"TORNADO EMERGENCY"

[Danny Neal]

I have to strongly disagree. The whole mission of the NWS is to save lives. If the emergency text works, they need to use it! What's the point of telling EMS, First responders, etc. there's a violent tornado coming - and not inform the public as well? The public has to know - that is the whole reason the NWS saved lives on May 4th. If something goes wrong, sure they'll be blamed. If something goes right, then they saved lives. I'll take 50 false alarms if 1 right one saves 1 life. If the forecaster knows there is a tornado emergency - a near 100% chance of a strong/violent tor hitting a populated area - and does NOT broadcast it, and lives are lost, that forecaster has no one to blame but himself.

I strongly agree with you, there is no question about that. I should have said they should contact every counties EM before the storm approaches and let them know the situation before the storm even strikes, so that when a tornado warning or emergency is issued they are already prepared for a first alert response. That way the last people being warned is the general public and even then they have 20-30 mins of time before the worst of the storm impacts them. Seems to me like the Tornado Emergency call made alot of those people who say "it aint gonna happen" really get down to cover on May 4th. Although it was issued about 10 mins before the tornado struck, that is still alot of time for you to get your loved ones and run to shelter. I hate to see this phrase overused however.

 

[Jeff Miller]

Looks like Wikipedia already has this as close to an "official product" as you can get, without it actually being one:

http://en.wikipedia.org/wiki/Tornado_Emergency

 

[Gabe Garfield]

Looks like Wikipedia already has this as close to an "official product" as you can get, without it actually being one:

http://en.wikipedia.org/wiki/Tornado_Emergency

Not to be pedantic, but Wikipedia can be edited by most anyone, so it does not offer much additional credibility to your argument (though I agree with it).

Gabe

 

[Doug Lee]

When I started reading this thread, I disagreed with Mike Smith. But after reading all the comments since then, my thinking has changed.

Now I am going to say something that might seem as controversial, but hear me out before you judge.

Warnings do not save lives.

People save their own lives, by taking shelter when a tornado or destructive storm approaches them. The warning is just a message that people should pay attention to what going around them and take action to protect themselves and their loved ones, co-workers, etc.

That then means the main question is: What should the message say to prompt people to act?

What Mike Umschied did worked, no dispute there. He deserves all the respect and accolades directed his way, and the same goes for everyone in the warning process. But the real question behind Mike Smith's concern is: Will it work next time? Will it work the 100th time?

I think, but I do not know for sure, that Mike Smith's real question/concern is: Aren't tornadoes already emergencies? If they aren't, why does the NWS issue warnings for them? Then shouldn't the public regard tornado warnings as emergency messages?

I also think another issue comes into play here: severe weather criteria for wind and hail. I'm kinda surprised that in 90+ posts no one mentioned it. The NWS issues a boatload and a half of SVR warnings, but then later I hear of very few damage reports confirming the need for a severe warning. This happens because the criteria for wind and hail are set too low. As a result, the public starts to ignore SVR warnings, and that effect sometimes carries over to TOR warnings.

I somewhat agree with Mike Smith that TOR warnings should avoid the word "emergency" because the emergency status of the situation is implied in the warning. But, if that's what it takes to get people to act, do it. Period. No brainer, case closed!

I believe a better approach would be to have some wording that indicates the scale of the danger presented by a particular storm. I would also like to see TOR warnings issued for specific towns or cities, even if a warning has already been issued for the county.

Now where did I put my flame-proof suit?

 

[jladue]

You know that the NWS warning forecasters are always adapting to their perceptions of public response. They know that tornado warnings have a 70% false alarm rate, if not more. That's the state of the science when no confirmed reports are available. When a confirmed tornado report comes in, the office typically puts the report out in a local storm report, in a severe weather statement, and eventually in the next tornado warning. The wording comes out and provides extra credibility to a warning through multiple products. We all know that. But what the research has shown is that people tend not to take action until they hear the threat from more than one source or confirmation. Typically a resident may get the warning on TV, and a friend calling. Or perhaps updates with enhanced wording provides the extra incentive. I've noticed the impact of the tornado emergency in Moore on two radio station jocks. You should've heard their voices become nervous when they read off that statement. When the forecasters in Norman were faced with live footage of a monster tornado that appeared not be anywhere close to roping out, they wanted to do something, anything, to provide enough urgency for people to realize the gravity of this situation. That is a situation for which most people would never have experienced in their collective memory. What the SOO at the Norman office came up with was putting the words "Tornado Emergency" into the severe weather statement. He was aware that a SVS statement didn't tone alert. The tornado warning was already tone alerting. Dave didn't waste time putting it out. Needless to say the response after the fact was overwhelmingly positive from their users.

