Thursday, July 2, 2015

Watch and Warning: errant communication?

During an interview for Wx Geeks, the subject of the confusion caused by the terminology of messages - namely watch and warning - came up as a topic.  Given the short time available during the show, I want to offer some more comments about that topic.  The claim is made by some that the similarity between the two words (what I refer to as the "wa-wa" problem) is the source of the confusion.  I don't pretend to understand how and why this confusion arises; after all, I'm a meteorologist, not a communications expert.  However, I find this premise to be pretty much ridiculous.  To my knowledge, no one has done any work to validate that the public's inability to distinguish between a watch and a warning lies exclusively or even primarily in the "wa-wa" arena.

Imagine we decided to coin new terms for the content of watches and warnings, calling them instead kenutenaries and chinkaderas, respectively.  Does anyone honestly believe that calling them tornado kenutenaries or severe thunderstorm chinkaderas would clear up the underlying problem?  I seriously doubt it.  So what do I attribute this problem of distinguishing watches from warnings?  Anything I'm about to say is pure speculation, of course;  I've done no studies and have no scientific basis for my ideas, but I do have decades of experience with the meteorology and our attempts to communicate its hazards.  What seems plausible to me is that many people in that great, faceless mass called "the public" are basically not interested in the weather very much, unless it's going to affect them directly and personally.  I understand that.  I appreciate that not everyone shares the passion of weather geeks when it comes to the atmosphere.  Not everyone is passionate about hockey, or pole dancing, or scrapbooking, or mathematics, either.  I get that.  Most of the topics limited numbers of people are passionate about don't involve events that can prove fatal to the general public. 

Here's the kicker regarding this widespread lack of interest in the weather (and geophysical hazards, in general):  it can rise up on occasion and kill you!  One would expect, naively, that knowing that risk would get most everyone's attention.  It seems clear this isn't the case.  If you're uninterested in the atmosphere, that doesn't protect you from its threats.  There's one very effective way to protect yourself from atmospheric hazards:  being prepared for them.  If the distinction between watch and warning is an important thing for you to recognize in order to take appropriate action (and it is!), whose responsibility is it to know that distinction beyond any doubt?  Yours!  Everyone's!!  We meteorologists can turn ourselves inside out and backwards trying to figure out how to wordsmith this difference so that no one could possibly misunderstand it, and still, there inevitably will be those who will, by personal choice, not make any effort, and so will remain confused and unable to articulate the difference.  After all, it could never be of concern to them, right?  Until it is.  Then those very same ignoramuses are quoted in the media after a weather disaster "We had no warning!" even when they did have a warning!

If I've learned anything in 40+ years as a meteorologist, it's that you can lead horses to water, but they won't necessarily drink it.  There will always be those whose lack of a sense of responsibility for their own safety will mean they have no clue about things, and certainly won't be sensible enough to plan for what is, after all, a rare event.  Yes, it's "normal" not to be hit by a tornado, so the so-called normalcy bias means people are reluctant to accept that something rare might actually affect them directly and personally.  That normalcy bias is reflected in their behavior when a tornado is in their vicinity:  They want to confirm that it's actually about to happen to them.  But when this complacent, it-will-never-happen-to-me attitude is confronted by an approaching tornado, the odds are good that such people will be ill-prepared and therefore only luck enables them to survive.  They'll be ready to blame anyone but themselves for their misfortune.

Yes, a significant fraction of folks, even in Oklahoma, don't know the difference between a watch and a warning, despite decades of attempts to educate them for their own safety.  Some horses always will refuse to drink, no matter what name by which we refer to the water. 

Yes, we should do the social science studies to learn in more detail why people choose to be ignorant and perhaps there is some verbiage we can use to make ourselves more clear to the public.  At some point, however, we must also recognize that there will never be a time when 100 percent of the public understands perfectly those weather hazard messages we're attempting to convey.


Gilbert said...
This comment has been removed by a blog administrator.
MileO said...

You're spot on with the... 'the average person is not actually interested in weather' - they're interested if it's going to rain or but not much else. Unless everyone is logged onto all these social media pages flogging warnings and watches, who do they believe and of how much!

