Just How Violent are the Tornadoes this Year?

[thewxmann]

The tornado death toll this year has been unprecedented, the greatest since 1953. So the public and the media have been asking, are the tornadoes simply more violent this year? This analysis should shed some light on the issue.

In my blog, I have written a short piece comparing the number of violent tornadoes this year to that of years from 1950-2010. Here's a chart of the number of EF3+ tornadoes since 1950:

Of note,

The violent tornado count to date in 2011 is 1.3 standard deviations above the annual 1950-2010 mean, and nearly 2.5 standard deviations above the 1975-2010 mean.

More in my blog post.

[OHSnow]

One thing to keep in mind, as Dr. Forbes of TWC (who was involved in some of the early Fujita scale rankings) has pointed out, is that tornado damage assessments have gradually became more fine-tuned as greater knowledge of engineering and the winds necessary to cause certain structural damage were gained. So some of the early years may not necessarily be comparable to the numbers for the 90s and 00s. By the early 90s, damage assessments were pretty much at the same level as today -- the enhanced Fujita scale merely altered the windspeeds to better reflect contemporary wisdom. But some of the 60s and 70s assessments may have been more liberal in handing out the F3, F4, or F5 moniker than they would be today.

The tornado death toll this year has been unprecedented, the greatest since 1953. So the public and the media have been asking, are the tornadoes simply more violent this year? This analysis should shed some light on the issue.

In my blog, I have written a short piece comparing the number of violent tornadoes this year to that of years from 1950-2010. Here's a chart of the number of EF3+ tornadoes since 1950:

Of note,

More in my blog post.

[AtticaFanatica]

The problem with statistics on (E)F scales is the observation error of the (E)F scale measurement. It's large, which skews any type of analysis probably to the point where it's difficult to glean anything useful. It's not a coincidence that most of the violent tornadoes that have struck have impacted, at one time or another, cities or towns.

[thewxmann]

The problem with statistics on (E)F scales is the observation error of the (E)F scale measurement. It's large, which skews any type of analysis probably to the point where it's difficult to glean anything useful. It's not a coincidence that most of the violent tornadoes that have struck have impacted, at one time or another, cities or towns.

While there is observation error (along with many other sources of error, such as long-lived violent tornadoes counted less than five short-lived ones), in general I think most truly violent tornadoes are, at least nowadays, rated EF3+, and most truly weak tornadoes are rated at most EF2. What I have observed is that most non-violent tornadoes that strike cities erk out an EF2 rating (e.g. La Crosse on 5/22/11, Atlanta on 4/14/08, Ft. Worth on 3/28/00), which many truly violent tornadoes manage to hit something and obtain an EF3 (Bowdle 5/22/07, the six EF3+ on 5/23/08).

I will admit however that rural tornadoes that reach violent strength for only a short time will likely get missed. So I'm not doubting that there are significant error bars. But I don't think that this study is pointless either. If a certain year is truly violent, it will produce enough powerful tornadoes to show up in the long-term averages, since on average, there's going to be a certain amount of tornadoes missed every year. Plus you can't ignore the EF4/5 rankings that I posted in the link, nor the large amount of debris balls we've seen compared to other years, nor the large amount of high CAPE/SRH environments we've seen with setups this year.

The point of this was, if someone were to ask me (like I saw in another forum), are the storms more violent this year? I'd say, take a look at this.

[thewxmann]

One thing to keep in mind, as Dr. Forbes of TWC (who was involved in some of the early Fujita scale rankings) has pointed out, is that tornado damage assessments have gradually became more fine-tuned as greater knowledge of engineering and the winds necessary to cause certain structural damage were gained. So some of the early years may not necessarily be comparable to the numbers for the 90s and 00s. By the early 90s, damage assessments were pretty much at the same level as today -- the enhanced Fujita scale merely altered the windspeeds to better reflect contemporary wisdom. But some of the 60s and 70s assessments may have been more liberal in handing out the F3, F4, or F5 moniker than they would be today.

Agreed, which is why I looked at the mean/SD of both 1975-2010 and 1950-2010. Earlier numbers were certainly inflated, but it's worthwhile to look at how this year stands even when factoring in those inflated numbers - we're still in the upper quartile of the past 60 years.

