Radar/GPS Analysis of 2007 Tornado Intercepts

Joined
May 1, 2004
Messages
3,417
Location
Springfield, IL
I took my GPS logs and archived Level 3 radar data and produced some animations of my tornado intercepts from this year. Its interesting to not only be able to see the entire chase from a radar prespective, but its also insightful as to what mistakes were made and how things could have been done differently. I've learned a lot from watching these animations, and I thought it would be valuable to share them here.

March 28 - Goodland KS (Multiple Tornadoes)
KGLD Animated GIF (2.5MB) Individual PNG Frames
Full Log

This chase was as close as I've gotten to perfect, and yet there was still room for improvement. Jerry and I ran around to three different storms before we settled on the right one. The running around cost us a tornado near Sharon Springs. I can't complain though, because this chase was unbelievably fantastic.

April 24 - Nickerson, KS (Weak Multiple Vortex Tornado)
KICT Animated GIF (3.5MB) Individual PNG Frames
Full Log

This was an incredible chase, not just because of the fantastic LP supercell structure, but also because we were 150 miles south of initiation, in a separate state, with storm motions against us, and we still managed to intercept. Lesson learned: We should have made our move north sooner after seeing better cu development along the dryline in KS and that OK was choking behind the linear mess.

May 4 - Lebanon, KS (Nocturnal Supercell and Rope Tornado)
KUEX Animated GIF (3.5MB) Individual PNG Frames - Chase Analysis
KDDC Animated GIF (4.6MB) Individual PNG Frames - Greensburg EF5
Full Log

I made two animations for this chase. One from KUEX wherer we wound up chasing, and one from KDDC where the Greensburg EF5 moved through. Was going north a mistake? I believe that the answer is not so obvious. At the time of initiation, the storm we were after was 70 miles to our north with a motion against us. What would become the Greensburg storm was more than 100 miles to our south. Although motion would have been in our favor, the added distance meant we would not have intercepted that much sooner. Also, as we departed, our northern storm was discrete and we had a visual on a very strong looking updraft tower. While we couldn't see the southern activity, it appeared multicellular on the radar. A more thorough investigation and scrutiny of the Mesoanaylsis might have helped push us south, although we probably wouldn't have caught the discrepencies that Jon Davies detailed in his analysis.

What if we had gone south? There were two possible routes. Straight south on 281 to Pratt, and then west would have been the wisest. However, to make better time we might have gone southwest on 56. This would have meant dodging the core and manuevering around the meso (and unbeknownst to us at the time, an EF5). Should something have gone wrong, (a flat tire, stuck in the mud, etc) the situation would have become extraordinarily dangerous. Also note that another cell went up north of the Greensburg storm before the Greensburg storm was fully mature. If we had taken the southwest option, that storm would have gone up right over our heads and we might have wound up chasing it instead of the Greensburg storm, ultimately seeing less than we caught near the Nebraska border.

May 4 was a succesful chase for us, but I do have mixed emotions about missing the Greensburg storm. It was a legendary event with a 2 mile wide wedge, satellite tornadoes, and other, separate torndaoes. Yet the devastation and death it caused would have been disheartening and ruined whatever joy we would have gotten from a succesful intercept or exhileration from feeling such power.

May 5 - Great Bend, KS (Stovepipe Tornado)
KDDC Animated GIF (4.6MB) Individual PNG Frames
Full Log

This was a stressful, yet rewarding chase with a low contrast tornado southwest of Great Bend. After our first wall cloud intercept we went north and west back to the dryline to setup up downstream while the storm we wound up intercepting our tornado on was already developing well south of there. We wound up making a huge loop and could have intercepted much, much earlier than we did. Another lesson learned from watching this animation: stay in the area of development, and be prepared to double back sooner. We also went north after this storm which turned out to be a mistake as it was gusting out and we had trouble keeping up. The wise move would have been to drop south down the line, catching the nocturnal tornadoes that many of you got.

Well, I hope you guys enjoyed the animations and my analysis. My friends have described the animations as looking like either a fun video game, or a drunken mouse in a maze.
 
Nice work !
Thanks for posting the animation, images and analysis.
 
Thanks, Skip. Very helpful and informative. On May 5, I made the opposite mistake from yours, staying too far south (in Woodward when the second batch of storms fired around Laverne) and then ending up playing catch-up. Your animation gives me a better picture of what was going on, not only with the storms I was chasing but also with other surrounding ones.
 
Thanks for the images Skip.

A few comments:
-I don't think you would've taken the cell just to the north; a few scans worth of loop and it would've been clear that was a left split.
-The cell in northern KS certainly did look good upon initation and later, both cells exhibited massive hooks. Looking at the radar loop, it almost seems the first cell there fired on a boundary, and then as it lifted north, it became elevated. Maybe that's why the northern KS cells failed to produce more prolifically?
 
Since no one has asked, how did you build these? I can gather that you imported archive data into GRL3, but how did you:

1. Get a GPS log in there
2: Did you still have to save screenshot by screenshot one at a time?

Very cool idea, thanks for posting it.
 
Thank you for the comments all.

I used Franson GPSGate to play back my NMEA logs that it records. The logger is one of the really nice features in v2.0 since you can play your log back in any application. It has some kinks in it, however, and handles errors in an amateurish way (poorly worded errors with poor grammar), but still has a great feature set for a very reasonable price.

I load in archived Level III data from the NCDC archive, 20 volumes at a time as thats the max GRLevel3 can loop through at once. I then match the GPSLog time to the radar's time and screen capture. I advance to the next volume in GRLevel3 and then play the GPS logger at 10x speed, screen capturing again when the times match, repeating until I've captured the entire volume.

Paint Shop Pro automates the capture process in that I can do multiple captures with one hotkey, capturing only the client area of the window. I then have it automatically save all of the captures, and run a batch script to further crop the captures and save them with descriptive names in chronological order. Jasc Animation Shop assembles them into an animated GIF.

The whole process takes a couple of hours, although I'm getting faster at it the more I do it. The most tedious process is having to pause to reload the next 20 volumes and then starting the logger playback again.
 
Skip, that is really cool! It's always nice to go back and see what you could change after the fact, hopefully it helps in the future!
 
Back
Top