Wind patterns—the physical manifestation of the general circulation of the atmosphere—have been known since people starting sailing long distances across the oceans. Some, such as the Viking Erik the Red, found out about persistent westerly winds the hard way during failed attempts to sail from Iceland to Greenland. Others, such as Christopher Columbus, found that easterly winds would carry them across the Atlantic if they sailed down the west coast of Africa before trying to head west. By the late 1600s, these wind systems had been plotted on sailing charts and were used to advantage by ships' captains. The first conceptual model of atmospheric circulation did not appear until 1735, when the English scholar George Hadley (1685-
1768) suggested that the conservation of angular momentum explained the easterly flowing trade winds. By the middle to late 19th century, the Americans William Ferrel (1817-91) and James Henry Coffin (1806-73) had both developed the familiar three-cell model. As the illustration shows, the model describes air rising at the equator and at 60° of latitude, sinking at 30° latitude and at the poles, and flowing along the surface from west to east between 30° and 60° of latitude, and easterly elsewhere. Their conceptual model showed what happened—not why it happened.
Scientists continued to work on this problem off and on, gathering observational information and attempting to make sense of it. The lack of upper-air observations meant that they were only able to see what was happening at the surface. An accurate description of the general circulation needed to include air movement throughout the atmosphere. Scientists would also need simultaneous measurements over a large area of the Earth's surface—an unlikely event given the limited availability of transportation, balloons, and instrument packets. The first attempt to gather a geographically smaller and yet massive amount of upper-air information took place in the mid-1930s, when Jacob Bjerknes and Erik Palmen arranged the launch of a "swarm" of 120 radiosondes from 11 European countries into a developing midlatitude cyclone. A similar event would not take place again until the International Geophysical Year.
By the mid-1940s, meteorologists were really no further ahead in determining the general atmospheric circulation. As the British meteo-
The three circulation cells in each hemisphere present a simple view of atmospheric motion that is made more complicated by topography and heat-transfer processes.
ATMOSPHERIC CIRCULATION PATTERN
Air rises /
© Infobase Publishing rologist Sir David Brunt (1886-1965) noted in 1944, the only way to figure out the wind pattern was to start with the temperature distribution, derive the corresponding pressure pattern, and only then draw a conclusion about the most logical wind circulation. Meteorologists were working on the problem; they just had no way of reaching a final decision on how the atmosphere transferred energy to support circulation. The problem might have remained unanswered except for the introduction of numerical weather prediction.
Original work in numerical weather prediction was not directly connected to the general circulation problem. Models and computers could not handle multiple layers and large geographical areas. The initial runs in the United States did not even cover all of North America—there was not enough memory or calculating power. The purpose of numerical weather prediction was to make a forecast for a defined geographic area, not for the entire Northern Hemisphere and certainly not for the whole globe. Meteorologists were sure that as computer size increased they would be able to do so. In the meantime, the meteorologist Norman A. Phillips (1923- ), who had worked with the Meteorology Project, decided to create a general circulation model over a geographical space approximately 6,200 miles by 3,700 miles (10,000 km by 6,000 km) and to integrate the defining equations every hour for as many days as he could—that is, until the model started producing nonsense. His atmospheric model was simpler than the real atmosphere. Phillips's "world" did not distinguish among land, water, and frozen ocean areas, and he started with a predefined initial atmosphere. Running for 12 computer hours, Phillips's model churned out charts 31 days into the future. The resulting maps provided a reasonable approximation of atmospheric behavior, showing that disturbances (that is, "storms") would be created, live, decay, and ultimately die over a several-day period. This success meant that unlike weather prediction, which only needed to be calculated for a few days, prediction of atmospheric changes over many years would be possible in the future. It would be possible to create not just weather models, but climate models.
Phillips's work, published in 1956, became the basis for all future climate models. Many such models run today. Significantly more complex than the first general circulation model, today's models include detailed topography, changing atmospheric chemistry, and atmospheric variables such as moisture content. Environmental policies based on the output from these models have the potential to shape the way Earth will look in the next century and beyond.
Was this article helpful?