Tornado Emergencies

[Mike Smith]

Hi Greg... appreciate your thoughts but I have one item upon which I would like more information.

As an aside (which might be a candidate for a split topic), I question the cited usage percentages (e.g., 2-3%) for NOAA Weather Radio

Earlier in this thread, I produced a link to one such study. There was a second scientific study with the same result in the same edition of WAF. There is a third study that was done by Eve Gruntfest's group -- sometimes cited by Liz -- with the same result.

OK, here's my question: Can you cite an independent, scientific study that shows the 2-3% to be significantly in error?

Mike

 

[Patrick Marsh]

Mike,

I don't think Greg is questioning those statistics. I think he is questioning if these studies include second hand information. Using Greg's example, does the does the study poll those students at school? If so, are those registered as getting their weather information from the NWR or from the principal?

While someone conducting a study can justify using either, the fact is that if the NWR wasn't being used, NONE (in the example put forth) would have received the warning without the NWR.

Not taking sides, just trying to offer clarification.

 

[Mike Smith]

Patrick,

The link to one of the studies is provided. You can read the paper and judge for yourself.

Best wishes,

Mike

 

[Greg Stumpf]

The link to one of the studies is provided. You can read the paper and judge for yourself.

Patrick is correct in his understanding of my question. I was just curious how "receive warnings from NWR" is actually defined in these studies.

I'll do my homework and check out those three papers.

 

[Patrick Marsh]

Patrick,

The link to one of the studies is provided. You can read the paper and judge for yourself.

Best wishes,

Mike

When the warning was issued, there were 190 people in the 65 surveyed homes that sustained F4 or F5 damage. Among the 65 households surveyed, television was the most commonly cited source of the tornado warning (89%), followed by a telephone call (37%), sirens (37%), AM/FM radio (25%), and National Oceanic and Atmospheric Administration weather radio (3%). Most of the households (55%) received the warning via more than one source. Time between receiving the warning and when the tornado struck was greater than 30 min for 57% of the households, 15–30 min for 20%, 5–15 min for 17%, and less than 5 min for 5%.

The study states that 3% got the warning information by NWR and it goes on to say that 55% got the warning information from multiple sources. The study doesn't say if the 3% includes these multiple sources in it's statistics, or if they only used the first source. The study also doesn't say what constituted "getting the warning information" nor did it say if respondents considered receiving "second hand information," such as mentioned in Greg's post, in answering the survey.

Furthermore, 65 households in the Oklahoma City Metro area is highly unrepresentative of the nation.I wouldn't classify this as a scientific study on where people get their warning information, but rather a study on how a specific subset of the population responded to a PARTICULAR event.

I'm not questioning the statistics that are most often cited. I question the methodology. In Greg's school example if the study you linked to talked to the students, we have no idea if the surveyors would have said the students received their warnings from the NWR or the principal. If credit is given to the principal, then the statistics of NWR users would be lower; however, I would argue that NO ONE would have received the warning without the NWR and the statistics are skewed. I think Greg's point is a valid.

Can you cite a scientific study that takes my concerns into consideration and still comes up with a 2-3% value?

 

[Glen Romine]

While the topic has drifted some, and I'll offer nothing to help, it may be worth noting that while NWR may not reach a very large audience DIRECTLY, the audience is likely much larger INDIRECTLY, such as through radio, television, and emergency management agencies most of which do receive warnings through NWR. The example from the Parson's plant a few yers ago is another excellent example where a private company relied on NWR warnings to alert their employees to take cover. Odds are low anyone at the plant was carrying their own NWR receiver that day, yet they all benefited from the broadcast warning. While in the digital age, many organizations may not use NWR as the primary source, it is a fairly reliable backup when the power/internet goes out for warning information (if they have it powered by batteries).

 

[Jeff Miller]

My opinions:

It has been well discussed in this thread the False Alarm Ratio is around 75% with Tornado Warnings. It is also well known that the majority of the public has been desensitized to Tornado Warnings in general, most will rush outside then take shelter in a basement, because the public is very well aware that there is a greater chance they will NOT get hit then there is a chance they WILL get hit.

Simply watch any tornado warning video and you will most likely find crowds outside watching then taking cover. This is typical American Warning response. Tornado Warnings do NOT have the same urgency response with the public that they used to. There is a certain "this cannot happen to me" sense with Joe Public these days.

Take now into effect what happens when a "Tornado Emergency" is issued. We know this is "enhanced wording", and not an official product. However, keep in mind again: the FAR ratio for TOR warnings is around 75%. the FAR ratio for Tornado Emergencies? I believe now it's around 25-30%, based on the few that have been issued. Also keep in mind that TV stations in Greensburg have been documented (I saw the YouTube video) to broadcast that the NWS has issued a Tornado Emergency for a specific town (i.e. Greensburg) - and the urgency level went up tenfold. With that said, how can this be a bad thing? An F5 ripped through Greensburg with a very minimal loss of life - no doubt thanks to enhanced urgency from enhanced wording.

Point 1: There is no way to reduce the Joe Public lethargy to Tornado Warnings, they expect it not to hit them. As was pointed out, it's not a guarantee.

