NSSL SWAT Case Study - 26 October 1996 Minnesota Low-Topped Supercells

A low-topped supercell outbreak occurred in Minnesota on this day. The storms occurred in a highly-sheared environment, moving toward the northeast at speeds up to 40 knots .  Despite the low-topped nature of these storms, their storm tops were typically some 6 to 8 miles downstream from their respective base.  Algorithm mesocyclone detections are challenged by this type of "severe lean." In addition, the time between elevation scans allows storms to move.  Hence, the vertical structure of a storm appears even more sheared as viewed by the radar.

It is difficult for the algorithm to vertically associate two widely separated circulations as being part of one mesocyclone.  It would be useful for the algorithms to adjust their vertical association techniques when the vertical shear is very strong, as in this case. The shear information could be obtained from the WSR-88D Velocity Azimuth Display (VAD) vetical wind profile, or from a Near-Storm Environment (NSE) algorithm reading in data from a mesoscale model.

Also, in many of the Minnesota storms, circulations were only visible on the lowest elevation scan.  Having only one 2D circulation brings up an entirely new set of problems.  Namely, the algorithm must have a criteria for determining when a 2D detection is strong enough to warrant classification as a 3D mesocyclone (Under non-low-top circumstances a 2D feature would not be considered a mesocyclone without vertical continuity).

Presented here are various WSR-88D radar images from the Minneapolis, MN radar (KMPX). Included in some of the WSR-88D images, there is output from NSSL's Mesocyclone Detection Algorithm (MDA).  A yellow circle depicts a mesocyclone.  The red-in-yellow circle depicts a mesocyclone whose base is at the lowest elevation scan(where it is a more likely tornado threat).

Description:  The Storm Initiation panel serves to point out the unique way in which this storm was formed.  Essentially, a rear flank downdraft(RFD) from a pre-existing mesocyclone provided a boundary to focus convection.  As seen in the first cross section, the storm tower forming on the RFD was not relatively tall, but there was already a good rotation signature at the lowest scan.  The rotation then ascended rapidly along with the storm tower.  Only six minutes later there was a mesocyclone , and only six minutes after that the mesocyclone was tornadic.  The MDA performed very well when tested on this quickly forming tornado, as seen in the Tornadogenesis loop. Description:  The storm can be seen collapsing in the Reflectivity Panel.  The top row of images in the panel follows the decrease of reflectivity at sweep 2, and the bottom row follows the decrease of reflectivity at sweep 1.  The middle frames represent the time of the tornado report.  Notice how a bounded weak echo region occured in the storm's center at the time of the tornado.

Judging from the Storm Relative Velocity Loop, it appears that this storm had a strong mid-level mesocyclone whose rotation was brought to the ground when the storm collapsed.  This would be why the shear signature is much stronger preceeding the time of the tornado.  The strongest shear had dropped below the view of the radar beam; in other words, Stearns Co. saw a descending mesocyclone versus the ascending mesocyclond seen near Cyrus.

Description:  On the Overview, the Insanti Co. storm is toward the lower right and is overlain by a mesocyclone detection.  Being the furthest south and east of any storm, the Insanti Co. storm had no competition for inflow and was located over the warmest, most humid surface air available.  These factors contributed to the strength and impressive radar representation of this low-topped supercell.

The Closer Inspection images are especially interesting.  The storm relative velocity portions of the two images illustrate the cyclonic flow surrounding this intense mesocyclone (red in yellow circle).

By matching damage reports to the radar data it appears that the tornado touched down around 2338 UTC and lasted to 2351 UTC. 

The case was presented at the 28th Conf. on Radar Meteorology as the paper entitled Doppler radar algorithm performance during a highly sheared tornado outbreak by E. Howieson, B. Grant, G. J. Stumpf, and D. W. Effertz.

Back to NSSL SWAT Case Study Table of Contents Page.