National Weather Service Announces Unfortunate Tornado Warning Experiment

[Keith Wilson]

Melbourne NWS in August, 2005, received praise for issuing a tornado warning for the 100 mph winds associated with the decaying eye of Hurricane Charley. It spread across the NWS and morphed into something unfortunate: Telling people in the path of Katrina to go to the lowest floor as a 30 ft. storm surge came in.

This I wholeheartedly agree with you on.
To a certain point I can understand their thinking since many folks in that area are transplants that may not be familiar with hurricanes, but the blanket tornado warning issued primarily for a high wind event is redundant when there is already a hurricane warning in effect. I live about 25 miles inland from the hurricane prone Gulf Coast. Overlapping hurricane/tornado warnings in that situation is a dangerous practice IMO, and since then it has been a common practice that I disagree with.
Within the eyewall of a strong hurricane, it can be difficult to distinguish between the hurricanes straight line wind damage from tornado or gustnado damage anyway.

 

[rdale]

Keith, that's why offices use the EWW product for that now...

 

[Jeff Duda]

There is a thread in here discussing the need for better advanced warning time, but that seemed to boil down to being impracticle right now due to the negative effects that earlier warning would have on the accuracy % of the larger warning area actually experiencing severe conditions.

That being said, if longer warning times aren't practical, then maybe more detailed information pertaining to the specific threat levels within a warning is the direction that needs to be explored. Given the events of last year, I would rather see them try this sooner than later and I hope that it experiences at least some measure of success to build on.
Granted, a better job could have been done getting the word out about the changes.

The wild card in the experiment is the accuracy of the information that they base the threat levels on.

You're touching on something very similar to what the Warn-on-Forecast project is all about. I agree with that methodology. I think a better way to improve the current warning system is to reduce the sizes of the polygon warnings and be much more specific about which areas are under the greatest threat. For example, instead of saying, "THIS TORNADO WILL IMPACT...BLANCHARD...NEWCASTLE...NORMAN...MOORE...OKLAHOMA CITY," say instead, "THIS TORNADO WILL IMPACT WESTERN NORMAN...ESPECIALLY AREAS WEST OF INTERSTATE 35...EASTERN MOORE...ESPECIALLY AREAS EAST OF INTERSTATE 35...SOUTHEASTERN OKLAHOMA CITY...ESPECIALLY AREAS BETWEEN SOONER RD...SE 149TH ST...INTERSTATE 240...AND ANDERSON RD."

I also think reducing the maximum possible warning length would help. I swear I've seen some TOR warnings issued for more than 60 minutes before, and plenty that last for longer than 45 minutes. While I'm still in favor of the longest possible lead time, I think how you represent that lead time is as important as how much lead time. If your warning text says, "TAKE COVER NOW," like pretty much all warnings do, that is not the way to go about having a long lead time in my opinion. I understand it's a blanket warning meant to cover all situations people could be in, but some people only need a minute or two to seek shelter; others may need significantly more time, though. However, if you're telling someone who is 15...20...30 minutes downstream from the area of greatest threat within a tornado warning to take shelter immediately, you should expect them not to follow your instructions. If you keep the warning lead time down, you will probably be more likely to get an immediate response from people. Obviously for situations like large gatherings (football games, indoor/outdoor concerts etc.) you will need as much lead time as possible, but those events are generally fewer and farther between, and I think a good manager of those events would already be in touch with a local meteorologist or a private company to get their warning. They wouldn't necessarily need one from the NWS. As long as you add areas downstream and clip areas upstream on a regular basis (say every 10 minutes) and keep the warning sized so that a person is never more than say 30 minutes away from the greatest threat, I think you'll see a better response from others.

Also, warnings should be narrowed so that people can stress that the polygon itself consists of the area under the greatest threat (however high it may be), and your chance of being impacted by the threat decreases the farther away you get from the polygon. I know current warnings already have that premise included, but I don't see it stressed very often, and I think the boxes are still too large. You don't need to TOR warn an entire supercell when the tornado threat only exists for the southern flank of the storm. And tornadoes don't wobble 10-15 miles back and forth on a path, yet some warnings are still 20-30 miles wide.

