New Takes on Historic Quakes

Two centuries on, scientists revisit the magnitudes of New Madrid’s biggest rumbles

In the autumn of 1811, the United States was barely 35 years old. The fledgling nation included only 17 states, all east of the Mississippi River, but it boasted a lot of new territory thanks to the Louisiana Purchase of 1803. Neither the buyers nor the sellers knew that the recent addition’s basement contained a seismic time bomb nearly ready to go off.

| Damage caused by rumbles near New Madrid in 1811–1812 was captured in photos a century later (images shown). quake points: USGA, background: Geoatlas/graph-ogre, Adapted by Janel Kiley, pictures: USGS photographic library
SANDY EXPLOSION Pressure from seismic shaking can cause sand buried beneath layers of silt and clay to liquefy and exude out at the surface (diagram at left). Many such sandblows are still visible in the New Madrid region (aerial shot of a blow near Blytheville, Ark., at right). from left: USGS (modified from Sims and Garvin, 1995), adapted by Janel Kiley; photo: Tish Tuttle/M. Tuttle & Associates
HAZARD ZONES Many people think just of the West Coast when discussing major quakes in the United States, but the New Madrid region and coastal South Carolina face similar threats. USGS

At around 2:15 a.m. on December 16, 1811, a series of massive earthquake pummeled what is now southeastern Missouri and northeastern Arkansas with ground motions so strong that trees snapped in two as they whipped back and forth. The landscape rose several meters in some areas and sank in others, changing the courses of creeks and waterways. During one of the quakes, even the mighty Mississippi was diverted; portions near the quake’s presumed epicenter flowed backward for at least several hours, and possibly a day or more. People felt the temblor as far away as New York state, and seismic vibes from an aftershock that struck at dawn traveled almost as far, reaching residents in Georgia and what would soon become the state of Louisiana.

Another quake of a similar size — maybe an aftershock, or maybe a separate quake along a different portion of the same fault zone — rumbled on January 23, 1812. The final major shaking in the series came about two weeks later, on February 7, when spreading seismic waves flung books from their shelves in Charleston, S.C., and rattled cups and saucers in Washington, D.C.

Scientists long considered these quakes along the New Madrid Seismic Zone — a zigzag-shaped set of faults named for the small town in Missouri near where the quakes were felt most powerfully — to be some of the strongest ever on the North American continent. After all, some scientists estimate that the area violently shaken by the three most energetic quakes was two to three times as large as that experiencing comparable ground motions from the magnitude 9.2 quake that slammed southern Alaska in 1964, and about 10 times the area similarly affected by the magnitude 7.8 quake that wrecked San Francisco in 1906.

As recently as the late 1970s, scientists estimated that the strongest of the New Madrid quakes may have been magnitude 8.75. But in the last decade or so, researchers have proposed that the New Madrid quakes were smaller, possibly much smaller. Debate about the size of these quakes rages in journal papers and in conference presentations, with some of the most recent arguments made in April at the annual meeting of the Seismological Society of America. This year’s get-together was held in Memphis to commemorate the bicentennial of the New Madrid quakes, which occurred about 160 kilometers north along the Mississippi.

“These quakes were felt over impressive distances,” says Susan Hough, a seismologist with the U.S. Geological Survey in Pasadena, Calif., and a proponent of substantially downgraded New Madrid strengths. Until recently, she notes, suggesting that the quakes had magnitudes around 7 or so was “a bit of heresy.”

To better estimate the quakes’ actual sizes, some scientists are tweaking models of how the Earth’s crust behaves beneath the sediment-smothered region of the central United States, as well as analyzing how surface sediments respond to strong shaking. Other researchers are digging into archives to uncover even more records of damage associated with the quakes.

Though the findings are primarily of historical interest, understanding the energies involved and their impacts could lead to building codes that better protect structures — and human lives — throughout the Midwest.

Rift of opinion

The New Madrid Seismic Zone is one of the most hazardous in the lower 48 states, a surprise to those who think only California faces risk of sizable earthquakes. California and the rest of the West Coast sit along the boundary between the North American tectonic plate and two other plates that jostle and scrape past its edge, a process creating seismic stress where hunks of crust lock together. But the Midwest lies far from any such plate boundary. While the causes behind many “intraplate” quakes remain a mystery, the New Madrid Seismic Zone is right over a weak spot in the underlying crust — an 80-kilometer-wide, 300-kilometer-long region where, around 600 million years ago, tectonic forces began but ultimately failed to rip the North American plate apart (SN: 5/29/93, p. 342).

Because sophisticated seismo­meters that can locate an earthquake and estimate its magnitude weren’t developed until the late 1800s, scientists have always been uncertain about the strengths of the New Madrid quakes. The best estimates, like those of many ancient temblors, have been made with the help of what’s called the Modified Mercalli Intensity Scale, which categorizes quakes based on their effects into one of 12 classes: Class I quakes are generally not felt except in very favorable conditions; Class VI events break windows and move heavy furniture; and Class XII shocks trigger landslides, toss objects into the air and destroy buildings and bridges. By mapping such effects across a wide area, researchers typically end up with a bull’s-eye letting them narrow down the epicenter of a quake and estimate its magnitude.