It's even more difficult with lightning education. I've been rampant on my own social media page giving thunderstorm information as it happens and with detailed lightning statistics of where the lightning is starting and where it is tracking, along with the type and kA amplitudes ( current of the lightning strokes for those non-weather geeks)

I've found that, for where I live in Darwin, storms and lightning are regular visitors during the months from October to April. Most 'weather' pages here simply copy the radar screen grab, add lots of arrows and text but in essence don't have a clue about the actual motion/character of the storm but send out 'Heads up!!' warnings as soon as the thing goes orange on the radar. Hence causing all kinds of panic ( expecially when they tout words as severe...going red! ) We don't have tornadoes or rampant supercells, so most of our storms are prolific lightning producers.

Over the last several years I've constantly been giving 'alerts' - not warnings nor watches - but alerts that in essence, alert the public of incoming or forming local storms that are producing lightning and to at least go inside ( pack up the kids bikes in the park and get home type scenarios) If the storm becomes severe warned then I simply copy/paste the bureau's warning to the page and don't add anything that would indicate panic, confusion or treating people like idiots. Give them the facts - don't gloss it over and never, ever make your own assumptions if you don't understand the thermodynamics and meteorology of thunderstorms. This also includes cyclone watches and warnings - I don't paraphrase - I simply share our bureau's links.

Education is a wonderful thing, and your are correct, Chuck. The minority is the weather community and those that are active on said pages - the majority is the public, business, infrastructure, government and councils.

One can't stop humans going about their daily tasks etc when storms arrive, but your topic echoes what I, and many others, are struggling to do - get more eyes and ears open!

Anonymous said...

The bottom line is you cannot teach common sense. NWS has got to reach a point where they acknowledge they have done all they can, their efforts are defensible and the rest lies squarely on the public's shoulders. A vast number of folks lack common sense (amongst other attributes) and NWS must realize they cannot overcome that.

Anonymous said...

Nobody is going to like this but I'm going to say it anyway. My view is that the SPC does a pretty good job with watches. I am satisfied that if the SPC tells me that there is an elevated risk of tornadoes in and around the watch area, they are correct.

With warnings, the NWS seems to be it's own worst enemy. Essays and writings by Dr. Doswell indicate that the radar operators don't get in trouble for false alarms and damaging the credibility of the weather service. However, if they fail to sound the alarm they can get in trouble. Review the radar data, find some rotation, and bang they may find themselves sweeping floors.

So, the NWS in my view pushes the science beyond what it is capable of, purposely maintains a high FAR and then blames the public when they don't believe the reports from chicken little that the sky is falling.

Pretty screwed up way to run things if you ask me.

My view of how things should be done is the opposite, issue a false report, spend a few weeks in jail to think about it.

Of course, an easy way around this problem is to simply stop lying. Don't issue a tornado warning without a tornado. It's not difficult.

Sound the alert and issue a tornado bulletin or update, or whatever, just not a warning unless you have ground truth. If the forecaster believes based on his experience and the data available to him that there is a low, medium, or high probability of a tornado, say exactly that.

Issue a "warning" only when it's on the ground.

After the fact, review the data, and verify that the forecaster was reasonably accurate. Yes there was a medium chance for a tornado based on available data.

All of a sudden, the NWS has credibility.

The only other part of this is to make the public aware of what the science can and can't do. Tell them flat out there will be the occasional tornado that is not forecasted.

As I see it, the problem is that the high false alarm rate means that people are only likely to take action after confirming a tornado warning.

See the official NIST report on Joplin as an example of this.

The NWS unfortunately is a government organization. That means at best it will be mediocre. Governments do nothing correctly. They're the biggest bunch of screw ups on the planet.

The NWS needs to get rid of everyone who has any political connection and needs to be free of government interference.

If the government has an issue with the NWS, take it to court. The NWS should only be saying what it can prove in court. Opinions stated by forecasters should be opinions that a majority of experts would agree with based on available data, thus they shouldn't get in trouble for that. If they miss a tornado, the reason should be that it did not show up on radar or give any other warning. Not the fault of the NWS.

The government should end up leaving the NWS alone because any court case should show that the problem was caused by a lack of funding. Not incompetence. Remove the bureaucrats and the government officials and your incompetence problem should go with it.