 

[Kelly Sugden]

You're touching on something very similar to what the Warn-on-Forecast project is all about

Warn on forecast is far away from actual real-time implementation. Using model forecast updraft helicity for a warning is not a good idea with the current state of NWP. I l like to add this is coming from a guy who likes the HRRR. I've been very impressed with the results. Anyway, BAMS had an interesting article on this recently but this is far away from actual usage.

 

[beaudodson]

I've read through the PDD several times now and unless I'm missing something, it very obviously states "3) Audience - The target audience for this product includes: national, state and local emergency managers; media partners; the private weather enterprise; government and military agencies." Nowhere does it state "the general public". As I read it, it was developed from the results of the recent service assessments. Certainty something needs to be done to "warn" the public and I see this as a step in the right direction. It appears that the "audience" needs to take this information and then disseminate the information to the public in such a manner as to encourage them to make the right decision regarding their own safety.

Mike - a question for you; Could you provide some examples that you think could improve upon what this PDD is trying to address?

And according to the PDD date, it was released on January 11, so yes, nearly 3 months. Does anyone know with certainty that the target "audience" was not provided this document back then and its now only coming to light outside of the targeted groups??

You are correct. It never states the audience is the public. I think that has been overlooked by some.

More information is great. When I read a warning the first thing I look for is anything that says confirmed tornado - confirmed damage - confirmed large hail - confirmed anything. I like the idea of more information - I like the idea of tags.

Will the public understand this? I have no idea. I don't think that is the purpose of the experiment.

The am concerned that the NWS has not communicated these changes to on-air meteorologists. I know local meteorologists (on-air) that have not been told about the changes and this impacts their markets. I hope that the NWS is going to contact those in charge of communicating their message. If not then this experiment will be a waste of time.

I am not sure (Rdale) that this is going to stop tornado warnings for squall lines - QLCS events. Why do you think that? I spoke with a local NWS employee a few days ago and that is not the impression I got from his statements. Maybe I misunderstood you.

There are pros and cons to this project/experiment. For one - I will be happy to receive as much information as possible from the warnings products (although some offices have always provided as much information as they possibly could).

The big question, for this particular topic, is how to the on-air meteorologists respond to these changes. Will they willingly participate - will they convey the tiered level warnings to the public? I know some won't. Some will. For the experiment to be a success everyone needs to be on-board.

I will be curious how one ramps up a warning and then ramps it back down. Tornadic storms can go from nothing to an EF5 in a matter of seconds/minutes (Joplin, MO). How does one issue a level 1 warning - then upgrade it to a level 3 warning - then downgrade it to a level 1 warning - all within one or two counties? How will that be broadcast on the NOAA Weather Radios?

I wish the NWS would stop issuing severe weather statements for tornado warnings - weather radios that are programmed just for tornado warnings do not activate for the follow up/updated tornado warnings. Maybe this new tiered system will address that problem. I know I am not alone in my wishing on that subject.

 

[beaudodson]

The "creep" I described has certainly occurred: Very few TE's since Greensburg have verified. Patrick Marsh is preparing a formal verification that I know he plans to post when it is complete.

It would be interesting if Patrick posted events where a tornado emergency SHOULD have been issued but was not.

Comparing tornado emergencies that busted and events that should have had a tornado emergency would give us a better idea of whether forecasters have enough confidence to know when to issue one and when not to issue one. This is part of the debate on this thread. Can forecasters make a call on that subject - can they make that call enough times to justify the tiered warnings.

 

[Jake Orosi]

The big question, for this particular topic, is how to the on-air meteorologists respond to these changes. Will they willingly participate - will they convey the tiered level warnings to the public? I know some won't. Some will. For the experiment to be a success everyone needs to be on-board.

I think that's a big part of the experiment. If given differentiated information, will broadcasters relay that difference to their audiences, or is "tornado warning" ultimately all they will hear?

For what it's worth, a lot of TV stations just scroll the actual language of the product across an on-screen banner, or interrupt the broadcast audio with audio of the NWR alert. Only a few broadcasters break in live and explain things, and often only when a metro area is under the gun. That's the impression I've gotten at least.

 

[beaudodson]

I think that's a big part of the experiment. If given differentiated information, will broadcasters relay that difference to their audiences, or is "tornado warning" ultimately all they will hear?

