Wednesday, June 15, 2011

False Alarms

Call it the emergency management question of the year: why, in this era of Doppler Radars, advanced forecasting tools, cells phones and the internet, why have more than 500 Americans died in tornadoes since January?

We've heard several theories on the subject. First, the global warming crowd chimed in, claiming that the rise in the earth's temperature is fueling more thunderstorms, which (in turn) means more tornadoes. Unfortunately for Brother Gore and his camp followers, spikes in severe thunderstorms and tornadoes are often associated with a cooler, La Nina pattern, like the one we're experiencing now.

In fact, some long-range forecasters (notably Joe Bastardi) believes we're returning to a pattern similar to what we saw in the early 1950s, which was marked by a series of deadly tornadoes in places like Flint, Michigan; Waco, Texas and Worcester, Massachusetts. Of course, tornado forecasting was in its infancy back then. The weather bureau (as it was known in those days) was reluctant to use the word "tornado" in its forecasts, fearing it would incite mass panic.

Obviously, the weather service has overcome its early hesitancy, and now issues hundreds of tornado watches and warnings each year. And maybe that's part of the problem. Broadcast meteorologist James Spann of ABC 33/40 in Birmingham has a lengthy post at his weather blog, criticizing the NWS for the high number of "false alarm" tornado warnings. A few salient paragraphs:

*I firmly believe apathy and complacency due to a high false alarm ratio over the years led to inaction in many cases that could have cost lives.

The FAR (false alarm ratio) for many NWS offices when it comes to tornado warnings is in the 80-90 percent category. I say this is simply not acceptable. Sure, the POD is excellent (probability of detection), but if most of the warnings are bad, then what good is a high POD?

I ask the NWS to consider stopping the use of tornado warnings when trying to catch small spin-ups within a squall line (or QLCS). These tornadoes rarely last more than a few minutes, and are next to impossible to detect in advance. And, in most cases, the greatest damage from a QLCS is from widespread damaging straight line winds, not tornadoes.

These kind of warnings force us to go on the air for 40-45 minutes, often after tornado signature has vanished from the radar. Sirens sound, the NOAA Weather Alarm goes off, severe weather apps on smart phones alert users. Getting these kind of warnings over and over and over again totally create an ocean of people that won’t be paying attention when a real tornado emergency is in progress.

I heard it over and over as people described their April 27 experience. “I hear those sirens all the time, and nothing ever happens”. The cry wolf syndrome is very real, and very dangerous.

*Too many people believe they should hear a siren before a tornado strikes."

I'm not a meteorologist, but as someone who's lived much of his life in Tornado Alley/Dixie Alley (and chased countless storms as a journalist), Mr. Spann raises very good points. Sirens are grossly inefficient as a warning system (as he notes in his blog); often, they sound across an entire county, when the storm is only threatening a fairly small geographic area. Many of us have heard the sirens sound on a sunny day, in conjunction with a tornado that's 30 miles distant, and moving away away from our neighborhood.

There's also the problem with power supplies; on several occasions, sirens failed to sound because the tornado knocked down electrical lines and the neighborhood was in the dark before the storm arrived. Residents waiting on the siren to take cover were surprised, and some paid for that delay with their lives. Spann believes that sirens should removed, and its hard to disagree with his reasoning. Without them, residents will be forced to rely on more accurate warning systems and (the theory goes) seek shelter sooner.

While NOAA weather radio is more reliable, it has similar problems, as Spann observes. The system needs a GPS upgrade so the NWS can deliver warnings to threatened areas, and not neighborhoods that are in the same county, but face no threat from the storm. Spann advocates greater reliance on social media and technology like Ustream (which words on any smart phone), or the iMap weather radio app, which delivers warnings to individuals inside the warning "polygon."

But there is a drawback to all of this. If you live in tornado country, you have more options than every for severe weather coverage and warnings. And collectively, we've reached the saturation point. For whatever reason, residents switch to cable channels that aren't carrying weather bulletins, surf away from weather-related websites, and push warning calls to their voice mail. The federal government wastes plenty of money on junk studies; how about investing in research that's actually worth the cost--Send teams to places like Oklahoma, Alabama and even Massachusetts to determines how many people in a warned area actually receive notice that a tornado is coming, and take necessary precautions. Compare that to the number who ignore the warnings and decide to press their luck.