Since then, I've seen the "tornado emergency" used for small tornado events that caused no damage. I've also seen large destructive tornadoes pass through populated places with no mention of an emergency, though the tornado was warned for. For the most part, when I've seen a tornado warning emergency SVS bulletin, there was a significant tornado causing damage in a populated area shortly thereafter. As for statistics to back up my claim, they need to be done. I'm interested to see myself.

All NWS forecasters are also always aware of their mission to save life and property. That said, Mike Smith correctly points out that that every life is important. Best intentions aside, population centers provide the greatest number of reports and invariably get the majority of warnings. The attention goes to the highest population centers because the greatest threat to life and property reside there. Therefore, I don't think it's very likely that a tornado emergency SVS would be issued for mainly rural regions regardless of whether or not it's right or wrong. There may be exceptions however.

Do I think a tornado emergency SVS is the best solution? I don't think so. Remember it was pulled out of thin air in an honest attempt to do something to save a few more lives. It's quite analogous to Max Mayfield calling the Mayor of New Orleans to say "this is the big one". Max didn't come up with that solution as a result of doing peer reviewed research. He just did it because he wouldn't have felt good not doing it. Intellectually, I think probablistic warnings would be the way to go. That is where the highest probability of tornado would be based on the strongest evidence (e.g., confirmed reports, radar signature, environment, nearest proximity to being hit), and weaker probabilities would go to less evidence (e.g., radar only indications, further downrange) that a place would get hit. But when it comes to inciting the best human response, I don't think we have anything more than instinct to guide us. I believe it's imperitive that we start merging sociological research into the mix to figure out what will be the right strategy for doing this.

BTW, if you're inclined, we teach a whole course on warning operations. The course touches on a lot of what's been discussed in this thread. And a lot of what's discussed is something that the course could also consider too. You can see parts of it at http://www.wdtb.noaa.gov/courses/awoc/index.html

 

[Joe Nield]

Warnings do not save lives.

People save their own lives, by taking shelter when a tornado or destructive storm approaches them.

Not only is that NOT controversial, it is exactly the truth.

Something the operational meteorology community learned from the 1965 Palm Sunday Outbreak is that the perfect forecast or warning means nothing if the people who need it do not receive it. That is a lesson we continue to learn and work towards today. Response is arguably the most important and least controllable of part of the integrated warning system chain.

We (NWS) are basically in the business of information provision for safety purposes. Hopefully, if we do our jobs well, we provide people with the information they need to stay safe.

Sociologists have found that everyone goes through a five step process when they are under a warning. First, they have to receive it. If that doesn't happen, process over. Second, they have to understand it. Third, they must verify it (turn on the TV or radio, go outside and look around). Fourth, they must personalize it, meaning that they must believe that they are under a threat. Only after all of those other four steps have taken place, will the fifth step occur: TAKE ACTION.

Research has found that this process takes place essentially every time a person is affected by a warning, regardless of previous false alarms.

In the end, all we can do as warning meteorologists is wave our arms and shout at people as clearly as we can in our products to try to convince them to take action.

Standard disclaimer applies. All opinions are mine alone.

 

[Damon Poole]

Remember, the phrase "Tornado Emergency" is enhanced wording used in a Tornado Warning or Severe Weather Statement to highlight imminent danger, used at the discretion of the local NWS office issuing the product. This is exactly the same context as the Particularly Dangerous Situation (PDS) phrase in Tornado and Severe Thunderstorm Watches. It's an enhancement to existing products, not a new type or level of warning product. It's not designed to cover all types of warning situations, only the most extreme and dangerous. When used properly, as we've seen, it obviously works brilliantly. I say if it works, do it!

 

[rdale]

Nobody said that it didn't do a good job in several high-profile cases... But since "TE" has been used for small tornadoes with minimal damage - how do you set the criteria? How do you keep TE from starting to mean nothing different than a "regular" warning?

I think what people are looking for is some sort of criteria -- if it's going to be issued, make it have the same meaning regardless of who's issuing it.

PDS boxes have set, numeric criteria they have to meet.

 

[Scott Overpeck]

I've been reading this thread for a while mainly because it has been interesting to read about how others respond to warnings and what a warning should mean. I think there have been some very good points made. Still, the main question that a NWS warning forecaster has to think about is, "How will Joe Public respond to this warning?" Will the public understand the threat?

