Forecasting the Unpredictable main content.

Forecasting the Unpredictable

The North Atlantic Oscillation’s positive-phase trend since the early 70s, with its mild winters over the northeast US and Northern Europe, has added fuel to the already potent debate about global climate change. Undeniable is that Northern Hemispheric temperatures are now at their warmest. And for the past 40 years, the rate of warming has been particularly brisk.

Scientists attribute a sizeable amount of Northern Hemisphere surface temperature increase to the NAO's mechanism. But they are concerned about global warming’s potential influence on this smaller-scale climate pattern. Many researchers agree that the accumulation of greenhouse gases is influencing, to some extent, the NAO's recent positive trend. “The precise mechanisms by which these changes occur is what’s up for debate,” says atmospheric scientist Jim Hurrell, a director at the National Center for Atmospheric Research.


Scientists aren’t completely clear what prompts the NAO’s seasonal and decadal trends. But several research hotspots are offering tantalizing clues. The hopes are that one of these avenues will eventually allow forecasting of future NAO phases, or reveal a “smoking gun” that explains global warming’s link to the NAO.

Making Models

To answer the big-picture questions about the NAO, climate scientists require computer models to tease apart random occurrences from true causes and effects. “When you want to study the complex interaction between the ocean, the atmosphere, and the land, it’s not possible to build a physical model,” says physical oceanographer Martin Visbeck of the Lamont-Doherty Earth Observatory at Columbia University. “We have to build computer analogs that obey the same physical laws that the climate does.” Visbeck has developed such a model used extensively to investigate the interplay between NAO and the ocean. 

If you’ve watched a predicted storm barrel across a video map on a weather forecast, you’ve seen a visual representation of a computer climate model. Visbeck describes the underlying equations as a mathematical version of a “Lego world”: a globe sectioned into countless interlocking bricks. Just as on Earth, each brick in the model obeys its own laws as determined by its spatial position and geophysical characteristics. Each computerized brick interacts with its neighbors in numerically describable ways. “There’s a chain reaction,” says Visbeck. “For example, if it’s warm in one box, then we know the next box downwind is going to warm up.”

Once the computer program is assembled, scientists can enter data—for example, temperatures—into this dynamic Lego machine and watch what comes out; how the model’s version of the atmosphere, oceans, or land responds to those temperatures for a particular time period. To test a model’s efficacy, it’s often run in reverse to simulate the climate for a number of prior years. The results are then back-checked against actual observations. If the model simulation seems solid, it might be able to be run forward to accurately forecast future events.

Martin Visbeck
Making Progress

Visbeck singles out three recent research areas where computer models have shed brighter light on the cause of the NAO’s seasonal and decadal variability. While the most obvious route to understanding causality seems like it’d be to uncover what drives the changes in the Icelandic and Azores pressure systems, that’s not where researchers have been looking.

“It’s hard to make a forecast for the pressure differences,” explains Visbeck. That’s because pressure systems are largely a product of highly mercurial atmospheric change. The atmosphere, as Visbeck explains, has a fairly lousy “memory.” Atmospheric events—air temperatures, currents, and the like—change their values and directions often, rarely repeating patterns. (Scientists call this concept “chaos.”)

Scientists have turned instead to more sluggish systems, ones that project patterns for the longer term, for links to NAO activity. Ultimately, these systems can drive atmospheric change, and thus have a better shot at being predictors.

The first example comes from a team at the Met Office, the UK’s meteorological service. The Met Office predicts the sign of the approaching winter NAO solely from ocean surface temperatures of the North Atlantic in the preceding May. By running their model of this relationship back fifty years, the team has correctly "back-casted" the phase of the actual winter NAO two times out of three. By operating their model forward, Met Office is attempting to predict the upcoming winter NAO. Their forecast for winter 2004-5? You guessed it—positive.

What’s the connection between ocean temperatures and this climate pattern? The NAO is intimately linked with the storm track over the North Atlantic, which affects how warm the water gets. By summer, the storms peter out, but the ocean can sustain its May temperature into the winter. By then, the seawater temperatures are influencing air pressures above—the starting point for the next NAO.

While the North Atlantic can stoke some seasonal atmospheric change, tropical seas, with their extra dose of solar warming—exert much more pronounced, long-term effects on global climate than northerly oceans. One example is El Niño: it originates in the tropics but affects midlatitude weather.  

Jim Hurrell knows this about the tropics. That’s why, in 2003, he plugged winter sea surface temperatures of the Indian Ocean into a global climate model. In response, the model reproduced the weather hallmarks of a positive NAO phase.

Hurrell’s research reflects a second breakthrough on what mechanisms may be governing NAO variability. Since 1950, average winter temperatures of the Indian Ocean have heated up about 0.6 of a degree, with an associated increase in rainfall. Hurrell says that this tropical change is likely a direct consequence of global warming.

Hurrell suspects that the mechanism linking the Indian Ocean with the far-off Atlantic has to do with tropical waters’ effect on the west-to-east path of storms in the North Pacific. “This storm track goes around the world, so it has a ‘downstream’ effect over the North Atlantic.” As this link gets investigated further, it may be able to be used to generate long-range NAO forecasts.

A third area of big-picture NAO research is looking at what’s affecting the atmosphere from the top down rather than the bottom up. It’s fairly well accepted these days that greenhouse gases are changing the chemistry of the stratosphere, the atmospheric layer above the troposphere. The troposphere is the layer in which we live and where the NAO goes about its business.

Scientists have proposed that the reduction of ozone and the buildup of greenhouse gases in the stratosphere have livened up the circulation in this sensitive atmospheric layer in recent years, such that its westerly flow has significantly strengthened. Some teams’ findings suggest that an intensified stratospheric circulation can provoke the troposphere below, nudging it toward a more positive NAO index. Although this coupling is highly debated, it brings up intriguing ideas: that if we could predict our anthropogenic influence on stratospheric chemistry over the next several decades, we might be able to likewise forecast the NAO’s behavior. And if we temper our influence, we could alleviate our potential effect on the NAO.

Making Forecasts

If scientists develop accurate NAO forecasts, business owners, government officials, and everyday citizens will benefit: Growing seasons and crop yields could be anticipated, ecosystems could be protected, and water and energy resources could be better managed and financed. Visbeck points to his work with several Norwegian energy companies as an example. Because hydroelectric energy prices directly correlate to the NAO phase that season (high NAO = more water flow = cheaper electricity), predicting the index would help the companies plan for their purchasing of surplus energy if needed.

“It's very rewarding for a scientist to be able to say what the future might hold,” says Visbeck. “But more importantly, predictability allows for more insightful stewardship of our planet, of our resources, of our whole population.”

Related Links

UK Met Office's winter NAO forecast

Weather Models: USA Today
General information on forecasting with weather and climate models.