El Reno Oklahoma tornado downgraded to EF3

[Bob Hartig]

Let's take a hypothetical solid-state tornado--mythical, a control-group tornado--with sustained winds of 250 mph. Mobile radar verifies that the winds remain that speed throughout its entire career. (Like I said, hypothetical.)

This tornado traverses nothing but open fields. It travels five miles and hits nothing before it dies, not even a solitary tree. It is rated EF-0.

Ten minutes later, the exact clone of this tornado forms at the edge of a large urban/suburban area, five miles of which it proceeds to devastate. Again, mobile radar shows sustained winds of 250 mph. Well-built homes are swept away. Rebar is snapped. The tornado is rated EF-5.

It is essentially the same tornado as the first. Nothing about it has changed except its location. In the urban area, there were plenty of DIs. In the open country, there were none, and therefore, presumably, no means of determining the tornado's true intensity.

Except there was a means: the mobile radar.

Once the ratings are assigned, along come the media, who invariably equate damage ratings with wind speeds in their reportage. "The tornado that hit the town has been rated an EF-5," they are told by an NWS official.
"Ah!" they say. "So the winds blew at over 200 mph?"
"That is correct. In fact, mobile radar verified winds of 250 mph.

"Now, about the previous tornado. That was an EF-0."
"Ah!" say the media. "So the winds blew between 65 and 85 mph."
"Weelll..."
One savvy reporter pipes up. "But mobile radar measured winds of 250 mph in that tornado too," she says.
"Nevertheless, the tornado was an EF-0."
At that point, if I were one of those reporters, my next question would be, "What the ----?"

I'm not advocating for making the media the determinant in these matters. But it makes no sense to create further confusion for them and then bash them when they don't get it right. They reflect the public, not the scientific community--a public who wants to know in simple terms. If the terms aren't simple enough, the public will do the simplifying with simplistic inaccuracy.

Perhaps a dual rating could be developed which sets actual damage next to damage potential, with wind speeds assigned to the latter. For example, the two tornadoes I've just described might be rated EF-0/5 and EF-5/5, respectively. That approach would answer two practical questions which I think lie at the heart of why the scale was devised in the first place: (1) What did this tornado do? and (2) What could it have done? I don't think that would be too hard for the media or the public to grasp. When actual wind speeds aren't known--which currently would be most of the time, but who knows how that will change--the second field would simply be left empty, thus: EF-3/? (or something similar).

I understand that there are arguments against what I'm suggesting; I could build my own case against it. It's certainly not as elegant as a single number, but the single number just doesn't seem to do the job. So here's something to consider as at least the germ of a solution in looking ahead. Because it seems to me that there is a need to look ahead and provide for future possibilities that are just beginning to emerge. Failure to make such provisions has already caught up with us.

 

[Rob H]

After seeing this argument linger for several months, I've come to the realization that everyone is very close to being on the same page. No one is arguing that tornadoes can only be measured by damage caused going forward. Everyone thinks it's a good idea to update the EF scale to incorporate measured winds when they're available. Everyone agrees that the flip-flopping is ridiculous, and will be even more so in a few years when El Reno gets its EF-5 rating back.

The real argument has always been whether or not measurements should provide ratings when matched with the wind speed estimates associated with DIs/DoDs on the EF scale *as is*. I've heard several valid arguments against this:

1. EF scale wind speed estimates have never been thoroughly validated, so measured winds probably don't match up with the DIs/DoDs perfectly
2. How low is low enough for measurements to be indicative of surface conditions? 100m off the ground? 250m? 500m?
3. There's been some recent work done that theorizes individual suction vortices may have 2-3x the wind speed as the mean wind flow. How does this affect damage?
4. There is location bias in that EF-2s and EF-3s in OK and KS might be upgraded to EF-5s, but storms in other parts of the country might not because of mobile doppler preferred "stomping grounds".
5. The EF rating is essentially worthless to the scientific community, but public policy is driven by it. We've already had one politican use the Moore and El Reno EF-5s in an attempt to cut climate research and privatize the weather industry. Any changes need to be well thought out.
6. It's an extremely small sample set. We're talking .1% of all tornadoes at best can have actual measurements. Just like the period that Graziulis back-rated, the period that Fujita rated, the modern age, and the EF age - we'll need to have a clean demarcation where policy has changed in case researchers are looking at ratings and associated tornadoes. It should be full bore as of this date, and none of this unsure wishy-washy back and forth that we have now.

I think everyone on both sides of this argument would be perfectly content with laying out a plan to address these topics (and others) and then either updating or replacing the EF scale.

