Do We Warn Too Much?

[gkrangers]

Who's to say a marginal radar situation won't result in a destructive tornado? Or go from "meh" to "Oh ****" in one scan?

I don't envy the NWS mets one bit, and respect them a great, great deal.

[baroclinic_instability]

I'm not saying that.

Not really sure what you were trying to say then. I will give you the benefit of the doubt, but it would be quite a disservice to all the scientists, researchers, and forecasters who are working and striving towards resolving the problem warning forecasters and meteorologists in general deal with every day.

..." I'd say many people don't want to admit a high FAR is a problem when the emphasis for many mets is making sure they don't miss..."

[isohume]

Recommendation 13—NWS meteorologists should work with communications experts and

behavioral experts to evaluate the impact of different lead times on reliability of warnings and on

human responses. This should include determining the correct balance between increasing lead

time without adversely increasing false alarms and creating “warning fatigue.”

I'd like to see the results of this study. Do you know what the status of the project is?

[CT Rain]

Not really sure what you were trying to say then. I will give you the benefit of the doubt, but it would be quite a disservice to all the scientists, researchers, and forecasters who are working and striving towards resolving the problem warning forecasters and meteorologists in general deal with every day.

..." I'd say many people don't want to admit a high FAR is a problem when the emphasis for many mets is making sure they don't "miss..."

Oh no doubt the research that is being done is great and extremely important. All I'm saying is that there are really no easy answers on how to fix the problem or change the current warning system. It's a lot easier to say the system we have isn't perfect but it's the best we can do than say the system is flawed and needs to be changed when the options aren't clear cut.

[CT Rain]

Who's to say a marginal radar situation won't result in a destructive tornado? Or go from "meh" to "Oh ****" in one scan?

I don't envy the NWS mets one bit, and respect them a great, great deal.

Did you read the blog post in post 1?

[JBG]

I'd argue warning fatigue is more dangerous than not warning marginal situations.

I tend to agree. While relating to hurricanes rather than tornados, I would argue that some of the horror relating to Katrina was warning fatigue.

[JoMo]

I'd like to see the results of this study. Do you know what the status of the project is?

I don't know. I just know of that service assessment, and how it matches up pretty well with what people said after the Joplin tornado.

[isohume]

I don't know. I just know of that service assessment, and how it matches up pretty well with what people said after the Joplin tornado.

The only reason I ask is it seems EM weighted. Meaning...it doesn't look like the NWS was collaborated with in the statement. The last sentence: "determining the correct balance between increasing lead time without adversely increasing false alarms" is baffling to me.

First off...the lead time has nothing to do with the FAR. If there is lead time, then you have an event...or a hit for the POD. If the warning turns out to be a hit on the FAR...then there is no lead time. The two are not related. Secondly...I am curious how they plan on increasing lead time in the first place.

[AtticaFanatica]

We all know that EF0/EF1 tornadoes are not "marginal" weather events and that missing one certainly can be a pretty big deal, right?

[Analog96]

I think the National Weather Service's implicit duty "to protect lives and property" implies that they need to overwarn, rather than underwarn. In other words, don't KILL the warnings by issuing a severe warning every time there's lightining. But, on the other hand, if it's borderline, issue the warning. The National Weather Service's mission statement seems to imply that it's better to issue a warning where there was "almost" verification, than to NOT issue a warning when there should have been one. As far as tornado warnings vs severe thunderstorm warnings, the general public pays more attention to tornado warnings.

[Chicago Storm]

We all know that EF0/EF1 tornadoes are not "marginal" weather events and that missing one certainly can be a pretty big deal, right?

Apparently not.

[Analog96]

Who's to say a marginal radar situation won't result in a destructive tornado? Or go from "meh" to "Oh ****" in one scan?

I don't envy the NWS mets one bit, and respect them a great, great deal.

From what I have seen, usually if it's "marginal in one scan", they won't pull the trigger and issue the warning, because it can easily fizzle on the next scan, and scans happen every 5-7 minutes now. If it's still there on the second scan, then they push the warning button.

[JoMo]

Got distracted while writing this, but below are some of my memories and how people reacted to the May 22nd tornado in Joplin:

1. Denial- People don't think it can happen to them, either for religious reasons (God won't let me get hit by a tornado) or because they don't want it to happen to them. I personally didn't think it would happen to me even though I was looking at radar, "Joplinmet" was pinpointing the location of the rotation on TV and when it become apparent that it might happen to me, I marginalized the threat, thinking it would only be 'elevated or 'weak'.