Point 2: Tornado Emergencies greatly increase urgency. If used properly, this will continue to be the case, and will save lives.

I thus believe Tornado Emergencies should stay, under the following guidelines:That the tornado is confirmed by as many reliable and trustworthy sources as possible to be on the ground, heading for a large population center, and the forecaster has reason to believe damage and loss of life is imminent or occuring, and there are little signs of the tornado losing strength or dissipating before impacting said population center.

I also believe that for Tornado Emergencies to stay, there should be a rigid guideline established by the NWS to ensure it is not issued unless all criteria and circumstaces are met, documented, and a case study follow up mandatory after the issuance of said Tornado Emergency for further analysis and refinements for further issuances down the road.

I also believe documented Tornado Emergencies that have failed, and this will happen, should also be carefully researched and analyzed for discrepencies that can be avoided in the future.

I don't see how this is even debatable, or why in my opinion, such a potentially significant life saving tool could be so negatively discussed by some. Since we've seen the results of OK and KS Tornado emergencies, lauded as life saving and a great step in warning procedures and public safety, I think it's ridiculous to throw water on something so greatly needed to curb the lethargy induced by high FAR on TOR warns.

 

[Jarrod Cook]

Great discussion guys...

Again, as a reminder...a tornado emergency is NOT an official product in any way. I keep seeing people using the terms "issuing a TE," etc. A TE is never issued.

Another point I want to add to the mix regarding the 2-3% statistic of NWR listeners during a severe wx event...I haven't read the report in detail, but one thing to keep in mind is these 2-3% of folks will only hear the enhanced wording if the NWR is actually being listened to (as someone mentioned earlier). This means that if the enhanced wording is included in the TOR, then yes, more folks will hear it. If it is used in a SVS (which seems to be the majority of cases), then less people will hear this on NWR. This is because most people only listen to the broadcast long enough to get their warning, then they turn it back to silent and wait for the next tone alert.

So the point is that the enhanced wording only does good when the media is able to pick up on the urgency it brings and they are able to portray said urgency to the public as best they can. I know this was the case in Greensburg, as I was watching the live streaming broadcast and it was clear the meteorologist on air had read the SVS and he indeed did mention that a "rare tornado emergency" was declared.

 

[Mike Smith]

I think Greg's point is a valid.

Can you cite a scientific study that takes my concerns into consideration and still comes up with a 2-3% value?

I have cited three (two taken in Oklahoma, one in Colorado) and since recalled a fourth (Kansas), Dr. Scott Lillibridge's from the Centers for Disease Control. All were in the 2-3% range. How many do you want? Respectfully, your response has been hypotheticals and anecdotes. Where is a study showing a different number?

Are you contending schools have no other way to get a warning? That is silly. There are storm sirens, TV, commercial weather companies, telephone warning services, parents calling the school, etc.

Assume there has been a school somewhere/sometime that got its warning solely from NWR. Would that even increase the market penetration from 3% to 3.1% nationally? If you counted all of these hypothetical instances it might be it up to, what, 5%?

When you contend the central U.S. is not typical, you are probably right -- the surveys probably overstate the NWR market penetration because of the concern about nighttime tornadoes that does not exist in, say, California or Vermont.

I dislike rehashing things that have been argued to death a number of months ago. While I know some don't like hearing it (and please don't shoot the messenger) the fact is that NWR has very low market penetration.

 

[Alex Lamers]

A couple of observations, mainly to echo what Greg said...

1. If the phrase "TORNADO EMERGENCY" is appended to the header of a tornado warning, that will reach many more users than just those listening to NOAA Weather Radio. Many people will read the text products online or through some sort of software program like WeatherBug or Interwarn, etc. etc. I don't know if there is a comprehensive study on how many people receive warning information via NWR, but the verbiage in the warning product will reach many more users than just NWR listeners. Also, let's not forget the growing number of people getting warning/weather information via a PDA or cell phone.

2. TV/Radio broadcasters can and do convey the fact that the warning contains such verbiage. I recall seeing this watching streaming coverage of the Greensburg, KS event.

The point I'm driving at here is that the phrase "tornado emergency" will most likely reach far more than 2-3% of the population, especially if TV broadcasters mention this...and the majority of people get their warning information via television.

 

[Craig Maire II]

Here's a question for Mike, if a large and potentially violent tornado were to be reported headed directly towards Dodge City etc. with the discussion that has gone on in this thread and with you stating in so many words (hope I got this right) that it was not a good idea now that you have had time to think about it (sending out a tornado emergency statement) would you or your peers still send out a Tornado Emergency? Or put enhanced wording in the tornado warning and severe weather statements? I'm just wondering and not trying to offend anyone etc.

 

[Patrick Marsh]

I have cited three (two taken in Oklahoma, one in Colorado) and since recalled a fourth (Kansas), Dr. Scott Lillibridge's from the Centers for Disease Control. All were in the 2-3% range. How many do you want? Respectfully, your response has been hypotheticals and anecdotes. Where is a study showing a different number?