 

[David Hoadley]

I wonder how this downgrading of the El Reno tornado and others, which apparently were once --and largely-- based on radar-determined wind speeds, will affect future grants to mobile radar operators? If their data is not considered conclusive, then what value does it have beyond experimental research?

 

[Dave Gallaher]

Perhaps this is too simplistic (and maybe too much trouble), but adding two characters to the existing EF designations could combine the two elements. If a V0 through V5 range was codified to represent radar-measured velocity at a fixed elevation range AGL, we could wind up with El Reno being designated as EF3V5, thus stating both damage and radar data in one acronym. This would simplify record keeping, at least.

 

[Michael Thomas]

Interesting discussion. Taking Doppler radar measurements aside, I wonder whether more information could be included within the EF scale. For example, using the classic example of tornado moving through an empty field and being designated an EF-0 - is this strictly true? Or is this best to record such a tornado as "equal to or greater than" EF-0 to indicate the lack of DI's? On the other hand, a tornado moving through a suburban area and only causing EF-0 damage in the presence of suitable DI's is simply an EF-0 tornado indicating confidence in this assignment.

 

[Rob H]

I wonder how this downgrading of the El Reno tornado and others, which apparently were once --and largely-- based on radar-determined wind speeds, will affect future grants to mobile radar operators? If their data is not considered conclusive, then what value does it have beyond experimental research?

Research, and digging into #1-3 in my post above, have tremendous value. Modifying the EF ratings for .1% of the tornadoes has much lower value, comparatively speaking. Also, I don't think anyone is saying mobile radar measurements are less accurate or conclusive than damage assessments using the EF scale. It's just kind of apples-to-oranges until the EF scale or procedures are formally updated.

Based on the above, I'd hope the future for mobile radar grants looks at least as good as it does today, if not better.

 

[Ebduncan]

I may not be a expert, but I have followed the weather my entire life.

The F scale, or the EF scale, to me are not accurate. The F scale is solely based on wind speed. While the EF scale is based on the damage caused. There is much more to a tornado than its wind speed. Much like Hurricanes the size also matters. A tornado which is over a mile wide will have wind speeds over a longer duration. This in turn will create more damage versus a tornado which isn't nearly as wide with the same wind speeds. You can have a EF0, or F1 tornado which is 2 miles wide have a much larger impact than a F4 or EF4 tornado which is only 100 yards wide. I think there should be a classification on size as well as wind speed which would best represent event the people experienced.

I don't understand why radar data cannot be used to classify a tornado. I understand that the wind speeds higher above ground level may not accurately describe wind speeds at ground level. There are many factors that come into wind speed at the ground and wind speed above the ground. Above the ground there is no restriction from trees, houses, and other objects to disrupt the wind. At the ground the wind is influenced by these factors, however, in some cases the wind speed can be amplified at ground level because it is being funneled. Strictly judging a tornado based on the damage can misrepresent the actual wind speeds due to these factors. Using wind speed data from radar and you assume those wind speeds actually reached the ground. This isn't conclusive in either case. The best case scenario would be to have radar wind speed data along with damage assessment at ground level.


Then again rating a tornado in the first place is more of a second hand thing. First hand your trying to warn the people in the path of the tornado so they can get to safety.

 

[Dave C]

Why don't hurricane ratings have this same problem? It seems hurricanes are pre categorized by wind speed alone without damage assessment (while some adjustment may take place later), and that severity scale generally serves the public just fine. Shouldn't that also be what the tornado scale is about - simplicity, and rapidly describing severity to the public? If I am ignorant about hurricane categorization and it IS as muddled as the EF scale, someone please enlighten me. I can't believe this much argument even exists over creating a simple indication of tornado severity - in simplicity, the F scale was much more useful. Seems like NWS is stuck on the "if a tree falls in the woods...." question and not focused enough on making information clear to the public and research community.

 

[Bob Hartig]

4. There is location bias in that EF-2s and EF-3s in OK and KS might be upgraded to EF-5s, but storms in other parts of the country might not because of mobile doppler preferred "stomping grounds".

6. It's an extremely small sample set. We're talking .1% of all tornadoes at best can have actual measurements. Just like the period that Graziulis back-rated, the period that Fujita rated, the modern age, and the EF age - we'll need to have a clean demarcation where policy has changed in case researchers are looking at ratings and associated tornadoes. It should be full bore as of this date, and none of this unsure wishy-washy back and forth that we have now.

I don't know why this topic has suddenly captured my attention the way it has these past few days. Now I can't seem to shut up about it.

