2010 Post Chase Radar and GPS Analysis

Joined
May 1, 2004
Messages
3,417
Location
Springfield, IL
I really learn a lot going back and comparing the radar against my GPS track after a chase. I thought other people might find it helpful as well. The following are animations created from some custom software I'm developing that shows level 2 reflectivity and my logged GPS position at the time. I've got Adam Lucio's log in on some of the animations as well for comparison. I've got all my tornado days of the year here, as well as some of my more interesting busts. I probably learn just as much if not more from the bust animations than the intercept ones. Its great seeing where you went wrong and how you could have done the chase better.

Tornado Intercepts:

April 22, 2010



Tornadoes we're between 5:15 - 5:25, a few random ones between 6:30 and 6:45, passing pretty close to a multivortex at 6:44.

The first video is a broad view and the second a zoomed in one with Adam's track as well. The ideal play would have been the southern storm to catch the Goodnight, TX tornado. I'm not sure what would have lead me to choose that storm over the northern other than maybe interference issues, but the northern one did produce twice and looked quite a bit beefier on the radar. A better play for the northern cell would have been to stay where I was originally instead of getting impatient driving back down to Clarendon, up to the storm, only to have to turn and come back the same way. I didn't miss the tornadoes, although i would have been closer had I stayed put.

May 22, 2010



Tornadoes between 6:15 and 7:05 with the picture in my avatar at about 6:35. Adam, Danny, and Ben got several more between 7:40 and 7:55.

This is about as close as I've come to the perfect intercept. We're on the storm at initiation, and stick pretty close until it goes permanently HP. I had great data thanks to Brandon, but wound up not using it at all. I don't think I looked at the radar once until we bailed on the storm. It was a great storm for visual chasing. Couple of improvements would have been not racing the hook north to US 12 and then east Bowdle. We crossed paths with the tornado a couple times there and it got a little hairy. Our road preferred road options weren't navigable and we decided to race the storm instead of falling back. The other improvement would have been hanging around for the wedge longer. Kevin's car started smoking and we were about to get cored by the RFD so we dropped south midway through the show.

The second animation shows the infamous field incident the best starting at about 7:40. You can see how the map depicts the road going through, when in reality it didn't, and then the turn south into the field as the hook passes overhead.

June 13, 2010



Tornado at approximately at 4:20

I didn't have data when I went into the OK panhandle so I assumed the storms were still orientated in a fairly linear fashion. I could tell there was at least a kink in the line to my south though where the Slapout, OK tornado developed, but I didn't know until afterwards that the storm was discrete and had a classic hook echo. The HP mode and being north of the hook made it look like a solid line. An improvement would have been to stay on the gravel roads I was on just prior to the tornado. These roads were in good condition, and I hit a huge convergence once I got on the highway that prevented me from stopping and getting decent video of the tornado.

June 17, 2010



Almost continuous tubage from 6:04 to 7:06 and then 7:45 to 8:02

We had no data on this chase (or very little). The radar isn't as good in this animation as Minneapolis is quite far from the storms and the radar had to look through quite a bit of precip. You can still get the general idea of what's going on, but the radar makes the storms look much sloppier and and HP than they actually were. Many times when it shows us under 50 dbz it was just lightly raining or not at all. Adam and I were caravanning on this chase and we got split up right at the first tornado, so its interesting to compare how the rest of the chase went as we split up.

July 14, 2010


Tornadoes between 2:30 and 2:40

I was fairly lucky in my initial placement on this day as I got the needle in the haystack tornado. I came in from the north during the intercept, however, so there is a lot of rain in my video and stills.

And now for some busts...

May 10, 2010


Busted due to a combination of being out of position and storm speeds.

With little capping in place, the upper level energy already well established, we should have been just downstream of the triple point (closer to Adam) at the start of this animation instead of in Tulsa. The storm we did wind up intercepting gave us a couple minutes worth of hazy rain free base before waving goodbye to us at warp speed. It apparently produced soon afterwards, but with the storm speeds, keeping up was not an option. We have caught the Wakita beast as it crossed 35 had we gone north on 35, not sure how much construction would have hampered us though, but it sounds like Twistex did the same thing until they lost their mesonet.

June 1, 2010

Classic screwed by Iowa chase.

