The year 1811 was the 35th year of the United States. Most Americans lived along the east coast, but the newly opened Louisiana Purchase opened up new opportunities. Along the Mississippi, new towns and cities were being established, serviced by barge traffic moving down the Ohio to the Mississippi, and then onward to New Orleans and the sea.
In 1811 a slave revolt rocked the Gulf Coast in January. The steamship Juliana made its first steam-powered voyage between New York City and Hoboken in October. American forces defeated Tecumseh at the Battle of Tippecanoe in November while elsewhere tensions grew between the new nation and its former colonial power. This would lead to war. The White House would eventually burn.
A brilliant comet graced the skies in the autumn.
But right now, it's December of 1811, and we're in the sleepy town of New Madrid, MO.
New Madrid had struggled. It was founded in 1788, when the territory was actually Spanish thus, its name. The territory reverted to France, and then was sold to the United States. Designed to capture all that traffic coming down the Ohio, it never really happened. Floods wrecked the place. Ste. Genevieve, a city to its north, ended up moving to higher, harder ground after a flood in 1785. This would be to its benefit in 1811 and 1812, as the city sustained very little damage despite the violence of the quakes. By 1811, the place was pretty quiet. Much of the territory was sparsely populated. Unbeknownst to them, the ground beneath their feet was about to give way.
Now, the Americans of 1811 were not ignorant of earthquakes. New England was rattled by several large earthquakes in the 17th and 18th centuries (a few exceeding M6 in 1638 and 1755, several M5-pluses, and one centered near Montreal in 1663 exceeding M7), and news fairly quickly traveled back and forth across the Atlantic and from the Spanish colonies in South America, where earthquakes were commonplace. If the city-building cultures of the Southeast and Midwest US had survived, they likely could have told stories of the times the ground shook, their temples sank as foul debris erupted from the ground, rivers ran backwards, and their villages collapsed but much of them were long gone by 1800. A comprehensive report published a century after the earthquakes in 1912 did make note of native legends, however.
Throughout the post I'll be using the term "M" for magnitude. I'm universally, unless otherwise indicated, referring to "moment-magnitude" and not the catchall "Richter Scale". This is also denoted as "Mw". Moment-magnitude is the preferred scale for large earthquakes for a variety of reasons. A comprehensive look at this can be read here.
On the 16th of December, at 2:15 am, strain that had been building up for perhaps half a millennium gave way. We know that this was the southern end of what we now call the New Madrid Seismic Zone, on a fault named the Cottonwood Grove Fault, which probably ruptured along its entire length. The epicenter was likely in northeast Arkansas. Creating an earthquake of approximately magnitude 7.2, the seismic waves rattled the entire eastern United States, causing damage at Louisville and Cincinnati, damaging New Madrid and utterly destroying villages to its southwest. Bells, however, did not ring in Boston---the quake likely wasn't even perceptible there. They did, however, in Charleston, South Carolina.
The quake likely caused a sieche on the Mississippi, which ran backwards for a time. At Little Prairie, liquefaction destroyed the village, and liquefaction up and down the valley caused the earth to spit up sand and other organic material. These features are still present today, and are a major key as to how we have an idea as to nature of this fault system's behavior over the last several thousand years.
It was the middle of the night; it was cold, the air stank of regurgitated organic material from deep within the earth, and the river was nuts, agitated from the quakes and full of water expelled from the earth by the shaking. This would have been terrifying by itself, but large aftershocks would continue through the night. An aftershock at sunrise that day likely had a magnitude similar to the first mainshock. Its epicenter is not as well known, but it may have been in northwest Tennessee, on a segment of what we now call the Reelfoot Fault. This large aftershock, perceptible across much of the Midwest and East, is why some consider this a sequence of 4, and not 3, earthquakes.
This was the beginning of eight weeks of constant shaking. In January, on the 23rd, the next large earthquake occurred. Likely close to an M7, Its epicenter is, like the dawn aftershock of December 16th, not as well known. It likely was to the northeast, perhaps as far north as south-central Illinois. Even more aftershocks rocked the area. The big one occurred on February 7th. This one caused the most disruption, and had the most fantastic effects.
