Assigning EF Scale Ratings Based on Radar Indicated Wind Speeds

[JoMo]

Since you all are talking about damage and the rating scale. I thought I'd post this in here as well:

 

The ASCE report for the Joplin tornado has been released. It found that most of the structures in the path were poorly constructed. 83% of the damage was EF-2, 14% of the damage was EF-3, and 4% of the damage was EF-4. There were no structures built to withstand an EF-5 in the path. Flying debris made the damage much worse. Maximum winds were still near 200-208 MPH. The ASCE recommended hurricane straps to tie the roof on, which is something all new homes in Joplin must have now.

 

http://www.joplinglobe.com/topstories/x120729257/Civil-engineers-release-study-of-Joplin-tornado-damage

[thewxmann]

A lot of behind the scenes wrangling going on the last few days that reach far beyond OUN. Would not be surprised if this is downgraded.

 

Would Rozel, Bennington, and Wichita also be downgraded?

[AtticaFanatica]

A lot of behind the scenes wrangling going on the last few days that reach far beyond OUN. Would not be surprised if this is downgraded.

Would Rozel, Bennington, and Wichita also be downgraded?

I don't know, but that would be one way to lessen the embarrassment for OUN.

[andyhb]

From Dr. U.

 

 

Directive NWSI 10-1604, Post-Storm Data Acquisition, requires us to use the EF scale, which is an impact-based rating assigned to a tornado after extensive investigation of the damage it caused. EF ratings are determined by observed damage rather than measured wind because we have no consistent way to measure wind speed for every tornado that occurs. Adhering to NWSI 10-1604 ensures we continue to use consistent methodology throughout the country for assigning EF ratings.

Tornado research is an exciting and rapidly evolving area of science, and we are able to capture more information about the character of tornadoes than ever before. We are updating Directive NWSI 10-1604 to allow the option of including this new information, when available, in the narrative of tornado summaries. The new policy will allow NWS to document available data that are scientifically valid and reliable without changing the objective and consistent EF assessment. Until the update is finalized into policy, WFOs should continue to follow current policy.

[tornadotony]

From Dr. U.

With all due respect to Dr. U. (and plenty is due), the EF-scale is not simply an impact-based scale.  If it were, we wouldn't have the lower-bound/expected/upper-bound ranges for each DOD.  For instance, if you have two houses swept away, the "impact" is the same, even if one house is secured by being bolted and strapped to its foundation and one has no anchoring.  In both cases, the house is swept away, but I don't think there's any disagreement that the force required for the same impact to both houses is vastly different or that the difference is to be considered in the EF-scale rating.  I'd venture that there's no doubt amongst people here that the unanchored house would likely barely qualify for an EF3 rating, whereas the bolted/strapped house would likely require EF5 consideration.  The Pandora's box of trying to estimate actual tornado intensity is already open.  I would hope in the future, if we can ascertain a better relationship between wind speeds and damage (a long time from now), some of these ratings that look like they'll be reversed here are given a second look.

 

Dr. Chuck Doswell's thoughts on this (which happen to fit my feelings very well): http://cadiiitalk.blogspot.com/2013/06/the-ef-scale-ratings-brouhaha.html

[andyhb]

Knew that would be good when I saw the word "brouhaha" in the link, and it was.

 

Also, this is basically right on target with what is being discussed here.

 

Denying the value of the most direct measurements of windspeed in tornadoes in order to maintain consistency with an inconsistent data set strikes me as silly.

[Stebo]

With all due respect to Dr. U. (and plenty is due), the EF-scale is not simply an impact-based scale.  If it were, we wouldn't have the lower-bound/expected/upper-bound ranges for each DOD.  For instance, if you have two houses swept away, the "impact" is the same, even if one house is secured by being bolted and strapped to its foundation and one has no anchoring.  In both cases, the house is swept away, but I don't think there's any disagreement that the force required for the same impact to both houses is vastly different or that the difference is to be considered in the EF-scale rating.  I'd venture that there's no doubt amongst people here that the unanchored house would likely barely qualify for an EF3 rating, whereas the bolted/strapped house would likely require EF5 consideration.  The Pandora's box of trying to estimate actual tornado intensity is already open.  I would hope in the future, if we can ascertain a better relationship between wind speeds and damage (a long time from now), some of these ratings that look like they'll be reversed here are given a second look.

 

Dr. Chuck Doswell's thoughts on this (which happen to fit my feelings very well): http://cadiiitalk.blogspot.com/2013/06/the-ef-scale-ratings-brouhaha.html

 

I completely agree with his and your thoughts on the situation. To me it isn't doing science justice to ignore the data that was collected for the sake of being consistent with older methods.

