Share this article

print logo

Don Paul: On busted forecasts and Winter Storm Stella

I choose to write on this topic at a time when my forecast and those of many of my local colleagues are doing just fine in the midst of our long-duration winter storm. The timing is not coincidental; I’m not as dumb as I look!

A busted forecast is a fuzzy phrase that purports to describe a forecast gone bad. Some laypeople call forecasts busted that are nowhere near my idea of a bust. For example, if I had forecast a range of 8 to 12 inches of snow for the metro area, and East Amherst got 14 inches while West Seneca got 7 inches, that would be an imperfect but not busted forecast. Nineteen inches in East Amherst and 2 inches in West Seneca? Well, then you may have something.

There appears to have been a major bust for the largest population center in North America in terms of snowfall amounts for this Nor’easter, colloquially known as Stella.

The basics of what happened in New York City are this: With the many computer models and ensembles/multiple runs of those models that run every few hours (depending on the model), most models as of Saturday had been keeping the center of the powerful storm just far enough out to sea that its circulation would not tap enough of the warmer air over the Atlantic to do anything more than produce very heavy, water-laden snow. During the weekend, the European model and its ensemble began to nudge the path of the low closer to the coastline, still keeping things all snow except for far eastern Long Island and Cape Cod, where sleet and rain could enter the picture.

Due to a certain lack of confidence, the National Weather Service New York Forecast Office waited until Monday afternoon to elevate a blizzard watch that went up Sunday afternoon to an actual blizzard warning. They were, of course, seeing the same uncertainty I was, and probably in greater detail. But by late Monday evening, I posted on my Facebook page: “Also, a number of models suggest NYC's projected snow totals are overdone due to a lengthy period of sleet holding accumulations down. Still, impact will be major there due to strong to High Winds, some coastal flooding, and heavy rate of snowfall before sleet interruption.”

The evidence was becoming better defined that the low would take an even more westerly path, allowing it to tap the warm air over the Gulfstream and to rush it in aloft over New York City. Here is one of the model runs I relied upon for that post. Note its depiction of precipitation change over New York City when you click on it frame by frame or animate it.

Of course, it was easy for me to speculate on a forecast area that I do not serve. But think of the dilemma for the National Weather Service's New York Forecast Office. They had only just issued the blizzard warning a few hours earlier, triggering massive emergency manager response, expense and deployment of manpower. Flip-flopping is bad business, and especially so in this case. To have just quickly acquiesced to sneaking suspicions and drop the warning would immediate raise questions about why it was issued in the first place.

At the time of issuance, the evidence was still pretty strong for a bona fide blizzard, with more uncertainty for eastern Long Island and even more for Boston. In fact, the blizzard warning verified for the interior suburbs to the north and west of New York City. Winds gusted to over 50 mph closer to the coast, and there were power outages. The intensity of the storm was as expected. But the temperatures a couple of thousand feet up were warmer than originally expected, and that made all the difference for places like Central Park and the airports.

I am certain the National Weather Service forecasters on duty who went home while the overnight shift replaced them did not sleep well, if at all. I am certain they did not like the looks of what they were seeing when I, an outsider who did not have the massive responsibility they had, was seeing the same things. Gulp. In the morning, as the reality unfolded, the blizzard warning was downgraded to a winter weather advisory.

I have empathy for these meteorologists. In 40 years of doing this, I have certainly had my fair share of busted forecasts.

Last week, the Weather Channel did a piece trying to rate the most difficult regions for weather forecasting. The Great Lakes came in high, as did the Northeast. For Western New York, it’s the agony of precise location of lake-effect bands and their motion, as well as intensity. Over time, we do much better than we did when I arrived in 1984. High-resolution computer models have been a great tool for forecasting wind direction. Even now, as I’ve written in the past, if the low-level wind forecast is off by just 5 degrees on the face of the compass in central and northern Erie County, you can easily warn — or not warn — the wrong 150,000 people.

When a snowstorm or blizzard forecast busts in the New York City area, a heck of a lot more than 150,000 people are affected. The stakes are much higher. More than 20 million people live there.

I grew up right across the river from Manhattan. The battle of the rain/snow line with Nor’easters was a nail-biter for most storms. And for a snow loving kid, I can’t tell you how many times I was bitterly disappointed when blinding snow suddenly began pinging on the glass and visibility drastically improved. I knew it was the beginning of the end for heavy snow, and this disappointment continued in college because New Brunswick, N.J., is less than an hour from NYC.

Yes, we’ve had incremental but tremendous improvement in forecasting abilities since I went to school. But meteorology will never be an exact science.

Allow me to use my tired analogy: the atmosphere is not Lego blocks that snap neatly into place. More busted forecasts will come. Despite the nonsensically inaccurate jokes about being wrong “half the time,” it’s not easy to get over a major bust, even when we know we routinely do a good job. Does the October Surprise ring a familiar note?

Busted forecasts leave psychological scar tissue, and we are fully aware of a tendency for many people to remember the busts more than they do routine accurate forecasts. I probably have the same tendency to malign economists unfairly. I suppose “weatherman” jokes are funnier.

It goes with our territory, but I can get a wee bit defensive when I KNOW I blew a forecast and people just have to point out what I already know.

Story topics:

There are no comments - be the first to comment