THE next time people on the East Coast experience a particularly mild winter, they may have sunspots to thank. In over a century of looking, the strongest link yet between the 11-year solar cycle and weather variations on Earth has recently been found: one that includes winter temperatures over much of northeastern North America.
The researcher who discovered the effect is Karin Labitzke of the Free University of Berlin, an expert on weather in the stratosphere. She and her collaborator, Harry van Loon of the National Center for Atmospheric Research in Boulder, Colo., will present the details of their findings to the American Geophysical Union at a meeting this week in San Francisco.
Dr. Labitzke made the discovery while studying an effect that equatorial winds have on weather at the North Pole. Researchers at the University of Washington had reported that the vortex of high-altitude winds - which swirl over the pole during winter - is stronger and colder when stratospheric winds at the equator blow from the west than when they blow from the east.
The West German scientist confirmed the polar effect and, in a 1982 paper, pointed out an intriguing fact: Despite the strengthening effect of western equatorial winds, the polar vortex broke down - causing mild weather - when the number of sunspots was near maximum.
Then, last February while attending a meeting in Washington, D.C., Labitzke got the idea of plotting the sunspot cycle against winter temperatures in the polar stratosphere during periods of west equatorial winds. As she did so, Labitzke found that the temperatures moved up and down in lock step with the number of sunspots. Atmospheric pressure at the North Pole behaves similarly (see graph).
When she and Dr. van Loon analyzed this relationship further, they discovered that it was the strongest correlation ever found between a terrestrial weather phenomenon and solar variability. ``I've been in this business for 40 years, and I've never seen anything like this before,'' van Loon comments.
Previously, one of the strongest candidates for a solar-cycle effect was the 22-year drought cycle in the Southwest United States that appears to persist in climatic reconstructions from tree rings that extend back to the early 1600s. A prehistoric possibility is a series of distinctive bands in rock outcrops in Adelaide, Australia. These vary regularly with an 11-year cycle.
Other apparent correlations between the solar cycle and terrestrial weather have failed to hold up. For instance, the water level of Lake Victoria in Africa varied with the sunspot number for two solar cycles, but then the apparent connection disappeared.
The problem is that purely statistical connections, as in the Lake Victoria case, cannot be trusted. Without a physical mechanism that can explain a mathematical correlation, scientists are reluctant to accept their validity. In the sunspot-weather case, it has not yet been possible to establish such a mechanism. The energy output of the sun appears to vary by only a tenth of a percent during the solar cycle. And that is not enough to influence Earth's weather directly.
Labitzke and van Loon do not claim to have identified such a mechanism, and the period of their data is relatively short, lasting only for 3.5 solar cycles, rather than the six cycles that would be considered completely convincing. The statistical relationship, however, is exceptionally strong.
The key to such strong correlations is the direction of high-altitude equatorial wind currents. Since the early 1950s, meteorologists have known that these winds switch from easterly to westerly every two to three years, a phenomenon known as the quasi-biennial oscillation, or QBO.
Scientists have found that when the QBO is in its westerly phase and sunspot numbers are high, surface temperatures and pressures are also higher than normal throughout an irregularly shaped area surrounding the North Pole. For instance, their calculations indicate that the average temperature in January and February at Cape Hatteras, N.C., during winters when the sunspot number is high is almost 3 degrees Celsius higher than when the number is low: the difference between a relatively mild and severe winter. And, although the correspondence is not as strong, the opposite holds true when the QBO is easterly.
Labitzke and van Loon calculate the odds at less than 222 to 1 that dividing the records into two series - corresponding to different phases of the QBO - could result in an accidental correlation as strong as what they have found.
More significantly, perhaps, their discovery suggests that the equator is the place to look for a possible mechanism that amplifies small variations in solar flux so that it affects terrestrial weather. They have noted that the tropopause, the boundary between the upper and lower atmosphere, seems to rise and fall over the equator in time with the solar cycle. So they reason that if changes in solar activity can directly affect tropical convection, the signal could spread throughout the atmosphere.