A New Algorithm for Automated Mesocyclone Detection

A well-developed intense mesocyclone appears as a pair of maximal and minimal patterns on Doppler radar. The WSR-88D system includes a typical detection algorithm for mesocyclones and tornadoes, which examines azimuthal wind shear in a Doppler radial velocity field (Stumpf et al., 1998). However, target mesocyclones are always associated with natural smaller-scale perturbations and larger-scale ambient flows, which often contaminate the patterns of mesocylones. Therefore, it is difficult to precisely estimate the rotation and divergence intensities from the radial velocity observational data. The purpose of this study is to develop a new algorithm for automated vortex detection of mesocyclones and tornadoes in a single Doppler radar. We suggest to use the two-dimensional continuous wavelet transforms as a preprocessor, to extract the particular spatial scale information from a wind field and to filter out the signal noise and largerscale flows. This algorithm consists of two parts. The first part is to detect the locations of mesocyclones and the second part is to estimate their vorticity and divergence. 2. NUMERICAL METHOD 2.1 Wavelet Transforms We apply a two-dimensional continuous wavelet transform to Doppler observational data instead of using original data to analyze vortex structures. A continuous wavelet transform is defined as follows.