Respectfully, I have never once disputed that number. I have called in to question HOW that number is reached. A quick look back through this thread shows that you have linked only one paper and that paper did NOT explain how that number was reached. I have offered "hyptheticals and anecdotals" that further my argument that without knowing HOW the numbers are reached the number is meaningless.

Are you contending schools have no other way to get a warning? That is silly. There are storm sirens, TV, commercial weather companies, telephone warning services, parents calling the school, etc.

Assume there has been a school somewhere/sometime that got its warning solely from NWR. Would that even increase the market penetration from 3% to 3.1% nationally? If you counted all of these hypothetical instances it might be it up to, what, 5%?

I know of several schools in just the city of Fort Smith, AR who received their warning information via NWR. In fact, there was a movement to put NWRs in EVERY school.

I have not said that the schools ONLY receive their information from NWRs. They may receive their information from multiple sources. My point is no study I've seen, including the one you linked, has stated how that situation is handled.

When you contend the central U.S. is not typical, you are probably right -- the surveys probably overstate the NWR market penetration because of the concern about nighttime tornadoes that does not exist in, say, California or Vermont.

Based on this, then the study you linked to did NOT overstate the NWR market. It only contained information on the 3 May 1999 tornado that was during the daylight. I think it would be interesting to see how many people would have been using NWR if that tornado occurred 9 hours later. Just because someone used a NWR on storm A on a particular day, doesn't NOT imply they don't ever use a NWR.

Any statistician will tell you that a sample size of 65 families in a community of over 500,000 people is absolutely too small to use. Did they accurately represent the elderly, the poor, etc. What about farmers who may have one? All I'm trying to say is that you can't based on a simple survey you can't accurately capture what the reach of the NWR is.

I dislike rehashing things that have been argued to death a number of months ago. While I know some don't like hearing it (and please don't shoot the messenger) the fact is that NWR has very low market penetration.

I won't argue the small penetration of the NWR into the weather information dissemination market. I have no evidence to support or contradict this. I WILL argue the adamant use of the 1-3%. I will continue to do so until I see a study that explains how this number is reached and the study is not conducted in the immediate aftermath of a particular event. If someone wants to put a numerical value on the number of people using NWR, then someone should conduct a poll across the country (or region in question) and should NOT directly reference a particular event.

As a note, two of the studies you cite come from the same event. This is hardly corroborating. Since I haven't read each paper in that edition of WAF due to time restraints, I haven't found the other reference, my initial thought is that the same data set MIGHT have been used by both papers. I haven't read the WAS*IS paper because I haven't been able to find it.

Again, all I'm saying is that without knowing HOW the 3% number is reached, it shouldn't be stated as fact. I will say that 3% of the residents affected by the 3 May 1999 tornado that hit Moore got their warning information from NWR.

 

[Mike Smith]

Here's a question for Mike, if a large and potentially violent tornado were to be reported headed directly towards Dodge City etc. with the discussion that has gone on in this thread and with you stating in so many words (hope I got this right) that it was not a good idea now that you have had time to think about it (sending out a tornado emergency statement) would you or your peers still send out a Tornado Emergency? Or put enhanced wording in the tornado warning and severe weather statements? I'm just wondering and not trying to offend anyone etc.

Are you asking Mike Umsheid or Mike Smith?

 

[Craig Maire II]

Sorry I should have clarified, this question was for Mike Smith. I may have made a mistake, does Mike Smith work for the NWS in Dodge?

 

[Mike Smith]

Patrick,

I am attempting to generate light rather than heat, which is why I suggested you look up the studies so you could satisfy yourself or point out what you believe is wrong with them. I don't have a problem with them. I think it is unreasonable to request that I type the results or do a detailed defense of those studies in a thread like this. But, to give you an overview (from memory):

Two looked at Moore-OKC, Scott's looked at Wichita-Andover. They interviewed actual victims of the storms. In each case, they asked people how they got their warning. I don't see anything wrong with that approach. In these three cases there is no issue about "sample size" because they attempted to ask every single person who was "hit" by the storm, including apartment dwellers, people in base housing at McConnell, AFB, etc. In the case of fatalities, they asked neighbors, relatives, etc.

In the Colorado study (2002), they asked two groups (one college students, the other non-college residents) how they would like to get flash flood warnings at night and got the same result. Note they asked students, one of the groups about which you have expressed concern.

While I do not know the sample size, I was able to find a couple of graphics summarizing results:

1. What are your weather information sources? (they could name as many as they wanted)

NWR
Students 4% Residents 3%

Since they could name as many as they wanted, the number using NWR exclusively would be smaller.

2. How do you prefer to get your warnings? (Asked by time of day)

NWR
2:30am 3%
11am 4%
5:30pm 4%

Keep in mind the Ft. Collins Flood was still in peoples' minds when the survey was taken and that there was and is considerable publicity regarding the threat of nocturnal flooding from Boulder Creek.

Boulder is a "NOAA" town. My sense is that if the numbers are this low in the Boulder Creek area as well as in Kansas and Oklahoma, it seems reasonable they are not higher elsewhere.

I'll stand by the 2-3% as a national average that is, if anything, probably generous.