2. Some kind of cue to personalize the threat - People want to see, or hear the tornado with their own eyes/ears to make sure they are going to be affected by it. I mean, who wants to sit in a dark, damp basement for 20-30 minutes every week or so? For me, it didn't become a reality until I heard it since I could not see it due to trees and houses and even then I wanted to believe that it was 'elevated' or 'weak'. I did attempt to see it out a west window but saw nothing but pitch black sky and probably a part of the wall cloud which had an orangish hue against the black sky. I only caught a glimpse.

3. Warning fatigue - We receive a lot of warnings in Joplin. A lot of the time 'nothing happens'. It also seems that if 'nothing happens' at the location the person is at, it also counts for the fatigue even though something may have happened at a different location. Basically, if they aren't personally affected by it, then to them, nothing happened.

4. Too busy in their own world - Some people are just 'too busy' to pay attention to what the weather is doing and some didn't 'hear' the sirens or check on the weather since they didn't care because 'nothing ever happens' or the 'weatherman is always wrong'.

5. Siren Confusion/Fatigue - Joplin tests the sirens every Monday at 10 AM if the weather is clear in Spring. This may have led some people to become used to them. Also, there is confusion on what exactly the sirens mean. Some people think that when the sirens stop, this is an 'all-clear' which isn't the case. For May 22nd, the sirens sounded with the first tornado warning which didn't cover the area that would eventually be hit. The sirens sounded again constantly once the tornado was on the ground, actually it had probably been on the ground for a minute or two before they actually sounded again. I know I heard the siren right after the sound had become audible. There was also an issue on May 24th. The sirens sounded for a tornado warning, then again sounded for 75+ MPH straight-line winds. Neither happened to me so I guess that was a 'bust'

6. People don't know where they are at - This is a huge thing. Some people do not know what county they are in, they don't know directions, they don't know what towns are near them. I watch "Joplinmet's" facebook during severe weather events since he keeps everyone updated on there and sometimes I get pretty.... frustrated.... at people not knowing where they are or what's around them. There are people asking if they will personally be affected by the storm that live about 2 miles east from an area that he says will be affected by it and that the storm is moving east.

I was lucky with the Joplin tornado and only caught the EF-1 edge which was enough to pull a small tree out of the ground and cause shingles to be ripped off the roof. I experienced what it was like to be one of these 'normal' people who had limited ability to find out weather information when I did not have power on May 24th when the sirens sounded again. I actually relied on the radio station more then NOAA WX RADIO since the radio station gave up to the minute reports and had a working radar and were relaying that information in real time.

[Riptide]

Got distracted while writing this, but below are some of my memories and how people reacted to the May 22nd tornado in Joplin:

1. Denial- People don't think it can happen to them, either for religious reasons (God won't let me get hit by a tornado) or because they don't want it to happen to them. I personally didn't think it would happen to me even though I was looking at radar, "Joplinmet" was pinpointing the location of the rotation on TV and when it become apparent that it might happen to me, I marginalized the threat, thinking it would only be 'elevated or 'weak'.

2. Some kind of cue to personalize the threat - People want to see, or hear the tornado with their own eyes/ears to make sure they are going to be affected by it. I mean, who wants to sit in a dark, damp basement for 20-30 minutes every week or so? For me, it didn't become a reality until I heard it since I could not see it due to trees and houses and even then I wanted to believe that it was 'elevated' or 'weak'. I did attempt to see it out a west window but saw nothing but pitch black sky and probably a part of the wall cloud which had an orangish hue against the black sky. I only caught a glimpse.

3. Warning fatigue - We receive a lot of warnings in Joplin. A lot of the time 'nothing happens'. It also seems that if 'nothing happens' at the location the person is at, it also counts for the fatigue even though something may have happened at a different location. Basically, if they aren't personally affected by it, then to them, nothing happened.

4. Too busy in their own world - Some people are just 'too busy' to pay attention to what the weather is doing and some didn't 'hear' the sirens or check on the weather since they didn't care because 'nothing ever happens' or the 'weatherman is always wrong'.

5. Siren Confusion/Fatigue - Joplin tests the sirens every Monday at 10 AM if the weather is clear in Spring. This may have led some people to become used to them. Also, there is confusion on what exactly the sirens mean. Some people think that when the sirens stop, this is an 'all-clear' which isn't the case. For May 22nd, the sirens sounded with the first tornado warning which didn't cover the area that would eventually be hit. The sirens sounded again constantly once the tornado was on the ground, actually it had probably been on the ground for a minute or two before they actually sounded again. I know I heard the siren right after the sound had become audible. There was also an issue on May 24th. The sirens sounded for a tornado warning, then again sounded for 75+ MPH straight-line winds. Neither happened to me so I guess that was a 'bust'

6. People don't know where they are at - This is a huge thing. Some people do not know what county they are in, they don't know directions, they don't know what towns are near them. I watch "Joplinmet's" facebook during severe weather events since he keeps everyone updated on there and sometimes I get pretty.... frustrated.... at people not knowing where they are or what's around them. There are people asking if they will personally be affected by the storm that live about 2 miles east from an area that he says will be affected by it and that the storm is moving east.

