For the last 900,000 years, mile-thick ice sheets have waxed and waned in the Northern Hemisphere with remarkable regularity – building over periods of about 100,000 years and retreating in the space of only a few thousand years, only to repeat the cycle.
Now, a team of scientists from Japan, the US, and Switzerland suggests that the North American continent is the breeding ground for these cycles. It's a region where climate and the ice's effect on the Earth's crust play off each other to draw out the length of a glacial cycle triggered by changes in solar radiation that come with changes in Earth's orbit.
This feedback between climate and ice becomes most dramatic at the end of the cycle, when an ice sheet that has bulldozed its way too far south and gotten too heavy for its own good meets up with a warming climate.
"When the ice sheets get to this huge state, they are very hard to keep that way," says Maureen Raymo, a paleoclimatologist at Columbia University's Lamont-Doherty Geophysical Observatory in Palisades, N.Y., and a member of the team conducting the study.
Once a change in Earth's orbital characteristics brings on the next warming event, "the whole system just goes into catastrophic collapse,” she explains. “It melts back a little, seas start to flow into the depressed crust, this floats the ice and melts it from below.”
In a tiny fraction of the time it takes to build continental-scale ice sheets, the sheets retreat to high-latitude havens atop Greenland and the northern reaches of the Canadian archipelago.
The study, led by Ayako Abe-Ouchi, a climate scientist at the University of Tokyo and the National Institute of Polar Research, resulted from a unique approach to modeling ice ages.
The study provides "good insights that clearly advance our understanding" of ice ages, notes Penn State University glaciologist Richard Alley, who was not involved in the study, in an e-mail. It also confirms aspects of ice ages that researchers have well in hand, he says.
The approach linked individual atmosphere, crust, and ice models in a way that needed only information on the amount of sunlight reaching Earth to generate ice-sheet behavior over the past 400,000 years that geologists have gleaned from more than a century of field studies.
Changes in the amount of solar radiation striking Earth come with changes in Earth's orbit occurring at intervals of 41,000, 23,000, and 19,000 years.
The study reaffirms that changes in the amount of summer sunlight striking northern high latitudes sets the process in motion. Indeed, changes to the shape of Earth's orbit over time, as well as long-term changes in the orientation of its axis, and their impact on solar radiation at high northern latitudes were the most significant astronomical influences in the team's simulations.
The team also weighed the relative contributions of changes in atmospheric concentrations of carbon dioxide, a greenhouse gas, to the 100,000-year glacial cycles.
While carbon dioxide decreased as the ice sheet expanded and cooled the climate and increased again as the climate warmed, CO2 levels did not determine the overall sequence of events during each 100,000-year cycle, the researchers concluded.
The simulations also showed how the interplay between orbital and climate influences kept North America's ice sheets growing, even as a significant but smaller ice sheet building in northern Eurasia waxed and waned based on orbital changes happening at 40,000- and 20,000-year intervals.
Looking closely at the last glacial cycle, which ended about 10,000 years ago, the team estimates that by the time the North American ice sheet reached what is now Kansas, its southernmost point, it held enough water to raise sea levels by 90 meters (295 feet). But the sheet also depressed the crust.
The team estimates that a 3,000 meter thick sheet would have depressed the crust by about 1,000 meters. This would have lowered the altitude of the top of the sheet, bringing it into contact with warmer air. In addition, as the sheet melted, a slowly rebounding crust would gather melt water along the sheet's edges, providing the lifting and melting from underneath that would accelerate its retreat.
This aspect of the work alone was worth the price of admission, suggests Shawn Marshall, a climate scientist at the University of Calgary in Alberta, Canada. He notes that previous modeling studies couldn't deliver the onset of ice sheets growth and advance without conditions colder and snowier than climate models predict for an Earth in the cold phase of its long-term orbital variations.
Getting out of an ice age was even tougher, he writes in an outsider's analysis of the study, set for publication in Thursday's issue of the journal Nature.
This is the first free-running model that appears to get the ice ages right without a lot of tweaking along the way to include factors like ocean circulation or the influence of dust on ice melting in order for the onset or retreat of ice sheets to happen at the right times, he notes.
The work still has room for improvement, he continues. The model treats some processes a bit too simplistically.
Still, the researchers "make a convincing case" that North America's shape and location on the globe, as well as the slow recovery of the crust as the ice begins to melt turn variations in solar radiation occurring in cycles measured in a few tens of thousands of years into a 100,000-year glacial cycle.
What makes North America so special?
"Maybe that's because the mountains deflect the polar jet stream farther south," says Dr. Raymo, referring to a high-altitude river of air that forms the boundary between cold polar air and warmer air to the south. Deep southward meanders in that river during modern winters can bring snow to regions where it's rare, such as the US Southeast.
Or perhaps during the last glacial maximum the ice sheet covering Scandinavia was vulnerable to warm Atlantic Ocean water, she adds.
"We don't know, in fact that's the next work we're doing – experiments to try to investigate that," she says.