Observing the NAO on a year to year basis, however, does not render a complete picture of how the anomaly affects the climate of the Atlantic basin. From week to week, the NAO flip-flops between positive and negative phases seemingly at random, sending good and bad weather intermittently to both southern and northern Europe. Yet, each winter the NAO almost always shows a predominantly negative or positive average for the year. When these yearly averages are put into an index and plotted next to one another, a clear pattern emerges. Since the 1960s, the entire index has overall been growing more positive.
 

One of the goals for atmospheric scientists studying the NAO has been to predict the sign and strength of the NAO from year to year and decade to decade. They could then warn European farmers of when they should plant their crops, alert Mediterranean resorts as to the amount of rain they are likely to receive, and generally forecast next year’s winter weather trends more accurately. Despite its regular appearance, the NAO is still too erratic to predict by looking at a chart of its history. “There are simply too many variables that go into these trends to make them easily predictable,” says Hurrell.

The only way scientists could forecast the dips and peaks of the NAO is if they first understood exactly what was causing the two pressure systems to vary relative to one another. Most agree that the high and low would develop on their own over the Atlantic and that they would change in strength from week to week or even month to month. The irregular sinusoidal pattern exhibited by the NAO, however, would require some type of climatic memory. For the NAO yearly averages to climb upward or downward over several consecutive winters, there would have to be some mechanism in the atmosphere or the ocean that keeps track of where the Azores high and the Icelandic low were the year before. But atmospheric currents change in temperature and density so rapidly over time that there is no way they could maintain a pattern into the spring and summer months after the low- and high-pressure systems break up. The current thinking is the NAO’s variation must be tied to the land or the ocean.

Several years ago scientists made a breakthrough when they confirmed through the use of computer models that part of this climatic memory driving the NAO lies in the deeper ocean temperatures of the Atlantic and changes in these temperatures are largely responsible for variations in the NAO. Mark Rodwell, a climate researcher at the Met Office in the United Kingdom, was one of the researchers who made the connection. Based on this earlier work, he is now using similar models to make forecasts on the sign of NAO nearly one year in advance.
 

“Though this is largely a statistical relationship, there is a reason behind our forecasts. The idea is that if you want to make a forecast for the winter, then you need to look at sea surface temperatures of the winter before that,” says Rodwell. The NAO is responsible for the path of strong storms that pass across the Atlantic, and these strong storms influence the temperatures of the ocean. By the spring of each year, the NAO has left a deep mark on the temperatures of the Atlantic. During the summer, these ocean temperatures are largely preserved because a relatively thin layer of water heated by the sun covers the ocean beneath like a thermal blanket. When the following winter rolls around, the warm layer is removed, revealing the sea temperatures from the previous spring, which in turn affect air pressure over the Atlantic and the next NAO.

To make their forecast, Rodwell obtains the average sea surface temperatures for the current May from satellite readings. May temperatures are the best indicators of ocean temperatures the following year, as winter storms have dwindled and the thin layer of summer water has yet to cover the ocean. Rodwell plugs these values into a computer simulation of the atmosphere and ocean over the Atlantic and runs the simulation forward through the next winter to obtain a forecast of the average values for the NAO for January through March of the following year. “Using the forecast, we’d expect to get the sign of the NAO with 66 percent accuracy. This is better than the 50 percent chance you’d have without any forecast at all,” says Rodwell. Unfortunately, this year, the model wasn’t quite on target. It predicted a slightly positive NAO to occur over the winter 2003.

 Mimicking Mother Nature
 The Highs and Lows of the NAO