Thursday, January 27, 2022

Meteorological Mystery: Newly Discovered Information Pertaining to the Udall and Greensburg Tornadoes

May 4, 2007, looking north: The Greensburg Tornado is exiting the 
city while the Trousdale Tornado touches down to the east.

In my book, Warnings: The True Story of How Science Tamed the Weather, I discussed the unusual similarities between the 1955, F-5 intensity, Udall, Kansas, Tornado and the 2007, EF-5 intensity, Greensburg Tornado. The former tornado killed 82 and the latter 11. Among the similarities were the radar echoes, as shown in this figure from the book. The Udall Supercell was also responsible for the F-5 Blackwell, OK tornado that occurred an hour earlier killing 20. 

Above is an illustration from Warnings. It is a superimposure of the Tinker AFB (TIK) radar on May 25, 1955 (black pen) and the Dodge City (DDC) NWS radar for May 4, 2007. The Udall supercell runs off the top of the tracing because it moves outside the range of the TIK radar. 

Earlier this week, on Twitter, Oakhurst_Wx made me aware of a 1956 Bulletin of the American Meteorological Society paper (BAMS) that features radar photos of the Blackwell-Udall Supercell. They came from a radar operated by Oklahoma A&M (now Oklahoma State) University in Stillwater (SWO). Because of this discovery of these photographs, I wish to offer two hypothesis that are not made in my book that I outline below. There is one factual error in the paper: the range rings' interval in the SWO radar photos are not a constant 10 nautical miles. The interval varies.

Similarities in the Radar Presentation of the Two Supercells
In 1955, the term "supercell" had not yet been coined and it is not mentioned in the BAMS paper. Instead, the supercell is called a "thunderstorm complex." The images below were from ten minutes before (arrow) to the time the tornado struck Blackwell. 
SWO radar images, click to enlarge.
The range of the radar's display appears to have changed between 9:19 and 9:24pm.
There is also an issue with the location of the echoes that is partly described below. 
Figure 15 shows the classic "hook echo" that was and is a signature of tornado potential. Some caution is needed when interpreting the images. The BAMS paper notes several technical issues with the radar's output. In the photos, the direction (azimuth) is incorrectly rotated 13° in a counterclockwise direction. In that figure (arrow), the tornado is on the ground south southwest of Blackwell. Figure 16 was when the tornado reached Blackwell at 9:24pm CST (per BAMS, 9:26pm in more modern National Weather Service references) at a true azran of 347° and 40 miles. 

In Figure 15, it appears a storm is developing southwest of the supercell. It strengthens and races northeast to merge with the Blackwell-Udall supercell. If it is an actual storm rather than a radar artifact, it grows at a remarkable rate in just five minutes Moments later, it begins to merge with the primary storm.  However, the paper mentions the SWO radar was "double pulsing" and to use caution. This caution may be well-founded in that the radar tracing does not indicate an echo merger and the paper's range for Udall was half of the actual range of 
Screen capture of the BAMS article
The "hollow core" was iso-echo contouring which was used
to depict areas of hail and the heaviest rain before color 
radar was invented in 1976.
90 miles with an azimuth of approximately 356 degrees stated in the excerpt above. The photo below was captured 4 to 6 minutes before the arrival of the tornado in Udall. When the Udall supercell (Image A) at 10:29pm (after being rotated clockwise) is compared to the Dodge City WSR-88D image at the time the tornado struck Greensburg (image B), the similarity is remarkable. 
Image A, 10:29pm radar, moments before Udall was struck
For non-meteorologists reading this, when an X-band radar beam passes through an intense supercell, much of the scattered energy is attenuated (diminished) as the pulse passes through the thunderstorm's torrential rain and hail and again when the scattered energy passes through again on its way back to the SWO radar antenna. So, in "real life," the protuberance (P) would be stronger than indicted here. The location of the hook echo (H) is inferred by me with confirmation in the text of the BAMS paper.

The Dodge City radar image (image B) at 9:54pm, as the tornado was passing through Greensburg, is below. 
Image B. Radar image as the Greensburg Tornado was in progress.
Focus on the lightest gray to compare shapes with A, above.
The color version of the Dodge City radar image (image B) is below. The labels P and H are as above.
My conclusion is that the incredible similarity between the Udall and Greensburg supercells continued to the time the respective cities were hit. 

The Supercell Beyond Udall
The BAMS paper includes a map that shows -- as in many other meteorological references -- the Udall Tornado making an abrupt right turn. (path map 1, below) and, possibly, causing scattered damage. The dashes indicate uncertainty and they take the tornado toward the Flint Hills town of Atlanta, Kansas. Those dashes were based on lesser damage in north Udall and some damage east or northeast of town, based on material I have reviewed. 