I agree with Mike Smith's premise that by virtue of a tornado warning it is an emergency. But as Jim Ladue pointed out, there is a 70% false alarm rate or more with tor warnings, so how does the public perceive the threat with such a false alarm rate? That is where we have to ask, "What wording in a warning will get the public to act to secure their lives and property?" Other than using "tornado emergency" in a SVS, the only thing the we (NWS) can do is confirm the threat. The warning will go from radar detected to confirmed assuming there are enough ground truth, and trusted ground truth of a tornado. Those reports hopefully will also communicate back to us the degree of the threat. As a warning forecaster, we are responsible for communicating that threat. With what we have now, that leaves special wording in an SVS or a new tornado warning if one expires. I think communicating that a tornado is confirmed on the ground, it is large and will cause a large degree of damage needs to be done, but in few words. That is where "tornado emergency" comes into play. In two words, I can better communicate the threat. Until polygon warnings and probabilistic warnings are better developed, that is what we are left with. But still, given a tornado warning with a probability of a tornado at 100% for Whatever City, I wonder how the public will react to that. Will that be enough for someone to take action or will heightened wording still be needed? Seems like we will deal with this question for a long time to come.

As for Doug Lee's question about warning criteria being too low - good luck with getting an answer. That is a box of worms better left closed. Just my 2 cents worth.

 

[Brian Stertz]

I think this Tornado Emergency thing really is a no brainer, and as with all things involved in the politics of NOAA/NWS it is likely to get bogged down in the basic checklists of how such an event will be designated for this higher awareness SVS level. Certainly if multiple chasers, spotters, LE officials, and general public can provide solid backing by visual obervations of the tornado and the damage it has done, this can surely be enough cause (if supported by undeniable radar backup) to present the Tornado Emergency awareness designation. This will present the public with heads up that something bad is approaching and that they need to be on the ready to make immediate decisions to protect their lives and get into shelter. In the case of this being a Friday night in Greensburg KS and a mile wide plus tornado is bearing down on the town, a Tornado Emergency was certainly in order, and sadly validated as this "brush hog" violent tornado tore across the town. Had this been in the 1940's/1950's or earlier the death toll probably would have been tremendous. The TOR Warning did it's job and the Tornado Emergency/SVS helped to support that this was in fact a very dangerous and life threatening tornado approaching in the night.

 

[Robbie Cox]

Not only is that NOT controversial, it is exactly the truth.

Something the operational meteorology community learned from the 1965 Palm Sunday Outbreak is that the perfect forecast or warning means nothing if the people who need it do not receive it. That is a lesson we continue to learn and work towards today. Response is arguably the most important and least controllable of part of the integrated warning system chain.

We (NWS) are basically in the business of information provision for safety purposes. Hopefully, if we do our jobs well, we provide people with the information they need to stay safe.

Sociologists have found that everyone goes through a five step process when they are under a warning. First, they have to receive it. If that doesn't happen, process over. Second, they have to understand it. Third, they must verify it (turn on the TV or radio, go outside and look around). Fourth, they must personalize it, meaning that they must believe that they are under a threat. Only after all of those other four steps have taken place, will the fifth step occur: TAKE ACTION.

Research has found that this process takes place essentially every time a person is affected by a warning, regardless of previous false alarms.

In the end, all we can do as warning meteorologists is wave our arms and shout at people as clearly as we can in our products to try to convince them to take action.

Standard disclaimer applies. All opinions are mine alone.

Joe,
Could the NWS implement the reverse 911 system to get the warning out to the most people? Just think how many more people would get the announcement if they got a phone call at the workplace or home making it more personnel of the danger coming.
Would this not eliminate the first 4 steps in the process and take action sooner ? NWS has done this with cell phone alerts.

Also with the broadcast flag bill going in effect in 2009 how will this affect the NWS getting it's services out to the most folks that can't afford to update to a converter box or afford cable or satelite ?

My problem as a asst.chief is by the time the weather services issues a warning teletypes it to the county then goes Thru every tone for each dept the information is already 6-10 minutes old and that's just to get the info not including time to respond to the darn thing .

 

[John Robinson]

Robbie raises some interesting points.