The previous two big quakes (not including the dawn aftershock of December 16th, which probably was reverse-faulting) were likely strike-slip earthquakes. Other then the liquefaction they caused and landslides along riverbanks, they generally left few major changes upon the landscape. The February 7th quake did. Occurring on the Reelfoot Fault which strikes northwest/southeast across Missouri, Kentucky, and Tennessee and crosses the Mississippi in three places, it raised a notable scarp that likely caused the reported appearance of waterfalls on the Mississippi, which then ran backwards in places, much like water does in front of a low-head dam. It damned a creek, creating a major lake. This earthquake's magnitude was perhaps M7.5. New Madrid, which had generally survived the previous quakes somewhat unscathed, was utterly destroyed as it subsided 10 feet. The February quake was just as widely perceptible (perhaps even more so) as the previous three large ones, disturbing much of the east from their sleep. It even cracked masonry as far away as South Carolina, which would itself be rocked 74 years later by its own M7 enigmatic quake. No one knows how many people may have perished during the sequence. The area was sparsely populated at the time, so perhaps less than 10.
One could probably argue that not until August 23, 2011 would an earthquake in the United States be so widely felt by so many although, that's as much function of population density as it is attenuation.
From then on, the sequence slowly died off, although two more significant earthquakes of magnitudes greater than 6 occurred in January 1843 (at Marked Tree, AR) and October 1895 (at Charleston, MO). Both locations are at opposite ends of the seismic zone. The new falls on the Mississippi eroded away. A fun footnote is Congress enacted the very first disaster relief bill for the 1811-1812 earthquakes. The bill, perhaps as a precursor to FEMA troubles in the mid 2000s, was a disaster in itself and for a time, New Madrid became synonymous with fraud. Eventually, the events were somewhat forgotten.
Now you may have noticed I did something. Perhaps you learned at some point that the enigma of the Mississippi Valley was that these quakes were super large. They exceeded magnitude 8, and perhaps approached magnitude 9. They were the biggest ever recorded in North America. They rang church bells in Boston. Absolutely none of this is true, but it adds to the mystique surrounding the sequence. As the events were forgotten, tall tales began to proliferate about them. However through decades of research we have a somewhat good handle on their magnitudes and some of the most recent research suggests some may not even be over M7 and the largest was M7.2. For this post, I chose the lower-middle ground.
The reasoning for the higher magnitudes were many of the observed effects. Very strong shaking was noted as far away as Louisville, and the presence of sandblows and the fault scarps indicated to early researchers that an extremely large series of earthquakes had to have occurred. However, as more things were learned, especially how different soils respond to shaking, magnitudes came down. This was even observed in 1811 and 1812 where one very observant resident of Louisville noted that the river valleys shook more than the hills. The reason the quakes were felt at huge distances has to do with the rigidity of the crust east of the Rockies. The ground simply transmits seismic energy better than in the West, where a comparable quake would not be felt as widely.
Some may say, well, that isn't so bad. An M7 isn't as bad as an M9. But Haiti's quake in January of 2010 was just that, M7, and perhaps 1 out of every 40 Haitians died because of it. An M6.3 last February did so much catastrophic damage to Christchurch, New Zealand that significant portions of the city and suburbs will have to be abandoned. Even August's "measly" M5.8 in central Virginia shut down a nuclear power plant, caused some scattered damage even as far as Baltimore, and could have caused serious destruction if it had been directly beneath Richmond or Washington D.C. An earthquake's magnitude isn't always an indication of mass death. In Japan, let's not forget that the vast majority of those who died in the March quake drowned in the tsunami and not from falling buildings.
What's more interesting is what seems to be an intense debate over how the United States quantifies the hazard in this region. Some say it's high. Others say not. Some say the cost is well-worth it. Others say perhaps, money could be better spent elsewhere, especially in this "age of austerity".
And of course, the ultimate question is "why here?"
Roughly 750 million years ago, the supercontinent Rodinia began to break up, in much the way Africa is today. Faults were created. For whatever reason, the rifting failed although Rodinia did break up. Eventually the continents came back together as Pangaea. This supercontinent also broke up. 200 million years ago, more rifting occurred, and this included lava eruptions, perhaps as the proto-North America moved over a hotspot that's now in the region of Bermuda. The rifting still failed but the weakness remained.