That type of research--perhaps more than anything else--may determine why some people make the fatal mistake of ignoring tornado warnings, or delaying action until the last possible moment. Bottom line: there's a certain amount of individual responsibility that (partially) determines who will survive the storm. You can deploy every bit of warning technology known to man and some people won't act until they hear the sound of an approaching tornado. By then, it's often too late.
ADDENDUM: James Spann is one of the best at his craft; we watched some of ABC 33/40's coverage on April 27th; it was masterful, and worthy of a local Emmy or even a Peabody Award. But broadcast mets must also ask if their wall-to-wall coverage adds to the overkill. Spann is correct when he notes that some of the non-stop coverage is unnecessary. But it's hard to find a news director in Tornado Alley who's willing to forego that level of coverage, realizing that ratings jump an average of 10% during a severe weather event with wall-to-wall coverage. And, if you're lucky enough to have a James Spann in front of the camera, the audience figures go even higher. We haven't seen the ratings from Birmingham on April 27th, but it's a sure bet that ABC 33/40 led the pack--and by a large margin.

That's the secret of TV tornado coverage. You throw out hours of programming (and commercials) to keep the public informed, but stations also create a "brand" that will bring viewers back, over and over again. Ask a TV news consultant about the value of a Gary England (KWTV, Oklahoma City), a Dave Brown (WMC-TV in Memphis) or James Spann in Birmingham. There's a reason for those non-stop "weather orgasms" on local TV and it's rooted (in part) in attracting viewers to a particular weatherman on a certain station.
While many of Mr. Spann's suggestions have merit, some are non-starters, for legal reasons. The NWS is deathly afraid of "missing" any tornado that might kill someone, or inflict serious property damage. The weather service office in Peachtree City, GA (Atlanta) made that mistake on March 20, 1998, failing to detect an F-2/F-3 storm that struck portions of Hall and White County around 6:30 am, local time. Warnings were finally issued, but only after the storm was already on the ground.

This wasn't the first time the weather service has made such a mistake, and it won't be the last--weather forecasting remains an inexact science. And while the courts have generally ruled that the NWS is immune from lawsuits triggered by its errors, it only takes one judge and jury to open the floodgates. That's one reason the wether service errs on the side of caution and issues many tornado warnings based strictly on Doppler radar indications, without confirmation of a funnel on the ground.


Nancy Reyes said...

the joke is that you can tell an Oklahoman because he goes outside during a thunderstorm. (i.e. to look for tornadoes).

In severe storms, you can't always hear the siren because of the noise, and unless you live in a city, you might not get the warning (as you noted). I always relied on the weather channel, but our satellite dish would go off with very heavy rain.

Many folks can't afford a safe room or backyard shelter: we used our neighbor's concrete shelter across the street...and getting my elderly husband there took five minutes...

But the real problem is trailer parks and those in cars...

Unknown said...

Boinky--Agreed; I live about 60 miles from Bertie County, NC, which was devastated by a killer tornado in early April. Twelve people died in the storm; if I'm not mistaken, all but one lived in mobile homes.

I'm also amazed at the number of people still seeking shelter under highway overpasses during tornadoes. The danger of doing that has been thoroughly documented; in fact, several people died in the 1999 Oklahoma outbreak because they took cover under an overpass. But I've seen traffic cam footage during recent severe weather events in Texas (and elsewhere) showing a traffic jam around overpasses as a tornado approached; a fair number of the motorists abandoned their cars, and climbed into the rafters of the overpass. Apparently, they don't realize that a twister passing through an underpass creates an enormous wind tunnel effect, and the debris is channeled into small areas, like the space between the top of the embankment and the overhead beams--where people try to take cover from the storm.

Bad habits die hard, and they sometimes kill the folks who refuse to break them.

taylormade said...

I knew people who were killed in the March 20, 1998 tornado that this article mentions. What I am hearing is that they admit that no warning was posted because of human error. If that's true, I think the victims of that storm had every right to sue the pants off of the National Weather Service. People who were well-loved are now gone- children, parents, grandparents, friends. The cost is much too high for something like this to happen again.