[Analog96]

The problem with statistics on (E)F scales is the observation error of the (E)F scale measurement. It's large, which skews any type of analysis probably to the point where it's difficult to glean anything useful. It's not a coincidence that most of the violent tornadoes that have struck have impacted, at one time or another, cities or towns.

I think this data COULD be useful if we could separate major from minor tornadoes, something like F3 and up. If a tornado was ranked at F3 or higher, it probably would have done some pretty good damage, regardless of when it was ranked.

[Analog96]

While there is observation error (along with many other sources of error, such as long-lived violent tornadoes counted less than five short-lived ones), in general I think most truly violent tornadoes are, at least nowadays, rated EF3+, and most truly weak tornadoes are rated at most EF2. What I have observed is that most non-violent tornadoes that strike cities erk out an EF2 rating (e.g. La Crosse on 5/22/11, Atlanta on 4/14/08, Ft. Worth on 3/28/00), which many truly violent tornadoes manage to hit something and obtain an EF3 (Bowdle 5/22/07, the six EF3+ on 5/23/08).

I will admit however that rural tornadoes that reach violent strength for only a short time will likely get missed. So I'm not doubting that there are significant error bars. But I don't think that this study is pointless either. If a certain year is truly violent, it will produce enough powerful tornadoes to show up in the long-term averages, since on average, there's going to be a certain amount of tornadoes missed every year. Plus you can't ignore the EF4/5 rankings that I posted in the link, nor the large amount of debris balls we've seen compared to other years, nor the large amount of high CAPE/SRH environments we've seen with setups.

The point of this was, if someone were to ask me (like I saw in another forum), are the storms more violent this year? I'd say, take a look at this.

You make a good point here. While they may have handed out higher rankings in tornadoes that hit urban areas in the older days, they were more areas then where nobody lived, or cared to report a tornado. So some of that "over reporting" may be evened out because of that.

[AtticaFanatica]

While there is observation error (along with many other sources of error, such as long-lived violent tornadoes counted less than five short-lived ones), in general I think most truly violent tornadoes are, at least nowadays, rated EF3+, and most truly weak tornadoes are rated at most EF2. What I have observed is that most non-violent tornadoes that strike cities erk out an EF2 rating (e.g. La Crosse on 5/22/11, Atlanta on 4/14/08, Ft. Worth on 3/28/00), which many truly violent tornadoes manage to hit something and obtain an EF3 (Bowdle 5/22/07, the six EF3+ on 5/23/08).

I will admit however that rural tornadoes that reach violent strength for only a short time will likely get missed. So I'm not doubting that there are significant error bars. But I don't think that this study is pointless either. If a certain year is truly violent, it will produce enough powerful tornadoes to show up in the long-term averages, since on average, there's going to be a certain amount of tornadoes missed every year. Plus you can't ignore the EF4/5 rankings that I posted in the link, nor the large amount of debris balls we've seen compared to other years, nor the large amount of high CAPE/SRH environments we've seen with setups.

The point of this was, if someone were to ask me (like I saw in another forum), are the storms more violent this year? I'd say, take a look at this.

Most of this is conjecture, and I would take issue with most of your claims.. Check out Figure 8 from this paper to get a sobering look at just how poor the damage evaluations likely are.

[thewxmann]

Most of this is conjecture, and I would take issue with most of your claims.. Check out Figure 8 from this paper to get a sobering look at just how poor the damage evaluations likely are.

Sure it's conjecture. This isn't a peer-reviewed paper or anything. But how else do you compare the violence of tornadoes between different years? What would be your methodology? I don't think it's a trivial question. Besides, the whole damage scale idea was originally based on a conjecture, that we could connect the Beaufort and the Mach Scales and assume a correlation between damage and wind speed when such engineering studies hadn't been done when the idea was devised (as far as I know). So conjecturing isn't necessarily worthless or without merit.

Regarding the paper, I am fulling aware that many "weak" tornadoes are underrated (and vice versa). But my point is we are making [admittedly, a very rough] assumption that the more violent a year - the more actual violent tornadoes nature spawns - the more (E)F3+ ratings will be earned across the country. It's not going to be perfect, but if we do that, we can make a crude estimate on how violent a year is.

