IE 11 is not supported. For an optimal experience visit our site on another browser.

Snow Letdown: How Do Weather Forecasts Go Wrong?

New York City was expecting a massive amount of snow, but it never materialized.

New Yorkers woke up expecting to be completely buried by a blizzard that hit the city on Monday night. Instead, they got around six to seven inches of powder.

That is a respectable amount of snow. It's not, however, the historic storm that some weather experts predicted would dump three feet on New York City — a warning that caused the city to practically shut down with subway service suspended and all road travel banned.

The massive snowfall predicted for New Jersey and Philadelphia also failed to materialize. But meteorologists were right about New England. Worcester, around 60 miles from Boston, was blanketed in 25 inches of snow on Tuesday morning.

"Swing and a miss… By about 50 miles to the east," Dan Zarrow, the chief meteorologist for Townsquare Media in New Jersey, wrote in a blog post Tuesday morning.

Perhaps the most extensive mea culpacame from Gary Szatkowski, a meteorologist at the National Weather Service's office Mount Holly, N.J., who apologized to New Jerseyans in a series of tweets:

Kate Bilo, a meterologist at CBS' Philadelphia affiliate, assured online critics that she and her colleagues are punishing themselves enough.

So how did weather experts get this one so wrong, at least in some areas?

With this storm, they knew there would be a pressure gradient, or "western wall of snow," where everything east would get hit hard and everything west would escape with relatively little snow, according to Greg Carbin, warning coordination meteorologist at the National Oceanic and Atmospheric Administration (NOAA).

The gradient during this storm was very narrow, measuring 50 to 150 miles across. That made it extremely difficult to predict what exact areas would fall on either side of the "wall of snow."

"The science of forecasting storms, while continually improving, still can be subject to error, especially if we're on the edge of the heavy precipitation shield," the US National Weather Service said in a statement. "Efforts, including research, are already underway to more easily communicate that forecast uncertainty."

Weather predictions for the United States mostly depend on data from two Geostationary Operational Environmental Satellite system (GOES) satellites.

They split up North America, one taking the western half, the other the eastern half. For 10 years, they have been faithfully sending down images that help meteorologists predict things like the current blizzard.

In 2016, they are slated to be replaced with two new GOES-R satellites. They will take images with much better resolution and will be able to sense parts of the electromagnetic spectrum that current satellites can't detect.

"There should be an increase in the quality of the forecasts due to the simple fact that you are increasing the quality of the data," Carbin told NBC News.

Still, we need to develop technology that can better process the deluge of data coming in from these satellites, he said.

"It's going to require greater bandwidth, it's going to require greater processing speed," Carbin said. "But technology tends to rise to the occasion in these situations."

NBC News' chief meteorologist Bill Karins pointed out forecasters are "only as good as the information we get" from weather computers, which are in turn dependent on observation systems from around the country.

"Until we spend more money and get more investment into those systems, we're always going to have forecast errors like this," Karins said.

But advancing technology probably still doesn't mean our great-great-grandchildren will know exactly what the weather will be at all times.

Ideally, meteorologists would have sensors and satellites tracking every fluctuation in the atmosphere and an incredibly powerful computer running a complete simulation of weather on Earth. Even in that scenario, Carbin said, it would be impossible to get forecasts right 100 percent of the time.

Think of the famous "butterfly effect." Essentially, even knowing all of the the initial conditions in a nonlinear system like the weather won't allow someone to know the outcome days later.

"There is a limit to predictability when it comes to forecasting and it’s always going to be there,” he said. “Human behavior behaves like that, the stock market behaves like that and weather behaves like that.”