As far as the forecaster is concerned, their responsibility and concern should be with the accuracy of their products. The propagation of those products is not their concern.

It's real simple, the NWS puts out an accurate and timely product. Some people won't listen. Sucks to be them. Don't lose sleep over those who don't listen, spend your time learning more about weather science.

Lisa MacArthur
Riverside RI.

Chuck Doswell said...


You're right - there certainly are things about your comments I don't like and with which I disagree. Your post is too long in part because you have way too much "white space" - please try to be more concise and avoid so much emptiness. Now, on to your comments:

The use of radar to issue tornado warnings is driven by several factors, one of which is lead time. When we wait until a tornado actually begins, the people at the start of its path get no warning lead time at all. Most people likely would find that to be unsatisfactory warning service. Moreover, there are places and times when we have at most only a few spotters, if any, so the tornado might go on for quite a while before a spotter report is called in and triggers a warning. A tornado warning is not a guarantee that a tornado will occur - it's that the tornado hazard threat has risen to a level where a warning is warranted in the opinion of the warning forecaster. Do I think the NWS is doing a completely satisfactory job? Not necessarily - their performance varies from office to office, and even from one forecaster to another in the same office. Overall, their performance is probably not up to the level it should be - for many reasons - but it's not for a lack of desire on the part of forecasters to do the best possible job.

As for "lying" - that's a complete mischaracterization of what forecasters do. Forecasts can never be perfect, but a busted forecast is not a lie. The situation is a lot more complex than you understand, which is natural for someone not familiar with the forecast process. The fact that forecasters are required to make forecasts beyond the capability associated with the current state of the science is not entirely their fault - users have demanded those forecast products and NWS management has given in to public pressure and mandated their staffs provide them, even if the forecasters know the science isn't there to make those forecasts reasonably accurate.

It remains to be shown the extent to which the high FARs create a credibility problem for the NWS. There are some indications it's not the necessarily the primary factor in people seeking confirmation. The jury is still out.

Having worked in the NWS and in a NOAA science lab, I believe your blanket condemnation of government agencies isn't valid. Are you aware that the NWS is a government agency? Some of your comments suggest you think the government is "interfering" with the NWS, whereas the NWS is part of the government. Most of the people with whom I worked were doing their level best to achieve the best possible results for the taxpayers. Your extremely negative opinion of government employees sound more like a political statement to me than something based on any evidence.

I agree fully that the primary concern for NWS forecasters should be their forecast timeliness and accuracy, but even if perfect forecasts could be issued, the forecasts must be as effective as possible. To be effective, forecasts must be received by users, understood by users, be credible to users, and be helpful in the process of users making decisions based on the forecast information. There's a place for social sciences in all this, but certainly the forecasters should not be diverting their attention into areas best handled by social scientists (working with the forecasters to improve forecast effectiveness).

Anonymous said...

Thanks for your thoughtful reply.

Trying to be brief.

I am only interested in changing the wording. Reserve the word warning for what is now a "tornado emergency." Use bulletin for anything less than ground truth. But continue to set off the alert. No change in lead time. Just more accurate wording.

I agree that they are not lying using the NWS definition of a tornado warning. Using my definition, they are. My definition doesn't matter.

I agree with you that the government employed meteorologists do as good a job as the system allows them too and that they are as good as anyone at their job.

I hope you're right about my opinion of government agencies being invalid. Unfortunately I can cite the government run weather bureau for incredible stupidity and incompetence when it comes to the history of tornado forecasting. Start in the late 1800's when they actually tried and had some success, then some government idiot kills the program, then they are forced back into tornado forecasting in the 50's thanks in part to Fawbush and Miller and leaked military forecasts. I think you once cited the same book I read in one of your essays. The history of government tornado forecasting was pretty bad. In my opinion only a government organization could get away with such incompetence and still survive. Mind you, I am sure that the meteorologists would have loved to work on the challenge and would have done a good job.

As far as the government interfering with the NWS, my recollection is that you wrote in one of your excellent essays about this issue. The government hacks on top interfering with the meteorologists. An example is that only a government hack would prefer their forecasters to forfeit their credibility in order to warn for any conceivable spin-up. The government hacks are trying to pushing the science to do more than it can. Then when nobody believes them and fails to take warnings seriously, they can come back and say "we warned you." But, in reality, they didn't. A warning has to come from a credible source. You lose that credibility and the warning no longer counts. Only a government run agency would operate like that. Anyone else would be put out of business.

