Skip to main content
U.S. flag

An official website of the United States government

The New Madrid Seismic Zone

  

Science of the New Madrid Seismic Zone

Overview

New Madrid Seismic Zone
New Madrid Seismic Zone - Quaternary Fault Localities. Earthquakes with magnitudes equal to or larger than 2.5 are shown by the yellow dots. (Public domain.)

When people think of earthquakes in the United States, they tend to think of the west coast. But earthquakes also happen in the eastern and central U.S. Until 2014, when the dramatic increase in earthquake rates gave Oklahoma the number one ranking in the conterminous U.S., the most seismically active area east of the Rocky Mountains was in the Mississippi Valley area known as the New Madrid seismic zone. Since 1974, seismometers, instruments that measure ground shaking, have recorded thousands of small to moderate earthquakes. The faults that produce earthquakes are not easy to see at the surface in the New Madrid region because they are eroded by river processes and deeply buried by river sediment. A map of earthquakes epicenters, however, reflects faulting at depth and shows that the earthquakes define several branches of the New Madrid seismic zone in northeastern Arkansas, southwestern Kentucky, southeastern Missouri, and northwestern Tennessee. Other relatively young faults, which are not necessarily associated with recent earthquakes, or the main seismicity trend in the New Madrid region, are shown in this map. It shows 20 localities where geologists have found and published their findings on faults or evidence of large earthquakes (from sand blows; see image to the right).

1811-1812 Earthquakes

In the winter of 1811 and 1812, the New Madrid seismic zone generated a sequence of earthquakes that lasted for several months and included three very large earthquakes estimated to be between magnitude 7 and 8. The three largest 1811-1812 earthquakes destroyed several settlements along the Mississippi River, caused minor structural damage as far away as Cincinnati, Ohio, and St. Louis, Missouri, and were felt as far away as Hartford, Connecticut, Charleston, South Carolina, and New Orleans, Louisiana. In the New Madrid region, the earthquakes dramatically affected the landscape. They caused bank failures along the Mississippi River, landslides along Chickasaw Bluffs in Kentucky and Tennessee, and uplift and subsidence of large tracts of land in the Mississippi River floodplain. One such uplift related to faulting near New Madrid, Missouri, temporarily forced the Mississippi River to flow backwards. In addition, the earthquakes liquefied subsurface sediment over a large area and at great distances resulting in ground fissuring and violent venting of water and sediment. One account of this phenomena stated that the Pemiscot Bayou "blew up for a distance of nearly fifty miles."

After the earthquake [of 1811-1812] moderated in violence, the country exhibited a melancholy aspect of chasms, of sand covering the earth, of trees thrown down, or lying at an angle of forty-five degrees, or split in the middle. The Little Prarie settlement was broken up. The Great Prarie settlement, one of the most flourishing before on the west bank of the Mississippi, was much diminished. New Madrid dwindled to insignificance and decay; the people trembling in their miserable hovels at the distant and melancholy rumbling of the approaching shocks.

See also: Detailed Summary of the 1811-1812 New Madrid Earthquake Sequence

Geology

Woodcut by Henry Howe
Woodcut by Henry Howe, from Historical Collections of the Great West (Cincinnati, 1854, p.239). (Public domain.)

The New Madrid seismic zone is located in the northern part of what has been called the Mississippi embayment. The Mississippi embayment is a broad trough filled with marine sedimentary rocks about 50-100 millions years old and river sediments less than 5 millions years old. The upper 30 meters of sediment within the embayment includes sand, silt, and clay deposited by the Mississippi, Ohio, St. Francis, and White Rivers and their tributaries over the past 60,000 years. Wisconsin valley train deposits formed during the glacial period from 10,000-60,000 years ago, and the Holocene meander belt deposits were laid down during the past 10,000 years.

The Mississippi embayment is underlain by Paleozoic sedimentary rocks up to 570 millions years old. The Paleozoic rocks are underlain by even older rocks that appear to have been deformed about 600 million years ago when the North American continent almost broke apart. During the process of continental rifting, a deep valley formed that is bounded by faults and known as the Reelfoot rift. The Reelfoot rift is identified today as a subsurface system of fractures and faults in the earth's crust. New Madrid seismicity is spatially associated with the Reelfoot rift and may be produced by movement on old faults in response to compressive stress related to plate motions.