[OceanStWx]

The purpose has always been to estimate the wind speeds in the tornado. Always. The need for DIs to infer intensity is an unfortunate necessity built upon the inherent small spatial scale and short time scale of the phenomenon.

The intensity database is essentially crap, it might generally be accurate to within +/- 1 EF scale. I don't understand the obsession with willfully keeping it that way in the name of consistency.

Let's put it this way: what scientific question can we better answer by closing our eyes, covering our ears, and telling everybody it was an EF3?

 

 

You many know better than I, but I seem to remember a presentation making the rounds from CSWR a few years ago that hypothesized something like most tornadoes fall around an EF2 strength based on radar sampling. Basically that many EF0 and EF1s are rated as such because of a lack of DIs.

 

Now there could be some sort of confirmation bias going on by only chasing the stronger supercells, thus stronger potential tornadoes. Or I could not correctly remember the presentation I saw.

[OceanStWx]

My first reaction to the upgrade, was if we have the data why not use it (at the time I hadn't even thought to read the NWS directive regarding storm damage surveys).

 

It does seem silly to ignore the added input for the sake of consistency.

[Ensō]

You many know better than I, but I seem to remember a presentation making the rounds from CSWR a few years ago that hypothesized something like most tornadoes fall around an EF2 strength based on radar sampling. Basically that many EF0 and EF1s are rated as such because of a lack of DIs.

 

Now there could be some sort of confirmation bias going on by only chasing the stronger supercells, thus stronger potential tornadoes. Or I could not correctly remember the presentation I saw.

 

I recall seeing a paper that came to that conclusion but I can't remember what it was. I'll see if i can find it, or maybe someone else will know.

[AtticaFanatica]

You many know better than I, but I seem to remember a presentation making the rounds from CSWR a few years ago that hypothesized something like most tornadoes fall around an EF2 strength based on radar sampling. Basically that many EF0 and EF1s are rated as such because of a lack of DIs.

 

Now there could be some sort of confirmation bias going on by only chasing the stronger supercells, thus stronger potential tornadoes. Or I could not correctly remember the presentation I saw.

 

 

I recall seeing a paper that came to that conclusion but I can't remember what it was. I'll see if i can find it, or maybe someone else will know.

Correct, tornadoes have a preferred intensity of EF2, quite different from what damage surveys tell us.

See Figs. 7 & 8 here: https://ams.confex.com/ams/pdfpapers/141821.pdf

I believe this work is in peer review currently.

[Ground Scouring]

True, SPC isn't involved directly in the assigning of ratings, however NWS HQ can dictate to the office what it can be rated and Dr U sent out an email to RDs a couple days ago reminding them that directives dictate that the EF-scale ratings be based on damage not radar data. He also re-iterated several times that the EF-5 is preliminary, hinting that the final rating may be based on damage (EF-3) with a note in Storm Data that higher winds were measured. 

Another point: how can the SPC decide whether or not there was true EF5 damage, particularly if further research allows more non-conventional EF5 markers--based upon scouring relative to soil type, mangling of vehicles, etc.--to be incorporated into the EF scale in the future? Now imagine the discovery that the El Reno tornado, after being downgraded, actually did produce EF5 indicators somewhere, even if not to structures themselves.

[Ensō]

 

 

Correct, tornadoes have a preferred intensity of EF2, quite different from what damage surveys tell us.

See Figs. 7 & 8 here: https://ams.confex.com/ams/pdfpapers/141821.pdf

I believe this work is in peer review currently.

 

That's the one, thanks Attica. I think we'd be surprised at the results if we somehow had the means to observe wind speeds for all tornadoes. Not to say that we'd have EF5s every week or anything, but I think it's pretty obvious damage surveys inherently underrate tornadoes, possibly by a pretty sizable margin.

[gymengineer]

I just read the thread more carefully about Ucellini's directive so I deleted the question I had:

In case anyone hadn't seen yet, the El Reno tornado is listed as EF3 as of 6/10 by the SPC.

[andyhb]

I just read the thread more carefully about Ucellini's directive so I deleted the question I had:

In case anyone hadn't seen yet, the El Reno tornado is listed as EF3 as of 6/10 by the SPC.

 

[Buckeye05]

Never thought I'd say this, but I am becoming a bit more skeptical of the EF5 rating for El Reno rating now. Max from extremeplanet.me did a thorough ground survey of the area and didn't find any scouring or vegetation damage indicative of EF5 surface winds. There was a lot of grass flattened and bent over, but none of it was scoured from the ground. Trees in the area sustained minimal debarking. Compare it to Moore from a few weeks earlier, which resulted in complete denuding/debarking of trees, and scoured the grass to the point where only mud remained. Both tornadoes occurred in the same general geographic area.