[AtticaFanatica]

Sure it's conjecture. This isn't a peer-reviewed paper or anything. But how else do you compare the violence of tornadoes between different years? What would be your methodology? I don't think it's a trivial question. Besides, the whole damage scale idea was originally based on a conjecture, that we could connect the Beaufort and the Mach Scales and assume a correlation between damage and wind speed when such engineering studies hadn't been done when the idea was devised (as far as I know). So conjecturing isn't necessarily worthless or without merit.

Sometimes there's just not a feasible way to do something accurately. My methodology would be to not do it, not that I don't think it was an interesting read or anything.

Regarding the paper, I am fulling aware that many "weak" tornadoes are underrated. But my point is we are making [admittedly, a very rough] assumption that the more violent a year - the more actual violent tornadoes nature spawns - the more (E)F3+ ratings will be earned across the country. It's not going to be perfect, but if we do that, we can make a crude estimate on how violent a year is.

The Doppler climatology shows considerable error in assessing weak, strong, and violent tornadoes. The number of violent tornadoes is going to be relatively low, both in actual number and in relative frequency. As a result, this leads to small sample sizes of violent tornadoes in a given year. The Doppler climatology shows ~15 actual violent tornadoes and 6-7 surveyed violent tornadoes. I don't believe the combination of small sample sizes (both in actual violent tornadoes and in surveyed violent tornadoes) and large errors (in this particular study, it was ~50%) really leave many options to accurately assess violent tornadoes on a year-to-year basis.

[Gorky]

I don't think it's all about how violent the tornadoes are this year although there have obviously been some very noticeable outbreaks. La Nina years tend to have an eastward bias as far as tornado locations go from what I can see (Troughs ejecting too quick to allow for suitable moisture return out west and shifting the threat east. The troughs plunging further south into Dixie Alley more frequently). Population density is several times higher in the Eastern US than the central plains so there is much more to damage. The Joplin tornado happening out on the plains more often than not would be forgotten about a few days later as the few farms buildings it would likely hit would not make the news. We had lots of tornadoes last year that had you placed over a town like Joplin would have done as much or more damage - imagine Bowdle with it's slow forward speed for instance. Those tornadoes stayed out of the CBDs of large towns.Yazoo City 1 mile north... Conger tornado plowing through downtown Albert Lea instead? They'd have been big disaster like we had this year but we got lucky, and have been for some time. What I'm trying to get at is we have lots of violent tornadoes every year, it's just that we've been incredibly unlucky that almost all the big wedges, all the large outbreaks have been in very populated areas. This is not the norm. Compare with 2004 which was another huge year for tornado numbers (and I'm not convinced this year will beat it unless we get another active hurricane season) but most avoided major populations with only 36 deaths by year end.

[blackjack123]

I mean what if you had a mile-wide EF3 tornado, with estimated winds of 150-160mph, take a 30 mile track right down through the heart of New York, during the middle of rush hour traffic. An EF3 tornado may not be the most powerful tornado but I am pretty certain you would have a very large death toll.

[mcalvert]

I mean what if you had a mile-wide EF3 tornado, with estimated winds of 150-160mph, take a 30 mile track right down through the heart of New York, during the middle of rush hour traffic. An EF3 tornado may not be the most powerful tornado but I am pretty certain you would have a very large death toll.

In a superdense urban area like Manhattan where you have a lot of relatively substantial buildings, you'd not only get a lot of damage from the tornado and its debris, but the inflow would be channeled through the streets (the only surface level conduits available). Those winds would be extremely damaging themselves and would add to the tornado's debris cloud because they would be carrying all manners of projectiles with them.

