Again! Fatal Tornado Strikes Perryton Without Sufficient Warning

Shockingly and sickeningly, another fatal tornado -- one that virtually certainly would have been effectively warned of 15 years ago -- has ravaged another community. 

An Investigative Report

© 2023 Mike Smith Enterprises, LLC

Bumped to Top of Blog Monday Afternoon

Please see update at bottom. 

Since the fatal, unwarned May 13 tornado in Laguna Heights, TX, members of the National Weather Service (NWS) have privately reached out to me and explained they believe the root cause of these issues is inadequate radar and storm warning strategy training. I'll have more on that below. 

For ease of reading, here is a map of the northeast Texas Panhandle. As of this time the tornado has been rated "at least EF-2" (final rating EF-3) on the Enhanced Fujita Scale. The death toll is three and more than 50 hospitalized. 

The Perryton Missed Warning Opportunity

Earlier in the day, NWS's Storm Prediction Center had forecast an enhanced risk of tornadoes in the Perryton area. That should have been the signal to the meteorologists in Amarillo, who have warning responsibility for Perryton, to place the radar on 80-second "tornado mode" (my term) when thunderstorms developed. Unfortunately, the radar was left on 240-second "precipitation mode," the type of setting that harks back to the middle 1990's when these radars were first installed. When the radar is in 80-second mode, developing tornadoes can be detected much more quickly. It should be routine procedure at every Weather Service office (WFO), when tornadoes are forecasted the radar is put into tornado mode.

Keeping in mind I issued tornado and other storm warnings from 1971 to 2018, I would have issued a tornado warning for Ochiltree County and Perryton at 4:44pm CDT based on rapidly increasing rotation that was moving from southwest to northeast. Please see below.

The left panel is the "reflectivity" (precipitation intensity) data that you see on television. A weak echo area in a situation like this can mean that rotation is spinning the rain and hail away from the center of rotation; meaning a tornado is present or is forming. On the right is the rotation at that time. 

Evidently, the NWS in Amarillo saw the same thing at about the same time. But, instead of issuing a tornado warning, they took an existing severe thunderstorm warning that warned of "minor damage" (see below) after "impact."

But, at the end of the message, they added the words, "Tornado Possible." Unfortunately, that does not cause the sirens to sound and the television stations to go into full tornado mode. The fact is that people do not take shelter when "minor damage" is forecasted to occur. 

The next radar data I'd like to bring to your attention is from 4:46pm. The first circulation, #1 in the upper right image, is slipping past the city just to the north.

However, a second (#2), broader, circulation has formed just west of the city. [Ignore the bright orange area. It is a radar artifact called a "side lobe."] Troublingly, there may have already been a tornado at this point. Coincident with #2, the lower left image shows what may be a lowered "correlation coefficient" -- which is a sign of lofted debris. I'm not sure at this point. But, in my mind, there is no question a tornado warning should have been out for Perryton at this time. 

At 4:59, the first circulation is passing to the north while #2 is a classic pre-tornado signature (right frame). Dry air is feeding into the storm at about 8,000 feet (left frame) which is also favorable. Finally, until this time, the parent thunderstorm was moving northeast. Here, it makes a sharp right turn -- another strong sign of tornado formation. 

While I don't know the exact time the tornado touched down because of the radar's 240-second data interval, I imagine it was on the ground or close at this point. It was moving east at 5:02pm. The hail to the north of the rotation was in a classic tornadic supercell configuration to the north of the tornado.

There was still no tornado warning. 

At 5:06pm, the time of the tornado warning, there is a tornado "couplet" (upper right frame, bright green next to bright red) over the north part of Perryton -- a sure sign of a tornado in progress. The spectrum width (lower left) showed highly disturbed air (white). It was at this time, finally, a tornado warning was issued... only after the tornado was moments from tearing up the town. There is video of the tornado synced with its damage path here. 

At 5:10pm on the Doppler wind data (upper right), we have a radar signature of what was a strong tornado with the with the bright green and yellow next to each other. 

The tornado continued to the east and moved out of Perryton by the next frame. 

Radar Operation

That the radar was being operated by an inexperienced meteorologist is further indicated by the "range folding" (sometime caused "purple haze" by meteorologists) that was allowed to creep in to the velocity data (upper right frame) in the last two data periods. The "pulse repetition rate" can be adjusted by the operator to keep that from occurring. 

Had the radar been operated properly, at 80-second data intervals, other meteorologists and emergency managers may have had the opportunity to issue their own warning to the people of Perryton earlier than the National Weather Service's which may have given some people the opportunity to seek shelter. When the NWS mis-operates the radar, and there is no other radar in the vicinity, everyone loses. 

In the 1990's and early aughts, the NWS had a classroom radar training operation set up in Norman, Oklahoma. All of their meteorologists spent four weeks learning how to operate their new radars (same ones they are using now) and how to issue warnings. That training is long gone and, as my generation retires, it is gone forever. One of the meteorologists told me, "A new meteorologist is told to sit next to a senior meteorologist and watch him issue warnings for an hour. The meteorologist-in-charge checks the box saying the new person has had radar training!"

What Is the NWS's Warning Strategy?

I don't know any more. I haven't a clue. Their stated, written goal is a "lead-time" (the interval between the warning and the touchdown of the tornado) of 13 minutes. But, at least to me, it doesn't appear they are even attempting to achieve that type of lead-time. The real tragedy is 15 years ago, this tornado would almost certainly been warned of effectively! NWS tornado warnings are regressing!

I would have liked to wait on this report until the damage survey was complete and other data could be gathered. I will likely revise it when those are completed. 

But, I decided to rush this out because tornadoes may occur tomorrow and I'm hoping that exposing these issues may help with tomorrow's warning operations. 

There were major warning issues with the 2011 Joplin Tornado and 161 souls lost their lives. Since, things have only gotten worse. Right now, I believe -- absent major, urgent change -- another Joplin is inevitable. 

Update Monday, June 19.  I received a call this afternoon from a meteorologist (who strongly requested confidentiality) who was on the northeast side of Perryton when the tornado formed and moved across the town. 

He said the analysis of the first small scale low pressure system (meteorologists call it a "mesocyclone") that I label #1, above, is correct. The wind was from the northwest as the "meso" went by then suddenly shifted to the east and southeast as #2 strengthened and dropped the tornado. 

Given the meteorological situation, he says, the "National Weather Service should have been on 'triple red alert' and should have been operating the radar at 80-second intervals," as discussed above. "Synoptically, there were similarities to Joplin in how fast the tornado formed."

He told me he never heard a tornado siren. There is also a post on a website from a different individual complaining about the sirens sounding in nearby Booker, TX but -- evidently -- not in Perryton. 

My Comments: The person who called me has a sterling reputation as a severe storms meteorologist. The bottom line is that the situation is even worse than I thought -- and I thought it was getting to be pretty bad!