 

El Reno

 

Moore

[AtticaFanatica]

There are no DIs for ground scouring or vegetative damage, so there is no such thing as evidence of EF5 winds using those fields, at least scientifically. All past evidence is anecdotal. The reason for that is ground scouring and vegetative damage likely are influenced by a whole host of factors, including vertical winds in the tornado, moisture content, wind residence time, the type of vegetation, dirt, asphalt, etc. and probably a lot of others.

It's one thing to compare actual measured velocities to DIs that are based on wind science engineering studies. It's another to compare them with a vague idea of what we perceive tornadoes of a certain intensity *should* do.

[Buckeye05]

There is no DI, but intense ground scouring is one of those commonly seen hallmarks of EF5 damage, and a rather consistent one at that. I know many will disagree, but I believe it is definitely something that should be weighed more heavily, and is of more significance than some of give it credit for. I absolutely agree that moisture content and soil composition should be factored in when looking at scouring, and one should bear in mind that central OK soil is slightly more prone to scouring than in other places with with denser soils like Mississippi. Both Moore tornadoes and the 2011 El Reno tornado occurred in this same general part of the state, and all three produced extreme scouring. The 2013 El Reno event did not for whatever reason. That, in my opinion, should at least be taken into consideration.

[AtticaFanatica]

As I said, there is no scientific evidence relating horizontal wind speeds in tornadoes to vegetation or other surface types right at the surface (z = 0 or very close to 0). This is not the result of people being lazy, it's the result of the complexity of the problem of a tornado interacting with the surface (I encourage you to pick a Dave Lewellen paper, say Lewellen and Lewellen 2008, and read it to better grasp this fact). Without any evidence, I'm not sure what you want taken into consideration. Tornadoes cannot be evaluated in a certain manner because you think something is significant with nothing but some conviction and a handful of memorable examples.

[Ground Scouring]

There is no DI, but intense ground scouring is one of those commonly seen hallmarks of EF5 damage, and a rather consistent one at that. I know many will disagree, but I believe it is definitely something that should be weighed more heavily, and is of more significance than some of give it credit for. I absolutely agree that moisture content and soil composition should be factored in when looking at scouring, and one should bear in mind that central OK soil is slightly more prone to scouring than in other places with with denser soils like Mississippi. Both Moore tornadoes and the 2011 El Reno tornado occurred in this same general part of the state, and all three produced extreme scouring. The 2013 El Reno event did not for whatever reason. That, in my opinion, should at least be taken into consideration.

Personally, I think that wheat is often a more resilient plant than is often acknowledged. The images which Max posted show the matted density and length of both the stalk and root structures. To a certain extent, those factors may have blunted the visibility of the EF5 winds.

I just read the thread more carefully about Ucellini's directive so I deleted the question I had:

In case anyone hadn't seen yet, the El Reno tornado is listed as EF3 as of 6/10 by the SPC.

 

Please post the source. Thanks!

[andyhb]

Article on tornado intensity estimation from SPC (Edwards, LaDue et al).
 

http://www.spc.noaa.gov/publications/edwards/ef-scale.pdf

[gymengineer]

Personally, I think that wheat is often a more resilient plant than is often acknowledged. The images which Max posted show the matted density and length of both the stalk and root structures. To a certain extent, those factors may have blunted the visibility of the EF5 winds.

Please post the source. Thanks!

http://www.spc.noaa.gov/climo/torn/fatalmap.php

[gymengineer]

With all due respect to Dr. U. (and plenty is due), the EF-scale is not simply an impact-based scale.  If it were, we wouldn't have the lower-bound/expected/upper-bound ranges for each DOD.  For instance, if you have two houses swept away, the "impact" is the same, even if one house is secured by being bolted and strapped to its foundation and one has no anchoring.  In both cases, the house is swept away, but I don't think there's any disagreement that the force required for the same impact to both houses is vastly different or that the difference is to be considered in the EF-scale rating.  I'd venture that there's no doubt amongst people here that the unanchored house would likely barely qualify for an EF3 rating, whereas the bolted/strapped house would likely require EF5 consideration.  The Pandora's box of trying to estimate actual tornado intensity is already open.  I would hope in the future, if we can ascertain a better relationship between wind speeds and damage (a long time from now), some of these ratings that look like they'll be reversed here are given a second look.