Over time, mass amounts of sediments covered the faults. Then the glaciers came. They didn't come down as far south as the Missouri-Tennessee border, but they added strain to the earth's crust. The removal of the glaciers roughly 10,000 years ago has caused a phenomenon called post-glacial rebound. Basically, this is the crust bouncing back up after being depressed by the weight of the glaciers. GPS measurements indicate that the New Madrid region, however, is not subject to it, as the areas that are still actively rising are north of the St. Lawrence River. Post-glacial rebound is considered one of the possible reasons the earth quakes in New England and the in the very active St. Lawrence River valley. Still, one proposal suggests that the quakes are the result of glacial outwash and erosion---the glaciers melted fairly quickly and the proto-Mississippi became an 80 mile wide raging torrent. This erosion removed many cubic feet of sediment, and the weight of this sediment leaving the region caused the crust below to flex,kind of like post-glacial rebound. The quakes experienced in the region (and there's considerable evidence for several sequences since at least 2350 BCE) are perhaps a result of that, as GPS measurements also seem to indicate the faults aren't moving in a fashion that'd substantially warp the earth's crust, and the fault system appears to have not been active before 64,000 years ago.
Another possibility relates to processes that occurred on the west coast. An oceanic plate subducted almost completely beneath North America---its remains are the Juan de Fuca and Cocos tectonic plates. A piece of it may still be floating about in the mantle, and it's somewhere beneath the Midwest.
Finally, all of North America, whether rising or subsiding, rebounding or not, is being pushed westward, generally. This is also adding to the strain.
That's the simple version and it's hardly complete. Obviously, what's going on beneath the center of North America is much more complicated and these complicated factors have led to an intense debate over the nature of this fault system. Is it a hazard? Or is it overhyped and full of misapplication of scarce funds? One group of scientists say yes to the former, and another says yes to the latter.
I recently finished Disaster Deferred by Dr. Seth Stein. Dr. Stein, who I talked about in my last diary ("How humans cause earthquakes, but likely, not some recent ones"), is a seismologist at Northwestern University. He's also rather controversial, as he is one of a group that believes that New Madrid may be shutting down for good.
That said, the book is highly readable. One part autobiography and one part lay science text (and according to some, one part soapbox), the book is, I think, rather accessible to readers who have a minimum of knowledge of geology. There's very little math (and it's well explained), and I think anyone with the slightest of interests in the area should pick this book up. Parts of it are actually funny.
This book won't fear monger at you. In fact, its main purpose is to do the exact opposite.
Dr. Stein takes a critical look at the validity of the USGS's seismic hazard maps, which currently place the middle Mississippi Valley in an area of significant seismic hazard. He does a great job going through much of the history and is very critical of the USGS's hazard mapping effort and FEMA, in addition to the Central US Earthquake Consortium. He all but accuses them of hyping up the hazard in the region to serve their own selfish interests (and this is where much of the controversy is).
The GPS data that he presents is compelling, but he describes GPS as magic. I suppose this is a small nit, since I don't believe in magic, but as science-fiction writers sometimes state, technology is magic if you don't know how it works. I would have liked, in this section, an explanation on how GPS works. He does give a great description on how GPS is used to track the growth of mountains, and how it's used to measure deformation across faults. Essentially, the GPS data collected from the New Madrid region over the last 20 years indicates that the region is building up no strain for a future earthquake. This means that, to Dr. Stein, there's no longer any threat. The USGS Hazard maps (which actually will be revised in 2014) need to be changed, FEMA and other emergency management groups need to lighten up, and resources should be spent elsewhere. His model for the region is that it's not like the San Andreas. It's more that quakes in this area are episodic and clustered, and they can migrate in time as the faults in the region interact in extremely complex ways. The quakes we still see today along the faults near New Madrid are part of a centuries-long aftershock sequence. This can be observed in other areas of the world, like China, where large intraplate earthquakes do occur, but generally don't repeat along the same faults for thousands of years.
I won't spoil anymore of the book and suggest you read it yourself, but his overall point is that seismic hazard is not the number one problem facing cities like Memphis and St. Louis. If there won't be another sequence of M7+ quakes for 5,000 years, if at all, why spend the money retrofitting buildings as both cities are actually doing. Give it a read. It's recommended highly by me. Do I necessarily agree with all of the conclusions Dr. Stein makes? No. I think there's great information in the book and quite a bit of it should get lots of science juice flowing, but I don't think we know enough just yet to say if New Madrid is shutting down.
States and communities in the Midwest will mark the 200th year since they rattled and rolled. In February, the Great Central US Shakeout will occur, with participants from almost a dozen states. They're preparing for a hazard that may be hundreds of years away. Or it may not be coming at all.