Having storms go up well to your east is never good, but the Iowa drivers and slow highways sealed our fate on this chase. Adam and I could do nothing but watch as the supercell kept just ahead of us, hooked, produced, and then died right as we finally caught up to it. We expected initiation further west in southeast NE or along the dryline in central NE, but instead it went up well to our east. The same thing happened to us on June 17, but luckily we were actually able to catch the storms that time. Monitoring the vis sat might have helped a little more on this chase.

June 10, 2010



Picked the wrong storm.

I don't really consider this chase a bust as I saw a couple of gorgeous supercells. It would have been nice catching the Last Chance tornadoes though. I committed to my target of southwest NE/southeast WY and when a storm fired just south of there in northern CO I was sold and stuck with the storm as it put down a nice wall cloud. Probably due to anvil blow off from the storms 100 miles to the south, it looked like it lost its low level instability and turned into an elevated, cold, outflow dominant storm. I raced down south, core punching the Last Chase storm, but missed the tornadoes by about a half hour. Adam did catch the tornadoes and you can see how his chase went in the second animation.

Hope some people find these animations helpful/interesting. Post your own radar/gps animation if you have them.
 
Real awesome software Skip! I really liked our Bowdle, SD animation, and seeing the gigantic hook and the 2 donut holes in the one scan on either side of us. We got real lucky.

Just think of how awesome this would be with phased array data ;)
 
Skip, you never cease to amaze me. This is freaking awesome! And for the infamous Bowdle Field Party, it really makes things more clear to those of us who were not there. That wasn't one of the faint roads on the map, it appeared as a darker line like a main side road. I could have made the same mistake easily.

I have a question: Does the GPS plot vs. the radar scan in any way take into account the delay in WSR88 radar in storm mode? In several of the videos it looked like you all drove right through the hook so I'm assuming not, but then again in the two CA videos I have there are some nice fairly closeup shots. It would be interesting to compare these loops to the videos. I just might do that later tonight.

If you decide to start selling this software, I'll buy. I'd love to be able to go back and critique what I've done (and what I've missed).
 
I have a question: Does the GPS plot vs. the radar scan in any way take into account the delay in WSR88 radar in storm mode? In several of the videos it looked like you all drove right through the hook so I'm assuming not, but then again in the two CA videos I have there are some nice fairly closeup shots. It would be interesting to compare these loops to the videos. I just might do that later tonight.

I'm not aware of this storm mode delay. Do you have some info handy on it? The timestamp is pulled from the radar file, and as far I can tell is pretty accurate. The sequences you see of driving through the hook are probably accurate. On 5/22 we drove right underneath the tornado producing region of the storm twice. On 4/22 the next exit off 40 west wasn't until I was underneath the wall cloud/hook. I would have liked to have avoided these situations, but ultimately we were ok. In fact I think you can safely be under the hook echo if you have the right roads and situational awareness. Since the radar is often seeing thousands of feet above the ground, the precip at that level doesn't directly correlate to what's happening on the ground. If you know where the tornado is, what direction its moving, and are out of the way, you should be safe even if the radar indicates you're in the hook (which would most likely be the RFD).
 
I felt like a complete nerd being able to reconcile the vehicle icon locations in the 6/17 animation with the scenes from MM in my head :(

Very, very cool!
 
I love watching these, especially when you watch Minnesota Mayhem and you see how very different the footage I got compared to what Danny and Skip got, yet when you look at the loop we were barely more than a mile apart. Its amazing what a difference positioning makes.

Watching the June 1st one is just sad...Oh Iowa...
 
Last edited by a moderator:
I'm not aware of this storm mode delay. Do you have some info handy on it? The timestamp is pulled from the radar file, and as far I can tell is pretty accurate. The sequences you see of driving through the hook are probably accurate. On 5/22 we drove right underneath the tornado producing region of the storm twice. On 4/22 the next exit off 40 west wasn't until I was underneath the wall cloud/hook. I would have liked to have avoided these situations, but ultimately we were ok. In fact I think you can safely be under the hook echo if you have the right roads and situational awareness. Since the radar is often seeing thousands of feet above the ground, the precip at that level doesn't directly correlate to what's happening on the ground. If you know where the tornado is, what direction its moving, and are out of the way, you should be safe even if the radar indicates you're in the hook (which would most likely be the RFD).

