Why iPhone and Android Weather Apps Are Freaking Out About Winter Storm Forecasts

A foot of snow in New York? Two feet? Well, it’s complicated.

On Wednesday, when cellphone weather apps began warning their users that a sprawling storm system could drop a foot or more of snow on New York City this weekend, a murmur of anticipation — or angst, depending on your point of view — spread across the region.

Similar scenes played out across the country this week as, seemingly by the hour, meteorologists shared newer and more intense forecasts for what the expansive storm could be expected to do as it barrels across the country this weekend.

Are those eye-popping forecasts wrong? Well, yes, and no.

Here’s how they come together — and why they differ from other forecasts you may be seeing.

When meteorologists at the National Weather Service create a forecast, they balance an arsenal of computer models; “ensembles,” or collections of dozens of simulations; and their own experience. They are even experimenting with using models that are powered by artificial intelligence. Your local television meteorologists may rely on the Weather Service’s guidance, or may add some of their own expertise when they put together the report for the 6 o’clock news.

In contrast, many weather apps show the results from a single forecast model, instead of the many that meteorologists will draw from in preparing their reports. The app pulls raw data and delivers it to the palm of your hand, but often without context that allows for the other possibilities. Sometimes, that ends up being the right one, and sometimes not.

Another complication that both the apps and the humans can struggle with on a storm like this: what happens as the precipitation hits the ground. There are rules of thumb that allow meteorologists to assume that a certain amount of rain falling high into the atmosphere will turn into a certain amount of snow, or freezing rain, or sleet, or just rain when it falls to the ground. But it can be especially difficult to predict that with certainty in advance, which is why forecasts often give such wide ranges for things like snowfall totals.

The most extreme forecasts, regardless of their source, are the ones that cause the most excitement, even as some meteorologists may be more cautious.

“Everything that catches attention is mostly nonsense,” said Eric Fisher, chief meteorologist for WBZ-TV in Boston and the author of a book chronicling major storms in New England.

Mr. Fisher cited viral snowfall maps that often spread widely on social media. “If you’re scrolling through your Instagram or your TikTok feed and you see something that’s big bright colors, and someone’s going crazy … that, unfortunately, gets the most attention these days,” he said.

In meteorology, the “gold standard” global computer models — the American (GFS) and the European (ECMWF) — often begin a week as rivals and end it as partners. This week was a classic case study in that convergence.

On Monday, the two models were picturing vastly different scenarios. The American model showed a sprawling, intense storm tracking well to the south, bringing freezing rain to the Gulf Coast and snow to the Deep South. The European model took a more northerly track but lacked the intensity of its American counterpart, suggesting rain for the Gulf Coast.

In somewhere like New York, the difference was stark: the European model teased Sunday snow for New York, while the American pushed the system safely out to sea.

The gap began to narrow on Tuesday, as the American model began to nudge its track slightly to the north. But it was on Wednesday afternoon that confidence among forecasters surged as the American model made a significant move north. For the first time, the models were in broad agreement on the storm’s initial stages for Friday and Saturday.

This alignment allowed meteorologists to turn their attention toward the Northeast; it was the first time the idea that half a foot or more of snow entered the conversation, though experts cautioned it was still too early for specific totals.

As of Thursday afternoon, the models have stayed largely in sync — but the eventual snowfall totals for Sunday into Monday still rest on subtle shifts in the storm’s path that won’t become clear until the storm arrives. The National Weather Service on Thursday said snowfall totals around New York will likely range from six to 12 inches, and forecasters were increasingly confident that over a foot of snow will pile up at some locations including across New York City.

So what’s a person to do?

Mr. Fisher advises that once you find a meteorologist or source that you feel is right most of the time and provides quality information, you should “stick with them” rather than constantly switching between outlets.

Related Content