Do old glaciers cause new earthquakes in New Madrid, Missouri?

By Mark Shwartz

The ghost of past glaciers may still rattle the American Midwest. During the last ice age 20,000 years ago, a gigantic ice sheet invaded North America, weighing down the hard upper crust of the continent for millions of years. Eventually, the glaciers melted. Freed from the heavy pressure of the ice sheet, North America slowly rose. This glacial rebound continues even today and triggers quakes in the New Madrid fault zone in Missouri, Kentucky, Arkansas and Tennessee, says Stanford geophysicist Mark Zoback.

"It's the only hypothesis we have that I think satisfies all the diverse observations," Zoback says, "especially that big earthquakes in this region appear to have been turned on during the past few thousand years."

His theory, which appeared in the February issue of the journal Geology, explains earthquakes that occur in the middle of the North American continent, far from the usual zones of high earthquake activity along tectonic plate boundaries, such as the San Andreas fault. Zoback's theory is also intriguing because of its possible application in explaining similar earthquakes along faults in plate interiors, including the Jan. 26 earthquake in Gujarat, India, that claimed more than 20,000 lives and left up to one million homeless.

The New Madrid fault is famous among seismologists. In 1811 and 1812, three of the largest earthquakes ever reported in the United States struck the fault and changed the course of the Mississippi River. Large aftershocks continued for many years. The tremors lasted so long and were so large that many of the frontier folk felt it was the end of the world. From 1811 to 1812, membership in the Methodist church increased by 50 percent in the earthquake zone, compared to 1 percent for the rest of the nation.

The cause of Midwestern earthquakes has long perplexed geologists. An answer may help scientists better predict whether earthquakes are likely to recur in the near future.

Most earthquakes occur at the edges of rigid crustal plates that float atop the Earth's more fluid interior, like crackers crowded on top of thick soup. These plates cause earthquakes when they slide under, over, into or past one another. For example, most California earthquakes happen as the Pacific plate slides underneath the North American plate along the San Andreas fault system.

New Madrid, however, has evaded explanation because it is right in the middle of the North American continental plate, far from the edges where earthquakes usually occur.

"Why earthquakes occur in the middle of plates and where they are likely to occur in the future is a fundamental outstanding problem," Zoback says. The New Madrid area sits squarely on top of what geologists call a "failed rift," where hot, molten rock underneath the Earth's crust once threatened to rise and push apart the North American continent. For some reason, the upwelling stopped and never successfully split the plate. But the process left its signature, Zoback says, creating a unique tension in the New Madrid portion of the North American plate.

Zoback points out that the magnitude 7.7 earthquake that devastated Gujarat, India, in January also occurred in a mid-plate fault right on top of a failed rift. Earthquakes that happen in these areas stymie geologists, for just the existence of a particularly stressed area doesn't cause an earthquake. Something has to trigger it.

Zoback and his former graduate student Balz Grollimund created a computer model showing that the removal of a heavy North American glacial sheet triggered earthquakes not underneath the glacier, but in the stressed New Madrid fault past the glacier's edge.

Twenty thousand years ago, the glacial ice sheet crept as far south as the middle of Illinois. The glacier did not extend all the way down to the New Madrid region of southeastern Missouri, but it was "big enough and thick enough" to strain the Earth several hundreds miles to the south, says Zoback. Essentially, at the edge of the glacier, the Earth bent, like a mattress will slope under a body's weight.

When the climate grew warmer and the ice melted, the ground rose from the released pressure and began to straighten out. It is still rising, Zoback says, for the soft, hot mantle underneath the crust "is this very viscous goo. When the load is removed, it takes a long time to respond."

The rising of the Earth following glaciation is a well-documented process. Crust is still lifting in Norway, Sweden and Canada's Hudson Bay. Zoback is now studying India to see if ancient glaciers may have triggered recent earthquakes in northern India as they did in New Madrid. To do so, he must determine the weight, thickness and area covered by the ice age glaciers. If the glaciers came close enough to the fault, as they did in New Madrid, they may have concentrated sufficient stress to give rise to earthquakes. Determining the causes of the quake in India likely will take many years.

If Zoback is right about New Madrid, he predicts that frequent big earthquakes will continue in New Madrid for the next 10,000 years, long enough for the Earth to bounce back fully from the glaciers. Right now, earthquakes tend to fall every 200 to 900 years. "It's been 200 already," Zoback notes.

In 1811, the edge of the Mississippi was still frontier land. Shocks from the New Madrid quake toppled trees and overturned flatboats more than anything else. But a similar earthquake today would have serious consequences. "There are nuclear power plants all over the central and eastern United States," Zoback says. And because the ground under New Madrid is old, cold and rigid, it spreads earthquakes well. In 1811, the New Madrid quake was felt in 27 states. Another big one in the region may affect metropolises relatively far from the fault.

Zoback hopes his theory about the lingering influence of past glaciers will give scientists a better tool for making predictions about the New Madrid fault. "Time will tell," he says, if his theory stands. But with a restless ghost in the ground, he may not have long to wait.