How to "Count" Tornadoes: Determining When One Ends and Another Begins

[Kyle Brittain]

So? How do we count tornadoes? I've heard some debate over time as to whether a tornado skipping along through the countryside counts as several tornadoes, or a single one.

If a tornado touches down, and then lifts up (no more ground circulation), but the condensation funnel continues to hang for time before touching down again, is this next touchdown considered a new tornado? (I could see this becoming tricky considering situations where there is a ground circulation with no condensation funnel aloft in sight). What if the condensation funnel disappears, then reappears, and touches down again? Would this now count as a new tornado (even though it may have originated from the same mesocyclone)? What about a cyclic supercell dropping several tornadoes, all from new mesocyclones, but from the same regenerating storm? (Most would probably agree that in this last case, each tornado would be counted individually).

I know this issue is super important for certain folks who like to brag about how many tubes they've bagged

How do you count tornadoes?

 

[Nick Wilkes]

If a tornado touches down, and then lifts up (no more ground circulation), but the condensation funnel continues to hang for time before touching down again, is this next touchdown considered a new tornado?

*Usually* that is considered one "skipping" tornado, as long as the gap between the touchdowns is less than two miles.

 

[Ethan Schisler]

I think for documentation purposes (storm data etc), it would be counted as each cycle in a supercell. For example, if a tornado lifts for several minutes, but the ground circulation remains, it is still the same tornado even if it touches back down again. Now as for how people count that for their "personal stats", that is up to them. In the case of a cyclic supercell, each cycle is a new tornado. For example, your first tornado is moving from south to north and you get another mesocyclone to develop to the east a couple miles and produce again, that is #2.

Although, if you have a situation where one mesocyclone occludes and the tornado totally lifts (IE no ground circulation), but it drops again 5 or 10 minutes later, I think you could make a case where that is a new tornado from the same occlusion. I can't see that occurring too often though, because as I said, cyclic supercells tend to be cyclic with NEW mesocyclones. If that makes sense.

Nevertheless, getting an *accurate* count on tornadoes during big days is always tough, because again, there is always dispute over where one ends and another begins. Personally for simplistic reasons, I think its best to count per tornado cycle, counting each time it (condensation funnel) touches down is kinda ridiculous and can lead to confusion in your statistics. Just my opinion

 

[Jason Boggs]

I've never been one to really count tornadoes, but if I did, I would use the technique that Ethan described. If a tornado lifts with the funnel still visible and then comes back down a few minutes later, I would not count that as 2 tornadoes. Now, if a tornado lifts, and the funnel dissipates completely, then another tornado forms, I would count that as 2 tornadoes.

 

[Dan Robinson]

Most of the time, if the circulation is still present in between condensation episodes, the tornado is still on the ground - at least minimally. That is, there is still a damaging circulation present there - in other words, the same tornado. I can't even think of any cases off hand where a single circulation truly weakened to "no tornado", then managed to restart. I'm sure it happens, it's just that I don't ever remember seeing one.

 

[TJKLECKNER]

I have always wondered how they kept track of tornado reports and actual tornadoes. In the heat of things, I know that often the same tornado gets reported in duplicate. I just wonder how often reports get exaggerated. "Today in Texas we had 20 tornado reports".

Note: I know later the weather service will scout the area and make a final and accurate report a few days later.

 

[Michael Towers]

I can't even think of any cases off hand where a single circulation truly weakened to "no tornado", then managed to restart. I'm sure it happens, it's just that I don't ever remember seeing one.

I’d argue the Rozel “tornado” was in fact two tornadoes and is an example of exactly what you described. While this event is officially listed as 1 tornado by the NCDC I’m nearly certain they are wrong and there were in fact two separate tornadoes. I’d normally tend to defer to the experts when it comes to documenting an event but I was there and I know there were at least a couple of minutes where there was absolutely no visual indication of a tornado. I have no record of the event from that time but the video below from Danny Neal covers that span and is about as convincing evidence I could find to confirm what I thought I witnessed (or didn’t witness) and verify my conclusion that there were two distinct tornadoes.

In the video you can see the tornado dissipate and any visual evidence all but disappear by the 16:45 mark. He stops shooting and at the 16:50 mark resumes and for the next two minutes there is absolutely no visual evidence of a tornado. No condensation funnel, maybe a little nib funnel from 17:30 to 17:45 but beyond that any funnel would have to have been obscured above the cloud base. If so did it produce violent circulation at the ground? Look at the rain curtains, they’re under the previous area of circulation yet show no signs of circulation, certainly not any violent circulations. You basically have a lowering, some light non-rotating precip and a few scud tags. Nothing evident to me (or apparently Danny and his crew) that would indicate the existence of a tornado. If I were to post those two minutes in a report as my video of a tornado I think the overwhelming response here would be…huh? People would either be scratching their heads, laughing to themselves, skewering me about the fundamentals of identifying a tornado or some combination of all of the above. At about 18:50 the funnel appears once again and in my opinion begins tornado two once the associated ground circulation develops.

