This past winter, weather forecasting got a major black eye in New York City. Words like “epic,” “crippling” and “historic” were used in describing what was thought to be the coming of a 100-year winter storm.

Lines ringed around blocks at supermarkets as the city stocked up on supplies for the doomsday scenario. Subways were ordered shut down. Stores, schools and airlines were closed in advance of the storm. And then — when the moment of truth arrived — it turned out to be typical New York City winter weather.

In the aftermath, there was an outcry from annoyed citizens about overhyped weather predictions. Government officials became defensive and weather forecasters took a lot of heat for not only missing the mark, but for their use of language that was too over the top.

The comedian Louis C.K., whose concert was cancelled at Madison Square Garden, told his fans they are calling this storm “historic” which…. Well I didn’t know you could call a thing historic if it hasn’t happened yet.”

Now that the winter season is over, the nation is still paying the price for the cold weather. The American economy barely grew at all in the first quarter of 2015. Extremely frigid winter in parts of the country is getting a major part of the blame.

The winter of 2015 has focused a light on the state of weather forecasting and the difficulty of getting it accurate. The forecast wasn’t entirely wrong. There was a huge storm. What was wrong is where the storm hit….and where it didn’t hit: New York City.

The devil was in the details. It’s a very fine line using different computer models from throughout the world to make a precise prediction of weather. When a city of eight million people is the target area, the implications of a miss are magnified.

The U.S. Senate Commerce, Science and Transportation Committee held a hearing on weather last month. One of those to testify was Dr. Jay Trobec, chief meteorologist at KELO-TV, the CBS affiliate in Sioux Falls, South Dakota. Trobec also teaches climatology at South Dakota State University and is commissioner of professional affairs for the American Meteorological Society.

“One thing that was clear on Capitol Hill is that everybody knows weather costs money these days,” said Trobec. “We are all impacted. Transportation, agriculture, tourism are all impacted. There is nothing in our society today that isn’t impacted by weather.”

Trobec told the senators that for all the fancy technology used to present the weather to audiences, one thing is paramount. “We must get the forecast right, because every day, every one of us is impacted directly or indirectly by the prediction.”

But, Trobec admitted, getting accurate forecasts is easier said than done. The United States once led the world in weather forecasting, but now that leadership has shifted to Europe. Last month, even the U.S. Air Force decided to fold its own forecasting unit and has signed on for weather information from UKMET (Unified Model of the United Kingdom Met Office) in the United Kingdom, a service more respected for accuracy than any in the United States.

However, the global leader in average accuracy for weather forecasting today is the European Centre for Medium-Range Weather Forecasts (ECMWF), based in Reading, England. It’s referred to as the “European Model” in the North America to distinguish it from the American Global Forecast System, or GFS.

In addition to the various algorithms used to predict weather, a key factor in determining accurate predictions is supercomputing power.

“If you look at all the biggest supercomputers in the world they are used by weather services for computer modeling,” said Trobec. “That’s true in Europe, Japan, Britain and around the globe. When you do computer modeling, you have to convert a complex series of factors about the atmosphere from every point on the Earth into ones and zeroes. It’s so expensive that only national weather services can afford to do that.

“We used to have the biggest computer in the United States,” he continued. “But now other countries have roared right past us. Who has the best? It changes week by week. But to be honest, in terms of weather modeling, we’ve let the Europeans get past us.”

In January, NOAA’s National Weather Service in the U.S. said it will increase its computing power tenfold by October to improve its forecasting. “The other countries almost immediately said their systems would be even bigger by then,” Trobec said. “We are in a kind of computer arms race.”

Another key difference between major weather services is that the United States makes its data available for free to anyone who wants it. The better foreign weather services charge substantial fees for the use of their data, even to media outlets in their own countries.

The reason the U.S. has fallen behind in weather forecasting, Trobec surmised, is the failure to spend the necessary money to keep up the computing race coupled with a brain drain of top people moving from the government to the private sector. “The private sector is peeling away some top talent for higher pay,” he said.

To regain U.S. leadership, he said, there has to be better cooperation between weather experts in the private, public and academic sectors. “Those three sectors have to be involved together to make U.S. weather forecasting the best it can be,” Trobec said. “And we have to get this arms race figured out. It’s crazy that our own models are not the best in the world.”

In the presentation of weather forecasts, the U.S. leads everywhere else in the world. “The weather graphics companies in the United States get the raw weather data and they add value to it by using their own expertise to find better ways to display it,” Trobec said.

“One of the big examples was the Baron VIPIR (Volumetric Imaging and Processing of Integrated Radar) system, which was the first to graphically display things like wind shear. They did all the calculations on the data and were able to display it. It became a monumental step in television weather. They have their own scientists who can pull data out and work with it to put it in a form that’s simple for viewers. Those companies do add value to the presentation of weather.”

In Europe, he said, while they may have the best forecasting models, their broadcast stations are not up to par with those in United States. The reason is those stations have to pay big dollars for weather data as well and most simply can’t afford it. And, Trobec added, many of the weather forecasts are taped earlier and are often several hours old by the time the public sees them.

“One of the things Baron (and others) are doing now is taking the radar data from the dual polarization (dual-pol) radar deployed by the National Weather Service and simplifying it for graphics display on television,” Trobec said. “So a little screen marker will automatically pop up where something is going on like tornado debris, very large hail or rain. Somebody watching television sees that marker and knows what’s going on right over them. That’s a big thing.”

The major weather technology players are WSI, the professional division of the Weather Company, the brand that owns The Weather Channel, weather.com and Weather Underground, AccuWeather and Baron Services. Other companies, like ChyronHego and Vizrt, create weather graphics gear using data from the National Weather Service or third party companies.

In April, AccuWeather and Baron announced a partnership, bringing AccuWeather’s display and digital information together with Baron’s weather intelligence and radar technology.

The Baron VIPER weather system for television stations is explained in this video.

AccuWeather is also utilizing Baron’s radar and severe weather products on its new 24/7, “All Weather, All the Time” AccuWeather Network. The network has incorporated Baron's severe weather intelligence in its live storm coverage, in addition to weather alerts, information and updates.

These companies use the data from the large supercomputers, along with their own, to shape their visual presentations. Some pay for the services, while smaller companies use the U.S. weather service at no cost.

However, Trobec warned, all the computer data and visual technology in the world is not going to result in the best television weather programming.

“A computer model is never going to tell you it’s going to rain or not rain. It will not tell you a tornado will occur or not occur,” Trobec said. “What it does is give you probabilities that the atmosphere will look like this or that. That’s why there is always a role for a human forecaster because we have to interpret the numbers that the computer gives us. That’s why we get such a variance.

“At KELO, we have both full WSI and Baron systems. I have another smaller weather computer at home for local conditions. I pick what I want to use on any given day based on the situation. Most stations are not that lucky. They tend to have one system or another. But it’s the meteorologist that separates one station’s weather from another. It takes experience to know what works best on any given day in a given part of the country. In the end, it’s the skill of the individual meteorologist that matters most.”