[jpeters3]

I don't think it's all about how violent the tornadoes are this year although there have obviously been some very noticeable outbreaks. La Nina years tend to have an eastward bias as far as tornado locations go from what I can see (Troughs ejecting too quick to allow for suitable moisture return out west and shifting the threat east. The troughs plunging further south into Dixie Alley more frequently). Population density is several times higher in the Eastern US than the central plains so there is much more to damage. The Joplin tornado happening out on the plains more often than not would be forgotten about a few days later as the few farms buildings it would likely hit would not make the news. We had lots of tornadoes last year that had you placed over a town like Joplin would have done as much or more damage - imagine Bowdle with it's slow forward speed for instance. Those tornadoes stayed out of the CBDs of large towns.Yazoo City 1 mile north... Conger tornado plowing through downtown Albert Lea instead? They'd have been big disaster like we had this year but we got lucky, and have been for some time. What I'm trying to get at is we have lots of violent tornadoes every year, it's just that we've been incredibly unlucky that almost all the big wedges, all the large outbreaks have been in very populated areas. This is not the norm. Compare with 2004 which was another huge year for tornado numbers (and I'm not convinced this year will beat it unless we get another active hurricane season) but most avoided major populations with only 36 deaths by year end.

I agree with you in the sense that we seem to have been very "unlucky" on several occasions this year with large/violent tornadoes seemingly biased toward populated, rather than rural areas; however, this year has been meteorologically spectacular as well, and I would conjecture that this year would receive significant meteorological recognition even if the death toll weren't anomalously high.

I don't think you can compare a year like 2011 with 2004 - perhaps 2004 will exceed 2011 in terms of sheer numbers, but in terms of the relative strength of tornadoes (again, partly just conjecture), this year has been nothing short of spectacular.

Take April 27th for instance - at least 6 tornadoes with path lengths > 70 miles in one day? 15 violent tornadoes in one day? This is unprecedented (aside from say palm sunday and super outbreak).

Yes, had bowdle occurred in Joplin, there is a good chance it would have produced similar damage and perhaps been rated EF-5, but where is the last-year's analogy (even rural) to El Rino-Piedmont? Hackelburg? Smithsville? Tuscaloosa-birmingham? or many of the numerous other high-end EF-4s this year?

I would argue that the meteorology, and not just "luck", is what has caused so many tornado deaths this year.

[AtticaFanatica]

but where is the last-year's analogy (even rural) to El Rino-Piedmont? Hackelburg? Smithsville? Tuscaloosa-birmingham? or many of the numerous other high-end EF-4s this year?

probably in Storm Data as weaker tornadoes

[jpeters3]

probably in Storm Data as weaker tornadoes

FIrst of all, these are all 70+ mile tornado tracks. The probability of an analogous long-tracked tornado last year having "missed" any significant structures is pretty low. The probability of numerous analogous long-tracked tornadoes "missing" structures and being rated weaker is even lower.

Look at the philadelphia MS tornado from 4-27 - this tornado did not track through very populated areas, but still left DIs worthy of a violent rating.

And the only even remotely analogous meteorological setup from last year is 5-10-10.

I guess what I'm saying is that the basis of this year's tornado "violence" has been anomalous meteorological conditions.

[tornadotony]

probably in Storm Data as weaker tornadoes

What's being completely and frustratingly lost in this discussion is that the EF-scale is primarily a damage assessment scale and secondarily a wind speed estimation scale. Thus, though there are years where numerous tornadoes that have the potential to produce EF3-EF5 damage do not get recorded as such due to the locations they hit, it's very safe to say this is an incredibly violent year because we rate damage first and an extraordinary amount of violent damage has occurred. And there is meteorological relevance to this because these statistics not only paint a fairly high (though admittedly we don't know how high) rate of EF3-EF5 activity, but they paint this activity over populated areas, which is largely a statement of the meteorological pattern we've been in for a good portion of the year. Of course, the damage estimations themselves are not perfect, but it's fairly safe to say that the best possible efforts to accurately log this year have been undertaken and, given the extraordinary difference between this year and most past year with regards to EF3-EF5 tornado occurrences, this year is remarkable.

[jpeters3]

Of course, the damage estimations themselves are not perfect, but it's fairly safe to say that the best possible efforts to accurately log this year have been undertaken and, given the extraordinary difference between this year and most past year with regards to EF3-EF5 tornado occurrences, this year is remarkable.

Very well stated. I think all of us will acknowledge that the EF scale is inherently flawed, as would be any other attempt to classify tornadoes in terms of their intensity; however, to argue that this year has had so many more more violent rated tornadoes by sheer luck is a more than a bit of a stretch. I think that we can safely assume that, more often than not, an assigned EF scale rating is at least in the ballpark with the tornadoes actual intensity.

