Nanny Weather

There’s been a great deal of talk in recent months about lies told by politicians and journalists. The media has been filled with accusations — some verified, some not, many of them petty, but some touching on issues as vital as national security. Not only are the interlocutors under heavy scrutiny — even the fact-checkers are suspect!

However, at times, some take comfort in the knowledge that not all segments of society have been infected with disrespect for truth. At least in the non-political, technical areas of government and science, where there is less reason to manipulate or manufacture facts, where the vote-getting emotions do not hold sway, they have felt on firmer ground.

Yet that assumption, too, has repeatedly been debunked. As it turns out, scientists have agendas as well.

But who would ever have thought that the National Weather Service would fall prey to the temptation to fudge the facts? What ulterior motive could there be that would impel a meteorologist to submit anything other than the most accurate possible forecast, when the person’s reputation depends on it? Nor do you need fact-checkers to tell us if it really was sunny, as they predicted. There’s just no way to spin a rainy day; nobody can tell us it wasn’t really raining.

So it was reasonably assumed that when the massive snowstorm that was forecast for New York City two weeks ago did not materialize, it was simply due to the fact that the predictive powers of meteorologists at the National Weather Service remain imperfect. Surely they did their best, but they just didn’t get it right this time. These are hard calls to make, after all — how much snow will fall exactly where.

Except that such was not the case, really. As of Monday afternoon before the storm, weather forecasters in the northeast had arrived at the conclusion that New York would not, in fact, be hit by 18 to 24 inches of snow, and already could foresee the eventual 7-inch outcome, with rain and sleet.

Yet, as Greg Carbin, chief of forecast operations at the Weather Prediction Center in suburban Maryland, confessed to The Associated Press, they decided not to make any last-minute correction to the official forecast.

Why? “Out of extreme caution we decided to stick with higher amounts,” he said. Carbin explained that they feared an overreaction among the public, who, upon hearing a revised prediction of only moderate snowfall, would fail to take precautions against the ice and sleet, which actually pose more of a danger on roadways than heavy snow.

Furthermore, it turns out that meteorologists, much as they work to protect their reputation for accuracy and reliability, are for the same reason wary of what is known in the profession as the “windshield wiper effect.” Sudden, dramatic changes in weather prediction undermine public confidence in future predictions, on whose proven accuracy people need to rely in the long run.

However noble the motive — willing to look bad on a blizzard that wasn’t, for the sake of public safety — the National Weather Service did not give the public its true estimate of the weather on the way.

Carbin was unapologetic; he believed they did the right thing. Well, he did hedge a little: “The nature of the beast is that there’s always uncertainty in every forecast and we have to get better at describing that,” he said.

The conservative commentator Ed Morrissey warned of the unwanted effects of such a policy: “Yes, keeping information away from people to maintain a panic mode certainly helped maintain ‘readiness and vigilance’ … this time. The next time NWS warns of an impending catastrophic weather pattern, will it get the same response? Or will their diminished credibility lead a lot more people to discount their warnings as yet another attempt to manipulate public response? … It’s as if no one ever read the story of the boy who cried wolf.”

In fact, just such a scenario played out with disastrous effect in the Red River flood in Grand Forks, North Dakota, in 1997. Anticipating unusually heavy runoff from the winter’s snows, the NSW predicted the river would crest at 49 feet; very high, but still not enough to top the levees, which had a capacity for 51 feet.

In the event, the Red River crested at 54 feet. The city was inundated, the entire population of 50,000 was forced to evacuate (just in time to avoid loss of life), 75 percent of the homes were damaged or destroyed, and cleanup costs ran to several billion dollars.

What the government meteorologists knew, but didn’t say, was that the margin of error in their forecast allowed for a 35 percent chance of a major flood. Had they informed the public of this beforehand, better precautions could have been taken, including reinforcing the levees with sandbags to a greater height, and a lot more people might have taken out flood insurance.

Forecasters later told researchers that they did not offer a more nuanced prediction because they feared that expression of uncertainty would result in lowered public confidence in their work.

If political pollsters can routinely acknowledge the margin of error in their predictions, so can meteorologists. There is no shame in admitting that the best you can do is to calculate a percentage of chances. The shame lies in hiding the truth in the belief that people can’t handle it.