Sunday, March 8, 2015

Weather Forecasting IS Very Accurate, Even If You Don't Believe That

Joe Queenan had some fun with what he said
are wildly inaccurate meteorologists. But contrary to
much public opinion, weather forecasting has
gotten really good. We just remember the
blown predictions better than the right ones.  
Meteorologists are sensitive to that old saw that they can keep their jobs even though they're wrong so often.

Part of the sensitivity is meteorologists are correct in their forecasts the overwhelming majority of the time.

Usually when they are wrong, it's because scientists don't fully understand the atmosphere yet, which gives the weather the ability to surprise even the most seasoned and smartest forecasters.

This issue came to prominence again thanks to a humor article making the rounds the past few days.

Writing in the Wall Street Journal, Joe Queenan suggests firing meteorologists whose forecasts  Queenan says were inaccurate during winter storms over the past couple of months.

Maybe we should fire Queenan if he's ever written something inaccurate?

Queenan's article was tongue in cheek, but it does revisit the familiar complaint that weather forecasts are "always" wrong.

It also seems he still hasn't quite gotten over predictions of a huge blizzard in New York City back in January, one that ended barely missing the Big Apple, but slamming New England, as expected.

Queenan writes, or rants, depending on your perspective, this:

"They say we'll get a dusting and we get an avalanche. They say that Snowmageddon or the Snowpocalypse or Snowgatory is roaring in from the Midwest, and it turns out to be a smattering of harmless flakes.

Schools close when they shouldn't, subways lock up in expectation of a blizzard that never arrives, people go on long drives when they should stay home. Meanwhile in Boston, theres no way to toss woefully inaccurate weather reports into the trash because nobody has been able to dig out the trash cans."

Um, like I said, the Boston blizzard was accurately forecast back in January, but whatever.
Many forecasters indicated that New York City risked
getting two feet of snow in a January blizzard,
but much less than that fell on the Big Apple.
Photo from Reuters.  

Queenan acknowledges that predicting the weather is difficult. But still doesn't cut meteorologists slack. He also seems to think older meteorologists are better than younger ones, which is bunk.

(By the way, Queenan is 64 years old, so he's no spring chicken.)

"Frankly, I think we need young weather people to take over the industry. Athletes are young. Snowplow guys are generally young. Couldn't we hire 12 year olds to hack their weay into esoteric computer models that really work and give us a head's up about what the cloud condensation nuclei are up to?

" how about a Manhattan Project where we train thousands of new weather people to replace the current crew? People with lots of energy and enthusiasm and, I would hope, brains?

And Queenan said this:

"Couldn't we say that any weather-personage who misforecasts more than three snowstorms in any winter season has to go home? Forever?"

Again, I get it. Queenan is pretty much being tongue-in-cheek.

The unfortunate reality, though, is nobody can perfectly, 100 percent accurately predict the weather. At least not yet. Forecasts have gotten much, much better over the past decade or so, and continue to improve.

How does one define a blown forecast? In January, predictions that the heaviest snow - two feet of it - would get into New York City were wrong. But that much snow came close, into central Long Island. Forecasts for the rest of New England in that late January storm were spot on.

Which meteorologists do we fire for that storm, then?

As Marshall Shepherd and Chuck Doswell pointed out in a Weather Underground blog post,  a more recent storm featured forecasts that Atlanta, Georgia would be within, but near the southern edge, of some heavy snow.

"....the rain/snow line was off by 20-30 miles. That is an amazing forecast from a meteorological perspective, but the public now has an expectation of pinpoint, backyard forecasts and timing as precise as that of the official 'world clock.'"

Weather forecasting has had some pretty big and life saving successes in recent history.  As Shepherd and Doswell point out, the general jist of Superstorm Sandy's fateful left turn into the U.S. East Coast was anticipated up to a week in advance. A big scary May, 2013 tornado outbreak in Oklahoma was forecast three days ahead of time.

Weather predictions can't be absolutely precise, or always right. The atmosphere is too chaotic for us to keep track of it that well. But nobody likes uncertainty. That's human nature.

It's really difficult when you get situations like a warm, humid summer afternoon when scattered thunderstorms pop up. It's usually pretty easy to predict the general weather for that kind of afternoon. Some places will get a thunderstorm in the forecasting area, others won't.

You also don't know much in advance exactly when a particular place will get a thunderstorm. At 1 p.m.? 3 p.m.? 5:30 p.m.? You never know.

Yet I've frequently had people ask me on, say, Monday, at exactly what time a thunderstorm will roll through, for example, Burlington, Vermont on Wednesday afternoon. "I need to know if it will be raining at 3:30 p.m."

They act disappointed when you say you don't know.

I guess the only thing to do is try to continue educating the public about how to interpet weather forecasts. Educate them as to what is reasonable to expect from meteorologists, and what not to.

One fairly complicated study shows weather forecasts for the next day are accurate about 95 percent of the time. Forecasts are still pretty good but not great five days out, and you should take a forecast for 10 days from now with a giant grain of salt.

Another interesting note about weather forecasting accuracy. People tend to remember blown forecasts when rain falls when it wasn't predicted. The public is generally more relaxed about forecasts that predict rain but the sun stayed out.

Some broadcast meteorologists have a "wet bias" according to a New York Times article. Consciously or unconsciously, some TV or other media meteorologists sometimes over-predict precipitation, perhaps to avoid incurring the wrath of viewers whose picnics were rained out.

As in any profession, there are good meteorologists and a few bad ones. But none of them are paid to be right half the time. They're almost always correct.

I'm not a meteorologist, personally. I'm just a journalist with a keen interest in the weather and climate. Of course I use the National Weather Service and other forecasting organizations, both public and private, as sources. I'll occasionally give forecasts based on the expertise of these organizations.

Almost without exception, I trust these sources. (There are some weather and climate 'experts' to use the term loosely, who are wildly dumb and irresponsible.  I don't generally use them in blog posts, except for the purposes of ridicule. You figure out pretty fast who's good and who's terrible.)

When there is a forecast that turns out to be really wrong, it behooves us all to explain why the prediction was incorrect. An educated weather watcher is a prepared one. They will also understand the continued limits on the understanding of the mysteries of weather and climate.

No comments:

Post a Comment