The October 9th, 2001 Cordell, Oklahoma Tornado -- An Environmental Study

So, you might ask why a meteorologist in Greenville, South Carolina (as of this writing) would do a review of a tornado that occurred in Oklahoma? A good question. I am quite interested in understanding tornadogenesis. As a meteorologist, one day I will have to issue warnings for large, damaging tornadoes, even here in the East. It is imperative that I be able to recognize an environment which could support these types of storms. And, as a person who likes to chase, it is important to know the conditions which favor tornado development on the smallest temporal and spatial scales, as there is nothing more annoying than being one supercell too far to the south when a tornado occurs. Really, I'm not much of a chaser, and the times I do go out, it is mostly to hone my skills as a meteorologist. Still, it is much easier to just go ahead and say "I chase" rather than spend 10 minutes explaining that I really do not chase.

Using the good old forecast funnel approach, I'll first include a couple synoptic scale maps. The only 12Z data I have is this rather vague 4-panel from Unisys, and a nice 500 mb map from the AMS DataStreme web page -- lets hope the ETA 3D-VAR caught the bad FFC sounding! Here are links to the 300 mb, 500 mb, 700 mb and 850 mb mandatory level plots from 00Z on the 10th, about 2 hours after the Cordell, OK tornado. These images are also from the AMS DataStreme web page.

There was a strong 500 mb short wave trough just west of the Texas Panhandle at 12Z. In fact, the trough was well east of the area by the time the tornadic storms formed during the afternoon of the 9th. What this trough served to do was push the dry line into extreme western Oklahoma, setting up strong convergence and resulting in a strong thermal and moisture gradient that looked like something you would expect to see in late May. The attending image shows the dry line, and in particular, the dry line bulge in southwest Oklahoma at 21Z, about the time the second batch of tornadic storms began to organize (see the table below for radar images of most of the tornadic storms this day). Click on the image for a larger view of the region at 21Z.

Notice how strong the surface winds were just ahead of the dry line, sustained at 30 kts. The flow was also backed to around 160 degrees at HBR, which was directly to the southeast of the two main supercells. Here are some other CAPS images from around 22Z:

• 500 Mb wind height, winds and vorticity
• Storm Relative winds
• Bulk Richardson Number
• surface dewpoint
• surface based CAPE

And here are the HBR surface observations for the afternoon and evening of October 9th:

   
   2 AM (7) Oct 10 71.1 (21.7) 64.9 (18.3) 29.87 (1011) S 19 
   1 AM (6) Oct 10 71.1 (21.7) 66.0 (18.9) 29.87 (1011) SSW 19 
   Midnight (5) Oct 10 72.0 (22.2) 66.0 (18.9) 29.86 (1011) S 23 
   11 PM (4) Oct 09 73.0 (22.8) 66.0 (18.9) 29.84 (1010) S 25 
   10 PM (3) Oct 09 75.0 (23.9) 64.9 (18.3) 29.82 (1009) S 23 
   9 PM (2) Oct 09 75.9 (24.4) 64.9 (18.3) 29.8 (1009) S 24 
   8 PM (1) Oct 09 75.0 (23.9) 66.0 (18.9) 29.78 (1008) SSE 19 
   7 PM (0) Oct 09 79.0 (26.1) 64.9 (18.3) 29.75 (1007) S 20 
   6 PM (23) Oct 09 80 (27) 66 (19) 29.73 (1006) S 21 thunder in the vicinity 
   5 PM (22) Oct 09 80 (27) 66 (19) 29.72 (1006) SSW 33 thunder in the vicinity    
   4 PM (21) Oct 09 78 (26) 66 (19) 29.73 (1006) SSE 31 thunder in the vicinity    
   3 PM (20) Oct 09 80 (27) 66 (19) 29.78 (1008) S 29 
   2 PM (19) Oct 09 75 (24) 66 (19) 29.82 (1009) S 28 
   1 PM (18) Oct 09 71 (22) 66 (19) 29.87 (1011) S 19 

These incredible surface winds resulted in a 90 degree shift in the wind just a few km above the ground. This type of right angle bend in a sounding is often (maybe always) associated with the strongest tornadoes. The accompanying CAPS sounding from Altus at 22Z shows the 90 degree turn. You will need to click on the image to get a good quality image. Again, this is in some ways a "perfect" sounding with the strong veering between 1 and 2 km, and a bit of mid-level weakness in the winds.

That's right, a weakness in the winds. To quote from Paul Markowski's graduate thesis (see references) "...there was some suggestion that storm relative winds in the 3-4 km layer were (statistically) weaker in the tornadic cases than in the non-tornadic cases" -pp120. However, Markowski also states "The finding herein of slightly weaker s-r winds in the 3-4 km layer should be viewed with the utmost caution, since there is no way of knowing the extent of the environment variability on each day... Furthermore, the difference of only 6.7 m/s (6.8 m/s) between the average 3 km SR winds (4 km SR winds) in the tornadic and non-tornadic cases may be small." --pp126-127.

Small, but far from insignificant I believe. Though he never out and gives a number, it appears that the s-r flow for tornadic cells averaged around 8 m/s for tornadic storms, and around 14 m/s for non-tornadic storms at 3-4 km. In looking at the LTS sounding, which is only a model sounding, albeit from a high-resolution model designed to forecast convection on the Plains, the storm relative wind at about 600 mb appears to have been on the order of 45 kts, or 23 m/s. The progged storm motion was from 240 degrees at 16 m/s, which puts the 3-4 km s-r flow at around 9 m/s, in line with the tornadic supercell numbers from Markowski. I will try to find the true supercell movements in the near future, but I suspect they were even a little slower.

I took several reflectivity images from the UCAR archive page. I hope they don't mind! What I have done is create a table of WSR-88D images valid for the time of almost every tornado occurrence in western Oklahoma on October 9th. The Cordell event is hightlighted in red. I skipped one at the end, and one obviously bogus report. Of course, a couple of these tornados were on the ground for nearly 30 minutes, so the times are in some ways rather arbitrary. The data come from SPC's preliminary reports. Oftern the exact location and timing of tornado events is not quite accurate in these reports, but they should be close enough. It appears that nearly every supercell in southwest Oklahoma on October 9th which hooked (i.e. had a strong RFD) produced either a strong tornado, or satellite tornadoes along the RDF/updraft interface.

As the evening wore on, the supercells continued east, but as the SPC event reports from the day show, the tornados were gradually replaced by hail events before the supercells dissipated completely. The reason for this can be seen by the 00Z OUN sounding. While this sounding was immediately downshear of the tornadic supercells, and indeed the supercells moved across the Oklahoma metro area shortly before dissipating, the LFC had become much higher as the low levels cooled. Thus the RFD's would were not as buoyant and tornado development did not occur.

Finally, the Vici and Purcell OK soundings showed similar wind speeds (as one would suspect) to the CAPS and RAOB soundings. These winds did appear to to a little stronger in the mid levels. However, what I find interesting is that there clearly was a mid level jet which was exiting the region about the time the tornadoes developed. In other words, the mid level winds were weakening at the time the tornados developed. Also, the region appeared to be in the left entrance region of this jet, typically an area of subsidence!

-Bryan McAvoy
10 October, 2001