Radar-Sampled versus Damage-Estimated Wind Speeds in Tornadoes

[Jeff Duda]

Introduced by Mike Smith in another thread, this recent publication presents an analysis of wind speeds in tornadoes estimated by mobile Doppler radar versus the assigned wind speeds using the EF scale. There is a discrepancy in the two!

Comparison of Tornado Damage Characteristics to Low-Altitude WSR-88D Radar Observations and Implications for Tornado Intensity Estimation

Abstract Given the obvious difficulties in directly sampling tornadic wind fields, ongoing work continues to improve estimates of near-ground wind speeds in tornadoes. This study builds upon a recently proposed empirical relationship between radar-observed velocities in the lowest 150 m above...

journals.ametsoc.org

This paper follows up previous papers by Karen Kosiba and Josh Wurman that also show discrepancies between radar-sampled wind speeds and damage-estimated wind speeds:

Wurman et al. (2021), PNAS: https://www.pnas.org/doi/full/10.1073/pnas.2021535118
Kosiba and Wurman (2023), Nature: The strongest winds in tornadoes are very near the ground - Communications Earth & Environment

 

[William Monfredo]

We've gotten conservative about categorizing the high-end tornadoes over the decades to be sure.
Back up to the 1970 Lubbock tornado. When one house was left standing and the one next door looked totally damaged, Fujita said that's because of tornadoes within a tornado, suction vortices. And wind-engineering types such as McDonald quickly pointed out...this house isn't constructed with as much care compared to that one. At the end of the day, both were correct; it's about the atmosphere and the construction.
Today, a mobile-radar meteorologist might also add, "I don't know, wait a while, I've got to standardize a formula to bring my radar-derived winds at a higher altitude a lot lower to the ground, but life's difficult beneath a radar beam, because tornadoes feature asymmetrical flow patterns; not all tornadoes behave the same. And I only go out at specific times when I want to, can't be everywhere, so I miss a whole lot of events!"
And where are we now? Warning, a little joke ahead, but true. What's the difference between an EF-4 & an EF-5 tornado? Oh, about 6 inches! Since if the anchor bolts are spaced a foot apart, you get limited to EF-4, but if placed at the engineered spacing of 1/2 a foot, then EF-5 is OK.
By the way, are the strongest winds in tornadoes really very near the ground? As a blanket statement: No. Friction still applies as does decades of work in photogrammetry. As a quick example, the Dallas, TX F-3 tornado-winds of April 2, 1957 (Monthly Weather Review, May '60, p 169.)

 

[Mike Smith]

the Dallas, TX tornado-winds of April 2, 1957 (Monthly Weather Review May '60, p 169.)

If I recall correctly, the tracers used in the above diagram came from the Walker Hoecker movie. You will find the MWR reference here: WIND SPEED AND AIR FLOW PATTERNS IN THE DALLAS TORNADO OF APRIL 2, 1957

A short segment is here:

. Walter's movie is the closeup at the start of the video.

 

[William Monfredo]

As some of you have too, I've reviewed articles for scientific, academic journals. My apologies in advance if I got distracted by a bad article.
When I first noticed The strongest winds in tornadoes... the Kosiba and Wurman (2023) Communications Earth and Environment article, I didn't say anything here, but I will now. I would not have approved that title among other aspects of it. I don't know who reviewed it; why is the structure so convoluted to the point where the methods are on the next-to-last and last pages?
Consider that phrase "very near the ground." We live in a postmodern age where words mean everything and nothing... all at once. Would it be too revealing to let people know straight up that the EF-scale's not really about their ~ 50ft to 500 ft above-radar-zone? In fact in some ways, this "small-sample" article argues against itself, its own purposes, and very existence quite well.
To their credit, these two understand they don't have 3-second wind-gusts (that wind engineers at damage surveys concern themselves with, and the authors choose not to care about.) From their article:

"Radar-measured wind speeds in this analysis represent a very short period, <<1 s, spatially averaged, wind measurements, not average wind speeds, 1–3 s, or other averaging period wind gusts as measured by stationary anemometers."

So, that's a problem, a hurdle of relevancy. Just how & where does a tornado huff & puff & blow a house down? By their own admission:

"Many, if not most, structures impacted by tornadoes suffer damage caused by winds at heights <15 m AGL. There are not yet sufficient tornado wind speed observations from <15 m ARL to extend the profile...to the standard meteorological observation height of 10m AGL. "
"Residential structures common in tornado-prone areas typically extend from the ground to only about 3 m (single-story manufactured homes), 5 m (single-story house) or 8 m (two-story houses), with first stories centered at about 2m AGL."
"Critically, the strength of winds in the 0–15 m AGL layer in the presence of man-made structures and trees are nearly impossible to sample by radar due to blockage from those same objects. Therefore, very near ground tornado Vg in built environments remain rarely sampled."

I can be obsessive, funny, sarcastic, serious. Let me leave you with this. In many ways, picture a car wreck on the highway, maybe a David Lynch movie, and a cop arrives on scene of the cleanup and says, "Nothing to see here, folks." Now, I need to look closer at the other two articles!

 

[Jeff Duda]

Great to hear from a fellow peer reviewer, William!

I have also seen some articles get published that I found disagreeable had I been a reviewer for it. Within the past year I reviewed an article (that subsequently was rejected) that was based on a lot of previous research by the same authors. When I checked out some of those articles I was dismayed at the poor quality of some of them. Had I been a reviewer, those papers would not have been published in those forms.