Monday, May 18, 2009

Kirksville, MO tornado environment on 5/13/09

Sorry I haven't had time to post any case studies for awhile. But with 3 deaths from tornadoes in northern Missouri last Wedensday 5/13/09, I decided to make time for a short analysis of that event.

The tornadoes were rain-wrapped at times (see the photos above), and the most intense damage was rated EF-2 by the National Weather Service. A good environment for generating significant tornadoes aided the supercell that produced the 3 tornadoes in sequence from near Milan to Kirksville.

The surface map at late afternoon showed a low (see surface map above) moving eastward across northern Missouri, with south-southeast winds just east of the low and a retreating outflow boundary from morning storms that had moved across the area. Farther southwest, surface winds had a westerly component. Storms on radar began to build at mid afternoon near the low and then southwestward in advance of a surface front. But the tornadic storm (see arrow on radar images above) remained near a focal point just east of the surface low, and could take advantage of southeasterly low-level flow and increased storm-relative helicity (SRH).

RUC model analysis soundings at Chillicothe (CDJ) and Kirksville (IRK) highlighted the dramatic difference in low-level shear (see the sounding plots above). At CDJ, although MLCAPE was quite large (near 3000 J/kg) and deep layer shear was favorable for supercells (around 40 kts), southwest winds made for a small/straight/unidirectional hodograph in low-levels with small SRH. In contrast, at IRK east of the surface low and outflow boundary, low-level wind shear was quite large as a result of southeast surface winds and stronger flow just above ground, with a looping hodograph and big SRH (>400 m2/s2!). Even though MLCAPE was less than 2/3rds that on the CDJ model sounding, the combination of very strong low-level shear and strong deep layer (> 50 kts) in the Kirksville area really made a difference!

Saddled with an important mid-afternoon meeting, Shawna and I were only able to make it to a supercell east-northeast of Chillicothe (see the cell southwest of the Kirksville supercell near CDJ on the 5:02 pm radar image above), which was frustrating. Although the Chillicothe storm had a lowering and some decent supercell structure, the roughly 50 mile distance between supercell locations certainly made for distinctively different storms and tornado potential.

- Jon Davies 5/18/09


Adam L said...

Interesting study. Proving that sometimes more shear is better than more cape. Although 1800 j/kg is certainly enough!

Jon Davies said...

Sounds like you did well in catching the tornadoes near Edina, Adam. With > 400 SRH and 50 kts of deep shear, 1800-2000 J/kg is more than enough, for sure!

Danny Neal said...

GREAT study Jon. It was fascinating watching the interaction and merging between 2 tornadic storms just west of Edina. As you noted it didn't take the 3000+ CAPE to squeeze out a dominant storm. The motions of those wall clouds were some of the most intense I have seen in some time. It was almost text book. I am glad we got the window to clearly see the tornado before it wrapped itself up. There really needs to be a radar near Quincy IL. We were lagging in data out there but for a post event summary it would be nice to get awesome radar data. Again.... thanks for typing this up. Did you detect gravity waves/horizontal convective rolls on any of your analysis??? Granted I am not great at picking these out via radar/satellite but I thought i saw some cloud streets associated with a gravity wave near Kirksville around 2-3 PM.

Jon Davies said...

Hi Danny:

Thanks for the comments!

The merging cell into the Kirkville storm near Edina was very interesting... it actually initiated southeast of Chillicothe out ahead of the front over 1-1/2 hrs earlier. That cell really struggled until it got pulled leftward and merged into the IRK storm. You asked about gravity waves/horizontal convective rolls. On satellite from the UCAR archive, one can see stratocumulus lines/rolls over central and parts of northeast MO in the 20z-21z time frame that may reflect the enhanced low level SRH east & northeast of the morning outflow boundary.

Very interesting case!

Jared Farrer said...

Very intresting study! I was very late getting to the storm but managed upon it around monticello. But by that time lack of light and incredible amounts of precip were obscuring views of any tornadic activity. You had to be close!

By the way Jon were do you get your sounding data. It looks like you are using the GDL site but I wanted to ask?

Scott said...
This comment has been removed by the author.
Scott said...

Great write-up, Jon. My decision to break toward the "north" storm early that day (vs. the Chillicothe storm) was solely based on the subtle boundary and 15kt SE wind at IRK. I was a little worried about the cooler air but knew that storm would provide me with the best tornado chances, and boy did it ever! I was able to catch the first tornado the storm produced (near Milan) which, as far as I know, no other chasers witnessed. After that I was able to witness the genesis of the Novinger/Kirksville tornado at relatively close range. Thanks for taking the time to do this write-up, I can never get enough post analysis to either reiterate my though processes or give myself a facepalm after a poor decision.