First, I'm afraid NWS' using something like Reverse 911 would present several problems. One would be trying to maintain a gigantic list of ever-changing phone numbers. Also, I suspect that there would be some objections, privacy-wise, to a government agency having access to everyone's phone number. Then, there would have to be a lot of phone lines available to make the required number of calls in a very short period of time. Next would be the reliability of the telephone service during a tornado outbreak. A previous tornado could easily knock out service to an area before a warning was issued for that particular area. Our WFO, in the middle of the state, would have to call all the way to three of the state's borders. As an example, earlier this year, the Dumas, AR, tornado knocked out all phone service, both land-line and cell, to that area. Then, later that same afternoon, another tornado warning was needed for the same area.

Regarding the switch to all-digital TV in 2009, that's certainly going to be a concern. It wasn't all that long ago that the local TV stations became available on satellite TV. Prior to that, people in the rural areas (a huge part of the state, in other words), couldn't see the warnings via the local TV stations.

For the problem of the first responders having access to the warnings in a timely manner, our WFO worked diligently for close to a year to gain access to the state's new digital radio system. Now, as soon as a warning is issued and it comes off our printer, it is announced by us on the assigned radio channel. The state police and the state emergency management agency are required to monitor this particular channel. And, the local police and fire departments are aware that the warnings will be broadcast. All they have to do is leave one of their radios tuned to that channel. This system has worked well enough such that two other NWS offices that serve other parts of the state are now trying to get onto the system. It's a little tougher for them, however, since they are outside the state's borders, but this is being actively worked on. Earlier this year, when the Dumas tornado knocked out all the phones in that area, we could easily talk directly to the Incident Commander, state police, and emergency managers via the radio when we needed to reach them. It worked great. The new radio system, by the way, was the state's answer to the interoperability problems that plagued NYC on 9/11.

 

[Rich Thompson]

The tornado emergency wording can be added at any time to a tornado warning or an SVS. The wording of those products is at the discretion of the warning forecaster, and should reliably and concisely reflect the threat to life and property. David Andra (OUN) and Mike U. made the call when faced with a once-in-a-career situation, and the heightened awareness seems to have helped.

However, as Rob D. and Mike S. have pointed out, what are the criteria for such wording? I was *in* the Great Bend tornado emergency and had serious reservations in real time, plus I've seen at least two other situations with that wording in W TX that didn't amount to much. PDS tornado watches are another flavor of an existing product, but they do have specific criteria for their verification: *multiple* EF2+ tornadoes, or at least one EF4+ tornado. Some sort of criteria should be set for the "tornado emergency" wording, or we'll have no way to know if they're being issued in the proper circumstances. So, we have to determine the "proper circumstances", and then verify these things each time they're issued.

Rich T.


p.s. Probabilistic warnings with appropriate wording in the text would serve the same purpose, and could even include the "tornado emergency" statement for specific situations.

 

[Jeff Miller]

Rich,

how long do you believe it will take to establish the criteria for the issuance of this - for lack of better words - enhanced tornado warning - with Tornado Emergency verbage? You mention that specific criteria needs to be developed to ensure uniform application and that it is being issued at the right time. Until that happens, there will be, as you said, issuances at the sole discretion of the forecaster. Fortunately we have some great forecasters out there with excellent judgement - but some will undoubtably be issued when not warranted.

 

[Rich Thompson]

Rich,

how long do you believe it will take to establish the criteria for the issuance of this - for lack of better words - enhanced tornado warning - with Tornado Emergency verbage? You mention that specific criteria needs to be developed to ensure uniform application and that it is being issued at the right time. Until that happens, there will be, as you said, issuances at the sole discretion of the forecaster. Fortunately we have some great forecasters out there with excellent judgement - but some will undoubtably be issued when not warranted.

Jeffrey,

It's tough to say how long it will take - it depends on whether or not anyone is interested in verifying the products. We would need to establish several "metrics" that are related to a "tornado emergency". Would it have to be a tornado of a particular size/intensity/longevity, and would the tornado have to hit a town of a particular size (area with a particular population density)? I don't have answers to these questions right now.

We could verify the tornado emergency in more subjective terms, where 5/3/99 and 5/4/07 were obvious successes, while 5/5/07 and others wouldn't be considered as successful. Whether the verification is purely subjective or semi-objective, we'd still have to establish what percentage of the issuances *should* be successful.

For PDS tornado watches, we're looking at high confidence in multiple strong/violent tornadoes occurring within the watch. If we're issuing for the right situations, we should see multiple strong/violent tornadoes occur in ~75% of the PDS tornado watches.

Rich T.