Kds86z
Member
- Messages
- 4,297
- Location
- Greencastle, PA
Rochelle was 50 Miles from my sister. My brother in law is an officer and was on the scene. To close!
-
Welcome to TalkWeather! We see you lurking around TalkWeather! Take the extra step and join us today to view attachments, see less ads and maybe even join the discussion. CLICK TO JOIN TALKWEATHER
Significant Tornado Events
- Thread starter locomusic01
- Start date
CheeselandSkies
Member
- Messages
- 3,840
- Location
- Madison, WI
Also on April 9, in 1953 (also in Illinois), the link between tornadoes and the "hook echo" radar reflectivity signature was first discovered.
Tornadoes of 1953 - Wikipedia
SpotlightForRareTornadoes
Member
- Messages
- 299
- Location
- Byron, CA
5 days ago was the Two-year anniversary of the Lewistown Chris Riske incident.
Don't let frustration get the best of you.
Don't let frustration get the best of you.
SpotlightForRareTornadoes
Member
- Messages
- 299
- Location
- Byron, CA
Tomorrow is Terrible Tuesday, 46 years later.
This video nearly aged like wine!
This video nearly aged like wine!
locomusic01
Member
- Messages
- 1,588
- Location
- Pennsylvania
Also a fairly long track (almost exactly 22 mi) for such a classic, slow-moving plains grinder. Undoubtedly not 100% perfect because I moved on before I got too deep into researching it, but it should be reasonably accurate (also missing several death markers):
It didn't cause a ton of damage outside of Leedey itself because the area was so sparsely populated, but it seems to have been very intense throughout most of its lifespan. Also probably wider at points than I have it here (between a quarter and half a mile), but I didn't have enough info or damage points to accurately estimate in most places outside of town.
Lake Martin EF4
Member
- Messages
- 935
- Location
- Oakland, Tennessee
OTD 10 years ago...
And 78 years ago...
stormstalker.wordpress.com
And 78 years ago...
April 9, 1947 – The Woodward Tornado
[Click photos to view larger versions.] The morning of April 9, 1947 dawned cool, breezy and decidedly gloomy across the Southern Plains. Thick fog descended like a blanket, reducing visibility to …
stormstalker.wordpress.com
Aaron Rider
Member
- Messages
- 324
- Location
- Pennsylvania
Of course that also means tonight is the anniversary of the disappearance of Joan Gay Croft, one of the most perplexing and creepy cold cases ever
This video seems to suggest she may have just died that night. The whole story her sister was told was that Joan was "taken to Oklahoma City", which sounds suspiciously like something an adult would tell a little child. Regardless, very creepy.
Decided to go a bit more into what Fujita's ground marking method was and what it was trying to do.
One feature of tornadoes is a transition between air flow in the boundary layer near the ground and that in the vortex above, which WS Lewellen termed the 'corner flow' region. Fluid dynamics ain't my field, but there are plenty of articles out there for those who do understand it. The upshot is that the near ground air flow in tornadoes changes direction rapidly from inwards to upwards. This seems to be part of why they're so destructive. It can generate high vertical velocities, so tornadoes easily lift objects in a manner that tropical cyclones and other extreme winds don't.
This abrupt change means that objects being carried in the flow near the ground can be left behind. The region of strongest corner flow isn't uniform due to turbulence etc. This means tornadoes have areas where debris are swept up and deposited. Some tornadoes have the right flow structure and encounter the right ground conditions (such as loose crop stubble) to leave behind debris piles that are more or less cycloidal, and it is those marks that various authors attempted to use to measure the wind speeds of tornadoes.
Regardless of the exact method, they involve assuming the region where a given mark is formed is compact enough that is can effectively conceived as 'drawing' it and that this is representative of the actual windspeed. Unfortunately the thesis I mentioned in my previous post doesn't clearly answer these questions. But one interesting thing it does conclude is that unlike what Fujita thought, you don't need subvortices to produce cycloidal marks, nor do the marks necessarily coincide with them when they are present. The methods also involve assuming that the marks are regular i.e. maintain the same relative distance from the tornado centre (though Fujita seems to have a way around this) which like some other proposed non-standard indictors assumes tornadoes are axisymmetric (they're not).