A more objective statistic that can be compared is overall tornado path lengths, which one could argue reflects on the capacity for an environment to support a sustained tornadic supercell to some extent (assuming the dynamics responsible for the behaviors of tornado families aren't entirely internal). This year has seen an unprecedented number of very-long tracked tornadoes, especially when compared to some of the other years in question.

[AtticaFanatica]

FIrst of all, these are all 70+ mile tornado tracks. The probability of an analogous long-tracked tornado last year having "missed" any significant structures is pretty low. The probability of numerous analogous long-tracked tornadoes "missing" structures and being rated weaker is even lower.

Look at the philadelphia MS tornado from 4-27 - this tornado did not track through very populated areas, but still left DIs worthy of a violent rating.

And the only even remotely analogous meteorological setup from last year is 5-10-10.

I guess what I'm saying is that the basis of this year's tornado "violence" has been anomalous meteorological conditions.

Tornadoes rated as violent have larger mean path lengths than those rated weaker (precisely because they have greater opportunity to hit structures) but given a long tornado path length, the probability that the tornado is violent is still very low (see, for example, Fig. 5 in Brooks (2004)). The number of tornadoes with long path lengths doesn't tell us anything about the number of actual violent tornadoes. In fact, this is basically my point, violent tornadoes with short path lengths are still violent tornadoes and are likely missed in numbers that are quite significant.

[AtticaFanatica]

What's being completely and frustratingly lost in this discussion is that the EF-scale is primarily a damage assessment scale and secondarily a wind speed estimation scale. Thus, though there are years where numerous tornadoes that have the potential to produce EF3-EF5 damage do not get recorded as such due to the locations they hit, it's very safe to say this is an incredibly violent year because we rate damage first and an extraordinary amount of violent damage has occurred. And there is meteorological relevance to this because these statistics not only paint a fairly high (though admittedly we don't know how high) rate of EF3-EF5 activity, but they paint this activity over populated areas, which is largely a statement of the meteorological pattern we've been in for a good portion of the year. Of course, the damage estimations themselves are not perfect, but it's fairly safe to say that the best possible efforts to accurately log this year have been undertaken and, given the extraordinary difference between this year and most past year with regards to EF3-EF5 tornado occurrences, this year is remarkable.

Right, it's remarkable in its societal impacts, absolutely. We simply don't know if it's remarkable in terms of the actual number of violent and/or EF-5 tornadoes.

[jpeters3]

Tornadoes rated as violent have larger mean path lengths than those rated weaker (precisely because they have greater opportunity to hit structures) but given a long tornado path length, the probability that the tornado is violent is still very low (see, for example, Fig. 5 in Brooks (2004)). The number of tornadoes with long path lengths doesn't tell us anything about the number of actual violent tornadoes. In fact, this is basically my point, violent tornadoes with short path lengths are still violent tornadoes and are likely missed in numbers that are quite significant.

I am not suggesting that path length should be correlated to intensity; however, path length is probably the most objective measure in the database. Therefore, a tornado season with an unprecedented number of long-tracked tornadoes would suggest a meteorological bias, since there is more accumulated time of mature tornado cyclones. This is reinforced by the anomalous damage assessment stats.

Look at this year so far compared to last year. It is unreasonable to attribute these differences to the deficiencies of the EF scale, and bias toward populated areas that happens to be persist through most of the season alone.

EF-3 EF-4 EF-5 Total

2010 31 13 0 44

2011 51 16 5 72

I would be interested to hear specific instances last year where you suspect that a violent tornado was missed due to inadequate DIs...

[AtticaFanatica]

I am not suggesting that path length should be correlated to intensity; however, path length is probably the most objective measure in the database.

Agreed.

Therefore, a tornado season with an unprecedented number of long-tracked tornadoes would suggest a meteorological bias, since there is more accumulated time of mature tornado cyclones.

Ok, but that's not a violent tornado bias, it's a long path length bias. As you said, the two should not be correlated.

This is reinforced by the anomalous damage assessment stats.

As it should be, longer track tornadoes have a greater probability to hit structures, so of course there going to be rated higher, that's the whole problem.