So, that's why I say, the government is interfering with it's own service. Let the meteorologists run it.

I agree fully with your last paragraph.

To sum up, I see the NWS as good people run by government hacks who somehow still manage to put out reasonably good products. As you have written in one of your essays, if you take shelter every time there is a warning, you're not going to spend that much time in a shelter.

If you follow NWS advice and pay attention to the weather, your chances of being killed by a tornado or lightning, or flash flooding, should be virtually zero.

Thanks again.

Lisa MacArthur
Riverside RI.

Chuck Doswell said...


One of the main points of my blog is that I don't believe the wording is all that big a factor in the public's lack of understanding regarding the distinction between watches and warnings.

The problem with many government agencies, certainly including the NWS, is their management. I make a clear distinction between the productive employees and their managers. It's not obvious to me that the private sector is any more capable of handling these issues than the NWS. They have the advantage of serving a limited clientele, who can make very specific requests for the kind of service they need. It's vastly more difficult to serve the whole "public" - everyone has their own specific needs and it's virtually impossible to address them all in any NWS product.

There are many things NWS management COULD do that would improve on NWS warning accuracy, but I want to reiterate: a warning does NOT mean that any specific user is necessarily going to be hit by a tornado - the science simply doesn't allow anyone to do that. It simply means the threat of a tornado in the warned area is high enough to trigger the forecaster to issue a warning. I personally would like to see the threat levels specified in the warning (as probabilities in some form or another). But that's another topic.

Anonymous said...

Thank you. I have read one of your essays on probabilistic warnings and I agree that it's the way to go.

Serving a limited clientele does make things easier.

I'm thankful for the SPC and their convective outlook products as well as for the doppler radar network and the Noaa Wx Radio as well as the people who make the system work. If you rely solely on the NWS you will be in good shape.

I like the fact that the SPC doesn't hype their products or the discussions. They just tell it like it is.

Stay cool.

Lisa MacArthur
Riverside RI.

Anonymous said...

I do not know where to start. I will make short. There should NOT be anything like Tornado Emergency (or the whole thing will go into social science and be studied by non focus groups for about 3 yrs). They are horrible in that they do NOT really equate to anything. They should study them about where and when the NWS is going to issue them because there not working out that good. I can go on about Tornado Emergencies but I will not.

"Magic Words" do not solve anything. The "social scientist" and NWS et al. are few and far between and are generally young. I have been saying for 20 yrs that really tornadoes kill mobile (manufactured) home occupants, elderly, handicapped, motor vehicle occupants, and somewhat non-English speakers. Magic words do not have a function of "normalcy bias" because they have "devastating" or "catastrophic" in the warning. And I think the words might do a little worse? The public that is faced by one or more of the things I listed before, really do not care if it is devastating or catastrophic. They maybe do not have anywhere to go to take shelter. Maybe the social scientist should get on the case to make then not drive when you should not, make shelter available, make them understand (especially when we are dealing with non-English speakers) with out using magic words and get the FAR down. Then they can look at the normalcy bias, which there are a multitude of excuses for one or more of the normalcy bias that are ingrained to the person and would be a tough nut to crack. The Watch/Warning program has been in place since to 1950s. Do not fool with it much. Do not throw the baby out with the bathwater! Focus on now getting vehicles to dive into flooded water which has been a failure even with the "turn around, do not drown." I am going to say that an increasing number of folks killed by a tornado know of the warning (and it is increasing with time). It is not a magic words, it is normalcy bias and ~half of the one's that die are going to not pay attention to the watch and warning and the other ~half wants to and they cannot get out of the way. The social scientist need to focus on the latter not the former.

Matthew Biddle

Anonymous said...