Fujita defined the translational speed as U and the tangential speed (speed of rotation) as V, with the radius between the mark and tornado centre as R. He defined the cycloid with the x axis as the tornado path for y = R sin ωt and x = Ut+R cos ωt (ω - angular velocity, t - time from initial x-axis crossing). Taking the rotation (turning) number for a mark c = ωt/2r, substituting it into the cycloid equation and differentiating gets dy/dx = (dy/dt)/(dx/dt) = (V cos 2πc)/(U–V cos 2πc) = (n cos 2πc)/(1–n sin 2πc) where n = V/U. A loop is created when n exceeds 1 (which I think would technically make it a prolate trochoid) whose width can be defined as the difference in x when the slope changes from ∞ to -∞, when 1–n sin 2πc = 0 or 2πc = sin^-1(1/n) . The time for this rotation angle is t = 2πc/ω = 2πcR/V and using the first equation obtains
. The width of the loop (w) is therefore 
. Finally, the relative loop width (W) can be defined 
which varies only with n.
Here calculating the wind speed seems simple provided the translational speed is known with reasonable accuracy, it's simply the tangential speed plus the translational speed, U+V or U+U*n.
However, Fujita continues. He describes a case "where a large number of suction marks appear, [so] it is very difficult to identify the loops produced by a single suction spot that has rotated more than once around the tornado center". It then moves to talking about loop shift (s) and calculating relative loop shift S = s/2R = 2πR/n = π/n, and that as a result s can be found from w. But I'm not sure why it's necessary to calculate S considering wind speed can be found from w only.
He then applies it to the Greentown tornado (L2). I had to read it a couple of times to understand it. By graphing out the estimated start and end times vs distance he shows that the speed of tornado family L was fairly steady, so takes the 62.5 mph as the tornado's translational speed (U). By examining the track he concludes the marks made seven revolutions around the centre in the period shown in the photo (presumably this is where s or S comes in?). He measures the distance between the loop tops and the estimated centre of the track. Then he groups the marks by the seven locations and uses the calculated speeds from w (I think to calculate and average time of revolution and from that the average tangential speed. This is the weakest part of the procedure for me, it seems to be for calculating an average. I'm not sure why it's necessary and it's not well explained, he talks about the variation in core size and the distance of the marks from the centre, but never explicitly says the averaging is to account for this. Finally he adds the translational speed on to arrive at a windspeed for the seven locations: 172, 176, 173, 180, 180, 173 and 166 mph.
(TBC)
One feature of tornadoes is a transition between air flow in the boundary layer near the ground and that in the vortex above, which WS Lewellen termed the 'corner flow' region. Fluid dynamics ain't my field, but there are plenty of articles out there for those who do understand it. The upshot is that the near ground air flow in tornadoes changes direction rapidly from inwards to upwards. This seems to be part of why they're so destructive. It can generate high vertical velocities, so tornadoes easily lift objects in a manner that tropical cyclones and other extreme winds don't.
This abrupt change means that objects being carried in the flow near the ground can be left behind. The region of strongest corner flow isn't uniform due to turbulence etc. This means tornadoes have areas where debris are swept up and deposited. Some tornadoes have the right flow structure and encounter the right ground conditions (such as loose crop stubble) to leave behind debris piles that are more or less cycloidal, and it is those marks that various authors attempted to use to measure the wind speeds of tornadoes.
Regardless of the exact method, they involve assuming the region where a given mark is formed is compact enough that is can effectively conceived as 'drawing' it and that this is representative of the actual windspeed. Unfortunately the thesis I mentioned in my previous post doesn't clearly answer these questions. But one interesting thing it does conclude is that unlike what Fujita thought, you don't need subvortices to produce cycloidal marks, nor do the marks necessarily coincide with them when they are present. The methods also involve assuming that the marks are regular i.e. maintain the same relative distance from the tornado centre (though Fujita seems to have a way around this) which like some other proposed non-standard indictors assumes tornadoes are axisymmetric (they're not).
