The Tornado Warning System Failure at Mayfield, Kentucky

An Investigative Report

© 2021 Mike Smith Enterprises, LLC

Since Friday's horrific tornado outbreak, a great deal of criticism has been leveled at the management of the candle factory owned by the Mayfield Consumer Products Company (MCPC). Criticism is warranted because sheltering should be carried out whenever a tornado warning is issued for a given location. 

But there is more to the story -- information that may help understand the thought process behind the decision not to shelter, or not to fully shelter, Friday evening when the violent EF-4 intensity tornado approached (I'm relying on the media on this topic). This was yet another situation demonstrating the urgent need to reform of the National Weather Service (NWS) and emergency management (EM) tornado warning processes across the nation. For those who are new to this blog, two of many other examples are here and here. 

In 2005, the NWS switched from using counties for tornado warnings to polygons that project the path of a forecasted tornado. At 5:28pm Friday, the NWS office in Paducah, KY issued a tornado warning for southeast Graves county for a storm of mild tornado potential located south of Mayfield. The storm was moving northeast. The polygon is shown below.

Plotted on top of the polygon (red) is the Doppler wind data (greens).
The location of rotation and direction of movement depicted in blue.

Note the official tornado warning polygon did not include the town of Mayfield. Yet, for reasons unknown, the NWS made the decision to insert Mayfield into the list of "locations impacted."

Local emergency management sounded the tornado sirens in Mayfield, even though it was not in the polygon. 

The illustration below confirms the rotation never approached Mayfield.

No tornado, large hail, or thunderstorm winds of 58 mph
or stronger were reported with this storm. The tornado warning was incorrect.

Confusion can result whenever there is a discontinuity between the a tornado graphic and text. Similar confusion has shown to be deadly in past tornadoes such as during the 2011 Joplin Tornado in which 161 died. 

Per the media, the siren activation caused people MCPC to send employees to shelter. From NBC News:

Employees were allowed to shelter for the first tornado warning. 

In Mayfield, the quality of the warning was manifestly terrible: there was no thunder, and it didn't even rain. No tornado occurred inside or outside of the polygon. When the violent tornado struck four hours later, it is unknown the extent to which the 5:30pm false alarm might have influenced some in the public not to seek shelter when the "real thing" arrived.

News reports indicate the situation at MCPC was confusing when the 9:30pm warning was issued:

Kyanna Parsons-Perez was in a storm shelter with other employees at [the] candle factory on Friday night when she felt wind, even though they were inside the building. 

So, at least some sheltering did occur at the candle plant. Again, my position is that sheltering should occur whenever a tornado warning is issued. But if there was sheltering hesitation, it is understandable after the bogus warning just four hours before. When sirens are sounding, the general public has no way of knowing whether a violent storm is approaching (9:30pm) or the warning is completely bogus (5:30pm). 

On Twitter, a number of people have asked why so many died. One could speculate that the earlier false alarm might have produced sheltering reluctance among Mayfield's general population. We just don't know. This is why the United States urgently needs a National Disaster Review Board, modeled after the extremely successful National Transportation Safety Board.

Both the inclusion of Mayfield in the earlier warning by the NWS and the decision to sound the sirens in the city were incorrect. Both indicate the lessons of Joplin have not been learned. The National Weather Service's own national statistics indicate the accuracy of their tornado warnings has decreased in the last decade. And, emergency managers continue to create more reasons to sound sirens. See example below.

In St. Louis, sirens are now sounded for severe thunderstorms
in addition to tornadoes: a recipe for more warning fatigue.

A new National Weather Service director will be appointed in the next few months. I am hoping that he or she will be sensitive to the many issues with the NWS's tornado warning program and will be determined to fix them. The taxpaying public deserves no less.

As to businesses, they should work with private sector meteorology consultants to create a plan unique to their locations and business requirements. Companies should then use commercial meteorologists to issue site-specific storm warnings that fit directly into their plan.

I am not a long-range forecaster. Some meteorologists who are believe that the relatively mild tornado seasons of 2020 and 2021 may not last into 2022. So, it is urgent to review your company's operations pertaining to tornadoes and make changes as we head into 2022.