Rob, per usual, you've offered some great thoughts. I think your point 6 speaks to your point 4: once things get hashed out (if and when that ever happens), there will need to be a clean demarcation, as you put it. That demarcation will have to accept regional bias as the price to pay for small, limited bursts of increased accuracy, and it will likewise have to accept that only a small number of tornadoes can presently be measured via remote sensing.

The thinking will have to be that it's better to have at least a few ratings that incorporate measured wind speeds--however haphazard and however focused on one geographic area those ratings and measurements may be--than it is to never, ever rise above a whole lot of uniformly debatable ratings which perpetuate the kind of discussion we've been having here. Over time, the regional bias may become more diffuse and the number of samplings increase. That would be my hope: that, given room to include better measurements and at least a few more truly accurate records, the number of those records would grow year by year--incrementally, no doubt, but small improvements are better than no improvements, and over time, the percentages will rise.

But it will take a decision to break with the status quo and do things differently. Obviously some legwork will be required in order to get there, but I think that's the mindset that will be necessary. At some point, the NWS will need to flip a definitive switch in order to move ahead.

 

[Dan Robinson]

It will be interesting to see how many more upgrades to 4 and 5 we'll see once the exclusivity of damage as a rating criteria is finally done away with. The number of 5s in recent history could triple. There have been many examples of high-end rural tornadoes rated on the low end of the scale. Here's just one example. It's interesting that this page makes reference to the fact that the tornado was sampled by one of the DOWs, but that these measurements could not be considered in the rating. And this was in the original F scale era, pre-EF scale.

http://www.srh.noaa.gov/lub/?n=events-2005-20050612

I also wonder if other methods such as photogrammetry could eventually be allowed. It seems like most significant/violent tornadoes are captured on video today, regardless of the location. While you can't see 'inside' of the tornado with video, it's possible to estimate the velocity of cloud material and debris on the edges.

 

[Mike Smith]

I also wonder if other methods such as photogrammetry could eventually be allowed.

There have been a number of tornadoes (Dallas, 1957; Great Bend, 1973) where extensive photogrammetric studies were done (see: http://journals.ametsoc.org/doi/pdf/10.1175/1520-0477(2001)082<0073:TTTAPA>2.3.CO;2 ). Again, the original Fujita Scale allowed for these. They would be an "estimated" wind speed with the lower case "f" nomenclature.

My opinion is that provided there are enough distinguishable cloud tags to do an adequate study, this is at least as accurate as estimates from damage.

 

[Kent Heiderich]

I like that answer Dan using "photogrammetry". Those suction & horizontal vortices could be measured. Only problem with that is what do we use at night, ir maybe?
Then there is the main question of the parent tornado. Maybe we need a new rating system that could include both in the future. It would appear as a EF rating/suction vortices.
I suppose a argument could be made to include average inflow winds & the rfd in smaller tornadoes that don't produce suction vortices. This is a very difficult equation to achieve in the weather community with the tools we have today.
Maybe somebody can make a "app" for this on our cell phones

 

[Michael Thomas]

I'm guessing "consistency" is meant in regard to climatology. Using mobile radar data would create a bias toward more strong/violent tornadoes in the Plains as compared to other regions because that's generally where the mobile radars are used.

4. There is location bias in that EF-2s and EF-3s in OK and KS might be upgraded to EF-5s, but storms in other parts of the country might not because of mobile doppler preferred "stomping grounds".

I would say that location bias is just something that has to be lived with - it already exists without the use of mobile dopplers. Here's a map of the population density of the US-
http://en.wikipedia.org/wiki/File:USA-2000-population-density.gif
Now a map of F4/EF4 and F5/EF5 tornadoes from 1950-2011-
http://www.ustornadoes.com/wp-conte...es-f4-ef4-and-f5-ef5-in-the-united-states.gif

There certainly has to be bias in the locations of F5/EF5 tornadoes. For example, Wisconsin (105.2 inhabitants per square mile) had 3 F5/EF5 tornadoes from 1950-2011 compared to 1 in South Dakota (10.86 inhabitants per square mile). Is this truly a reflection of the relative frequency of violent tornadoes between the two states or merely a reflection on the greater coverage of DI's in Wisconsin?

 

[A.Theunissen]

The future, not too far ahead will be about '' how to prevent the formation of a supercell '' . not in how to build better houses.

 

[rdale]

The future, not too far ahead will be about '' how to prevent the formation of a supercell '' . not in how to build better houses.

I'm not sure what that means, or how it applies to this discussion? You cannot prevent the formation of a supercell.