Fujita defined the translational speed as U and the tangential speed (speed of rotation) as V, with the radius between the mark and tornado centre as R. He defined the cycloid with the x axis as the tornado path for y = R sin ωt and x = Ut+R cos ωt (ω - angular velocity, t - time from initial x-axis crossing). Taking the rotation (turning) number for a mark c = ωt/2r, substituting it into the cycloid equation and differentiating gets dy/dx = (dy/dt)/(dx/dt) = (V cos 2πc)/(U–V cos 2πc) = (n cos 2πc)/(1–n sin 2πc) where n = V/U. A loop is created when n exceeds 1 (which I think would technically make it a prolate trochoid) whose width can be defined as the difference in x when the slope changes from ∞ to -∞, when 1–n sin 2πc = 0 or 2πc = sin^-1(1/n) . The time for this rotation angle is t = 2πc/ω = 2πcR/V and using the first equation obtains
. The width of the loop (w) is therefore . Finally, the relative loop width (W) can be defined which varies only with n.Here calculating the wind speed seems simple provided the translational speed is known with reasonable accuracy, it's simply the tangential speed plus the translational speed, U+V or U+U*n.
However, Fujita continues. He describes a case "where a large number of suction marks appear, [so] it is very difficult to identify the loops produced by a single suction spot that has rotated more than once around the tornado center". It then moves to talking about loop shift (s) and calculating relative loop shift S = s/2R = 2πR/n = π/n, and that as a result s can be found from w. But I'm not sure why it's necessary to calculate S considering wind speed can be found from w only.
He then applies it to the Greentown tornado (L2). I had to read it a couple of times to understand it. By graphing out the estimated start and end times vs distance he shows that the speed of tornado family L was fairly steady, so takes the 62.5 mph as the tornado's translational speed (U). By examining the track he concludes the marks made seven revolutions around the centre in the period shown in the photo (presumably this is where s or S comes in?). He measures the distance between the loop tops and the estimated centre of the track. Then he groups the marks by the seven locations and uses the calculated speeds from w (I think to calculate and average time of revolution and from that the average tangential speed. This is the weakest part of the procedure for me, it seems to be for calculating an average. I'm not sure why it's necessary and it's not well explained, he talks about the variation in core size and the distance of the marks from the centre, but never explicitly says the averaging is to account for this. Finally he adds the translational speed on to arrive at a windspeed for the seven locations: 172, 176, 173, 180, 180, 173 and 166 mph.
(TBC)
Last edited:
- Messages
- 630
- Location
- Columbus, OH
Fascinating post, haven’t seen Fujita’s direct calculations until now. I couldn’t remember what a prolate trochoidal shape was until I looked it up again and I’ve gotta say, I can definitely recall at least some cases where the ground scouring was definitely of that exact shape. I can recall many different and interesting shapes calved into the ground from scouring but that type of shape definitely stands out to me.
It is quite interesting to see that he attempts to calculate windspeed based on the spacing between the loops W in conjunction with the value w. I’m not entirely sure why either, the only thing I can think of is him incorporating some sort of phase shift* (to anyone else reading, see below if don’t know what this means) into the equation, but again, I can’t really think of a reason why that is necessary off the top of my head outside of calculating a sound value for U. This phase shift idea sounds a little weird too considering the fact that there’s multiple examples of ground scouring patterns overlapping with itself, so it can’t be the same part of the corner flow region in every instance. But modeling it this way is really smart because it gets around a lot of the extremely complex wind dynamics, obviously at the cost of some accuracy, but it’s probably the best we can do with classical mechanics outside of windspeed calcs from damage indicators.
