Below is Part IV of my series on extended-range weather forecasting taken from my unpublished manuscript INSIDE THE WEATHER CHANNEL.
FOR ENTERTAINMENT PURPOSES ONLY–Part IV
ADVENTURES IN EXTENDED-RANGE FORECASTING
Let’s go back to Super Bowl week 2006. On January 30, I wrote in my weekly outlook narrative:
Well, it must be Monday because I’m clueless as usual. The models are all hinting at a major storm for the eastern U.S. this weekend, but as far as details go, they’re all over the place. For instance, for Saturday evening, one model has an intense storm center over the upper Ohio Valley [which would have been bad news for the Super Bowl scheduled to be held in Detroit on Sunday], another has a storm off the Mid-Atlantic coast, and third places one just north of Maine. Thus, I can offer only glittering generalities (a term I recall from psych 101) at this point: look for heavy rain, high winds and a swath of heavy snow someplace over the eastern third of the country Saturday/Sunday. Depending on which model verifies, there could even be severe thunderstorms in the South.
On Tuesday, veteran long-range forecaster Dave Houtz working in the Global Forecast Center raised the white flag: “I give up,” he said in disgust, when the models, after briefly agreeing with one another late Monday, went their separate ways once more.
As I left work on Wednesday (the end of my workweek), I commented to Dave, “I’m still clueless. I have no idea what’s going to happen this weekend.” The model that on Monday had touted a big storm for the Ohio Valley over the weekend–and then backed off–was back at it, this time for the Northeast.
It turned out to be pretty much a tempest in a teapot. None of the models, prior to Thursday–ah ha, remember the Bernard Three-Day Theorem?–had a really good handle on what finally happened: a storm with more wind than anything else blew northward into southeast Canada. Detroit got a couple of inches of snow while up to nine inches coated parts of northern Michigan. Hardly an epic system. And there were no severe thunderstorms in the South.
Here’s another frustrating example. On Monday, November 17, 2008, I wrote:
After Thursday, I’m totally at a loss. Rarely have I seen the models so disparate. Often in such situations I can lean toward one solution or another, but I really don’t have any idea which way to tilt this week. Even the government forecasters have thrown up their hands.
Under the heading FRI/SAT/SUN I added this:
One model develops a strong storm that sweeps from the Great Plains to the Eastern Seaboard and is replete with everything from wintry precipitation to severe thunderstorms. The same model also predicts a return of Santa Ana winds to Southern California.
Another usually-reliable model indicates that something similar might happen, but that it wouldn’t initiate until Sunday.
Still other models suggest nothing like the previous scenario will take place. One has a weak “clipper” system racing eastward across the Great Lakes, while a couple of others hint at a flabby upper-air disturbance darting from the Great Plains to the Southeast.
A case of too much information, I guess. It may be another 48 hours before the models are able to hold hands on this one.
In the end, nothing major happened.
Photo: A crystal ball in action (no, those aren’t my hands).
But to the truth, sometimes predicting the weather more than a few days in advance is about as reliable as using a crystal ball to foretell the future.