Based on the previously unavailable radar data, I am offering a hypothesis that the damage east of Udall might have come from a third, perhaps eventually strong, tornado touching down just east of the city.

There is now good reason to believe path map 1 is not accurate in the sense of making an abrupt right turn. The map of the Blackwell and Udall paths (path map 1, below) shows the Blackwell Tornado making an abrupt left turn as it moved north. It would be unusual for the next tornado from that supercell, Udall, to go in the opposite direction. Note also the zig-zag in the path of the Udall storm in between Geuda Springs and Oxford. It is possible that the damage noted when the path was being mapped was due one or more satellite tornadoes, which we also didn't know existed in the middle 1950's (path map 3).
Path Map 1

Now, look at the detailed path map of the Greensburg Tornado and the next in its supercell's series, the Trousdale Tornado on path map 2 below. 
Path Map 2, National Weather Service. Compare to photo at the top of this article.
Like Blackwell, note that the Greensburg Tornado made a left turn that turned into a corkscrew path. Simultaneously, the Trousdale Tornado touched down on US 54/400 just east of Greensburg much like the Udall Tornado touched down near the curvature point in the Blackwell storm did 52 years later (path map 1).

If we return to the Stillwater (SWO) radar images, the supercell continued to move in its pre-Udall direction of slightly east of due north. 
SWO radar images, pre-rotation correction
Interestingly, the hook echo becomes much more distinct at 10:44pm, indicating the possibility of yet another tornado or a continuation of the Udall Tornado. This 1955 Blackwell--> Udall --> Third Strong Tornado would be similar to the 2007 Greensburg-->Trousdale--> Macksville (path map #3) three violent tornado sequence as, based on Doppler radar, they were similar in wind speeds to the infamous 1990 Bridge Creek-Moore Tornado as displayed on a WSR-88D. The official EF-ratings of Trousdale and Macksville were "only" EF-3 because they did not strike structures that were strong enough to rate a higher category. 
Path Map #3, National Weather Service
Greensburg (5), Trousdale (13) and Macksville (14) Tornado Path map. 
Please note the small "satellite tornadoes" (6-12). They were satellites
to the Greensburg Tornado's mesocyclone. 

If the tornado had continued on its existing path, it would have moved through the western edge of the Flint Hills between Douglass and Rose Hill, then north to between Andover and Augusta. These were sparsely populated areas. It is entirely possible that, once damage was found to the east, the meteorologists at the time didn't think to look farther north. 

Adding to the evidence is the BAMS article included the report of cancelled checks from Udall to the west of the town of Florence, Kansas (see below).
I know of no meteorological situation that would have carried the checks 62 miles to the north if the parent supercell had taken a turn to the east. 

As to how far the Udall or, more likely, a third tornado might have traveled, a rough guess is about 25 to 30 miles. The Blackwell and Udall tornadoes were on the ground for that distance. If that is correct, the tornado would have lifted near the location of today's Kansas Turnpike south southwest of the town of Towanda. It was then a very sparsely populated area and the tornado -- by then it was around 10:40 to 10:55 mph -- could have escaped detection. 

Where Was the Wichita Radar Data?
Much of this mystery would have been cleared up if the Wichita Weather Bureau radar had properly tracked the supercell. It did, in northern Oklahoma. However, go back to Image A. Given the orientation of the storm, it would have extended into the eastern suburbs of Wichita. Winds there were gusting to 45 mph with torrential rain and hail. That caused the WB radar to become completely attenuated (blotted out) based on my interviews with meteorologists Ellis Pike and Victor Philips who were working that night. That attenuation, which was not understood in 1955, caused the Weather Bureau to issue a misleading statement that "all warnings have been lifted." Many of the people of Udall retired for the night upon receiving that assurance. Eight-two people died in Kansas. 

Summary
I offer two new hypothesis about the devastating Kansas-Oklahoma F-5 tornadoes of May 25, 1955.
  •  The incredible similarity between the Udall and Greensburg supercells continued to the time the respective cities were hit. 
  • The damage east of Udall might have come from a third, eventually strong, tornado touching down just east of the city. It is also possible a rear-flank downdraft caused that damage and the Udall tornado itself continued on to the north. 
Please consider these results preliminary. I am going to continue investigating and will either report back or will submit a paper to a meteorological journal.

© 2022 Mike Smith Enterprises, LLC

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.