It is worth mentioning that this appears to be a model based purely in 2D Cartesian/Radial and somewhat neglects the diagonal upward component of motion at this corner flow region. Is this model taking the corner flow region as a single point, and is it marking that as the instantaneous transition point between horizontal-to-vertical wind? Would it be better to model this in cylindrical coordinates, at the cost of some sanity? On the surface, I like it’s definitely doable but it would certainly become very nontrivial. Forgive me if I’m misunderstanding this part. If I’m understanding this correctly, I imagine n > 1 in practically all cases for tornadoes outside of EF0-1 tornadoes, which typically do not scour the ground anyways.
* For laymen: a phase shift is defined as a displacement of a wave along its direction of propagation (where it’s going).
Yes I 100% agree. Of all the things to censor on a website about weather events, this one makes zero sense. Who cares if someone calls an event a bu$t during or after? They’ll appropriately be called out if they’re wrong.
Aaron Rider
Member
- Messages
- 324
- Location
- Pennsylvania
Notable events today in tornado history:
1979 - the monstrous Wichita Falls F4 highlighted a severe Red River Valley outbreak - this is absolutely one of the scariest looking twisters I've seen in photographs; I think most folks here are pretty familiar with this one
1929 - A strong outbreak ripped into Arkansas, in much of the same areas recently affected. The biggest monster tore a long, violent track through the countryside of Jackson County. Touching down just outside of Batesville, it eventually grew into nearly half a mile in size before utterly destroying the communities of Possum Trot (good old Arkansas...) and Sneed at incredible intensity. A bunch of denuded trees was all that was left of the school in Pleasant Valley.
Not knowing exactly at which angle the visible tornado was traveling, an engineer on the Missouri Pacific Railroad sped his train up to 75 MPH (which, needless to say, is pretty much HIGHBALLING in old railroad parlance) to successfully escape the vortex.
"The tornado was like judgment day - that's the only way I know to tell it." - Claude Cook, quoted in the Daily World of Helena, Arkansas
1979 - the monstrous Wichita Falls F4 highlighted a severe Red River Valley outbreak - this is absolutely one of the scariest looking twisters I've seen in photographs; I think most folks here are pretty familiar with this one
1929 - A strong outbreak ripped into Arkansas, in much of the same areas recently affected. The biggest monster tore a long, violent track through the countryside of Jackson County. Touching down just outside of Batesville, it eventually grew into nearly half a mile in size before utterly destroying the communities of Possum Trot (good old Arkansas...) and Sneed at incredible intensity. A bunch of denuded trees was all that was left of the school in Pleasant Valley.
Not knowing exactly at which angle the visible tornado was traveling, an engineer on the Missouri Pacific Railroad sped his train up to 75 MPH (which, needless to say, is pretty much HIGHBALLING in old railroad parlance) to successfully escape the vortex.
"The tornado was like judgment day - that's the only way I know to tell it." - Claude Cook, quoted in the Daily World of Helena, Arkansas
(from)
The only other explicit speed from Fujita I've found (so far) using this method is from an unpublished 1967 paper. He examined the 1967 Belvidere IL tornado near Woodstock and from two marks found 200 and 210 mph. He rated the 1990 Goessel tornado F5 using this method (as mentioned in his book) but doesn't give the exact numbers. Even though he regularly notes cycloidal ground markings I haven't found any more windspeed calculations.
When he first conceived of 'suction spots' he thought that the marks were created by convergence under the centre of a vortex. Interestingly, Zimmerman claims this is impossible, because a tornado with a low enough swirl ratio that the maximum corner flow is in the centre would be too weak to pick up anything, and that the effect of surface roughness means such low swirl ratios are impossible anyway, hence the maximum corner flow is always annular. When Fujita found that there were rotating suction vortices (with Greg Forbes apparently finding one in the Parker City tornado rotating 39 m/s) he found that adding 20-30 m/s would get higher speeds more in line with the F scale. But as mentioned before the necessary corner flow doesn't require multiple vortices (found using simulations far more complex than those available to Fujita).
