On the value of “call-to-action” statements in severe weather warnings


Chuck Doswell

Posted: 20 October 2009 Updated: 01 May 2011: some minor changes

This is my opinion. If you wish to communicate your opinion regarding this topic, you can contact me at cdoswell at earthlink.net - either use the email hyperlink or cut and paste after replacing _at_ with @). However, if you're not willing to have your comments posted here, along with my response, don't waste my time or yours.

The notion that tornado warnings should contain one or more "call-to-action" (hereinafter CTA) statement(s) has become quite widespread of late, perhaps in part because of a growing concern for how the recipients of severe weather warnings respond to the message of approaching hazardous weather. I suppose that the underlying intent is that a CTA within a warning should increase the frequency of some desired response, although it's not evident just what the desired response by each recipient should be. That is, the situation can vary from one recipient to another, so that there likely is no "one size fits all" message we can distribute within a warning that would offer personal advice, tailored to the needs of each individual.

I take it as axiomatic that for a severe weather warning, as is the case for any meteorological message, the value of that information to the recipients depends on them:

  1. receiving that message,
  2. understanding the information contained within the message,
  3. believing that the information within the message is accurate and relevant to them,
  4. knowing what the information contained within the message means,
  5. deciding whether or not to take action based on that information,
  6. taking the correct action.

Severe weather warnings typically are issued for finite areas that encompass a diverse group of recipients. That diversity is a major factor in making CTAs something of a problem. What sort of advice to the recipients can be offered in a brief warning? There isn't much time within a short-fuse severe weather warning to cover all the possible situations in which the recipients might find themselves. A warning that included a wide enough variety of alternatives to cover all possible actions (depending on the personal circumstances of each of the recipients) would become unacceptably lengthy. Therefore, any particular piece of advice regarding actions to take can apply only to some fraction of the recipients. Other people, needing to take different actions, won't be told what they need to do. So how does a forecaster choose what to say, and to whom? Forecasters are educated and trained to make weather forecasting decisions, not to make such decisions for the recipients of weather warnings. I suppose someone could tell the recipients to "take their tornado precautions" but that seems pretty vague. Just what are someone's "tornado precautions"? It depends on the their situation, of course.

The notion of the credibility of the source for the warning is a critical issue. It's not always clear the extent to which people trust the public weather service warnings. Many people who receive their information from TV have a great deal of trust in the weather broadcaster of their choice, whether or not such trust is actually warranted.

It’s my opinion that a forecaster should confine the content in a weather warning to the weather and not engage in telling warning recipients what to do. Those recipients find themselves in various kinds of situations and whatever a forecaster might have to offer to some segment of the recipients will be irrelevant or even possibly harmful to other segments. For private sector forecasters who have clients with very specifically-defined situations, of course, a warning could be very precise about including information about what its recipients should be doing. But a private sector weather broadcaster addressing warnings to "the public" (very close to the same audience as the public sector forecasters address, since the NWS does very little to disseminate its warnings, except to the very limited audience which happens to have "NOAA" weather radios) has the same problem as a public sector forecaster. Weather forecasters should forecast the weather, not make forecasts of societal impacts or tell their users what to do!

Presumably, the goal of CTAs is to address points #3-5, above. Is it worthwhile to attempt to offer advice to some limited fraction of the recipients of a warning at the very last minute? It seems to me that the time to help people learn what to do and how to make such a decision needs to be long before the threatening weather even develops. The public should be taught what are appropriate things to consider when confronted with a threatening weather situation. That learning should be done months and years before having to confront the threat. It strikes me as doing very little for the public in general to isolate some group amongst the diverse recipients and tell them what to do in the last few minutes before the weather hits them. We need to be involved in preparedness programs of substance, not last-minute advice to those in the path of a dangerous storm.

It’s evident from the limited sociological studies that most recipients of weather warnings are not doing what the NWS is telling them to do. Most of them don’t seek immediate shelter just because they’re told to do so in a warning message. Does a CTA have much of an impact on that? I have no evidence, to date, but I also believe there’s not much evidence to suggest that the CTA does have the putative desired effect. It’s my understanding that recipients of severe weather warnings generally seek additional information to confirm that announced threat before they take any protective action. Information-seeking to confirm a weather threat can take various forms: going outside to look at the sky, calling neighbors or relatives, seeing what their favorite weather broadcaster is saying on TV, etc. I suppose the “normalcy bias” means that most warning recipients seek to establish the reality of the threat because such threats definitely aren't "normal" in their lives. Severe weather hazards are relatively rare events and only affect people occasionally, even in the most severe weather-prone areas of the world. Most people have been in warnings before, but nothing happened to them. Is this effort to confirm the threat a reaction to the so-called “cry wolf” syndrome? It seems quite possible, but we have yet to establish much about the impact of previous perceived false alarms. What do people consider a false alarm? If a tornado passes nearby but doesn't affect them personally, do they or don't they consider that a false alarm?

In any case, it seems to me that warning recipients aren’t doing what the NWS is telling them to do, despite the CTAs, and are only marginally more likely to do so as a result of modifications to CTAs. Further, we presume to know what people should be doing, when the reality is that we, as warning forecasters, know little or nothing about what constitutes the best thing(s) to do for each and every citizen in the path of a storm.

A person living in a mobile home out in the country should make decisions about what to do that would differ substantially from someone living in a steel-reinforced concrete home. The more vulnerable one is to a weather hazard, and the longer it takes to get to an adequate place of shelter, the earlier in the process one should begin to take actions. But those are personal choices by the warning recipients. How could CTAs provide the correct information for all the people with diverse needs within the threatened areas and convince everyone to respond without delay? I just don’t see how they can be of much help, and they waste valuable time in a warning message, when the goal should be to provide the most useful and detailed weather information.

If the tasks associated with preparing communities for hazardous weather have been done properly in the months and years before the event is imminent, then the clear and simple expression of a warning forecaster’s best judgment about the impending weather situation should suffice. Obviously, if someone needs to be told what to do at the last minute, the effectiveness of that message is going to be reduced. It seems unlikely that adding the CTAs to warnings is something that wil offset that lack of preparation to any significant degree.

The NWS can and should be doing everything they can to promote the process of becoming prepared for severe weather. People need to have a clear picture of what their options are in the situations where they might encounter severe weather: at home, in their car, at work, asleep in their beds, at school, at play, on vacation, ... virtually anywhere and any time. If they have an understanding of how to minimize the hazard to themselves in all these diverse situations, then the only things they need to negotiate the path 1-5 above are:

The recipients should know, in advance, what to do and when to do it because they've been taught how to use that weather information in the situations where they might encounter severe weather hazards. Last-second orders are unlikely to be effective. If we consider the simple facts about who is dying in tornadoes, for example, they aren't the people who have good options - they are the people who didn't get the message in the first place, or didn't believe it would happen to them, or had no place to go in the time they had. For such recipients, then no amount of CTA statements in the warning will suffice for them to take useful action to protect themselves!