Yesterday was a good example of the NWS handling a severe weather situation very well. I had been following the convective outlooks so I knew what to expect but if I hadn't, about 5am I received emails from Accuweather and that gave me the text of a special weather statement put out by the SPC. Even if I hadn't been paying attention, I would have seen those statements and would have been advised that it was going to be a day of watches and warning. On NWR the NWS explains both terms plainly when they use them. When they issue a watch, what a watch is is explained quite well. When they issue a warning, that warning is explained very well. The convective outlook said that the main threat was wind and hail. I do not recall any tornado warnings but some of the severe tstorm warnings did mention that there was weak rotation and a slight chance of a tornado. I was very pleased with them handling that way. Overall, I think they did a great job. Since the terms watches and warnings are so clearly explained on NWR (weather radio) and I believe in the text of the watches and warnings as well, there appears to be no merit to any confusion as to what the terms mean since they remind you every time they use them. Hats off to the SPC and the NWS. I feel that I was very well informed yesterday and that the accuracy of the watches and warnings was quite good. I was south of the storms so I didn't get any. However, I do believe that when they issued the watch for my area, they were correct that the potential was there. None of the severe thunderstorm warnings were for my area. Checking the warnings against the radar indicates that they seemed right on target. I didn't have to put up with any hype or any commercials. Lisa MacArthur Riverside RI.

Anonymous said...

Matthew, as far as tornado emergencies go, I think it is the result of forecasters who care deeply about saving lives and feel that they need to find a way to communicate that the situation at hand is not a standard warning. I don't agree that they are meaningless. We know that 3 out of 4 tornado warnings are false but we have no statistics on "tornado emergencies" as far as I know. I would like to think that if someone is going to use the term "tornado emergency" there is a 100 percent chance thus there would be no false alarms. If you read the text that goes along with a watch or a warning, you don't need to know ahead of time what they mean. The text explains it so that a child could understand what they are. So, I really think that it's a moot point. People have a right to be stupid and not listen to the warnings. All I care about is the accuracy of the watches and warnings and that they be worded reasonably well. The NWS and SPC meet those goals pretty well right now. I'm not going to lose any sleep over those who pay no attention and get killed. Subscribe via email to get watches and warnings to your email and your phone. That way you won't be caught by surprise. Recommend a weather radio as well but your phone may be the only tool that you need. Then read Dr. Doswell's essays. That's all you need to know. Chuck has laid out virtually any tornado related topic in terms so plain that anyone can understand them. Follow his advice. As far as those who don't speak English, good reason to learn English. If I were to move to Germany I'd learn German. As for shelters, if you have your own house there is no excuse for not having one. They're not that expensive. If you're in an apartment then you really a plan and if you follow Doswellian wisdom on the subject, you'll be ok. Inconvenienced if you have to drive to a shelter but tough bananas, it goes with not having a shelter. Why so many people in Joplin, in tornado alley, not that far from OKC, did not have safe rooms is beyond me. I'm talking about people who own their own homes. Did they not remember the 1971 Joplin killer tornado? I know that the chances of any one house getting hit in tornado alley are pretty slim but they're not zero. Spend a few thousand bucks and put in a safe room. Then if you get the warning, go sit in the saferoom, monitor the situation and when it's over, come on out. Lisa MacArthur Riverside RI.

Anonymous said...

Reading the NIST report on Joplin as well as the NWS report on Joplin it appears that the reason for the high number of deaths had nothing to do with the WAWA debate. The government cited the high FAR which made people want "high intensity" clues before they took action. Unfortunately because of the fact that there are so many mines in Joplin there are few basements so most people had to rely on above ground shelter. Unfortunately that wasn't good enough. The majority of people killed were in buildings. and the majority of those were in residential buildings. Interestingly enough there was a wooden apartment building that was near St Johns but it was built properly with hurricane strapping and had a continuous load path from the roof to the ground and the damage wasn't very heavy, they didn't lose their roof and no one was killed at that apartment building. Amazing what happens when you build properly. Lisa MacArthur Riverside RI.

Anonymous said...

