Mar 8, 2012; 3:09 PM ET
When AccuWeather.com releases its seasonal forecasts, people ask how meteorologists can predict the weather so far in advance. What are these long-range forecasts based on? A great deal of research.
Paul Pastelok, leader of the AccuWeather.com Long-Range Forecasting Team, said that the team forecasts a season after studying the complex interaction of current ocean, land and atmosphere conditions around the globe.
For weather enthusiasts, here's a list of some of the weather indices used by the AccuWeather team: ENSO, PNA, NAO, AO, PDO, MJO, ocean temperatures, soil temperatures and analog years.
La Niña and El Niño
The weather patterns across this fluid, ocean planet are highly interconnected. For example, Pacific Ocean temperatures (La Niña/El Niño) are a key factor in forecasting weather patterns.
"We look at past La Niña and El Niño seasons, looking at the strength and trend of those El Niños or La Niñas," Pastelok said.
A La Niña occurs when sea surface temperatures across the equatorial central and eastern Pacific are below normal. El Niño occurs when these sea surface temperatures are above normal. The greater these temperatures depart from normal, the stronger the La Niña or El Niño.
Both phenomena have a significant influence on the jet stream and overall weather patterns across the globe. While
no two El Niños or La Niñas are exactly the same, general trends have been observed in the influence they have on the weather in the U.S. and elsewhere throughout the world.
Analog Years
Long-range forecasters also look for past years in which similar trends occurred. Meteorologists call these years "analog years". An analog year is a past year where there were similar temperatures in the ocean and pressures in the atmosphere to the current conditions.
"No year is ever the same, but you can get the general trends and patterns," Pastelok said.
Models are only part of the long-range forecasting recipe. Long-range forecasters also study recent trends in the weather, looking at temperature, rainfall, drought, snow cover and more. Trends can be recent months, year-to-date or even over many years.
"Models are projected in the future. But if they're not capturing what's going on now, they can't predict what will happen in the future properly," AccuWeather meteorologist Meghan Evans said.
How a Long-Range Forecast Can Bust
There are big risks when a forecaster makes a long-range forecast. The conditions of the atmosphere and the oceans constantly change. An example of a major oceanic change might be a sudden flip from La Niña to El Niño.
"[The oceans and the atmosphere are] not steady state," AccuWeather meteorologist Jim Andrews said. "If the changes are significant... then it can throw everything out of whack."
Non-events are also problems for long-range forecasters, like if a forecast blocking pattern doesn't happen.
"Either way," Andrews said, "you've hung your forecast on certain conditions and to have them happen or not happen can easily send the forecast astray."
Meteorologists can measure the accuracy of the long-range forecasts to see if it was good. Meteorologists gauge if temperatures were in fact above or below normal and whether precipitation and snowfall were above or below normal. General trends can be verified.
However, how people perceive the season often carries more weight than the numbers.
"It's so difficult because it boils down to perceptions," AccuWeather.com Expert Senior Meteorologist Bob Larson said.
For example, with winter, "you can predict a milder-than-average winter with less snow than average and have it verify," Larson said. "But if you have a big Thanksgiving snowstorm, a Christmas blizzard and a New Year's Eve ice storm, that's what people remember. They remember those big storms, not the weather in between, and think it was a terrible winter. Their minds drift to memorable events."