I wonder what the survey members found that led them to conclude it was one continuous tornado. Did the damage path not exhibit any appreciable spatial separation? Did the genesis of tornado two occur so close to the location of the termination of tornado one that it appeared as one continuous track? This event was witnessed by tons of chasers yet as far as I’ve seen nobody has questioned the accuracy of the assessment that it was one tornado. Listen to the guys in the video, Danny is a well respected veteran chaser and for those two minutes they’re talking about a separate area of circulation and not the tornado that should have been ongoing at that time. Maybe I’m missing something, maybe someone else has a vantage point that shows the continuation of debris at the ground for those two minutes but if so I’d be surprised. At first I thought perhaps the report was correct, they’re the experts and maybe I missed something. Maybe that’s what a lot of others thought as well, defer to the experts, it must have been an ongoing tornado. But I was there and I didn’t see it. Once I saw Danny’s video it confirmed what I thought as the event actually happened, that tornado two really was tornado two. I may be proven wrong but based on my own witnessing of the event and Danny’s video I have to conclude that these were two distinct, individual tornadoes that occurred in close spatiotemporal proximity pendant from the same parent circulation. If someone can prove me wrong I’ll gladly accept the current classification and go forward with a lot of questions about what I may or may not be seeing out there.

 

[Matt Salo]

As far as personal stats going, I don't even worry about counts any more really. The first 5 years or so in the early mid-90's it was cool to track counts, but it started getting old. After 20-30+ I stopped caring about counts. Personal tornadoes counts do nothing for me these days. I just want to see the damn tornadoes, even if only for a moment.

Other than that, I feel like if the same funnel bounces along a path but never fully dissipates, it should be considered 1 tornado until it fully dissipates and recycles. The problem with this though is that unless there is video evidence of said bouncing of a single funnel, and the only evidence is ground damage survey, this can be inconsistent.

 

[Dan Robinson]

I can't say for certain about the Rozel case, since I was on my way to the new meso when that second rope appeared - but my understanding is that the circulation was there all along. It was well into occlusion at that point, and with the wet fields, it would have been hard to make out a ground circulation unless one was viewing it super-close. The second tornado (the one at Sanford) did a similar deal where it vanished nearly completely after the occlusion was done, then a rope segment re-appeared in mid-air with literally no cloud material associated with the remnant updraft above it. In the Sanford case, though, the ground circulation was kicking up more of a debris cloud post-occlusion so that it never had the appearance of dissipating. Since it's common for an occluded/roping-out tornado to have segments re-appear, that was what I'd *assumed* happened with Rozel. Again, I didn't see that part of it, so I could be completely wrong.

 

[Alex Heimberger]

Was right there (Rozel) & don't have full account of this topic. Recorded video until I thought the tornadic rotation was done. If I remember correctly, it didn't take long for the second funnel to appear. I'm posting on our FB what I recorded around that time. All sped up aorund 5x. Don't get me started on the video quality. Only video I've uploaded to FB Page. https://www.facebook.com/TwistedAddictionStormChasing

 

[James Wilson]

I've never been one to really count tornadoes, but if I did, I would use the technique that Ethan described. If a tornado lifts with the funnel still visible and then comes back down a few minutes later, I would not count that as 2 tornadoes. Now, if a tornado lifts, and the funnel dissipates completely, then another tornado forms, I would count that as 2 tornadoes.

Exactly how I count them ... I saw 23 tornadoes in 2015 counting that way if I went with the every touch method I would probably have quite a bit more but it would not be accurate or honest IMHO.

 

[John Farley]

I wondered that about Rozel, too. It looked to me like it totally disappeared and then a new funnel formed, but I assumed at the time that the survey called it one tornado, not two, because either someone closer than me saw a circulation on the ground during the time a funnel was not visible or the damage path was continuous (which of course could happen with two tornadoes if the second one formed right where the first one ended). Another chase day with similar ambiguities was Trinidad, CO, June 6, 2014. There were definitely two cycles of the supercell that each produced a large tornado, but before the large tornado with the second cycle, there was a narrow funnel that went partway down and kicked up dust for a minute or so, then it disappeared and there was nothing visible for a good five minutes. Then the large tornado formed. I was too far away to tell if there was any ground circulation during the time no funnel was visible, and there was some disagreement among other chasers who saw it on whether it was one tornado or two with the second cycle; the NWS counted it as two tornadoes total, i.e. one with each cycle. But it seems to me that if there really were 5 minutes between ground circulations with no evidence of a tornado during that time, then it would be more accurate to say there were two separate tornadoes with the second cycle.