Look at this year so far compared to last year. It is unreasonable to attribute these differences to the deficiencies of the EF scale, and bias toward populated areas that happens to be persist through most of the season alone.

EF-3 EF-4 EF-5 Total

2010 31 13 0 44

2011 51 16 5 72

Why are we now including EF-3 tornadoes? And since when have we been discussing one year comparisons with 2010? That wasn't the argument (at least I don't think it was).

I would be interested to hear specific instances last year where you suspect that a violent tornado was missed due to inadequate DIs...

This is silly, one cannot accurately judge a tornado by its path length, width, or appearance. If it doesn't hit structures or it only hits structures determined to withstand a certain windspeed, the odds are good (as the radar climatology supports) that it is going to be underrated. I can't give specific examples because there's no way to do so, that's the problem.

[jpeters3]

Ok, but that's not a violent tornado bias, it's a long path length bias. As you said, the two should not be correlated.

Right, but the fact that this year has seen an anomalously high number of very-long-tracked tornadoes suggests a meteorological bias. It is not unreasonable to suspect that, given the fact that the damage surveys overwhelmingly support it, that this year has been meteorologically biased toward stronger tornadoes as well.

]As it should be, longer track tornadoes have a greater probability to hit structures, so of course there going to be rated higher, that's the whole problem.

good point, but this is again operating under the assumption that the EF scale is completely useless, which is an unreasonable assumption. There are non-structure related DIs, such as tree damage, ground/pavement scouring.

Why are we now including EF-3 tornadoes? And since when have we been discussing one year comparisons with 2010? That wasn't the argument (at least I don't think it was).

fine, forget EF-3 tornadoes (though, if this year were biased toward stronger tornadoes, we'd expect to see more EF-3 ratings as well, which is clearly the case). Forget 2010, compare to any year.

This is silly, one cannot accurately judge a tornado by its path length, width, or appearance. If it doesn't hit structures or it only hits structures determined to withstand a certain windspeed, the odds are good (as the radar climatology supports) that it is going to be underrated. I can't give specific examples because there's no way to do so, that's the problem.

why can't we judge by these characteristics? Again, I have not suggested that they are correlated to intensity - rather, that they suggest a meteorological bias.It is silly to assume, in light of statistics that seem to support it, that there is absolutely no bias toward stronger tornadoes this year.

[jpeters3]

This is completely subjective, but I cannot remember another year where I saw more videos of large wedge/stovepipe tornadoes that showed incredible, apparently violent motion. I cannot remember another year where I have seen this many strong tornado vortex signatures, with incredible gtg velocities.

Not to mention the number of incredible meteorological setups that we have seen so far.

[mcalvert]

We know that there is a general causal relationship between certain thermodynamic and wind profiles and the incidence of strong tornadoes. If there wasn't, the SPC would be out of business. While there isn't a 100% correlation between the combinations of high CAPE, shear and certain forcing mechanisms such as mesoscale boundaries, we definitely know that the probabilities of violent tornadoes forming are higher when these ingredients are in plentiful supply.

The values of some of the indices that have been developed to quantify the potential for tornadoes reached almost unheard levels on April 27. Even considering that there might have been some equally strong tornadoes in other years that were underestimated due to their sparsely populated and vegetated paths, the environment on 4/27/11 was extremely rare. It's not unreasonable to believe that the resulting tornadoes were also.

The tornadoes that hit Cullman and Tuscaloosa, AL were the most violent looking tornadoes I've ever seen. They looked like a bag full of kids hyped up on Pixi sticks and Jolt cola. They had multiple funnels at times that were moving both vertically and horizontally. They looked like they were so full of energy that they couldn't contain it within a nice, simple, uniform condensation funnel. Just unbelievable.

Note: As the tornado outbreak was progressing in Alabama, James Spann and Jason Simpson were flabbergasted at the WSI Significant Tornado Index values that they were seeing. They thought prior to the event that the maximum value of the index was 10.0. Multiple tornadoes exceeded that value, including one that tripped a 17.3 on the index.

[AtticaFanatica]

Right, but the fact that this year has seen an anomalously high number of very-long-tracked tornadoes suggests a meteorological bias. It is not unreasonable to suspect that, given the fact that the damage surveys overwhelmingly support it, that this year has been meteorologically biased toward stronger tornadoes as well.