I was lucky with the Joplin tornado and only caught the EF-1 edge which was enough to pull a small tree out of the ground and cause shingles to be ripped off the roof. I experienced what it was like to be one of these 'normal' people who had limited ability to find out weather information when I did not have power on May 24th when the sirens sounded again. I actually relied on the radio station more then NOAA WX RADIO since the radio station gave up to the minute reports and had a working radar and were relaying that information in real time.

Thanks for summarizing the fallacies of America.

[Stebo]

Erring on the side of caution is a much better approach than not. You can't make a judgment call per a radar if a marginal looking couplet would yield or not yield. Warning fatigue happens sometimes just due to the fact of a hyperactive pattern, sometimes its just unavoidable. Sadly people sometimes don't take warnings serious because the storm doesn't hit their particular locale. That doesn't mean it didn't hit Joe Smith's house up the road or across town. JoMo's comments are a great example why you should warn marginal or not, because it won't matter either way with some people but for everyone else it could potentially save their lives.

[isohume]

Erring on the side of caution is a much better approach than not. You can't make a judgment call per a radar if a marginal looking couplet would yield or not yield. Warning fatigue happens sometimes just due to the fact of a hyperactive pattern, sometimes its just unavoidable. Sadly people sometimes don't take warnings serious because the storm doesn't hit their particular locale. That doesn't mean it didn't hit Joe Smith's house up the road or across town. JoMo's comments are a great example why you should warn marginal or not, because it won't matter either way with some people but for everyone else it could potentially save their lives.

This sums up the some things pretty well. Folks forget or don't know how big in areal coverage the polygons are compared to the size of the tornado. There's a magnitude of difference and that is another weakness of radar/tor science.

[Ginx snewx]

This sums up the some things pretty well. Folks forget or don't know how big in areal coverage the polygons are compared to the size of the tornado. There's a magnitude of difference and that is another weakness of radar/tor science.

A weakness in relay to the public.

[isohume]

[/b]

A weakness in relay to the public.

Yeah that's what I said. All part of the weakness in the tornado warning methodology...or tornado warning science.

[BadgerWXman]

lol @ the nws mets asking for proof of warning fatigue... just lol... I'm convinced that's arguing for the sake of arguing because there's no way you could be that out of touch with reality... right?

[AtticaFanatica]

Let's be clear here. Based on the current science, a drop in the FAR will also drop the POD. If you advocate that, then fine (and I think Jon makes some good points in essentially advocating exactly that). But that's what they're dealing with.

[AtticaFanatica]

Of course, I'd also argue that many of the EF0/EF1 tornadoes listed in those statistics aren't actually EF0/EF1 tornadoes, so I'm skeptical of their accuracy despite the fact the graphs look completely logical.

[isohume]

lol @ the nws mets asking for proof of warning fatigue... just lol... I'm convinced that's arguing for the sake of arguing because there's no way you could be that out of touch with reality... right?

For some reason I'd like to see results of an actual sociological study as opposed to hearing EM's and others use the term freely and loosely. Especially when they give it much weight in relationship to lead time and FAR. I'm not saying something such as "warning fatigue" doesn't exist...I just want to see some actual numbers presented from a scientific study. Something to gauge and keep in mind the probable magnitude of. That's not asking a whole lot. lol

[Southland Wx]

Warning fatigue is real. There's no question about that. We are overwarned (and not just in meteorology...as a society, we're overwarned about everything).

The problem is multi-fold, and as has been stated there is no singular correct answer to this issue.

Just some points...

1.) The NWS mission, as has been stated, is in line with the idea that overwarning is better than not warning at all. This is an acceptable mission...it's not wrong...this is their job.

2.) A FAR of 75% is completely too high...there's no question about that. I'm first to defend the NWS in everything, but 75% is very disappointing. That being said, there is next to nothing that can be done to improve that other than advances in radar technology or interpretation. The reason the FAR is so high is because the NWS is doing their job. That's really the bottom line.

3.) Meteorology and sociology have crossed paths this year. There is a sociological element to this entire process that we know exists, but we don't know how prevalent it is and how serious it is. People are overwarned for everything.