I know it doesn't really matter much, but it is interesting to discuss. FWIW, I have seen some NWS damage surveys that talked about skipping tornadoes, but others that counted each damage segment as separate, even when they likely came from the same cycle.

 

[Michael Towers]

and with the wet fields, it would have been hard to make out a ground circulation unless one was viewing it super-close

or the damage path was continuous (which of course could happen with two tornadoes if the second one formed right where the first one ended)

I think the above might explain the determination of a single tornado while it was in fact two distinct tornadoes…two separate tornadoes happen in quick succession, there may be a short gap between the two paths but if so it may be considered due to the wet terrain masking any evidence. The conclusion is made that the event is one tornado when in fact it was two.

Was right there (Rozel) & don't have full account of this topic. Recorded video until I thought the tornadic rotation was done

I’d still like to see additional footage taken during the interval but like both of us I bet most chasers stopped filming at the conclusion of tornado one because they thought the tornado had ended.

What if the condensation funnel disappears, then reappears, and touches down again?

This is what I believe happened at Rozel but I only count it as two tornadoes due to the duration of time between funnels. If the cycle as you described happened quickly I’d assume it to be one continuous tornado.

What about a cyclic supercell dropping several tornadoes, all from new mesocyclones, but from the same regenerating storm?

Absolutely each would be a distinct tornado...but the counting may get lost in the delirium!

 

[Todd Krause]

I've never been one to really count tornadoes, but if I did, I would use the technique that Ethan described. If a tornado lifts with the funnel still visible and then comes back down a few minutes later, I would not count that as 2 tornadoes. Now, if a tornado lifts, and the funnel dissipates completely, then another tornado forms, I would count that as 2 tornadoes.

I certainly agree with that (assuming "few=4 minutes", but even that has a bit of wiggle room, since the paragraph below says "generally" for the 2 mile/4 minute rule).

Anyway, the NWS Storm Data Directive describes this. The link is here: http://www.nws.noaa.gov/directives/sym/pd01016005curr.pdf, with the tornado instructions on pages 74-86. There is a paragraph about how to count separate tornadoes on pages 77-78. Since the directive is lengthy, I'm copying/pasting the most pertinent paragraph here (although I'm not including section 7.40.9, which has some related info).

7.40.7 Determining Path Length and Width
“The tornado path length generally excludes sections without surface damage/disturbance, unless other evidence of the touchdown (e.g., a trained spotter report, videotape of the tornado over a plowed field, etc.) is available. The excluded section will generally not exceed 2 continuous miles or 4 consecutive minutes of travel time; otherwise, the path will be categorized as separate Tornado events. The beginning and ending locations of the excluded sections should be described as accurately as possible in the event narrative. In some cases, careful analysis and eyewitness descriptions will determine if separate tornadoes actually occurred within 2 miles or 4 minutes. Refer to Section 7.40.9 for related information. Use the event narrative to describe whether a tornado “skipped” or was continuous in these types of cases.”

For our office, video plays a big role in trying to figure this out. A spotter report can sometimes shed some light, but it's obviously best to get descriptions from multiple spotters. Reconciling all this can be challenging at times, especially when damage survey results are not straightforward. In those cases, we try to find as much chaser video as possible from various angles and form a coherent picture of what happened. Also, for cyclic tornadogenesis, we record them as separate tornadoes. And finally, if our office has missed video that shows something different, we're willing to update our information.

Todd
NWS Minneapolis/Chanhassen MN

 

[Michael Towers]

“The tornado path length generally excludes sections without surface damage/disturbance, unless other evidence of the touchdown (e.g., a trained spotter report, videotape of the tornado over a plowed field, etc.) is available. The excluded section will generally not exceed 2 continuous miles or 4 consecutive minutes of travel time; otherwise, the path will be categorized as separate Tornado events.

That would likely explain the assessment at Rozel as the distance of any gap would have been well within two miles. Perhaps a review of the above video reconciled with the survey would result in a change to the finding. Regardless, thanks for providing clarity for what is likely the basis of the one tornado determination.