Improvement of the WSR-88D Mesocyclone Algorithm

Abstract

The build 9 Weather Surveillance Radar-1988/Doppler mesocyclone algorithm (B9MA) is designed to locate mesocyclones [rotating thunderstorm updrafts with diameters between 1.8 and 9.2 km (1?5 n mi)]. Because there is less than a one to one correspondence between tornadoes and mesocyclones, the B9MA alerts forecasters when it detects circulations that meet its criteria but tornadoes may not be observed. On the other hand, some tornadoes are not accompanied by large-scale meoscyclonic circulations. Weather radars cannot resolve small-scale tornadic circulations, and the B9MA may fail to alert forecasters that a tornado is present. The B9MA is only one of many tools that forecasters should use to predict tornado formation. This paper describes limitations of the B9MA and how to improve its performance. The correlation between algorithmic detections and severe weather occurrence may be optimized under the premise that storms with strong, deep rotations are more likely to be associated with severe weather. Some tornadic circulations missed by the B9MA can be detected when the value of threshold pattern vector, a B9MA adaptable parameter, is lowered. When a rotational strength filter is added to the program logic, some algorithm detections of nontornadic mesocyclones can be eliminated.