How much warning of danger is too much?

As I write this (Monday afternoon - 31 Jan 2011), the midwestern US is under the threat of a major winter storm. This event has been anticipated for several days, thanks to the increasingly accurate numerical weather prediction (NWP) models that are the major foundation of weather forecasting these days. Amongst my meteorological acquaintances, there's been widespread grumbling about the "hype" associated with public statements regarding this impending event. Recent history shows a number of events that were "hyped" that turned out to be not much. A colleague has already written an extensive blog regarding the challenges associated with forecasting such situations. I have little to add to his excellent and detailed assessments of this situation.

However, I do want to comment about the issue of "hype" associated with weather warnings (which are forecasts, of course). First, I'll offer some background.

The local TV stations here in the Oklahoma City metropolitan area have been engaging in a decades-long "weather war" for ratings supremacy in this weather-conscious market. TV broadcasters here get much more air time than in most (if not all) other TV markets and it's been demonstrated (in the OKC market) that the most popular forecaster helps his station get the best ratings, on the average. One mechanism for this is their efforts to outdo each other to show how they're helping the public, which includes making almost any reasonably strong thunderstorm approaching the metro area sound like the imminent apocalypse. I've long felt that they have used the weather in this region as a mechanism for self-promotion in the battle for ratings, with service to the public a distant second in their priorities. In fact, I long ago quit watching the local stations (except when they're airing live tornado videos!), out of sheer disgust at their vacuous self-promotion. Of course, that's the industry standard in the media - ratings always trump service.

So much for the background. Now, I hear from some meteorologists I know that the warnings about the potentially serious impacts of this impending winter storm have been contributing to the "hype" associated with forecasting. I have two problems with such comments:

1) Any winter storm carries with it some significant hazards to human life (and property). That simply cannot be denied. Now, if we could forecast the weather perfectly (i.e., with absolute certainty) then the issue of what people choose to do (or not do) in response to those forecasts would be entirely in the hands of the users of the forecast information. I recall seeing news coverage in the past of, say, large numbers of cars abandoned on Interstate highways in a blizzard, even when the storms producing the blizzard were well-forecast. Apparently, those travelers either didn't receive the message, didn't understand the message, didn't know what to do about the message, or simply ignored the message for some reason (including not believing it, or not believing it applied to them). None of those options are in the hands of the forecasters, who are just trying to put out the best forecast they can. Perhaps some might blame the forecasters for these stranded travelers because the forecasters didn't word their forecast products strongly enough! Apparently, some forecast users need more than just a winter storm warning to take the threat seriously!! Forecasters are not to blame for this, however.

2) The seriousness of the wording in a forecast is likely related to forecaster confidence in the possibility for a major weather event. Some forecasters are more confident than others in any given situation, and disagreements about the confidence one might want to put into a forecast intended for public users of weather information are common. In the best of all possible worlds, forecasters would be "calibrated" so that their uncertainties wouldn't be widely different in a given weather situation. But we don't live in that best of all possible worlds - opinions about the likelihood of a given event will vary. If a forecaster is overconfident frequently, then s/he needs to re-calibrate, which necessarily involves doing a meaningful verification of the forecasts and learning how to use that feedback to achieve proper "calibration". Having said all this, if a forecaster is confident that, say, a major winter storm event is likely, why not spread the word about that in advance? Is that "hype" or just being honest about one's meteorological assessment of the situation? Should a forecaster suppress information just for the sake of not being accused of contributing to weather anxiety? It might not happen as forecast, but isn't it better to have prepared for the potential and have it turn out you didn't need it, than to make no preparations and then discover you should have? Better safe than sorry, it seems to me!! Err on the side of caution, if erring is inevitable. [I'm avoiding the subject of expressing uncertainty in the form of probabilistic forecasts. See here for some discussion.]

This brings up the subject of "panic" in relation to weather forecasts. For many decades, tornado forecasting was forbidden because some bureaucrats decided that the public would panic at the mere mention of the word "tornado" in a weather forecast. When the bureaucrats in the Weather Bureau were forced to being issuing tornado forecasts, no panic ever ensued. The notion that panic will result from telling the honest truth about weather expectations has no factual basis whatsoever. If foreknowledge of an impending winter storm causes people to rush to stores in order to stock up on items they might need for a period of paralysis in the wake of a storm, that isn't "panic" - it's common sense!!

As I write this, there's a debate ongoing about the need to add qualifiers to winter storm warnings - something along the lines of adding the "particularly dangerous situation" phrase to a winter storm warning. Personally, I see no reason for, or value to, such a step. If someone doesn't already know that a winter storm can be dangerous, it's not at all evident that adding such a phrase is going to make a difference for them. Of course, no one knows much about how the public sees such things because we have yet to do the hard work to validate some of these ideas. I could be all wet about this - adding such a phrase might make a huge difference. But I want to be shown some solid evidence before I change my mind.

We can never prevent some people from doing stupid things. There always will be those who will put themselves in danger deliberately, for reasons of their own. There always will be those who choose not to accept any personal responsibility for their own safety. There always will be those who refuse to make any preparations for weather hazards, under any circumstances. And I agree that we meteorologists can't simply put our best forecast out there and hope for the best when it comes to people being able to use our forecasts successfully. I've talked about this elsewhere. Public sector weather forecasts (including those by media forecasters) should be crafted to be of the greatest possible value to the greatest number of users, but we meteorologists are not knowledgeable about how to do that. We need factual information if we are to re-formulate forecast products to match that goal. Not guesswork, not opinions, and not speculation (without evidence).