The other published contemporary application I've seen of this method was by Ernest Agee, who used it for the West Lafayette tornado in 1976, in a region of F3-F4 damage. He uses it slightly differently and more directly. He concluded (from his own photogrammetry of Parker City) that the vortices moved much quicker relative to the centre on the left hand side of the tornado than on the right. He takes windspeeds calculated from direct measurements of s as averages and those from w as maximums. That he does find a substantial difference in the values of n between them has the interesting implication that the tracks are not regular cycloids/trochoids, but are somewhat distorted.
Adding in the translational velocity, the highest windspeed calculated is 83.1 m/s, or 186 mph. One notable thing is that the highest calculated widspeeds happen to be where the damage was most intense as well.
Finally, a few days ago I found that this poster presented at last year's Conference on Severe Local Storms, with some of the usual suspects when it comes to non-conventional ratings methods:
An interesting thing is that estimated windspeed based on ground markings fall consistently with the windspeed ranges implied by the EF scale, and if anything towards the lower end of that. The poster above uses it to back EF assessment obtained ratings. This is in contrast to the data obtained from radar which tends to show significantly higher winds even at low levels. Hence you have this position obtained from a few unconventional indicators above vs the Wurman/Lyza position that the EF scale significantly underestimates the actual windspeed based on radar data. Of course there are limitations to those methods and Lombardo's treefall method produces lower speeds than, say, Karstens'.
A potential confounder is that even if the ground marks are representative, we don't know how they reflect windspeed at the 10-metre height. The profile is somewhat contentious, because we don't have direct measurements. It can be modelled but aside from sensitivity to roughness, slip conditions etc. you don't see papers presenting the wind profile at, say, a centrimetric scale across the lowest metre (would need a lot of computing power). Zimmerman uses a 1-metre vertical resolution. Since the debris and deposition features are only a few centimetres high this could be important.
The only other explicit speed from Fujita I've found (so far) using this method is from an unpublished 1967 paper. He examined the 1967 Belvidere IL tornado near Woodstock and from two marks found 200 and 210 mph. He rated the 1990 Goessel tornado F5 using this method (as mentioned in his book) but doesn't give the exact numbers. Even though he regularly notes cycloidal ground markings I haven't found any more windspeed calculations.
When he first conceived of 'suction spots' he thought that the marks were created by convergence under the centre of a vortex. Interestingly, Zimmerman claims this is impossible, because a tornado with a low enough swirl ratio that the maximum corner flow is in the centre would be too weak to pick up anything, and that the effect of surface roughness means such low swirl ratios are impossible anyway, hence the maximum corner flow is always annular. When Fujita found that there were rotating suction vortices (with Greg Forbes apparently finding one in the Parker City tornado rotating 39 m/s) he found that adding 20-30 m/s would get higher speeds more in line with the F scale. But as mentioned before the necessary corner flow doesn't require multiple vortices (found using simulations far more complex than those available to Fujita).
The other published contemporary application I've seen of this method was by Ernest Agee, who used it for the West Lafayette tornado in 1976, in a region of F3-F4 damage. He uses it slightly differently and more directly. He concluded (from his own photogrammetry of Parker City) that the vortices moved much quicker relative to the centre on the left hand side of the tornado than on the right. He takes windspeeds calculated from direct measurements of s as averages and those from w as maximums. That he does find a substantial difference in the values of n between them has the interesting implication that the tracks are not regular cycloids/trochoids, but are somewhat distorted.
Adding in the translational velocity, the highest windspeed calculated is 83.1 m/s, or 186 mph. One notable thing is that the highest calculated widspeeds happen to be where the damage was most intense as well.
