wpostServer: http://css.washingtonpost.com/wpost2

Most Read: Local

Posted at 02:35 PM ET, 06/13/2012

Overcoming the tornado false alarm problem


A Thursday, June 7, 2012 photo provided by Andrew Kniss shows a funnel cloud, seen from Kniss's car on Highway 34, near Wheatland, Wyoming. (Andrew Kniss - AP)
Last week, in the aftermath of the June 1 tornado outbreak in Maryland, I argued that the National Weather Service needs to confront the tornado false alarm problem. 75 percent of tornado warnings are false alarms in the U.S. False alarms can lead people to take tornado warnings less seriously.

Link: 11 tornadoes confirmed by National Weather Service

My commentary generated a thoughtful discussion on how to address the issue. Readers put forward some great suggestions.

I especially appreciated the comment left by WJLA lead meteorologist Bob Ryan . It’s long, but insightful and worth reading. Here goes (edited for formatting, punctuation) - and I’ve bolded some key points:

This is a critical question and discussion. I had someone Tweet to me that Friday they were evacuated to the basement of the National Building Museum because of a tornado warning. The warning WAS NOT for downtown DC. TORNADO WARNING sets off all sorts of alarms and fears and decisons, some of which might result in more injury than from an actual small whirwind.

If NWS offices in Missouri and Kansas, after the tragic outbreaks of last year, can issue apocalyptic warnings for a possible extreme event with wording such as - “EXTREMELY DANGEROUS TORNADO WITH COMPLETE DEVASTATION LIKELY. ... SEEK SHELTER NOW! ... MOBILE HOMES AND OUTBUILDINGS WILL OFFER NO SHELTER FROM THIS TORNADO — ABANDON THEM IMMEDIATELY” - with today’s technology and science we can, on the other side, issue wording that conveys the degree of danger and risk. Not just “tornado warning” without anything else.

Example: “Doppler radar is detecting a strong thunderstorm likely producing a small whirlwind or small tornado. Winds of 50-85 mph are possible within the warned area. Stay inside and away from windows”

Or “radar and spotter observations show a DANGEROUS [emphasis mine] tornado likely producing major damage. Head to a storm shelter, bathroom or basement in the warning area.”

Or “ A LIFE THREATENING TORNADO [again my emphasis] is heading toward. . . seek a safe room, basement in a very protected area or underground shelter. This will be a devastating tornado”

We do have better skill now at seeing the likely strength of tornadoes. Beating a false alrm rate of 70% by verifying a short small EF-0 in the woods should not be a goal. Helping people make the best decision, such as not falling down the stairs at the National Building Museum for no warning or not getting into a serious traffic accident because all the driver hears is “TORNADO WARNING” IS a goal.

A “correct” forecast lacking good communication which results in a bad decision or unecessary action is a useless forecast. There are winter storms and big winter storms. There are small whirls, detected more and more, and the real McCoy tornadoes.

We all have a role in effectively communicating the real danger, beyond just yelling TORNADO WARNING so everyone will make the best decision. Some studies have shown 60% of the U.S. population still get confused between “watch” and “warning”. The science has made great advances we can make similar advances in communicating weather risk and danger. “TORNADO WARNING” with 70% false alarm rate needs fixing

Bob re-published this comment on his own blog and interestingly, Harold Brooks - presumably the same Harold Brooks who is a researcher at NOAA’s Storm Prediction Center - countered:

I don’t think there’s any evidence that the NWS can issue apocalyptic warnings with today’s science and technology with any skill.

So if the science can’t yet reliably distinguish tornado intensities and “tier warnings” so to speak, what can we do? Should we give up and just stick with the same system?

CWG commenter bdamis , who writes preparedness and response plans for the District, proposed the following - compatible with Bob Ryan’s recommendations:

We essentially need a way to create more common language that can describe the science behind a developing storm and subsequent tornado. And this information should not only be a part of our risk communication efforts but it should be shared with the public in an ongoing fashion, not just minutes before a storm. Information is power. Educate the public.

I think both Bob Ryan’s and Bdamis’ perspectives strongly argue for the National Weather Service to pilot or test some new warning messages and perhaps terminology AND ramp up education efforts.

In a professional discussion forum, I wrote the following, which I re-publish here:

My point in the [original blog post] is that we just need meteorologists and social scientists to critically look at this, and do some testing of alternative approaches . . . . It’s possible the current binary warning/no warning system is the best system. But it’s been the same system in place for years and to not explore new approaches, and to learn by trying/doing means we won’t ever know if something more effective is possible.

Your comments welcome...

By  |  02:35 PM ET, 06/13/2012

Categories:  Latest, U.S. Weather, Thunderstorms, Media

 
Read what others are saying
     

    © 2011 The Washington Post Company