 

Dr. Chuck Doswell's thoughts on this (which happen to fit my feelings very well): http://cadiiitalk.blogspot.com/2013/06/the-ef-scale-ratings-brouhaha.html

One thing from Doswell's thoughts that jumps out as a bit odd to me-- he seems to me to be downplaying a bit the actual importance of translating radar measured winds to an acceptable 10-m surface wind. 

In hurricanes, the importance of this idea has long been settled-- you do not take flight level winds and apply a formula down to get 10-m winds because the wind profile varies so much from storm to storm, and during the life of the storm itself (e.g., is it weakening? rapidly intensifying? expanding wind field?). Likewise, you do not accept above-10-m anemometer readings verbatim ever, and certainly any inclusion of non-standard reading into helping determine intensity requires a rather extensive and tortured analysis (e.g. Claudette, 2003). 

 

So, Doswell allows a crack open in the door for the argument that these Doppler readings shouldn't be used quite yet-- based on the state of the science. But, if we're trying to get the science right, isn't nailing down the relationship pretty vital before using the Doppler data in rating the tornado?

[CT Rain]

One thing from Doswell's thoughts that jumps out as a bit odd to me-- he seems to me to be downplaying a bit the actual importance of translating radar measured winds to an acceptable 10-m surface wind. 

In hurricanes, the importance of this idea has long been settled-- you do not take flight level winds and apply a formula down to get 10-m winds because the wind profile varies so much from storm to storm, and during the life of the storm itself (e.g., is it weakening? rapidly intensifying? expanding wind field?). Likewise, you do not accept above-10-m anemometer readings verbatim ever, and certainly any inclusion of non-standard reading into helping determine intensity requires a rather extensive and tortured analysis (e.g. Claudette, 2003). 

 

So, Doswell allows a crack open in the door for the argument that these Doppler readings shouldn't be used quite yet-- based on the state of the science. But, if we're trying to get the science right, isn't nailing down the relationship pretty vital before using the Doppler data in rating the tornado?

 

I was wondering this as well. If Attica could chime in that would be great. How settled (or not) is the science of translation of winds at a certain altitude down to 10 meters in tornado? Do we know if this can vary significantly storm to storm? I

[AtticaFanatica]

It's not "settled", but the current research shows the difference between the winds at the lowest radar levels and surface winds are not as large as originally thought, or in some cases are basically nonexistent (see, for example, Wurman et al. (2013)). This is essentially a nonissue, in my opinion, as long as you're within 100 m or so (being conservative) of the surface.

[gymengineer]

It's not "settled", but the current research shows the difference between the winds at the lowest radar levels and surface winds are not as large as originally thought, or in some cases are basically nonexistent (see, for example, Wurman et al. (2013)). This is essentially a nonissue, in my opinion, as long as you're within 100 m or so (being conservative) of the surface.

Which Wurman et al article are you referring to?

 

An 'early online release' article from the Bulletin of the American Meteorological Association Society authored by Wurman, et al, is entitled "In-Situ, Doppler Radar and Video Observations of the Interior Structure of a Tornado and Wind-Damage Relationship." It looks to be have been on track to be published in the journal in 2013.

In this article, the concluding sentence is: "Additional radar and in situ observations in tornadoes in the region < 30 m AGL (Kosiba and Wurman 2013) will be critical in refining knowledge concerning low level tornado structure, and the relationship between wind speed, direction, and duration and observed damage."

 

That's hardly a ringing endorsement for using the Doppler data operationally *right now* to rate tornadoes.

 

But, I take it that you were referring to another article? 

[gymengineer]

To me, a good analogy to this debate is the use of SMFR for determining hurricane intensities. What we know is that the NHC will certainly weigh the SMFR data much higher than flight level wind data, but they take into account other data (*including* the flight level winds, pressure, whether the storm is intensifying or weakening, anemometer data, even building damage) to arrive at their best guess for a hurricane's landfall intensity. 

 

How is using SMFR data for hurricanes any *less* accurate than using Doppler data for tornadoes? And if it's not much less accurate, than why should those ranking tornadoes be comfortable right now just using one piece of data for arriving at the ranking? 

[AtticaFanatica]

It's not "settled", but the current research shows the difference between the winds at the lowest radar levels and surface winds are not as large as originally thought, or in some cases are basically nonexistent (see, for example, Wurman et al. (2013)). This is essentially a nonissue, in my opinion, as long as you're within 100 m or so (being conservative) of the surface.

Which Wurman et al article are you referring to?

An 'early online release' article from the Bulletin of the American Meteorological Association Society authored by Wurman, et al, is entitled "In-Situ, Doppler Radar and Video Observations of the Interior Structure of a Tornado and Wind-Damage Relationship." It looks to be have been on track to be published in the journal in 2013.

