Doswell on tiered warnings

[Ian]

good read.. love doswell, would sometimes like a rant filter for him but still.

http://cadiiitalk.blogspot.com/2012/04/magic-bullet-for-tornado-warnings.html

[Wow]

Speaks a lot of common sense regarding the warning system. This is all a bad attempt to look like they are fixing something which can't be.. or doesn't need to be in its current state.

[Hoosier]

He makes some good points. I was a little surprised to see the new warning system go public so quickly, although it is only being used by certain WFO's right now. There are some situations that the tiered warning system likely won't be able to address very well (a rapidly intensifying tornado moving into a populated area) but overall I guess I am willing to give the benefit of the doubt. It will take years to evaluate the success or failure of this tiered system.

[isohume]

For once I actually agree with Doswell. There was no need for the change, even tho it's "only" experimental right now. The general public, including some wx weenies here, doesn't understand the difference btw a watch, warning and an outlook. Introducing various levels of tor warnings has the capacity to cause more confusion and harm, while doing negligible extra good or adding any practical utility to the public.

Warnings were in place for last year's outbreaks with plenty of lead time. It was dumb luck that highly populated areas were hit by such strong tors. A scenario which may not happen again for another 30 years...or maybe tomorrow. It's a typical "knee-jerk" reaction by the NWS to make it look like they are being proactive. Everyone has their own tolerance of when to take action in a given situation and they always will. That can never be changed. It's all based on one's personal experience and history with weather.

[JoMo]

He expects too much from the public. He expects that they will know what to do and he expects them to be aware of the weather situation. Sadly, that's not going to happen without some kind of stepped-up severe weather education.

The NWS will always fight with an 'optimism bias', people that believe there is a protective shield around their house/city, 'it always goes to the north/south of us' or it only 'happens in other places but not here'.

Some people need that extra *HEY IF YOU DON'T GET UNDERGROUND YOU'RE GOING TO DIE* wording. It makes them take notice since it describes the seriousness of the situation at hand.

Should the NWS not have issued the famous 'dire warning' when it came to Katrina?

I mean, don't get me wrong, it would be great if everyone took a tornado warning seriously and took the appropriate shelter instead of thinking that a tornado warning means that you should go outside and stand on your front porch but that isn't what happens in reality.

[donsutherland1]

In his blog entry, Chuck Doswell writes:

My long-time friend and colleague, retired NWS operational forecaster Alan R. Moller made me aware of the inadequacy of this situation: even a perfect forecast is useless if the forecast information:

• Fails to reach its intended users
  
• Is not understood by the intended users
  
• Is not believed by the intended users
  
• Is of no value in helping intended users to make critical decisions
  

In other words, responsibility for forecasts extends well beyond the forecasters themselves. Forecasters cannot control all those variables. Attempting to do so would only divert effort from the forecasting problem itself and, with the potential risk, of reduced forecasting accuracy.

First, reach is, in part, a function of technology and also, in part, a function of the intended users being tuned-in so to speak. Technology has made relevant and timely information almost universally accessible at the time it is needed. Failure to receive relevant and timely information has largely become a user issue today. Problems ranging from sleeping through weather alarms (for those who have them) to simply not paying sufficient attention can lead to such an outcome. That outcome is not a forecasting issue.

The second issue is both a forecasting and educational issue. Forecasters can make messages concise and well-worded to maximize potential public understanding. However, advancing understanding is also a function of educating and informing the general public. In other words, it is a "literacy" challenge so to speak. Technology has made the relevant background information perhaps as close to universally accessible as possible. Watches and warnings have clear definitions. Almost anyone who takes a minimal amount of time to browse the NWS website can understand the distinctions. There is an implicit obligation for users of such information to make the minimal effort to "educate" themselves. If they choose not to do so, that's not really a forecasting problem, even if the temptation to further simplify message language arises. There are limits to how far one can simplify things before yielding a counterproductive outcome For example, oversimplification could drain the messages of meaning and urgency, leading to a suboptimal response to those forecasts.

The third issue is a credibility issue. If forecasts are accurate--and today's forecasts are enormously accurate--they should properly be viewed as credible. That some users, for complex reasons, might choose to disbelieve a forecast, does not mean that the forecast lacked credibility. To bolster credibility, measures of forecasting accuracy could be widely disseminated (beyond the more technical avenues mainly utilized today). In the private sector, a rigorous methodology that assessed accuracy could create an opportunity for differentiation and greater profits.