Several problems afflict earlier, 8-plus magnitude estimates, though. For one thing, Hough noted in an editorial in the March/April Seismological Research Letters, the New Madrid area was sparsely populated at the time, and relatively few residents chronicled their experiences in journals, letters or other accounts that have survived.

Some accounts weren’t even considered reliable at the time: In 1814, Samuel Mitchill, a U.S. politician of the era who studied the statements of people living in hard-hit areas, reported to the Literary and Philosophical Society of New York that “much exaggeration was interwoven with some of the narratives. Some, indeed, were tinctured with fable and burlesque.”

Filling the blanks

Kent Moran, a historian at the University of Memphis’ Center for Earthquake Research and Information, has helped compile more than 600 accounts of the New Madrid quakes. Nearly two centuries after the events, fresh finds are still surfacing, he said at the seismology meeting. While the most reliable accounts come from official files kept by insurance companies and from documents generated by city and county governments, many of the newly found passages come from the archives of small newspapers in the towns that peppered the region. Moran has most recently uncovered newspaper reports from locales as far-flung as Pensacola, Fla. (not previously known to have been shaken by the quakes), as well as Charleston, S.C., and Louisiana. Still, he noted, “there are huge dead zones in the ‘felt reporting,’ ” most notably in areas to the west of the quakes, which were nearly uninhabited at the time.

Yet more reports from blank spots on the map, even if they are accurate, won’t solve some of the problems with existing data, Hough suggests in Seismological Research Letters. For instance, the damage caused by a New Madrid quake wasn’t always correlated with distance from its epicenter, making estimating magnitude quite tricky. In St. Louis, about 300 kilometers north of one quake’s presumed epicenter, ground motions split open a few stone homes and toppled a few chimneys. But in Ste. Geneviève, Mo., about 75 kilometers closer to the source of shaking, the quake was felt but reportedly caused no damage, Hough notes.

The key difference in damage suffered, she suggests, was the type of terrain underlying the cities: While St. Louis was built on floodplain sediments, the town of Ste. Geneviève had been moved back from the Mississippi River — presumably to higher, more stable ground — after a flood inundated the town’s former home in the late 1700s. Similarly, homes and structures along the Ohio River, as far as 800 kilometers from the quakes’ epicenters, suffered damage one or two classes higher on the Modified Mercalli scale than those located away from the river.

By analyzing a sample of reported earthquake intensities and then taking into account the types of terrain from which those accounts came, Hough suggested at the meeting that the large New Madrid quakes had magnitudes of around 7.0. In a detailed analysis reported online March 25 in the Journal of Geophysical Research–Solid Earth, Hough and USGS colleague Morgan Page pegged the main shock and aftershock of December 16, 1811, at about magnitudes 6.8 and 6.6; the January 23 temblor came in at magnitude 6.9, and the February 7 shock was the strongest at around magnitude 7.2.

In terms of energy released by the quakes, that’s a substantial downgrade; even the largest of the quakes would have released less than half a percent of the energy of a magnitude 8.75 quake.

The scope and extent of damage spawned by the New Madrid quakes indicates that the amplification of ground motions in loose sediments can turn even a moderate quake into a major menace. “The New Madrid quakes didn’t have to be large to be a threat to the mid-continent,” Hough says.

But separate work comparing the effects of the New Madrid quakes with those of two other large intraplate quakes suggests that the New Madrid quakes were substantially stronger than Hough’s estimate. Chris Cramer, a seismologist at the Center for Earthquake Research and Information, and colleague Oliver Boyd with the USGS in Memphis, considered the effects of two other intraplate quakes, the magnitude 7.2 Grand Banks quake that struck beneath the Atlantic Ocean south of Newfoundland in November 1929 and the magnitude 7.6 that devastated Bhuj, India, in January 2001.

Cramer and Boyd focused on damage and other effects at sites far from the quakes’ epicenters. At great distances from the source, the magnitude is most strongly correlated with damage. Other factors, such as the depth of the quake and how the fault ruptured, play a minimal role, Cramer reported at the meeting.

Cramer and Boyd estimate that the first of the New Madrid quakes was around magnitude 7.6, and its aftershock was somewhat less than magnitude 7.2. The January 23 quake fell somewhere between magnitude 7.2 and 7.6, and the last quake in the series, on February 7, was the strongest and exceeded magnitude 7.6. These estimates are generally about a half-magnitude higher than those figured by Hough and Page, meaning the quakes would have carried more than five times the energy.