Your statement regarding the fact that a lack of interest in the weather does not exempt you from it's impacts has unfortunately been proven true in Lancaster NH recently. A father and his little girl were killed in a circus tent collapse caused by a severe thunderstorm. The area was under an SVR and there was 15 minutes or so of warning. So, who is responsible for this tragedy? I agree with Dr. Doswell that we are ultimately responsible for our own safety. The child in case was too young to have any say. The father should have known a day in advance, maybe two, that thunderstorms were possibly on that day. By that morning, he should have learned that they were very likely. He should have known that being in big tent is a bad place to be when there is lightning around. He should have guessed that the circus company would not likely go to any great effort to secure their tent against strong winds. There was about 15 minutes of warning which he could have or should have received via his phone. At the first sign of a thunderstorm he should have gotten out of the tent and sought shelter in either a sturdy building or his car. The clowns who ran the circus had access to the same information. They should have instructed people to leave the tent because of the lightning threat and possible wind threat. They should also have erected their tent properly so that 60 mph or so of wind wouldn't have knocked it over. Calling this an act of god makes about as much sense as saying that if god wanted us to fly he would have given us tickets. That would be a whitewash. My question for Dr. Doswell is whether or not we need any laws to deal with this situation? Should those running an event have to monitor the weather? Should they be required to have a workable severe weather plan? Thank you for your input. Lisa MacArthur, Riverside RI.

Chuck Doswell said...


I have no expertise about what should and shouldn't be done by way of legislation to deal with venue weather hazards (or construction practices). That's in the arena of politics.

What I have learned from a lawyer familiar with the issue of venues and weather is that the legal responsibility for what happens to customers depends on a deceptively simple test question: If there's a reasonable expectation of a threat, then the venue is responsible for taking steps to mitigate that threat.

A few high-dollar lawsuits going against the venue operators and I think any irresponsible venue operators will have to at least try do something. There are responsible venue operators who are already taking steps to deal with weather threats, but this is a complex issue, involving lots of $$. How do you get hundreds of thousands of people to safety at a giant outdoor event like Bonnaroo in TN?

Anonymous said...

Thank you Chuck, what you learned from the lawyer is very interesting. As for the small circus it would be too burdensome for them to monitor the weather and if severe weather approaches, get everyone out of the tent and probably off the grounds, have them go back to their cars. Bonnaroo is a real challenge. The figure I saw was I think 90k people this year. The event is held rain or shine, no refunds for any reason. I didn't see any sturdy buildings, most people appear to be in tents and RVs. Some stay in nearby motels. The odds of a tornado impact are small, even smaller for a violent tornado, but they are not zero. It's appears to me that lightning is the greatest weather threat, other than the heat. Strong winds may be after that. I don't see where the Bonaroo people could do anything at all. As an attendee I would walk to have reasonably quick access to a car. Not necessarily to escape but for protection against lightning and strong winds. The car should not be upset by winds much less than 120 or 130 mph according to a wind tunnel study that I read. If you can park where nothing is going to fall on you, a car can mitigate the lightning threat as well as keep you safe in reasonably high winds. Beyond that, check the convective outlook every night and in the morning and if the next day looks favorable, leave and turn it into a storm chasing day. 90k people don't generally do anything fast so I don't see where the Bonaroo people can do much except hope for the best. The only mention of weather on their website seemed to be about the heat, keeping hydrated, and that it was a rain or shine event. An attendee should look over the grounds for a location that offers the most protection in the very unlikely event of a tornado. With 90k people in the open even a normal thunderstorm may injure several, some fatally. Lisa MacArthur Riverside RI.

Anonymous said...

If the SPC issued a PDS warning for the area of Manchester TN, Bonaroo could order everyone to leave. I don't know what they would do with the people who came by bus, were dropped off, or otherwise don't have any transportation and are staying in rented tents. Then you have to ask the question, do you spread everyone out increasing the risk that someone will be impacted, or do you you present as small a target as possible and hope for the best? For a lot of people, they're basically stuck there for the four days. The odds are very low so you can justify the risk. You can justify not having shelters available. I don't know what good it would do if you have 15 minutes warning? If you can't get to shelter in 15 minutes, the warning doesn't help much. Ironically, up in Lancaster NH, the circus people did have the ability to clear the area within the 15 minutes of warning that was given by the NWS. Bonaroo would have a hard time doing much of anything in 15 minutes as far as I can tell. What's your opinion? Lisa MacArthur Riverside RI

Chuck Doswell said...


Our "conversation" via this somewhat clumsy mechanism has become excessively long, in my opinion. It would be much simpler to do this via emails - if you don't have it, it's easy to find in my Web essays.