Geologic and seismotectonic model of the New Madrid region
Geologic and seismotectonic model of the New Madrid region (modified from Braile et. al., 1984).(Public domain.)

Liquefaction

The most obvious effects of the 1811-1812 earthquakes are the large sandy deposits, known as sand blows, resulting from eruption of water and sand to the ground surface. This phenomenon called earthquake-induced liquefaction is the process by which water-saturated, sandy sediment temporarily loses its strength due to the buildup of water pressure in the pores between sand grains as seismic waves pass through the sediment. If the pore-water pressure increases to the point that it equals the weight of the overlying soil, the sediment liquefies and behaves as a fluid. The resulting slurry of water and sediment tends to flow towards the ground surface along cracks and other weaknesses. Overlying soil "floating" on liquefied sediments moves down even gentle slopes, causing fissuring and lateral and vertical displacements. This type of landslide known as lateral spreading is commonly responsible for damage to infrastructure (bridges, roads, buildings) during major earthquakes.

During the 1811 and 1812 earthquakes, liquefaction and resulting lateral spreading was severe and widespread. Sand blows formed over an extremely large area about 10,400 square kilometers. Effects of liquefaction extended about 200 km northeast of the New Madrid seismic zone in White County, Illinois, 240 km to the north-northwest near St. Louis, Missouri, and 250 km to the south near the mouth of the Arkansas River. In the New Madrid region, sand blows can still be seen on the surface today. In the past, the sand blows were attributed to the 1811-1812 earthquakes. We now know that some of the sand blows pre-date 1811 and formed as the result of prehistoric New Madrid earthquakes.

photo of sand blow on top and illustration of layers and fissure on bottom
Photograph and schematic cross-section illustrating earthquake-induced liquefaction and formation of sand dikes and sand blows. The photo was taken on February 14, 2016 after the Christchurch, New Zealand earthquake. (modified from the original) (Credit: Martin Luff. Public domain.)

In the New Madrid seismic zone, many sand blows appear as light-colored sandy patches in plowed fields. Flood deposits bury other sand blows. Viewed from above, sand blow have circular, elliptical, and linear shapes and can range up to tens of meters in width and hundreds of meters in length. Viewed in cross-section or in excavations and riverbanks, sand blows commonly take the form of large lenses 1 to 2 m in thickness. Sand blows composed of several layers that fine upward from coarse sand to silt and capped by clay probably formed as a result of multiple earthquakes. Sand blows usually contain clasts, pieces of underlying deposits and soil horizons ripped from the dike walls as the liquefied sand erupted to the surface.

Archaeology

The lower Mississippi River Valley was a fertile homeland to Native Americans from about 9500 B.C. to 1670 A.D. The presence of Native Americans is still evident today in the occasional mound not yet destroyed by modern agricultural practices and the abundant potsherds, lithic tools and points, and bone fragments found in plowed fields and river and ditch cutbanks. Most artifacts encountered during studies of New Madrid sand blows are from the Woodland and Mississippian cultures, which thrived from about 200 B.C. to 1000 A.D. and 800 to 1670 A.D., respectively. Both cultural periods are subdivided into early, middle, and late intervals. Woodland ceramics are characterized by grog (ground up potsherds or fired clay) and sand tempering; whereas, Mississippian ceramics are characterized by shell tempering.

Aerial photograph of sand blow deposits
Aerial photograph showing light-colored patches that are sand blow deposits near Lepanto, Arkansas (from U.S. Department of Agriculture, January 26, 1964). Many sand blows formed above scroll bars of Pemiscot Bayou, also known as Left Hand Chute of Little River.n (Public domain.)
Photograph of some diagnostic artifact types in New Madrid region
Photograph of some diagnostic artifact types in New Madrid region: 1, Campbell Appliqué; 2, Bell Plain; 3, Nodena Elliptical point; 4, Nodena Banks variety point; 5, Parkin Punctate; 6, Madison point; 7, Varney Red Filmed; 8, Barnes Cord Marked; 9, daub with wattle impression. (Photo by Martitia Tuttle, NEHRP-funded research. Public domain.)

