Old idea may solve conundrum.

If the past is truly the key to the future, then researchers interested in forecasting the expected greenhouse warning must understand how carbon dioxide gas participated in the geologically recent ice ages.

Climate experts have known for more than a decade that atmospheric concentrations of this heat-trapping gas fluctuated wildly from warm periods to cold ones. But scientists still lack a satisfactory theory to explain how levels of carbon dioxide could vary so drastically as the ice sheets waxed and waned. Two researchers are now dusting off a previously discarded idea that they say may solve this problem.

In 1980, bubbles of ancient air recovered from glacial ice taught scientists that the atmospheric concentration of carbon dioxide sank to 200 parts per million during the last ice age and then rose to roughly 280 parts per million after the ice age. The swings in the greenhouse gas are important because they help keep Earth in a deep freeze during glacial periods and help warm the climate between ice ages.

Bradley N. Opdyke and James C. G. Walker of the University of Michigan in Ann Arbor are now reviving a 10-year-old theory to explain the variations in carbon dioxide concentrations. This scenario focuses on the interplay between changing sea levels and the growth of calcium carbonate coral reefs on the continental shelves.

During glacial periods, the expansion of massive ice sheets on the continents pulled water from the ocean, lowering the global sea level. This process exposed carbonate reefs, causing them to erode through a chemical process that pulls carbon dioxide from the atmosphere and deposits it in the deep sea, according to the theory.

When the ice sheets melted at the end of the glacial period, sea levels rose and coral reefs regrew on the continental shelves. The growth of reefs released carbon dioxide into the sea water. From there, the gas escaped into the air.

Wolfgang H. Berger of the Scripps Institution of Oceanography in La Jolla, Calif., originally proposed the reef hypothesis in 1982. Yet he and other scientists thought this process could account for only a small fraction of the carbon dioxide changes between glacial and interglacial times.

In the August GEOLOGY, Opdyke and Walker suggest that reefs may play a much greater role than anyone realized. "What we're arguing is that this may be the dominant, first-order effect in terms of the historic carbon dioxide changes that we see over the last 100,000 years," says Opdyke.

Previous analyses missed the importance of the reef hypothesis because they underestimated the amount of carbonate currently being deposited on the continental shelves by coral, says Opdyke. He and Walker present evidence that reef growth is depositing double what others had assumed.

Plugging the new numbers into a simple model of the global carbon cycle, they found that reefs could raise and lower atmospheric concentrations of carbon dioxide by almost 80 parts per million - roughly the same spread as that seen in the air bubble record.

Berger says the reef hypothesis will probably not explain the full carbon dioxide changes. "There is a tendency to look for a silver bullet," he says. "We would all like to find one factor that does more than two-thirds of the job so that we don't have to worry about five factors.... My feeling is that it's not going to turn out to be one mechanism that explains it."

Other factors proposed to explain the carbon dioxide swings are increased growth of photosynthetic ocean plants and a reorganization of ocean currents.