TORNADOSYMP Tornado Vortex and Wind Speed Detection by X-Band Phased Array Radar

The Phased Array Weather Radar (PAWR) network, e.g. the Shanghai PAWR network, usually consists of 3 to 7 X-band phased array radars. Each radar achieves fast scanning by using digital multi-beam technology and a fan-thin beaming method to finish a volume scan within 30 s, with a radial resolution of 30 m and a max-range of 45 km. The PAWR network has shown advantages in detecting the structure, evolution, and dynamics of local severe storms in better detail than the operational S-band radar (WSR-88D). By using a unique group synchronization scanning strategy, it can effectively reduce the observation time difference between adjacent subarrays to 5 seconds and ensure the accuracy of the retrieved 3D Doppler wind field, especially for fast moving small vortices.

Tropical cyclone tornadoes are a major type of tornado in China, e.g., over 50% of tornadoes during 2005-2016 in South China (Foshan) were spawned in a TC environment. However, compared to classic supercell tornadoes, TC related tornadoes often exhibit the characteristics of mini-supercells. The horizontal scale of mesocyclones in mini-supercells is only 2-4 km or even smaller, with vorticity limited to below 4 km above sea level, which are rarely detect by S-band radar. A mini-supercell with a tornado in Shanghai was detected by using PWAR but missed by the S-band WSR-88D on Jul 6, 2020. The positive and negative radial velocity pair structure and evolution process revealed that the vortex originated near the ground and then reached up to 2.0 km. Only 2-3 minutes before the tornado touched down, the contracted and stretched tornado vortex dropped rapidly from as high as 1.5-2.0 km to near surface.

A new algorithm for micro-scale vortex detection has been developed using high spacial-temporal PWAR data and 3-D retrieved wind data. Compared to the WSR-88D Mesocyclone Detection Algorithm (MDA), this algorithm is not only specifically configured with the horizontal and vertical thresholds of vortex scale for the detection of smaller scale vortices, but also designed to avoid reducing the performance of vortex detection due to the excessive number of noise points in high resolution PWAR data.

The main improvement strategies include: 1) improving one-dimensional vector recognition based on noise containment in 1D vector search; 2) improving the calculation strategy of multi-threshold layer-by-layer filtering in 2D feature filtering; and 3) using an optimal similarity screening technique based on the tree structure in 3D feature clustering. Compared to traditional algorithms, this algorithm can improve the capture ability of small-scale vortex detection in high spatial resolution data and the continuity of vortex tracking in high temporal resolution.

This algorithm can identify the maximum inbound / outbound wind speeds, core diameter, azimuth shear, and vorticity of micro-scale vortices. These parameters can be used to characterize rotational properties and strength of micro-scale systems (<2km), especially in the low levels, which is very helpful for predicting the possibility of a tornado in advance. Three mini-supercells embedded in the outside rain bands of Typhoon ‘Chaba’ were detected by the Guangdong Foshan PWAR network on Jul. 4, 2022. According to the detected mesocyclones (15:07-15:57) and TVSs (15:23-15:39), possible lead time of the 15:32 tornado could be 9-15 minutes.