4.) The media does not help this. Meteorologists on TV generally do a good job. The producers and management do not do anyone a service by overhyping for ratings. That's as much a part of the problem as anything. See: Houston TV leading into TS Don....see: Major broadcast networks during any significant weather event... see: certain meteorologists that publish comments and get them spread throughout the community...and honestly, see: global warming...both sides. It's not weather, but it's related enough that we warn even after weather events that this is only the beginning. Hype and exaggeration is overwarning...and it contributes to fatigue of warning even more....and possibly moreso than the NWS overwarning.

5.) The NWS has far too many warnings, but on the same lines, they don't have enough. The problem is... what people understand differs from what the NWS and others want them to understand. We still haven't overcome the watch vs. warning debate. Having a full suite of these is too much for people to comprehend. At the same time, I am of the belief that there needs to be tiered tornado warnings. The general TOR for a spin up in a line or borderline rogue cell. But then the Tornado Emergency, which I know gets used in some places...which is the more targeted warning and stronger worded one...for specific communities when we know they're going to likely be impacted. Some reform of the warnings I think is fair.

6.) People don't know geography. That's a failure in teaching. You need to know where you are and where you are in relation to the things around you...and what those things around you are. There's no simple solution to this part of the issue, but teaching basic geography is something that needs to continue beyond grade school apparently.

7.) And lastly, OEMs. Mike Smith of AccuWeather has done a lot of good work on analyzing and bringing this problem some attention. Often when there are warnings...even with the polygons now, they fire off sirens for the entire county, even if only part of it is impacted. This further roils people and makes the problem worse. Where tornado sirens exist, there really needs to be reform. EMs try and budget for uncertainty...which the NWS is already budgeting for in the polygon. It defeats the purpose of having a polygon.

It's a multi-fold problem and in order to resolve it, everyone needs to be brought to the table to share their side of the story and their concerns. No one expects perfection, but we can do a whole hell of a lot better than we are now.

[LongRanger]

One possible way to address Warning Fatique:

1) study to determine the quantity at which Warning Fatigue sets in on average for the intended audience... is that one Warning per month? per week? per day?

2) from that data calculate the average number of Warnings/year before fatigue

3) adjust each region's Warning criteria so that during an average year no more than that number of Warnings may be expected to be issued

[AtticaFanatica]

One possible way to address Warning Fatique:

1) study to determine the quantity at which Warning Fatigue sets in on average for the intended audience... is that one Warning per month? per week? per day?

2) from that data calculate the average number of Warnings/year before fatigue

3) adjust each region's Warning criteria so that during an average year no more than that number of Warnings may be expected to be issued

You're completely removing the meteorology at that point; forecasters shouldn't have what number warning it is in the back of their mind when they think there's a large tornado moving toward a populated area.

[The Great Zo]

2.) A FAR of 75% is completely too high...there's no question about that. I'm first to defend the NWS in everything, but 75% is very disappointing. That being said, there is next to nothing that can be done to improve that other than advances in radar technology or interpretation. The reason the FAR is so high is because the NWS is doing their job. That's really the bottom line.

You're mostly correct in this assertion -- sans adjustments in tornado warning philosophy, of course.

Even with advances in scientific interpretation and radar data, I am not sure that FAR gains (losses) would be significant. Since the NWS has to get the warning out ahead of the touchdown, a meteorologist is always going to have to be ready to issue the warning in advance of a clear-cut signal. I would argue that while advanced radar/scientific knowledge may lead to more indications of when a tornado is likely occurring (POD), it would also lead to more indications of when a tornado is capable of occurring but does not (FAR).

[JoMo]

The FAR at SGF is 85% and it seems they issue a tornado warning way too often in situations that don't warrant it. But like already stated, nobody wants to be responsible for an un-warned EF-0 tornado touching down and killing someone due to a tree or branch falling on their car.

[LongRanger]

You're completely removing the meteorology at that point; forecasters shouldn't have what number warning it is in the back of their mind when they think there's a large tornado moving toward a populated area.

I agree, no warning quantity limits for the mets, instead they'd be referencing criteria just as they do now. My approach means the criteria will be adjusted to an appropriate level as revealed by the study in step 1.

[AtticaFanatica]

I agree, no warning quantity limits for the mets, instead they'd be referencing criteria just as they do now. My approach means the criteria will be adjusted to an appropriate level as revealed by the study in step 1.

Ok, so then you're ok with the POD going down? Because that's the inevitable result.

[Analog96]

The FAR at SGF is 85% and it seems they issue a tornado warning way too often in situations that don't warrant it. But like already stated, nobody wants to be responsible for an un-warned EF-0 tornado touching down and killing someone due to a tree or branch falling on their car.

If it was totally unwarned, I agree. However, if there was already a severe thunderstorm warning in effect, any severe thunerstorm can topple a tree.