National Weather Service Tornado Warning Crisis

The National Weather Service's Issues With Tornado Warnings Have Reached the Crisis Level.

Inadequate North Dakota Warning

While the National Weather Service's forecast for tornadoes in North Dakota yesterday was excellent, the tornado warning for the Bismarck area last night left a great deal to be desired. People could easily have been killed by the late tornado warning.

The National Weather Service (NWS) radar showed both a hook echo by 10:19 (left) and strong rotation on the Doppler wind display (right) by 10:20. There is no question a tornado warning should have been issued by this time. The hook had characteristics of a debris ball -- making a tornado quite likely. 

Unfortunately, the tornado warning wasn't issued for another six minutes. 

Meanwhile, chasers and static cameras (see below) were capturing and transmitting images of the tornado.

This static camera shows the tornado in the right side of the image. We don't know the exact starting time of the tornado. 

Making things worse, the original tornado warning did not properly display!

Not only didn't the tornado warning display on RadarScope (the software I use), it didn't show up on Twitter/X -- which means it likely did not trigger automated radio stations, TV stations' automatic crawl generators and other communications systems. It finally displayed at 10:41 when the warning was updated. 

While the necessity of a tornado warning was completely evident at 10:19pm, had I been working the warning desk at WeatherData, Inc. there's little question a warning would have been issued at 10:11pm. Not only was there a hook echo, the computer-generated rotation data was more than sufficient for a warning at that time (data below). 

The NWS's published goal is 13 minutes of advance warning for tornadoes. Clearly, their efforts last night failed to meet that goal in spite of adequate radar indications. 

Major False Alarm Near Philadelphia June 19

June 19, the National Weather Service warned far northern Delaware and far southeast Pennsylvania of a tornado in these words:

At 5:02pm, a large and extremely dangerous tornado was located ... 7 miles north of Newark, moving northeast at 40 mph. This is a PARTICULARLY DANGEROUS SITUATION. TAKE COVER NOW!  [the capitalization and exclamation point are how the NWS phrased the warning].

Only one problem: There was no tornado. At all. Radar showed no rotation in the area. There was no "hook echo," right-mover or other indications of a tornado. In spite of the "particularly dangerous situation," et cetera, a warning never should have been issued. 

Here's the radar at the time the tornado warning was issued:

I have plotted the supposed location of the tornado 7 miles north of Newark, Delaware. At left, no sign of a supercell (the type of thunderstorm which produces "large and dangerous" tornadoes) and no sign of a hook echo. At right is the Doppler wind display. There is no rotation -- at all!

So, the people of the Philadelphia radio and television market and the region as a whole were subjected to a high-end tornado warning for no reason whatsoever. These avoidable false alarms cause some people, over time, to be less likely to take shelter when the "real thing" presents itself. 

The NWS's inability to consistently issue accurate and timely tornado warnings has reached the crisis level. One must ask the question: Where is the top management of the National Weather Service, or NOAA or the Department of Commerce? 

What is so troubling is that the NWS ~15 to 20 years ago had none of these issues. Their tornado warnings were of high quality and improving. Then Joplin occurred. It seems to, institutionally, have knocked their warning program off course. 

While no one perished in North Dakota last night, that was good luck and nothing more. Many of these unwarned tornadoes have been fatal. Unless something -- major -- is done, it is only a matter of time until we have another Joplin (2011) where the NWS's poor warning quality contributed to the loss of 161 lives.