Skip, what I was referring to is the scan time of the radar. It's what, around 4-6 minutes in storm mode and about twice that in free air mode? The time it takes to complete the scan is what I was referring to. But I wasn't thinking about the timestamps, I was thinking that this is where you were as you were viewing the radar. It was a brain fart on my part. Just act like I never said a word... :eek:
 
Ah, yeah, I'm not sure how long each scan takes. The radar does all of the tilts within that 4-5 minutes too, so I'm guessing each tilt takes less than a minute. I'm sure some radar guru knows. It seems accurate enough for these purposes though.
 
Ah, yeah, I'm not sure how long each scan takes. The radar does all of the tilts within that 4-5 minutes too, so I'm guessing each tilt takes less than a minute. I'm sure some radar guru knows. It seems accurate enough for these purposes though.

I'm no guru, but if you look at GR3 the VST is the volume scan time. It'll be different for each tilt, each higher tilt a little later than the one below it. The composite reflectivity is the one that lags since it waits until all the tilts are scanned before it is produced.
 
Nicely done, Skip. Where's the Blu-ray version of MM with this app PIP and synced up as an alternate angle though? Seriously, that would be awesome :D
 
Here's one more for you guys.

June 5, 2010


Adam and Danny saw a couple of nice tornadoes from a distance starting at about 8:10. It goes to show that with the right roads and terrain you can catch east moving storms from the west. I'm not even going to bother rendering my track from this chase as I was further west chasing the garbage in Iowa.
 
I hate that day..........

Don't feel too bad guys, I was 30 minutes late to the Bowdle wedge and only caught up in time to see the later tornado north of Roscoe and even missed the next one near Ipswitch. Been kicking myself since for having driven 15 hours to come up short by so little.

On 6-5 however, Brad Emel and I managed somehow to keep ourselves from going for the Des Moins area crap storms, and sat in Ottumwa until initiation of the supercell that spawned Yates City/Elmwood tornado. My avatar can give you an idea of how I feel about that day.

One chasers ultimate victory can be another mans 'just missed it', or worse yet 'total bust'.
 
Yeah, 6-5 was uhm. painful. Even though we saw 2 tornadoes from 8 miles away.... seeing Brad, Scott, Kurts, Bobs, etc, etc, etc video.............

baby-crying.jpg
 
I <3 6/5 :)

What would be sweet is if this was a web app that would accept gpx/kml files and spit out a custom animation!

I'd love to extend the capabilities of this software to do that. There are a few links in the process that would be quite tricky to do programmatically. The first is getting the radar data. I'm not sure if I can automatically interface with the NCDC HAS system to obtain the radar data. The next would be encoding the rendered frames into a video format that Youtube (or something similar) can use. It looks like Youtube has an API that would work for uploading the encoded files, but interfacing with and understanding these APIs is going to take a lot of effort. Right now I'm using Sony Vegas to encode the video, and it wouldn't be fair to the user if I said, "Ok, here's 5000 PNGs... use your video editor to encode them into video."

The turnaround time from when you upload your GPX data to when you have an animation on Youtube might be over a day as well, as it takes awhile for NCDC to send the requested data, rendering out the individual frames takes hours, and the resulting animation can be hundreds of megabytes and takes quite a while to upload to Youtube. Some serious server horsepower would be needed to fully automate the process or the wait for animations would be incredibly long, especially if there are multiple animation requests.

I'll start poking around though and see how far I get. If there are any developers on here who would like to help out with this project please let me know and maybe we can pool our resources.
 
Good day all,

Skip, you did an awesome job at putting these together. It compliments your work on the storm-lapse project / DVD as well, as many of these chases are in that as well!

Keep up the good work.
 
On 5/22 we drove right underneath the tornado producing region of the storm twice.



Skip on this video at 0:10 in time you (or I think it's you) drive up out of that gap where the tornado is crossing the road. I also have a Hollingshead DVD and some other earlier footage of the Bowdle event where I saw you even closer to the action. I don't have that footage on YT, but I can vouch for seeing your van pretty close to the tornado. My wife was a little nervous in that video as that was the first time she saw a tornado on the ground up close and personal. She saw rope tornadoes and so forth before, but nothing like that Bowdle wedge.

I really like this software you have put together Skip and like you said, it's probably a very good learning tool to see how you reacted in-situ.
 
Nice video Mark!

Wes - We definitely live in the hook region so it would not surprise me to see us punch the hook (5/10, 5/22, 6/17) to name some that stick out right off the bat. lol Skip is a great guy to know and have on your side! The possibilities are endless with him and his software skills.
 
Back
Top