(2 of 4)
Meeting of Masses. But contemporary weathermen base their forecasts on ever-evolving knowledge about how the world's weather machine works. They know that cool air generally brings dry weather and clear skies and that warm air ushers in clouds. They have also learned what happens when air masses of different temperatures meet. A warm, moist air mass, being lighter, tends to ride up and over a cold mass, bringing long, warm rains. A moving mass of cold air, being heavier, tends to plow into a warm air mass like the blade of a bulldozer, triggering brief, violent precipitation at the front, or leading edge, and leaving clear skies in its wake.
Weather forecasting has improved in recent decades. Satellites now track storms from high above the earth. Monitoring stations provide agencies such as the National Weather Service with thousands of readings daily, recording air and sea-surface temperatures, barometric pressures, relative humidity and wind velocities and directions. Sophisticated computers enable weathermen to perform in minutes calculations that might otherwise take months.
With all these aids, short-term forecasting is relatively reliable. For example, a meteorologist can see that a mass of cold, dry air is moving down from Canada at 960 kilometers (600 miles) a day. He can check its direction, figure out how long it will take for its front to collide with, say, a warm, moist air mass sitting over New York and tell about when and where the resulting rains will begin (see chart). Weathermen can make three-day forecasts with reasonable accuracy.
Gambler's Hunch. But beyond three days, the meteorologist's batting average falls rapidly. Five-day forecasts are considerably less accurate; a 30-day forecast is usually little better than a gambler's hunch. "It's like playing dice," says Warren Washington of the National Center for Atmospheric Research in Boulder, Colo. "The odds are that a 30-day forecast will be better than a purely random guess—but not much better." Still, both Jerome Namias of California's Scripps Institution of Oceanography and Donald Oilman, director of the National Weather Service's Long Range Prediction Group, using advanced theories and masses of data, last fall accurately forecast this winter's weird weather patterns. The U.S. Weather Service's 30-day forecast for the period ending January 15, however, was only partially correct. The Service accurately predicted below-normal temperatures for New England and upstate New York, but incorrectly foresaw near-normal temperatures for the New York City area and much of the South and Midwest. It predicted below-normal precipitation for most of the northern half of the country—not the snows that paralyzed Buffalo.
Weathermen acknowledge that there are still enormous gaps in the meteorological data that they need for more precise predictions. There are huge areas of the Southern Hemisphere, for example, for which no sea-surface temperature, wind direction and velocity or barometric pressure readings are available.
More work must also be done on developing computer models of the weather. But no matter how advanced the tools and theories, most meteorologists concede that certain aspects of atmospheric behavior simply cannot be predicted.