Finally, a few days ago I found that this poster presented at last year's Conference on Severe Local Storms, with some of the usual suspects when it comes to non-conventional ratings methods:
An interesting thing is that estimated windspeed based on ground markings fall consistently with the windspeed ranges implied by the EF scale, and if anything towards the lower end of that. The poster above uses it to back EF assessment obtained ratings. This is in contrast to the data obtained from radar which tends to show significantly higher winds even at low levels. Hence you have this position obtained from a few unconventional indicators above vs the Wurman/Lyza position that the EF scale significantly underestimates the actual windspeed based on radar data. Of course there are limitations to those methods and Lombardo's treefall method produces lower speeds than, say, Karstens'.
A potential confounder is that even if the ground marks are representative, we don't know how they reflect windspeed at the 10-metre height. The profile is somewhat contentious, because we don't have direct measurements. It can be modelled but aside from sensitivity to roughness, slip conditions etc. you don't see papers presenting the wind profile at, say, a centrimetric scale across the lowest metre (would need a lot of computing power). Zimmerman uses a 1-metre vertical resolution. Since the debris and deposition features are only a few centimetres high this could be important.
Last edited:
Lake Martin EF4
Member
- Messages
- 935
- Location
- Oakland, Tennessee
OTD 46 years ago...
On another note, I want to bring this post by Marshall up:
I really do wonder what convinced him to think that Vernon was an F5 candidate. I'm not dismissing it or anything, I genuinely want to know more about it.
On another note, I want to bring this post by Marshall up:
I really do wonder what convinced him to think that Vernon was an F5 candidate. I'm not dismissing it or anything, I genuinely want to know more about it.
I do think the calculation of wind speeds using the cycloidal swath method can serve as a valuable supplement. However, as you rightly pointed out, its greatest limitation lies in the fact that it almost exclusively captures near-surface wind speeds, likely at heights ranging from just one centimeter to ten centimeters. We know that wind speed at ground level (height zero) is zero, and near-surface winds are significantly affected by friction, which reasonably explains why this method yields lower values compared to mobile radar measurements—a point supported by some research findings (Lewellen 1999).
Interestingly, we now know that at least a portion of a tornado's wind speeds can be measured at very low altitudes (Wurman 2013), such as between 5 to 20 meters, which happens to be the height range specified by the Enhanced Fujita (EF) scale. However, mobile radars typically cannot scan winds below this altitude, while the cycloidal swath method reflects wind speeds even closer to the ground. This highlights just how challenging it is to determine tornado wind speeds.
Lake Martin EF4
Member
- Messages
- 935
- Location
- Oakland, Tennessee
A few days to perhaps consider:
5/19/10 (30% sigtor, max rating was EF1)
4/27/11 (the northern mod risk in TN/KY/IN/OH only)
4/27/14 (30% sigtor resulted in a single EF5 and nothing else before the high risk was discontinued. Only 18 tornadoes overall)
12/23/15 (major outbreak in Alabama expected, Alabama basically got nothing at all though a likely EF5 did hit Holly Springs on this day)
4/2/17 (30% sigtor, max rating was EF2)
4/5/17 (30% sigtor, basically the exact same result as 4/2/17)
5/18/17 (30% sigtor, rain basically snuffed out the entire High Risk area just as the favorable conditions arrived)
3/17/21 (45% sigtor, only a few EF2s resulted - this was brought up by CheeselandSkies)
And before 2010 (note: not all were High Risks, but most were):
5/2/99 (major outbreak expected, only a few tornadoes happened, then Bridge Creek happened the next day)
6/5/99 (High Risk, highest rating was F1)
12/23/02 (even Wikipedia calls this a bu$t!)
4/11/05 (SPC basically predicted Super Tuesday 2008 3 years early, 3 F0s formed)
4/6/06 (I think I recall this being mentioned as an example of a potent setup busting on Stormtrack)
4/7/06 (the 60% sigtor day, though it can be argued it verified from a numbers standpoint, just not a violent tornadoes standpoint)
4/13/07 (DFW high risk, only a few weak tornadoes confirmed)
4/24/07 (literally none of the tornadoes happened in the High Risk)
6/5/08 (30% tornado risk completely killed by blowoff, one chaser apparently called it the bu$t of the century)
4/26/09 (spawned the Roll tornado and very little else)
I'll likely add more later
Last edited:
CheeselandSkies
Member
- Messages
- 3,840
- Location
- Madison, WI
High risk "bu$ts" occurred with regularity in the '90s and '00s when they were both issued more frequently (IMO SPC's batting average during this era was still pretty darn good) and NWP wasn't as advanced as it is now (and hi-res CAMs not even really a thing). Nowadays, because the category is so rarely used, it makes the underperformers stand out even more.