For what it's worth, a lot of TV stations just scroll the actual language of the product across an on-screen banner, or interrupt the broadcast audio with audio of the NWR alert. Only a few broadcasters break in live and explain things, and often only when a metro area is under the gun. That's the impression I've gotten at least.

In addition to this (and I see this being a problem) will on-air broadcasters be unified in their approach? Or more likely - will one station tell their audience one thing and another station tell their audience another. I can already answer that question for my area - they won't be unified. This will cause further confusion and problems.

BUT - like someone else said - is that the problem of the National Weather Service?

 

[rdale]

I am not sure (Rdale) that this is going to stop tornado warnings for squall lines - QLCS events. Why do you think that?

Because that is what the directive says

For squall lines with embedded brief circulations, "Finally, the option to include a tag line which explicitly states "TORNADO POSSIBLE" will be available."

You can't do that for a tornado warning, since the low end states "This warning type will be selected for cases in which there is credible evidence of a tornado."

Nobody would claim that a one-scan 60kt GtG shear which disappears in 4 minutes is "credible evidence" - which is why QLCS eddies go into the SVR category now (like they were for 50 years prior to hi-res 88D.)

I will be curious how one ramps up a warning and then ramps it back down. Tornadic storms can go from nothing to an EF5 in a matter of seconds/minutes (Joplin, MO). How does one issue a level 1 warning - then upgrade it to a level 3 warning

Pretty simple. Issue a new Tornado Warning.

then downgrade it to a level 1 warning

You provide updates in the SVS. For major supercells, if they are at the point where a massive wedge is on the ground, is the information flow fast enough that it'll end (and the NWS will know) within the typical 15-30 minute tornado warning? Not likely...

 

[rdale]

will one station tell their audience one thing and another station tell their audience another.

They already do that now... When I would be on-air and the cross town station had a radio DJ reading the SVS statements, I guarantee you we were saying different things about the storm structure and severity. That didn't cause trouble then (well, maybe for the other station )

 

[beaudodson]

Not saying you are wrong - but I have some doubts that the NWS will stop issuing tornado warnings for these squall line spin-ups. I have already been told as much by some meteorologists. I hope you are right - I would love to see less tornado warnings.

And, yes - I understand what you are saying that it says on the paper.

I guess we will see how it goes. Hopefully someone will giving a running tab on all these warnings and see how the level 1, level 2, and level 3 warnings panned out.

I will certainly be curious to see how it goes.

 

[beaudodson]

They already do that now... When I would be on-air and the cross town station had a radio DJ reading the SVS statements, I guarantee you we were saying different things about the storm structure and severity. That didn't cause trouble then (well, maybe for the other station )

Yes they do that around here, as well. There are three different policies between ABC/CBS/NBC on how they are to handle tornado warnings. One station has to stay on the air for the entire warning. If they reissue the warning then they have to stay on the air for that much more time. They are not allowed to stop wall to wall coverage until the tornado warning expires. This is a television station. No matter how weak the storm or whether there has been any confirmation or not. It is painful, at times, to watch them try to come up with something to say for 30-45 minutes straight. Especially when there is only one warning in effect.

Another station only goes on the air when they think something important is going on - confirmed tornado.

The other station is sort of in between the two extremes.

 

[rdale]

Not saying you are wrong - but I have some doubts that the NWS will stop issuing tornado warnings for these squall line spin-ups. I have already been told as much by some meteorologists. I hope you are right - I would love to see less tornado warnings. And, yes - I understand what you are saying that it says on the paper.

Well hopefully the NWS takes a stand with forecasters who blatantly disregard policy, especially one as obvious as this.

 

[TrumanBoyer]

Noting the date that this goes into effect...

TORNADO EMERGENCY declared for downtown Oklahoma City !!!!!

April Fools !!!!

-T

 

[beaudodson]

Well hopefully the NWS takes a stand with forecasters who blatantly disregard policy, especially one as obvious as this.

Where do you see (in the PDF posted) that the NWS won't be issuing tornado warnings for squall line events (QLCS events)? I am not seeing that - at least not clearly.

Are you saying that these offices, that are participating in this experiement, are going to issue severe thunderstorm warnings with a statement that says - tornadoes possible - instead of issuing a tornado warning?

Just wanting to clarify that.

Thanks