by Mike Smith

In spite of better meteorological technology than ever and more raw scientific knowledge about storms, we are seeing a serious regression in a vital government program: the National Weather Service’s tornado warning program.

Tornadoes have been a bane of living in the United States since pre-Colonial times. In the late 19th Century, the Army Signal Corps attempted to create a tornado forecasting service. In spite of some signs of progress, it was shut down because tornado forecasts would, allegedly, “cause panic.” It was said that more people would die from panic induced by the forecasts than would be killed by the tornadoes.

In the 1950’s, the Weather Bureau – forerunner of today’s National Weather Service – was dragged, largely by outside events, into the tornado forecast and warning business. The Bureau achieved significant success in forecasting but, largely because of lack of adequate tools, was less successful with tornado warnings (the short term “take cover now!” messages).

That changed with better-trained storm chasers and spotters, combined with the National Weather Service’s NEXRAD – a national network of Doppler radars installed in the 1990’s. Those radars and the intense, month-long classroom training required of every meteorologist for their operation, led to unprecedented tornado warning success.

Research by Dr. Kevin Simmons demonstrates that 13 to 15 minutes of “lead time” (the interval of time from when a tornado warning is issued to then the tornado arrives) is ideal. From 2005 to 2011, National Weather Service tornado warnings averaged 13.3 minutes and tornadoes were detected in advance 73.3% of the time. At that same time, the radars were being “dual-polarized” to allow detection of tornado’s lofted debris for better tracking. Plus, the new generation of GOES weather satellites, the first that could sense lightning rates (which are sometimes very useful in determining in advance which thunderstorms will go severe or tornadic) was in operation. All of this should have resulted in new levels of tornado warning accuracy.

They did not. The quality of tornado warnings is deteriorating at an alarming rate!

I have been tracking this for the last dozen years. I wrote a piece for The Washington Post in May, 2021, which documented this trend.

By then, the tornado warning deterioration was well underway.

How have things changed since 2020?  We don’t know. The NWS’s tornado warning accuracy statistics used to be out in the open. Now, they are behind a login and password.

Before going further, allow me to stipulate: some tornadoes are not “warnable.” This can be because they are brief, because of problems with technology, or because of our incomplete knowledge of tornado science. These are not the focus of my concern.

The tragic fact is the Weather Service is missing strong tornadoes that are obvious on radar and, in a few cases, even after they are reported by trusted ground spotters and chasers.

The National Weather Service and local emergency management botched the warning of the May 22, 2011, Joplin Tornado which killed 161 people – by far the worst death toll in the tornado warning era. I researched and wrote a book about it: [link]

At first, it seemed Joplin was an isolated event. Now, the tornado warning misses are coming at an accelerated rate. And, the NWS is missing tornadoes across the nation, from New Jersey to Colorado and from Texas to Florida. Here are just some of the poorly-warned tornadoes:

2021

Indiana: [link]
Pennsylvania: [link]
Wisconsin: [link]
Oklahoma: [link]
Kentucky: [link]

2022

Iowa: [link]
Michigan (fatal): [link]
Kansas City (MO and KS): [link]

2023

Florida [link]
Virginia [link]
Texas (Dallas): [link]
Texas (fatal, Laguna Heights) [link]
Texas (fatal, Perryton) [link]
Texas (fatal, Matador) [link]
Colorado [link]

Remember: this list represents only some of the obvious tornado misses in the past ten years. I can provide more to anyone who wants to see them.

I readily admit I don’t know all of the reasons for alarming downward tornado warning quality trend — which is spreading like a cancer across the National Weather Service.

My educated suppositions:

The retirement of meteorologists born in the 50’s, 60’s and 70’s who not only were trained in using NEXRAD but learned to issue tornado warnings from early radar indicators such as hook echoes and right-moving thunderstorms. That experience cannot be replaced.
The four-week in-class radar and storm warning training for National Weather Service meteorologists has been discontinued. Two retired NWS meteorologists, both requesting confidentiality, recently told me that radar training is woefully insufficient in some cases.
Also playing a role is a misguided attempt to cut false alarms without the science needed to do so. NWS tornado false alarms have indeed been cut by 2%. But that is at the expense of issuing quality warnings when a tornado actually exists. The “probability of detection” (a warning out before a tornado touches down) has dropped by a whopping 24 percent!

In the past, the National Weather Service used to do “service assessments” to supposedly assess the quality of the service it provided during particular disasters. These were staffed by NWS, NOAA and related agency personnel. As you would suspect, they rarely found significant fault. Federal agencies investigating themselves is far less than ideal. Tornado-related service assessments have been fewer in recent years as any type of in-depth assessment would force the NWS to acknowledge these issues.

In my opinion, the only way to fix the tornado warning program and related issues is to create an independent National Disaster Review Board (NDRB) modeled after the hugely successful National Transportation Safety Board (NTSB). The meteorology part of the NDRB mission would be:

Investigate major weather forecast and warning failures. In addition to tornadoes, it would for example include events such as the Tennessee flash flood of August 23, 2021.
 The Board would recommend improvements.
The NDRB would also take over daily validation of the National Weather Service’s storm warnings and storm forecasts.

The National Disaster Review Board would study disaster response not just from the NWS but would also study FEMA, the Red Cross, local and state emergency management and other entities involved in a particular disaster.

To continue the status quo is to guarantee more lives are unnecessarily lost and that more mega-disasters like Joplin will occur.