About Thursday Night's Tornado Warning False Alarm in Little Rock

The people of Little Rock were rudely awakened by tornado sirens, weather radios, and smartphones screaming from triggered WEA (wireless emergency alerts) warnings for no good reason Thursday night because the radar was misinterpreted and thought to be showing a tornado hazard.

Let's begin by showing how a tornado signature on Doppler wind data should appear. This is during the 2011 Joplin Tornado. This is "base velocity" data, which is the raw wind direction and speeds as sensed by the radar. 

Regardless of the intensity of the tornado, the strongest winds should touch (as the yellow and dark blue do in this case) or be in close proximity. Now, let's review the warnings in central Arkansas.

First Tornado Warning

A tornado warning was issued at 11:52pm Thursday night. 

Below is the Twitter/X version I tweeted at 11:55pm. I had been watching the storms west of Shreveport and so immediately retweeted the warning without doing my own analysis. 

Below is the radar at 11:51 on which the warning was based. With a very brief glance, it looks like a tornado situation -- but, it wasn't. 
At left is the reflectivity data, the type of radar you see on television weathercasts
and at right is the "storm relative velocity" which is the raw wind velocity
corrected for the movement of the storm. There is also some noise cleanup. Circle
is the supposed location of the tornado.

There are two problems with inferring the presence of a tornado from those images. There was no "hook" echo nor had the storm made a "right turn" -- both indicators of a tornado on reflectivity. 

At first glance of the "storm relative velocity" (upper right) there appears to be sufficient rotation -- but, there actually wasn't. Below is the "base velocity" -- which is the pure wind data -- and it explains why.
There is little in the base velocity data which would indicate a tornado. Compare this to the Joplin image (above) and they look nothing alike. The dark blue = strongest northeast winds (much of which were caused by the storm's very fast movement to the northeast) are (as these things go) distant from the winds flowing to the southwest (circled). As this is an apples-to-apples  data comparison with Joplin, you can see the two look nothing alike. I suspect the "storm relative velocity" misled the meteorologists. 

This type of wind velocity display error is known as "side lobes." That's when the relatively wide radar beam detects stronger rotation higher up in the storm but incorrectly displays it as near the ground. The rotation, at least in part, is displayed in clear air.

The storm continued to move northeast. However, not only didn't produce a tornado, there were no reports of funnel clouds or wall clouds (the latter a precursor to a tornado). 

Tornado Warning for Little Rock
The storm never looked tornadic as it approached Metro Little Rock but the NWS issued a tornado warning for the city at 12:15am Friday.

Because, by this time, I was confident this was a radar misinterpretation, I made it clear this was the NWS's warning, not mine, when I retweeted it.

The radar data upon which the tornado warning was based it below. It is even less impressive than the data immediately before the previous warning. 
There is nothing about the reflectivity data (left) that indicates a tornado. Ditto the storm relative velocity data (right). The data is extremely noisy and there is no "center" of rotation. 

When we look at the base velocity data, there is no center of rotation and the noise is quite apparent. 

But it is the next data interval (12:17am) is conclusive this is a tornado warning false alarm.  This is a close-up on the area of weak "rotation. "
At left is the base velocity at 12:15am. While there is broad, weak rotation (evidenced by the pixels of red inside the gray) to me it is not close to being sufficient for a tornado warning. It is noisy data and, again, the stronger northeast winds are not near the interface with the southwest winds. 

We now go to the 12:17am data, at right. The blue has changed to red! What had been wind blowing toward the radar is displayed as winds away from the radar! Meteorologically impossible. Clearly this was a false echo (side lobes). At this point, it is absolutely clear it was side lobes and the tornado warning should have been cancelled. People should have been allowed to go back to sleep. 

Natural Disaster Review Board

A little more than a week ago, a plug-door on a Alaska Airlines 737-MAX-9 flew off in-flight while departing Portland, OR. The plane depressurized and other hazards suddenly developed. Yet, the plane made a safe landing with no serious injuries. One of the reasons for that happy outcome was the work of the National Transportation Safety Board (NTSB) in the field of "crew resource management." There had been many major airline accidents where the crew not performing as a team was a factor. The "human factors" experts with the NTSB worked with the Federal Aviation Administration and the airlines and solved that issue. So, when the plug-door in the passenger compartment flew off, explosive decompression occurred and the cockpit door flew open, the captain and her first officer performed flawlessly. That probably would not have occurred 30 years ago. The New York Times has a terrific article on this topic.

How does that apply to weather science and tornado warnings? While I don't know the specifics of the staffing at the Little Rock NWS Thursday night, in general there are at least two qualified meteorologists (usually more) during a tornado watch. I have to wonder what they were doing in the way of sharing radar data and the analysis of the same. A Natural Disaster Review Board (NDRB) could independently analyze the NWS, gather data from local television meteorologists, AccuWeather and other participants in this tornado warning. The NDRB's human factors experts could develop better ways the NWS and the larger meteorological community performs in situations like these. 

My Tornado Warning Work

I am confident a few of you are asking, "Mike, why didn't you tell your readers this was actually a false alarm, instead merely pointing this was strictly an NWS warning?" and that's a fair question.

Back in the 80's, a human factors group in Colorado learned that, in order to get people to respond in tornado warnings, flash flood warnings, blizzard warnings and other life-threatening situations, it is important that the meteorological community be seen as a "team." For example, when the tornado sirens sound, residents of the area (at night) turn on the television and immediately start "channel changing." If station A says, "Go to the basement, now!!!" and Station B says, "This is no big deal," people are simply confused. It is always my goal to bring light, rather than confusion, to these situations. So, I made the decision to tweet the warning but make sure it was clear it was the NWS saying a tornado threatened. 

Did I do the right thing? I'm not sure. That is why we need a Natural Disaster Review Board!

© 2024 Mike Smith Enterprises, LLC


  1. Often, meteorologists will only use the SRV product, while not checking the BV velocity data. While not an issue most of the time, sometimes errors like these are made, although I would prefer a false alarm over a unwarned tornado like the on in NC last week.


Post a Comment

Popular posts from this blog

Hilary's Forecast Path Shifts West; Updated 9:20am PDT

Dangerous Travel Conditions - People Reportedly Stranded

Dangerous Tornado Situation Developing Tuesday and Tuesday Night