In this article, the concluding sentence is: "Additional radar and in situ observations in tornadoes in the region < 30 m AGL (Kosiba and Wurman 2013) will be critical in refining knowledge concerning low level tornado structure, and the relationship between wind speed, direction, and duration and observed damage."

That's hardly a ringing endorsement for using the Doppler data operationally *right now* to rate tornadoes.

Digging further into the trove of Kosiba, Wurman articles, there's another one from late 2012 entitled: "Comparison of Mobile-Radar Measurements of Tornado Intensity with Corresponding WSR-88D Measurements." This was another candidate to the article you are citing.

From the conclusion section of this article: "The suggestion from this work is that the WSR-88D can potentially be used in isolation to estimate low-level tornado intensity; this possibility is supported by the independent work of Thompson et al. (2012), and ongoing research at the Warning Decision Training Branch of the National Weather Service (J. LaDue 2011, personal communication). The proposed estimation is by way of a linear regression model, and application of this model is relevant only once a tornado is already confirmed. The particular model presented herein is based on a dataset of tornadic events that range in intensity and radar distance, but this dataset is relatively small. More events will be incorporated into the statistical analysis as they become available so as to provide a more robust relationship."

Again, that's hardly a ringing endorsement for using the Doppler data *right now* in operationally rating the tornado.

Help me out by letting me know which article of Wurman you are referring to....

Wurman et al. 2013, which is the former. Sorry for the confusion. The other article is pretty interesting too, though, if you get a chance. If you look at the direct comparisons between radar and TIV, they're pretty similar (much more so than I ever would have guessed for sure).

Ringing endorsements would obviously be nice, but we're dealing with the best, or probably least bad, of several choices in estimating tornado intensity. This is about as well as you're going to do, seeing as surveys are woefully inaccurate, particularly over open country. More obs are definitely necessary for rigorous comparisons, as I alluded to above. The available science supports the radar choice, in my opinion. You could reduce the radar obs by almost 50% (far far higher than any experimental results) in this case and still have EF5 winds. I believe this is much more defensible scientific evidence than that obtained by judging winds over an open field where there is essentially no way to judge wind speeds.

[AtticaFanatica]

To me, a good analogy to this debate is the use of SMFR for determining hurricane intensities. What we know is that the NHC will certainly weigh the SMFR data much higher than flight level wind data, but they take into account other data (*including* the flight level winds, pressure, whether the storm is intensifying or weakening, anemometer data, even building damage) to arrive at their best guess for a hurricane's landfall intensity.

How is using SMFR data for hurricanes any *less* accurate than using Doppler data for tornadoes? And if it's not much less accurate, than why should those ranking tornadoes be comfortable right now just using one piece of data for arriving at the ranking?

SMFR is a passive remote sensing instrument that makes several assumptions about the local environment to even get the measurement of winds in the first place (hurricane folks can comment more accurately). Weather radar is an active remote sensing instrument in which there's little to no argument that what's being measured, assuming adequate SNR, etc., are direct radial velocity measurements within the given resolution volume. The only issue that I think has scientific basis is the height issue, and the evidence so far points to that also being a relatively minor exercise.

But let's, for the sake of argument, accept that premise. I don't see why one would be more comfortable with a damage survey done over open country, which has biases that everyone knows about and admits to, over actual scientific measurements.

[gymengineer]

SMFR is a passive remote sensing instrument that makes several assumptions about the local environment to even get the measurement of winds in the first place (hurricane folks can comment more accurately). Weather radar is an active remote sensing instrument in which there's little to no argument that what's being measured, assuming adequate SNR, etc., are direct radial velocity measurements within the given resolution volume. The only issue that I think has scientific basis is the height issue, and the evidence so far points to that also being a relatively minor exercise. But let's, for the sake of argument, accept that premise. I don't see why one would be more comfortable with a damage survey done over open country, which has biases that everyone knows about and admits to, over actual scientific measurements.

I agree with this sentence. I'm just saying, at this point, I don't see the need to throw out either data or only choose to use one piece of data. I understand that obviously the NHC only has to analyze the intensity of 10-25 cyclones each season as compared to thousands of tornadoes, so they can afford to take months to compare all available data and come up with a best guess of intensity. That's not practical per se for forecast offices.... but the spirit behind it can still apply-- use all data, including building engineer surveys (like after Joplin), and weight some data more than others. I guess I don't see the point of, given that the research is still cutting edge on "the height issue," completely dismissing the work of the engineers either.