I pointed this out in my prior diary but one hypothesis is that quakes in stable continental regions---New Madrid or Virginia or Massachusetts or Quebec or Pennsylvania or Oklahoma--- are episodic and migrating. So while the New Madrid region may be shutting down, seismicity may have moved elsewhere. Perhaps Oklahoma? I'd argue that perhaps it's too soon to downgrade the region's disaster making potential.
I think a better question is in how we prepare for disaster. I do agree with Dr. Stein's point he makes that perhaps California-style construction is inappropriate in the Midwest. If the quakes haven't shut down, then we have almost two, perhaps three centuries until they may repeat again. Most buildings and infrastructure may long be replaced by then, but as the Japanese learned the hard way, people forget. An entire millennium passed before the Japan Trench ruptured again, for example, and even tsunami caused by earlier quakes were forgotten.
On the other hand, geologists have discovered an almost 5,000 year record of sequences of large earthquakes. The last sequences occurred perhaps around the year 1450, around the year 900, and around the year 300, and perhaps 1100 BCE. This suggests an average of 500 years between quakes, with a great deal of uncertainty.
Quakes occurred elsewhere in the Midwest as well. Southwest Oklahoma was rocked by possibly two >M6 quakes roughly 1,300 years ago along the Meers Fault. There's evidence of large earthquakes along the Illinois-Indiana border in the Wabash River Valley. A fault system was discovered in recent years near Memphis and others near St. Louis. Perhaps the GPS data, which absolutely do show no movement, isn't as compelling as thought
Also, an earthquake of M6, which is still very possible in the area, can still cause catastrophic damage.
I think we'll have to wait and see. Perhaps we'll have to wait and see for 10,000 years. We know a lot more than we did 200 years ago, but still, this area is an enigma wrapped in a mystery.
I consulted a great many sources in putting this together, much of which is publically available on the internet. Google Scholar was a huge resource at getting articles that I didn't have access to. Much thanks to those scientists who keep their own preprints and copies of studies they submit to journals. Studies used to put this post together are (and most can be found on Google Scholar in full):
Stein, S. and M. Liu, 2009, Long aftershock sequences within continents and implications for earthquake hazard assessment
Hough, S.E. and S. Martin, 2002, Magnitude estimates of two large aftershocks of the 16 December 1811 New Madrid earthquake
Tuttle, M.P., H. Al-Shukri and H. Mahdi, 2006, Very large earthquakes centered
southwest of the New Madrid seismic zone 5,000-7,000 years ago
Tuttle, M.P, E.S. Schweig III, J. Campbell, P.M. Thomas, J.D. Sims and R.H. Lafferty
III, 2005, Evidence for New Madrid earthquakes in A.D. 300 and 2350 B.C
Calais, E., A. M. Freed, R. Van Arsdale, and S. Stein, 2010, Triggering of New Madrid
seismicity by late-Pleistocene erosion
Independent Expert Panel on New Madrid Seismic Zone Earthquake Hazards which can be read here.
The USGS's Quaternary Fault database
On the fictional side, Walter J. Williams penned a novel titled The Rift. Depicting a repeat of the 1811-1812 quakes (although written prior to recent research, it uses the higher magnitudes), it's a rather good disaster novel. It even has a nuclear power plant in peril (there are 14 or 15 in the Mississippi Valley) and a concentration camp created by white supremacists taking advantage of the breakdown in civil authority. Another author (who I've not read) has written a series titled the 7.9 Scenario. I personally think both scenarios are unlikely within my lifetime, and probably those of any in the next several generations.
Of course, a great deal of information comes from Disaster Deferred. In addition, information about New Madrid comes from After the Earth Quakes: Elastic Rebound on an Urban Planet by Susan Hough and Roger Bilham. Drs Stein, Hough, and Bilham all maintain excellent webpages:
Studies on GPS and continental deformation can be read at Eric Calais's excellent website and at Andy Newman's website. Another somewhat random site I found is here. Despite its disorganization there's a metric button of information, and the sites owner appears to keep up with all of the latest research.
NOVA created a program (one vastly better than the ones produced by The Weather Channel and the "History" Channel, and those I won't link to) on the subject.
Lastly, New Madrid 2011 is a site set up to commemorate the 200th anniversary of when the earth opened up and spent eight terrifying weeks shaking, rattling, and rolling their town.
And I now leave the floor to you.