Although there are uncertainties regarding their age ranges, certain pottery and point types, as well as plant remains, are considered diagnostic of various cultural periods. For example, Bell Plain, Campbell Appliqué, and Parkin Punctate pottery and Nodena points are diagnostic of the Late Mississippian period; Old Town Red pottery and Madison points are diagnostic of the Middle Mississippian period; Varney Red Filmed pottery is diagnostic of the Early Mississippian period; and Barnes pottery and Table Rock stemmed points are diagnostic of the Late Woodland period. Zea maize, or corn, became dominant in the Native American diet about 1000 to 1050 A.D. and is as an important temporal marker in the region.

Archaeology has played an important role in recognizing and dating prehistoric earthquake-induced liquefaction features in the New Madrid region. Sand blows found below Native American mounds and occupation horizons no doubt formed prior to 1811 because few Native Americans lived in the area after the 17th Century. Diagnostic artifacts found in association with sand blows provide a preliminary estimate of the age of the causative earthquake. Detailed investigations can further constrain the age of the event. For example, artifacts in an occupation horizon buried by a sand blow can provide an estimate of the maximum age of the liquefaction feature; whereas, artifacts in an horizon developed in the top of a sand blow can provide an estimate of its minimum age. Similarly, plant remains and other organics found in cultural horizons can be used to date associated sand blows. Radiocarbon dating of plant remains is the most commonly used dating technique in paleoseismology. It is preferable to have radiocarbon dates from both overlying and underlying horizons to bracket the age of the sand blow.

Paleoseismology

Log of trench wall at Dodd site, Missouri
Log of trench wall at Dodd site near Steele, Missouri, where sand blow and two associated sand dikes are exposed. The pre-event ground surface was displaced downward by 70 to 80 cm between the two sand dikes. Late Mississippian ceramic artifacts found above and below sand blow suggest that it formed between 1400 and 1670 A.D. Radiocarbon dating of charcoal in the soil horizon buried by the sand blow indicates that it formed after 1290 A.D. Radiocarbon dating of a corn kernel collected from a wall trench dug into the top of the sand blow indicates that it formed before 1460 A.D. Therefore, the estimated age of the sand blow is 1290-1460 A.D. (Public domain.)

Paleoseismology is the study of the timing, location, and magnitude of prehistoric earthquakes preserved in the geologic record. Knowledge of the pattern of earthquakes in a region and over long periods of time helps to understand the long-term behavior of faults and seismic zones and is used to forecast the future likelihood of damaging earthquakes. In eastern North America, where near-surface faulting is uncommon or difficult to identify, paleoseismology often employs liquefaction features to learn about prehistoric earthquakes. Earthquake-induced liquefaction features are distinctive and form as the result of strong ground shaking.

Liquefaction features include sand blows, dikes, and sills. Sand blows are deposits that form on the ground surface as the result of venting of water and sand. Sand dikes are sediment-filled cracks through which water and sand flowed. Sand sills usually take the form of lenses intruded below clay layers and are connected to sand dikes. Most large earthquakes around the world have induced liquefaction.

Over the past decade, paleoseismic studies have begun to unravel the earthquake history of the New Madrid seismic zone. Studies focusing on earthquake-induced liquefaction features utilized archaeology and radiocarbon dating to estimate the ages of liquefaction features, and thus, the timing of the earthquakes that caused them. In this way, sand blows across the New Madrid region were found to have formed during earthquakes about 1450 A.D., 900 A.D., 300 A.D., and 2350 B.C.

Sand blow deposit and related feeder dike exposed in excavation
Photograph of sand blow deposit and related feeder dike exposed in excavation. Sand blow buries soil that was at ground surface at time of event. Sand dike fills fissure that formed in soil. For scale, shovel blade is 20 cm wide. (Credit: Martitia Tuttle. Public domain.)

In addition, the size and spatial distributions of historic and sand blows that formed about 1450 A.D. and 900 A.D. were determined to be strikingly similar to each other, suggesting that the prehistoric earthquakes had similar locations and magnitudes to the 1811-1812 earthquakes. Furthermore, sand blows attributed to the 1450 A.D., 900 A.D., and 2350 B.C. earthquakes are composed of multiple, fining upward layers similar in thickness to those that formed in 1811-1812. These observations support the interpretation that the prehistoric events were similar in location and magnitude to the 1811-1812 earthquakes and also suggests that they too were earthquake sequences. Paleoseismic studies concluded that the New Madrid seismic zone generated magnitude 7 to 8 earthquakes about every 500 years during the past 1,200 years.

