (NMDA) for the WSR-88D network that when completed, will be proposed to replace the current legacy MDA that was developed in the late 1990s. While the legacy MDA utilized raw radial velocity fields to find and analyze storm-scale circulations, the NMDA uses velocity-derived azimuthal shear (AzShear), in conjunction with velocity-derived shear diameter and half-velocity fields. AzShear is created using a linear least-squares derivative (LLSD) technique to compute velocity gradients across a radar beam for a neighborhood of gates. This produces a rotation estimation field that effectively mitigates a vast majority of the noise and extreme outliers that are present in raw radial velocity data, which can cause circulation identification errors in the legacy MDA. Although the internal workings of the legacy MDA are being replaced, the NMDA is maintaining the existing visual display and output attributes of the legacy MDA familiar to many NWS forecasters. This presentation will describe the methods used by the NMDA for both the 2D and 3D detection of storm-scale circulations, as well as those used to track and derive trend information for each identified circulation. A comparison of performance will also be shown between the NMDA and the legacy MDA across a multitude of severe weather event types, providing a first look of this newly developed WSR-88D algorithm.