Violent long-tracked tornadoes are more likely to be rated correctly than short-tracked violent tornadoes because they are more likely to hit structures. My point is that having more long-tracked violent tornadoes doesn't necessarily mean there have been more total violent tornadoes.

good point, but this is again operating under the assumption that the EF scale is completely useless, which is an unreasonable assumption. There are non-structure related DIs, such as tree damage, ground/pavement scouring

Not completely useless, but with significant error, as the radar climatology shows nicely. The small number of violent tornadoes on an annual basis combined with a roughly 50% error in rating means it's difficult to tell an actual increase in violent tornadoes from a rating or sampling bias.

why can't we judge by these characteristics? Again, I have not suggested that they are correlated to intensity - rather, that they suggest a meteorological bias.

Ok, I guess I'm confused. What meteorological bias are you talking about? Not a bias toward stronger tornadoes but toward bigger/wider tornadoes?

It is silly to assume, in light of statistics that seem to support it, that there is absolutely no bias toward stronger tornadoes this year.

Nah, I'm saying the statistics that support it have error bars that completely overwhelm any perceived signal. I've shown evidence in support of this.

[jpeters3]

Ok, I guess I'm confused. What meteorological bias are you talking about? Not a bias toward stronger tornadoes but toward bigger/wider tornadoes?

Yes a bias toward bigger/wider/longer tracked tornadoes. I would argue that anomalously long tornado tracks suggests that there is meteorological bias that favors more sustained tornado cyclones.

So there is evidence outside of mere EF scale ratings that this year is meteorologically distinct - on top of this, there is the greater number of violent rated tornadoes this year, which suggests that this meteorological bias may also favor stronger tornadoes.

Also, the only real evidence presented in this thread has been the wurman study. While this certainly raises some questions, it is by no statically robust enough for confident conclusions regarding the EF scale. The radar-sampled tornado data set is very small (especially with respect to violent tornadoes), and there are the usual uncertainties regarding the comparisons between dopler reported wind velocities and actual ground velocities.

The reality is that the best data we have for comparison of tornado intensity strongly suggests that there have been more violent tornadoes this year, and more high end violent tornadoes this year. The synoptic scale setups - especially April 27 - have been relatively unprecedented as well.

[jpeters3]

an interesting thought after reading the Wurman study:

It is suggested by Figs 7/8 that the number of weak (EF-0) tornadoes is grossly over estimated. Perhaps weak tornadoes do indeed briefly attain wind speeds worthy of strong/violent classification, so that it is more likely the duration of strong wind speed through a tornado life cycle that differentiates what we commonly perceive as weak and strong tornadoes via F/EF scale ratings; i.e. "weak" (low EF rating) tornadoes are only briefly achieve strong/violent winds, while "strong" (high EF rating) tornadoes sustain strong/violent winds, and subsequently have an increased probability of inflicting strong/violent damage to a structure.

The rating system is still differentiating important information here: a sustained "strong" tornado poses a greater danger to life/property than a tornado that is briefly "strong". What probability differentiates the "sustained" and "weak" cases is a combination of internal dynamics (i.e. structural re-arrangements within the supercell itself) and external forcing (i.e. a "favorable environment"). The maximum velcoity difference in the Wurman study would not adequately differentiate between the "brief" and "sustained" cases since it only takes into account the maximum achieved velocity difference - not the duration of this maximum velocity.

So what may differentiate 2011 from other years, assuming internal processes are stochastic but state dependent (i.e. their behavior in a probabilistic sense does ultimately depend on external forcing to some extent), is that the external forcing favors sustained "strong/violent" winds, ultimately increasing the probability that man-made structures and life/property will experience these winds.

So if this is true, does it mean 2011 tornadoes are more violent? Depends on how you interpret "violent" - in terms of maximum velocities attained by tornadoes, maybe not so much as the duration for which these maximum velocities are sustained. It still, I would argue, makes 2011 a meteorologically distinct year.

Edit: I think this might be what you are getting at Attica? Either way, good discussion.

[AtticaFanatica]

Yes a bias toward bigger/wider/longer tracked tornadoes. I would argue that anomalously long tornado tracks suggests that there is meteorological bias that favors more sustained tornado cyclones.