Sand dike and sill exposed in drainage ditch in southeastern Missouri
Photograph of sand dike and sill exposed in drainage ditch in southeastern Missouri. Sand dike intruded weathered sand; sill emplaced below weathered clay. Layering within the dike and sill indicate that they formed during two or more events. For scale, knife is 8 cm long. (Credit: Martitia Tuttle. Public domain.)

 

Earthquake chronology for New Madrid seismic zone
Earthquake chronology for New Madrid seismic zone from dating and correlation of liquefaction features at sites (listed at top) along NE-SW transect. Some sites show age estimates for more than one feature related to different events (e.g., Eaker 2 and L2). Inferred timing of events is shown with colored bands. (Public domain.)

FAQ for Seismic Hazards in the Central U.S.

What is the estimate of the recurrence interval for 1811-1812 type earthquakes?

Paleoseismic (geologic) studies conducted over the last 20 years have shown that sequences of earthquakes of comparable size to that in 1811-1812 have occurred at least twice before, in approximately 900 and 1450 AD. This implies a recurrence interval of about 500 years.

Given this and other new information, can one estimate the probability of damaging earthquakes in the New Madrid seismic zone?

We have learned a tremendous amount about the New Madrid seismic zone since 1985. One of the things we have learned is that coming up with probabilities is much more difficult than we used to think. If we use the data on historical seismicity combined with the new information on recurrence of large earthquakes, and make the same assumptions that go into the National Seismic Hazard maps, we would estimate a 25-40% chance of a magnitude 6.0 and greater earthquake in the next 50 years and about a 7-10% probability of a repeat of the 1811-1812 earthquakes in the same time period.

However, it is VERY important to note that these estimates alone do not include information about WHERE the earthquakes might occur and therefore what shaking might affect any given location. More useful are the estimates of the likely amount of ground shaking that can be expected, contained in the National Seismic Hazard maps. The ground shaking estimated accounts for both the likely ranges of recurrence intervals and locations.

Does everyone within USGS agree on the cause and effects of a future New Madrid earthquake?

No one knows what causes New Madrid earthquakes. However, there are ideas that are being researched. Although there is great uncertainty regarding the cause of earthquakes, scientists generally do agree on what happens when they do occur, that is, the likely levels of ground shaking associated with the waves earthquakes emit. These levels are reflected in the National Seismic Hazard Maps, which represent the products of a long consensus building process. These maps also account for the uncertainties in our understanding.

Differences of opinion within the research community invariably will arise. Generally these are not critical to people outside the research arena. When they are, the USGS sometimes has held workshops to try to come to a consensus and at other times has announced our own internal consensus. Generally, we have met with the CUSEC State Geologists and been able to come to agreement at least between the State Surveys and the USGS, as well as many other scientists. In most situations, the State Surveys are the ones responsible to the State Governors and the USGS works closely with them.

What is the potential for a large New Madrid earthquake triggering an earthquake in the Wabash Valley?

All we know is that this has not happened in the past few 1811-1812-sized New Madrid earthquake sequences.

After a major earthquake in the New Madrid or Wabash Valley seismic zone, what changes to the landscape would we most likely see?

Deformation of the land surface directly over a fault that moves may manifest as very localized uplift or subsidence, or lateral distortions of up to several meters (for a very large earthquake). Shaking can cause ground failure of various types, including liquefaction and landsliding. These would have significant effect on the landscape in terms of damming streams, spewing sand and mud into fields, and causing areas near bluffs and rivers to slide and form a broken up surface.

Can you explain liquefaction? What conditions would increase or decrease the amount of liquefaction?

Liquefaction occurs when loose, sandy, water saturated soils are strongly shaken. The soils lose their capacity to bear any weight and can flow like a liquid. This process is accompanied by high pore water pressures that can force sand, water, and mud upward, often forming the signature sand blows of the New Madrid seismic zone. Many factors affect how susceptible materials are to liquefaction, but some of the most important requirements are the degree of water saturation, the size of the grains, and how well cemented they are.

After the 1811/1812 earthquakes there were reports that the Mississippi River flowed backward. Can you explain this phenomenon and what is it called?

One of the 1812 earthquakes occurred on a fault that actually crossed the river three times. The uplift along this fault formed a scarp or cliff that caused both a dam and waterfalls at different locations. The damming of the river would have temporarily backed the river up, which may account for the descriptions of the river boat pilots.

For More Information

Was this page helpful?