I know he’s looked at as sort of the patron saint/meme of the Wx Community, but doesn’t Broyles have an extremely good batting average on high risks? Last time I checked, it was like near 80% (11/14).
Aaron Rider
Member
- Messages
- 324
- Location
- Pennsylvania
Obviously today is the 60th anniversary of the Palm Sunday super outbreak.
1 note on this: According to a Twitter post a while back, Trey from Convective Chronicles was planning on releasing a video about this outbreak. Hopefully he had the time!
Can anyone think of any other significant outbreaks with the combination of 1) relatively low CAPE values but 2) absolutely preposterously high bulk shear?
For those who haven't heard this before, here is the terrifying audio from the church in Alto, Indiana that was destroyed by the Russiaville-Greentown tornado:
And here is a 17-year old video I recently found where a grandfather gives his grandson a brief tour of the destruction around Sheridan, Indiana:
1 note on this: According to a Twitter post a while back, Trey from Convective Chronicles was planning on releasing a video about this outbreak. Hopefully he had the time!
Can anyone think of any other significant outbreaks with the combination of 1) relatively low CAPE values but 2) absolutely preposterously high bulk shear?
For those who haven't heard this before, here is the terrifying audio from the church in Alto, Indiana that was destroyed by the Russiaville-Greentown tornado:
And here is a 17-year old video I recently found where a grandfather gives his grandson a brief tour of the destruction around Sheridan, Indiana:
Aaron Rider
Member
- Messages
- 324
- Location
- Pennsylvania
You know, his list really isn't that bad. But, in the comments, he says that Rochelle's contextual damage did not impress him. I don't know much about Coleridge '03 or certainly Knox City '53, but I do know nothing they did was MORE impressive than Rochelle. Ditto Picher, which many here actually think should have been EF5 too.
To be honest, I've never studied the Shawnee, OK tornado from 2013 either. That's interesting.
All I have to say to his comments about Rochelle is, As I said elsewhere about Pittsfield and Sunnyside in '65, seriously, what more does a tornado have to do to earn the highest rating? Rochelle did everything it takes. This is not some Moshannon or Bassfield "destroyed entire forests but didn't hit anything major in terms of human housing" or Mayfield "where it did destroy a good home, contextual damage was juuuuuust lacking enough you could argue against EF5" deal.
Put it this way. Even people who (very persuasively) argue that EF4 was appropriate argue equally firmly that Rochelle was clearly EF5.
As for Vernon 1979, here's an article about it: https://www.vernonrecord.com/2019/04/11/terrible-tuesday-remembered-the-story-of-a-photograph/
Despite my negative opinion of his opinion of Rochelle, let me see if I can reach out to him about Vernon. I briefly tried to find old newspaper articles on that one but failed. Most attention seems to have been given, of course, to the Wichita Falls F4. Knox City '53 is also foreign to me.
Palm Sunday 1965 is just so anomalous to me. Kind of fitting in with that favorable upper Midwest pattern during the mid 19th century. When you look at the annals of tornado history, you see history doesn’t repeat but it does rhyme. You can see super outbreak type events that generally are similar (4/3/74, 4/27/11, 1932, Enigma) and events with more than 10+ violent tornados.
Then there’s Palm Sunday 1965. From the blizzard storm like jet core, geographical location, and extremely high rate of violent tornados (which are probably undercounted judging by research). I’m not saying it’s a one off event, but there’s no real close comparison throughout recorded tornado history due to how new our records are. The amount of violent tornados that day is staggering