Makes sense. The work done on storm cycling supports the idea that consistently strong RFD outflows favor longer-lived meso's (and therefore tornadoes) while weaker RFD outflows tend to see meso's that cycle very quickly. What environmental aspects are responsible for that is not known.

So there is evidence outside of mere EF scale ratings that this year is meteorologically distinct - on top of this, there is the greater number of violent rated tornadoes this year, which suggests that this meteorological bias may also favor stronger tornadoes.

Maybe, but that's a pretty tenuous relationship without any clear evidence.

Also, the only real evidence presented in this thread has been the wurman study. While this certainly raises some questions, it is by no statically robust enough for confident conclusions regarding the EF scale.

It certainly is statistically robust. A climatology of actual radial wind measurements in 70 tornadoes, 50+ of them with obs within 500 m of the ground is, in fact, the best evidence we've ever had to evaluate the EF scale.

The radar-sampled tornado data set is very small (especially with respect to violent tornadoes), and there are the usual uncertainties regarding the comparisons between dopler reported wind velocities and actual ground velocities.

The reduction in wind speeds at the surface, based on in situ tornado measurements, is small.

The reality is that the best data we have for comparison of tornado intensity strongly suggests that there have been more violent tornadoes this year, and more high end violent tornadoes this year. The synoptic scale setups - especially April 27 - have been relatively unprecedented as well.

No, the best evidence we have available, actual wind speed measurements in tornadoes, suggests that the errors in surveyed tornadoes are simply too high to accurately gauge how a small sample set, like violent tornadoes, changes from year-to-year and to determine whether this year has been more violent than others.

[AtticaFanatica]

an interesting thought after reading the Wurman study:

It is suggested by Figs 7/8 that the number of weak (EF-0) tornadoes is grossly over estimated. Perhaps weak tornadoes do indeed briefly attain wind speeds worthy of strong/violent classification, so that it is more likely the duration of strong wind speed through a tornado life cycle that differentiates what we commonly perceive as weak and strong tornadoes via F/EF scale ratings; i.e. "weak" (low EF rating) tornadoes are only briefly achieve strong/violent winds, while "strong" (high EF rating) tornadoes sustain strong/violent winds, and subsequently have an increased probability of inflicting strong/violent damage to a structure.

I think some of them definitely do and the radar climatology shows the preferred strength of tornadoes is EF-2 rather than EF-0

The rating system is still differentiating important information here: a sustained "strong" tornado poses a greater danger to life/property than a tornado that is briefly "strong". What probability differentiates the "sustained" and "weak" cases is a combination of internal dynamics (i.e. structural re-arrangements within the supercell itself) and external forcing (i.e. a "favorable environment"). The maximum velcoity difference in the Wurman study would not adequately differentiate between the "brief" and "sustained" cases since it only takes into account the maximum achieved velocity difference - not the duration of this maximum velocity.

I agree that the length of time a tornado is strong or violent has a lot to do with its threat to life and property. However, a surveyed tornado suffers from the same weakness as the radar study in that a tornado is rated at its highest damage indicator. Most violent tornadoes are actually only violent for brief periods of time if you go by the damage surveys (probably because there's usually not long paths of structures), so it's difficult to know what fraction of the time violent tornadoes are actually violent. So I'm not sure that it really helps us that much, though I see what you're saying.

So what may differentiate 2011 from other years, assuming internal processes are stochastic but state dependent (i.e. their behavior in a probabilistic sense does ultimately depend on external forcing to some extent), is that the external forcing favors sustained "strong/violent" winds, ultimately increasing the probability that man-made structures and life/property will experience these winds.

So if this is true, does it mean 2011 tornadoes are more violent? Depends on how you interpret "violent" - in terms of maximum velocities attained by tornadoes, maybe not so much as the duration for which these maximum velocities are sustained. It still, I would argue, makes 2011 a meteorologically distinct year.

Edit: I think this might be what you are getting at Attica? Either way, good discussion.

I think we'd need to know more about the relative frequency of violent winds in the violent tornadoes. Unfortunately long tornado path lengths mean long distances where there's just not very much around to adequately evaluate the wind speeds.