Dupree Tornado - July 28th, 2008

A moderate risk of severe weather was predicted by the SPC (mainly for wind) on July 28th, brought about by a low pressure system developing in western South Dakota, and adjacent dry line and warm front, abundant moisture, high levels of instability, and decent shear. In fact, CAPE values above 3000 J/kg were present in central South Dakota by the time storms began to fire in the afternoon. The following images show a tornado which developed just south of Dupree, SD.

Above is the base reflectivity image of the Dupree tornado at 4:40 pm. An obvious hook echo is present on the radar just to the southeast of the Faith, SD METAR location. I have added a few features to the radar image including arrows indicating the wind flow, a green solid line indicating the slight risk of severe weather, a tan line indicating a moderate risk of severe weather, and a green dashed line indicating the dry line / boundary in the region.

The storm developed just north of the boundary which stretched from northwest to southeast through central South Dakota. This boundary was an obvious dryline with dewpoints in Philip, SD of 57F and 70F in Pierre, SD (shown in the yellow circles on the image below). A very moist, unstable atmosphere was present to the east, which a drier, more stable atmosphere was to the west. CAPE values along the boundary were approaching 1000 J/kg, helicity was near 50 m2/s2, and the significant tornado composite was approaching 1 to 2. As a shortwave trough entered western South Dakota, the storm near Dupree began to develop and quickly began to rotate. In fact, from storm initiation to tornado was about only 30 minutes. The reason for this quick development? The ingestion of horizontal vorticity produced by the dry line boundary indicated on the reflectivity image above.

On the storm relative image above, a clear couplet is seen in the storm, indicating the location of the tornado. At this time, the storm was producing a reported half mile wide tornado. The difference in dewpoint is indicated in the yellow circles.

The storm quickly received a tornado warning, which was confirmed after only 5 minutes. The storm continued to rotate, showing a strong couplet, for an additional 15 minutes (total storm length at this point, 50 minutes).

By 5:05 pm, the rear flank downdraft (RFD) had pushed out considerably to the southwest, south, and southeast (indicated by the white arrows). The southeast motion pushed the RFD into the inflow of the storm, quickly taking away the storm's moisture inflow source. Thus, even though a hook echo is still evident on the image above, it is likely that the tornado was retreating into the cloud at this point. This was also the time that the nearest storm chasers were finally starting to get closer to the storm. Sadly for them, the storm was gone 3 frames from this time period. Yes, the storm completely collapsed once the RFD took over the inflow region! By 5:25 pm the storm was gone.

An additional view of the RFD (also shown by the northeast wind in Faith, SD) using storm relative velocity:

So what can be learned from this? The ingestion of horizontal vorticity, or boundaries, can quickly result in a severe thunderstorm capable of producing a tornado. Also, when the RFD takes over the inflow region, storm development can quickly be stunted. The Dupree tornado is an excellent example of both of these phenomena.