Finally, as far as utility is involved, the usefulness of a forecast depends on the impact were the forecast acted upon correctly. In other words, when forecasters provide information that, if acted upon correctly, can save lives or help limit property damage, they have carried out their responsibility. If users choose to ignore the information, that does not mean that the forecasts lacked utility. It means that the users chose an alternative. All alternatives come with their proverbial opportunity costs. No matter how steep those costs, those costs have no relevance whatsoever on the utility of a forecast. Likewise, if a user did not know how to respond to a forecast, that's also a separate matter. More efforts at education and awareness might be required and responsibility for those efforts can extend far beyond the forecasting community. Such an outcome would not affect the utility of the forecast.

In the end, the expression "it takes a village" almost certainly applies to the broad value chain involved in forecasting. Meteorologists are just one link in that value chain. They cannot and should not be expected to assume all the responsibilities, some of which also reside with forecast users.

[johnc]

Are there any studies that indicate what percentage of the population receives the NWS text verbatim? My impression is that if people do receive a warning, it tends to be from a second-hand source such as the media or another person. The tiered warning system may better equip the media people with the information and urgency they need to communicate the message better to viewers and listeners. Though, IME, local TV stations probably have significantly better mets on staff than they used to.

Just speaking of this area here... Only one of the six radio stations on my car presets actually breaks in for EAS warnings. Perhaps that's only because the general sense of this area's population is that severe weather isn't dangerous. And maybe our region doesn't need to take severe weather as seriously as other parts of the country (?). But motorists -- and there's a lot of them around here -- are particularly susceptible to severe weather. You're moving so you can drive into a bad situation before you even know it, and can't get out easily. There's no quick access to shelter. And if the radio isn't going to inform you of warnings, and you don't have access to TV or the internet, how would you even know?

Also the TV break-in situation is pretty bad around here, except for the local stations. The cable and some of the other station EAS break-ins are kind of eschewing NWR broadcasts for their own little voice-overs. It's typically something simple like, "The NWS has issued [whatever] for [whoever] until [whenever]. End of transmission." Like gee, thanks. Now the viewer has to go out on his own to get further information, and a lot of people won't even be bothered to check until there's some kind of "visual confirmation" that a bad storm is in progress. When NWR is used, it's not unusual to have the broadcast cut-off in mid-statement or whatever. There just isn't much of an urgency to take what the NWS gives and pass it on faithfully.

Again, that's just this area. I'm sure the other parts of the country are much more sophisticated, where severe weather is more consequential.

[Hoosier]

Looking at the stats on the SPC website, it is startling to see how many deaths occur in mobile homes.  45 out of 63 fatalities this year have happened in mobile homes.  Lots of people were killed in more sturdy structures (well, more sturdy than mobile homes) in 2011 but there were still over 100 deaths in mobile homes.  That seems like an area that would be easiest to improve upon.  It's tough to do much in cases where people are killed in well built structures/underground.

[JoMo]

Looking at the stats on the SPC website, it is startling to see how many deaths occur in mobile homes. 45 out of 63 fatalities this year have happened in mobile homes. Lots of people were killed in more sturdy structures (well, more sturdy than mobile homes) in 2011 but there were still over 100 deaths in mobile homes. That seems like an area that would be easiest to improve upon. It's tough to do much in cases where people are killed in well built structures/underground.

Yeah, mobile homes are deathtraps. There's been wording and education for years that state how easily a mobile home is flipped in a storm or tornado but people still don't think it will happen to them. I know there's been talk of legislation requiring mobile home parks to have storm shelters.

[MGorse]

For once I actually agree with Doswell. There was no need for the change, even tho it's "only" experimental right now. The general public, including some wx weenies here, doesn't understand the difference btw a watch, warning and an outlook. Introducing various levels of tor warnings has the capacity to cause more confusion and harm, while doing negligible extra good or adding any practical utility to the public.

Warnings were in place for last year's outbreaks with plenty of lead time. It was dumb luck that highly populated areas were hit by such strong tors. A scenario which may not happen again for another 30 years...or maybe tomorrow. It's a typical "knee-jerk" reaction by the NWS to make it look like they are being proactive. Everyone has their own tolerance of when to take action in a given situation and they always will. That can never be changed. It's all based on one's personal experience and history with weather.

It will be interesting to see what comments/feedback come in following this experimental phase of these types of warnings.

