Virginia Beach EF-3: Another Inadequately Warned Tornado

It has not been a good weekend for the victims of strong tornadoes in Florida and Virginia. And, it was not a good weekend for the National Weather Service's (NWS) tornado warning program. 

We discuss the failure to issue a warning ahead of the EF-2 Palm Beach Gardens Tornado (PBG) in the post immediately below. 

USA Today

The Virginia Beach tornado has been rated an EF-3. The NWS issued a tornado warning just one minute before the storm touched down. When measured against the NWS's published goal of 13 minutes, this was another poor performance. Nor, was there a tornado or even a severe thunderstorm watch. To the people of Virginia Beach, this tornado came out of the blue. That shouldn't happen.

While this was not as straightforward as the PBG tornado, the data clearly showed enough for a tornado warning to be issued. I'm going to provide the data as I would have analyzed it. 

5:35pm EDT Velocity Data

The Wakefield NWS radar was well-positioned to survey this storm. The velocity data showed significant rotation in the circled area. 

Had I been on the warning desk, I would have been tempted to warn at this point but I likely would have held off to make sure the rotation was persistent enough to produce a tornado.

5:36pm Satellite Data

The bright green area over Virginia Beach is the jump.

The lightning suddenly "jumped" at 5:36. That makes the storm strongly suspicious. 

5:37pm Radar Data

This is an almost sure sign of a tornado: gate-to-gate rotation (where the colors are directly next to each other). If I had been working the warning desk (and I professionally warned of tornadoes from 1971 to 2018), this would instantly have prompted me to issue a tornado warning. 

So, I would have issued a tornado warning at 5:37pm. Per the NWS, the tornado touched down at 5:48pm -- that would have provided ten minutes of advance warning. 

But, the situation is even worse than these two misses.

Florida meteorologist Ed Vallee posted this, as a comment on my LinkedIn article, this morning: 

KMLB = Melbourne, Florida

In other words, not only weren't these two highly damaging tornadoes in densely populated areas inadequately warned, the rate of false tornado alarms continues unacceptably high. 

This was not the case a dozen years ago. 

From 2005-2011, the NWS was routinely attaining the 13 minutes of lead time. The average lead time was 13.3 minutes. The "probability of detection" (the likelihood that a tornado warning is issued before a tornado touches down) was 24% higher than today's! Science is supposed to advance not regress. You can find my 2011 Washington Post article on this topic here. 

You are probably thinking, "Has Mike contacted NOAA and NWS?" Yes!

All of the people with whom I have communicated are in a state of denial. In spite of these well-documented cases, the NWS thinks this these are isolated EF-0's and, maybe, the occasional EF-1. Instead, there are far too many EF-2's and EF-3's being missed.

As I have said, over and over, another Joplin -- where faulty warnings led to 161 deaths, the worst in the tornado warning era -- is entirely possible until this is fixed. But, the NWS shows no sign of acknowledging the problem, let alone fixing it. 

It may be time to call your congressperson and, ultimately, write the White House. We need a National Disaster Review Board modeled after the National Transportation Safety Board. Otherwise, I do not see the situation improving.