Using the MYRORSS Database to Create a Rotating Storms Climatology

Beginning in 2012, CIMMS and NSSL have been processing the entirety of the WSR-88D archive through the Multi-Radar, Multi-Sensor (MRMS) framework in order to generate a database of MRMS products for use in the development of climatologies for severe weather hazards and to help develop new warning guidance products for inclusion in the Forecasting a Continuum of Environmental Threats (FACETs). This dataset, called the Multi-Year Reanalysis of Remotely Sensed Storms (MYRORSS), from 1998 through 2011 with ongoing quality control efforts on the velocity-based products and a second round of processing beginning on more recent years.

The MRMS framework includes a linear least-squares derivatives (LLSD)-based algorithm that calculates the along- and cross-beam derivatives of radar variables. When applied to the Doppler velocity field, the cross-beam derivative provides an estimation of the rotational shear, which in the MRMS framework is named “AzShear.” The individual time steps of the AzShear product can be accumulated to provide a rotation track product, which shows the tracks of rotating storms.

Two major quality control problems exist within the rotation track product: 1) noise from poor velocity dealiasing, and 2) quasi-linear features, such as gust fronts, can oversaturate the signal from true rotational signatures. In order to reduce the noise, the dealiasing algorithm within the MYRORSS workflow was updated to the latest operational algorithm used on the WSR-88D network. Further, a multiple-hypothesis tracking (MHT) technique is employed when creating the rotation tracks in order to remove spurious AzShear features that are too small in size and/or magnitude, or not temporally coherent. However, the problem of the quasi-linear features remains. Since the MHT method depends upon the segmentation of the AzShear data, quantities such as aspect ratio are calculated for each cluster resulting from the segmentation. This presentation will explore removing features by the aspect ratio magnitude and the associated impacts from this on the resulting climatology.

This presentation will explore the challenges in generating a robust climatology of derivative products from a large WSR-88D dataset. The AzShear features will be compared to operational Mesocyclone Detection Algorithm (MDA) saved within the Severe Weather Data Inventory (SWDI) and stored MDA output at CIMMS/NSSL to determine the best method of processing to retain supercells. Initial findings of this climatology will also be discussed.