Also the TV break-in situation is pretty bad around here, except for the local stations. The cable and some of the other station EAS break-ins are kind of eschewing NWR broadcasts for their own little voice-overs. It's typically something simple like, "The NWS has issued [whatever] for [whoever] until [whenever]. End of transmission." Like gee, thanks. Now the viewer has to go out on his own to get further information, and a lot of people won't even be bothered to check until there's some kind of "visual confirmation" that a bad storm is in progress. When NWR is used, it's not unusual to have the broadcast cut-off in mid-statement or whatever. There just isn't much of an urgency to take what the NWS gives and pass it on faithfully.

Does this happen fairly often? The broadcast should not cut-off in mid-statement unless a new warning needs to play which then interrupts the current broadcast. I just want to make sure our NWR broadcasts are not just cutting off whenever.

[johnc]

Does this happen fairly often? The broadcast should not cut-off in mid-statement unless a new warning needs to play which then interrupts the current broadcast. I just want to make sure our NWR broadcasts are not just cutting off whenever.

Just to clear that up, I was referring to one of the TV stations (MindTV) that used to broadcast the NWR statement ("Tom voice") with a scroller. They've since replaced that with their own voice message starting this year.

The actual NWR broadcast is perfectly fine.

[MGorse]

Just to clear that up, I was referring to one of the TV stations (MindTV) that used to broadcast the NWR statement ("Tom voice") with a scroller. They've since replaced that with their own voice message starting this year.

The actual NWR broadcast is perfectly fine.

Ah, okay. Thanks for the clarification.

[Amped]

Sorry there is no way 1% chance of an EF0 thats probably not on the ground deserves the same warning as an a tornado with a 200kt GTG and huge debris ball. The current system yields too many false alarms that suprise the public but not the NWS who knew the thing had low damage potential even if it reached the ground. So I disagree with his rant, nobody ever said it'd be perfect, and nobody can convince me all tornados should be treated the same when 1 EF5 causes more damage than 1000 EF0s.

[isohume]

Sorry there is no way 1% chance of an EF0 thats probably not on the ground deserves the same warning as an a tornado with a 200kt GTG and huge debris ball. The current system yields too many false alarms that suprise the public but not the NWS who knew the thing had low damage potential even if it reached the ground. So I disagree with his rant, nobody ever said it'd be perfect, and nobody can convince me all tornados should be treated the same when 1 EF5 causes more damage than 1000 EF0s.

Certain radar sigs and reports already get different wording outside the experiment...I'm not sure I get your point. Also, how can you tell the difference btw an EF0 and EF1/EF2 tors on the radar...specifically with non/meso tors? What about broken-line snados? Distance from radome?

[andyhb]

As Stebo mentioned, the wording for a debris ball/TDS has been changed to something like "NWS Doppler Radar confirmed a tornado near x location".

[Amped]

Certain radar sigs and reports already get different wording outside the experiment...I'm not sure I get your point. Also, how can you tell the difference btw an EF0 and EF1/EF2 tors on the radar...specifically with non/meso tors? What about broken-line snados? Distance from radome?

You can't tell exactly but you can see if there's a strong couplet in a good envirnment vs a weak couplet in a marginal envirnment. So probably using the SPCs assesment of probablity for EF2+ tornados and the stregnth of the couplet you can determine what level of warning should be applied. It's a little subjective, but it still beats a binary all or nothing warning.

[isohume]

You can't tell exactly but you can see if there's a strong couplet in a good envirnment vs a weak couplet in a marginal envirnment. So probably using the SPCs assesment of probablity for EF2+ tornados and the stregnth of the couplet you can determine what level of warning should be applied. It's a little subjective, but it still beats a binary all or nothing warning.

Using a specific el-scan of "strength of couplet" wont always differentiate EF0s from EF1/EF2s on the ground, and operationally there is little time to investigate gate to gate shear thoroughly before issuance anyway. In any case, EF0s can and do create damage and injure or kill. It depends on where they hit. A tor warning should be taken seriously as is. If someone waits for an "enhanced" warning, it could be too late.

[Amped]

Using a specific el-scan of "strength of couplet" wont always differentiate EF0s from EF1/EF2s on the ground, and operationally there is little time to investigate gate to gate shear thoroughly before issuance anyway. In any case, EF0s can and do create damage and injure or kill. It depends on where they hit. A tor warning should be taken seriously as is. If someone waits for an "enhanced" warning, it could be too late.

Yeah it may make things difficult. But Ithink this system still has the option for offices to use mid level warnings all the time for "Legacy support" and high uncertainty situations.

[isohume]

Yeah it may make things difficult. But Ithink this system still has the option for offices to use mid level warnings all the time for "Legacy support" and high uncertainty situations.

Yeah the radar operator has always been able to add additional information to a TOR as needed. That wont change I'm sure.