Thar she blows

Others are gaining clues to the New Madrid quakes’ magnitudes by looking at the geological scars left behind. One of the most dangerous effects of a major quake is liquefaction, a process in which the pressures created by seismic waves temporarily turn moist, poorly drained sediments — especially those bearing great weights, such as from buildings — into something akin to quicksand. Liquefaction played a large role in devastating San Francisco’s Marina District, largely built on landfill used to reclaim former wetlands, during the 1989 Loma Prieta quake. More recently, during the megaquake that slammed Japan in March, ground motions that lasted several minutes triggered liquefaction that damaged homes, buildings and infrastructure such as roads, port facilities and buried gas and water lines.

During the New Madrid quakes, of course, there was little large-scale infrastructure to be damaged. But soils in the area are largely made of thick layers of dense, river-deposited silt and mud interleaved with layers of water-saturated sand, conditions that primed the ground to liquefy. When the seismic shaking commenced, the weight of overlying mud pressurized aquifers, causing massive geysers of sandy water to spew onto the surface.

Deposits left by these “sandblows” or “sand boils” were often huge, says Thomas Holzer of the USGS in Menlo Park, Calif. While some measured a few meters across, one sand boil covered at least 136 acres, he noted at the April meeting. And because the sand is typically so different from the silty soil in color and texture, the features can easily be discerned at ground level and in satellite images, despite decades of plowing and other agricultural activity on the fertile floodplains.

Previous studies suggest that about 11,000 square kilometers surrounding the New Madrid Seismic Zone were at least partially covered with sand boils. In large parts of eastern Arkansas and the Missouri bootheel, more than 25 percent of the ground is covered with sand. By studying the extent of liquefaction that had occurred at more than 250 sand boils in the region, Holzer and his colleagues estimated that magnitudes of the New Madrid quakes were at least in the mid-7s. “Magnitudes below 7 just can’t generate the extent of liquefaction that we see,” he said at the meeting.

The widespread signs of liquefaction suggest that underlying sediments are prone to strongly amplify ground motions, boosting the threat to any structure that might be built on the agricultural land in the future.

Future shock

In 1815, the U.S. Congress passed its first disaster relief act, to aid victims of the New Madrid quakes, appropriating $50,000 — which, in today’s dollars, amounts to a little less than $600,000. Damages from a modern-day New Madrid quake would dwarf that figure.

A magnitude 6.4 to 6.9 quake at the southern end of the New Madrid Seismic Zone, near Memphis, could cause damage and economic losses to private property and businesses of between $80 billion and $130 billion, Mary Lou Zoback, a geologist with Risk Management Solutions in Newark, Calif., reported at the April meeting. A magnitude 7.7 located on the worst possible spot in the New Madrid Seismic Zone could trigger losses exceeding $250 billion, she said.

Although the earthquakes of 1811 and 1812 are the largest known along the New Madrid Seismic Zone, plenty of smaller shocks have slammed the region too. In January 1843, a quake centered near New Madrid cracked chimneys and walls in Memphis, and was felt across an area exceeding 1 million square kilometers. The largest quake in the area since 1812, an October 1895 shock with an estimated magnitude of 6.6, inflicted damage in towns from Cairo, Ill., to Memphis, and was felt in 23 states and parts of southern Canada.

USGS scientists have estimated that the chances of a magnitude 6 or larger quake occurring along the New Madrid Seismic Zone in any 50-year interval are between 28 and 46 percent. There’s a roughly 5 percent chance of having a magnitude 7 or larger quake in that same interval, the scientists say. Other teams looking at the geologic record of the region, including trenches cut through sediments in areas affected by the quakes of 1811 and 1812, figure the average time between New Madrid–sized series of quakes is around 500 years.

Though no one knows exactly when a big quake will shake the region, seismometers are prepped to pin down a magnitude. And if studies into the New Madrid quakes succeed in providing clues, people will know what kind of damage to expect.

Recollections of rumblings: Records from news­papers and journals provide some of the best evidence for the damage caused by the 1811–1812 New Madrid quakes. Below are eyewitness accounts from a University of Memphis collection.

Eliza Bryan, New Madrid, 1816: “On the 16th of December, 1811, about two o’clock, A.M., we were visited by a violent shock of an earthquake, accompanied by a very awful noise resembling loud but distant thunder, but more hoarse and vibrating, which was followed in a few minutes by the complete saturation of the atmosphere, with sulphurous vapor, causing total darkness.”

Daniel Bedinger, Mississippi River, 1812: “Many acres of land in a body (as was discovered on the approach of day) had sunk to a level with the surface of the river, and some much lower leaving only the tops of the trees above water. Where the banks did not immediately tumble in, vast rents or fissures were made in the earth to an extent unknown. Some of these fissures received the waters of the river and other let those of the neighboring lakes and ponds, with no inconsiderable roarings.”

William Leigh Pierce, Mississippi River, 1812: “So complete and general had been the convulsion, that a tremulous motion was communicated to the very leaves on the surface of the earth. A few yards from the spot where we lay, the body of a large oak was snapped in two, and the falling part precipitated to the margin of the river; the trees of the forest shook like rushes: the alarming clattering of their branches, may be compared to the effect which was produced by a severe wind passing through a large cane brake.”

Sid Perkins is a freelance writer based in Crossville, Tenn.