Monday, June 21, 2021

Friday: Another Dangerous Missed Tornado Warning

I'm very pleased to report the local National Weather Service did an outstanding job with the tornado warning for the Chicago suburbs last night. That tornado was of sufficient intensity that people might very well have been killed but for the tornado warning, especially since the tornado occurred in darkness.

Unfortunately, effective warnings of powerful tornadoes are not always available these days. 

Damaged home in Jay County Indiana
Photo: National Weather Service

May 21, the Washington Post’s Capital Weather Gang (CWG) published an article I had written pertaining to the regression in tornado warning quality since the 2011 Joplin Tornado. The article is here. I made four primary points:


-       Errors by the National Weather Service (NWS) and local emergency management lead to 161 deaths in Joplin on May 22, 2011. Weather science has never fully examined what went wrong that day and how we can ensure that type of tragedy does not recur. 

-       NWS tornado warning quality has declined since 2011. This is evidenced by decreases in probability of detection (i.e., that a warning will be issued before a tornado touches down) and a 40% decrease in “lead time” – the amount of time from when a warning is issued to when the tornado touches down. 

-       Given a myriad of improved science and technology, combined with the advent of “partial tornado warnings” since 2011, the reasons for the decline are not entirely clear.

-       The solution to the decline and a myriad of other problems related to disasters in the United States is a “National Disaster Review Board.”


In the CWG article, I cited just three of a number of of tornado warning failures so far this year. 


It is unfortunate to have to report that, on Friday (18th), another serious tornado occurred without adequate warning – exactly the type of consequence feared if the status quo is allowed to continue. 


strong EF-2 intensity tornado occurred in Jay County of eastern Indiana starting at 3:50pm EDT. It nearly immediately destroyed homes and felled a 100 ft. communications tower in its first few minutes. It made an abrupt right turn and continued to cause destruction. Fortunately, no one was killed. 


There was neither a tornado watch nor warning in effect when the tornado touched down. 


The tornado should have been warned of by 3:47pm EDT (above). Unfortunately, a tornado warning was not issued until 3:54 – four minutes after the tornado was on the ground doing damage.


It is hard to understand the watch and warning failures in this case:

  • The rotational winds triggered the "tornado vortex signature" and displayed it for at least two frames,
  • The atmospheric environment was quite favorable for tornadoes (as measured by the flavors of the SIGTOR index),
  • Weather satellite imagery depicted favorable conditions,
  • A tornado had been reported (also without a tornado warning being issued) from the parent supercell thunderstorm approximately a half-hour before,
  • Individuals independent of the NWS were tweeting about the tornado threat before a tornado warning was issued.

Friday's was not the only damaging tornado without adequate warning to occur since the CWG article was published.

The NWS' comment there was "no evidence of a tornado" is especially dismaying. Matthew Cappucci outlines the strong meteorological evidence in his tweet. And, this was another case where the NWS office was contacted by third parties urging them to consider issuing a tornado warning. 

Below is some of the damage. Fortunately, it was cattle killed rather than people.

It is fortunate the New York and Indiana tornadoes did not occur in densely populated areas. 


Dr. Harold Brooks replied to my CWG concerns by referring back to a National Weather Service policy, put in place after Joplin, to attempt to tamp down on false alarms. Dr. Brooks published a comment (below my CWG article) where he discusses that, in return for one unwarned (E)F-3 intensity tornado every three years, we now have 90 fewer false tornado warnings yearly. 


This may be true, but it hardly accounts for these recent warning failures in New York and Indiana, situations where the evidence a tornado was forming was obvious. 


Brooks did not discuss "partial" tornado warnings (PTW) nor the likely effect of those making recent tornado warning statistics look better than they actually are. But, putting PTWs aside, let’s examine his unwarned F-3/90 false alarms/yr tradeoff a little more closely: 90 false alarms per year represents a reduction in false alarms of just three percent! That decrease is imperceptible to the public. Worse, F-3 tornadoes kill 26% of the time. Using the figures pertaining to fatal EF-3, -2 tornadoes in Dr. Brooks’ 2019 paper on tornado fatalities, one person is dies annually in return for that tiny decrease in false alarms. 


Perhaps the increase in the warning system’s credibility from, say, a 40% decrease in false alarms would be worth it as it would make the warning system more credible. One might surmise that with additional credibility, more people take tornado warnings more seriously -- and they would react accordingly -- resulting in a net saving of lives. 


But for a mere 3% false alarm reduction, this seems like a very bad trade. 


As to the proposed National Disaster Review Board, on June 15 The Weather Channel released an interview with FEMA Administrator Deanne Criswell. It was conducted by Jim Cantore. The video is below. The interview discusses the considerable criticism FEMA has attracted for its allegedly weak response in the Lake Charles area after it was attacked by two hurricanes in 2020. FEMA has long had a reputation for focusing more on public relations than fast and effective disaster recovery, example here .


A National Disaster Review Board, made up of genuine, non-political experts, could determine whether FEMA and other government and non-government players in the disaster prevention and recovery space are doing the best they can with available resources. If not, it could make concrete suggestions for improvements as the NTSB so effectively does for aviation. 


Many in weather science have a strong affinity for the NWS. Some of them, forecasters and researchers, have their salaries paid, in part or in full, by NWS and NOAA. All that is great.


However, in the final analysis, the NWS works for, and is funded by, the people of the United States. If the NWS wants to build a “weather ready nation,” it is incumbent on them to make their storm warnings the best they can be. Yet, the post-Joplin problems continue and I don't see anything that is going to fix them.

That is why an independent, outside organization is needed to give the agency advice as to how to get back on the right track. I urge Congress and the Biden Administration to move quickly on this proposal.

No comments:

Post a Comment