New Madrid Quakes May Not Have Been So Colossal

In the winter of 1811-12, four major earthquakes shook what is today southeastern Missouri. The first of the temblors flattened the small town of New Madrid and, according to some eyewitness accounts, temporarily reversed the course of the Mississippi River. Other reports over the years have claimed that the quakes could be felt as far away as the U.S. East Coast, where they rattled furniture in Washington, D.C., and rang church bells in Charleston, South Carolina. Together, the New Madrid quakes have become known as one of the most powerful natural events in U.S. history.

But were they really so powerful? A new examination of the historical and geological evidence cuts back the scientific assessment of the quakes significantly, and it firmly debunks the notion that the temblors were in the same league as recent massive quakes, such as the 1906 San Francisco quake.

That examination, by seismologist Susan Hough of the U.S. Geological Survey (USGS) in Pasadena, California, began somewhat accidentally. In 1999, she says, she had been reading some of the historical accounts of the New Madrid quakes and was struck by one account that clearly described what earthquake scientists call site amplification. That's when areas such as river valleys like the Mississippi, which are underlain by loose soil, shake much more violently than areas grounded on bedrock. When that happens, a quake of, say, magnitude 7 might seem like an 8 or higher.

Hough then checked more of the accounts and post-event studies and found that not one of them had taken site amplification into account. Eventually, she examined all available sources of information about the New Madrid quakes. She gathered photographs of the landscape that showed how the temblors had distorted the ground. She also read the eyewitness and newspaper accounts and compared those accounts with reports of more recent quakes whose intensity had been calibrated by instruments. Then she shared what she had gathered with four other earthquake experts. Based on her own assessments of the evidence and those by the other scientists, Hough estimated magnitudes of 6.8 to 7.0 for the largest of the New Madrid quakes. The previous consensus—including Hough's own research in the 1990s—had placed the events at about 7.2. The difference in magnitude is not insignificant: A 6.8 quake is only one-fourth as powerful as a 7.2. Hough reported her findings last week at the Seismological Society of America's annual meeting in Portland, Oregon.

Still, the downgrade doesn't change the fact that the U.S. midcontinent is quite seismically active. For example, USGS has recorded about 4000 minor quakes in the area since 1974.

The lingering question is the source of the geological stress in the Midwest. It's probably not plate tectonics, says Hough, but rather the continuing rebound of Earth's crust from the weight of the last glacial period, which ended about 11,000 years ago. The stress predicted from this process can produce quakes up to magnitude 7.0 about every 500 to 1000 years, she says, enough to inflict damage on the nearby cities of St. Louis, Missouri, and Memphis, Tennessee, but nothing on the order of the destruction that has come to be popularly associated with the New Madrid events.

Hough's findings about the quake magnitudes "make a lot of sense," says geophysicist Seth Stein of Northwestern University in Evanston, Illinois, and "they still fit the historical eyewitness accounts." But he's skeptical that glacial rebound is causing the New Madrid activity. "We've measured the rebound all over the central U.S. with high-precision GPS," Stein says, "and it's very small in the New Madrid area." He says the rebound effect is much bigger where the ice sheets actually had overlain the continent, "so it's hard to see why it should cause earthquakes so far to the south."

Seismologist Chris Cramer of the University of Memphis in Tennessee, says Hough's conclusions will need to be tested. Six significant events have occurred in the New Madrid seismic area in 200 years, including magnitude-6 quakes in 1843 and 1895. Plate tectonics may not be involved, he says, but "there's something else going on there that we don't understand."