Overview of the New Madrid seismic zone (NMSZ) in the Mississippi River valley with faults (black), earthquakes (yellow stars), and locations of sackungen features (red dots). Click on image to view larger version.

The central U.S. is not the usual location for large earthquakes in the U.S., but if you were living near the Mississippi River between Memphis and St Louis in 1811-1812, you may have begged to differ. Three large earthquakes in the New Madrid seismic zone between December 1811 and February of 1812 shook a huge area from the mid-west to the east coast. Historical accounts, as well as many scientific studies, have established their approximate locations (see map) and magnitudes (~M7+).

Details of past earthquakes are difficult to come by in this area where evidence of shaking has long since been eroded or covered by sediments from the Mississippi River, or modified by farming. Older earthquakes are even harder to pin down, but trenching and liquefaction studies have documented evidence for three earthquakes in the past 1100 years, occurring roughly 550 years apart, with a possible fourth earthquake 4400 years ago. Extending this earthquake chronology would help scientists understand the long-term pattern of earthquakes in this region, as constraints on the timing of past earthquakes are essential for estimating the future earthquake hazard.

Examples of sackung mapping (locations in Fig. 1b). a) Sackungen manifest as troughs, scarps, fissures, and benches on ridge crests and upper ridge flanks. b) Sackung features abruptly end at a farming boundary; therefore, sackung mapping is limited to locations where extensive anthropogenic modification is absent. c) Ridge cross-section (location in 2a) showing schematic representation of one possible sackung failure geometry at depth (red lines). Geologic section and location of failure surface based on Jibson and Keefer (1993). Click on image to view larger version.

Thus, when new high-resolution lidar datasets (Open Topography and Tennessee GIS) became available in 2010-2012 in northwestern Tennessee, USGS scientists looked to this new source of data for clues of past earthquakes. They were intrigued to discover indications of geologic cracks and scarps, called sackungen, on the tops of bluffs near the Reelfoot fault in the New Madrid seismic zone (One is a sackung, many are sackungen.) Sackungen have been observed in many places around the world and have been attributed to deglaciation (the melting of glaciers), slumping due to gravity, or strong shaking from earthquakes. These processes usually occur in previously glaciated or steep and mountainous terrain, in contrast to the glacier-free low-elevation bluffs in western Tennessee. Since the sackungen observed on the lidar only occurred on the ridges near the Reelfoot fault, even though similar bluffs exist throughout Tennessee, the scientists suspected these features were caused by earthquake shaking.

Careful mapping of 1762 sackung features showed them to be generally <1.5 m (<5 ft) high, 5 to 335 m (16 to 1100 ft) long, and as wide as ~5 m (16 ft). They are present only within 15 km (9 mi) of the Reelfoot reverse fault, and 85% are within 5 km (3 mi) of the fault. They appear on both sides of the fault, but are dominantly on the hanging wall (the side that "hangs over" the fault), which typically shakes more during an earthquake on a reverse fault. The sackungen are visible in the topmost sediments which have been dated to between 11,000 and 30,000 years old, which puts a maximum limit on the sackungen ages. From 11,000 years ago to present, there has been no deposition, just changes to those sediments due to deformation or erosion.

The amount of shaking required to create the sackungen in this location was estimated to be about 0.33g (1/3 the force of gravity). Since the shaking from the February 23, 1812 event on the Reelfoot fault has been estimated to be 1.4-1.5g, that earthquake, and presumably others before it, are believed to be large enough to have caused the sackungen.

Schematic locations, spatial densities, map distribution, and orientations of coseismic sackungen based on the New Madrid seismic zone. Reverse faulting produces ridgetop sackungen with higher densities and at greater distances from the fault trace in the hanging wall than in the footwall. Strike-slip faulting produces ridgetop sackungen having equal densities and distances on both sides of the fault, with a smaller overall distance from the fault trace than a reverse fault. Strong fault-normal shaking and/or static shear strain produces fault-parallel sackungen with fault-normal spreading for reverse and strike-slip faults. Grey shading darkens with increased deformation/static shear strain.

(a) Detailed hillshaded digital elevation model (DEM) of the Paw Paw trench site, derived from 0.5-m lidar. Sackungen indicated by white arrows. Outline of trench excavation depicted by white rectangle. Contour interval 1 m. Location and look direction of field photographs (Figure 4) indicated by black balloon shapes with inset arrows. Primary sackung and corresponding shear zone (SZ1) and subsidiary eastern and western features (SZ2 and SZ3, respectively), exposed by trench are labeled. (b) Topographic profile across the trench site, showing the morphology of the sackungen intersected by the trench. Schematic illustration of shears in the subsurface indicated by red lines. Click on image to view larger version.

To further investigate the possibility of earthquake-produced sackungen, USGS scientists dug a trench across an especially well-preserved sackung to date the geologic features. They found evidence of four episodes of sackung movement, with the oldest episode occurring about ~4400 years ago and the most recent episode being the 1811-1812 earthquake sequence. The sackungen were identified by colluvial deposits, or loose piles of sediments that had fallen off the top of the bluffs into the cracks. If the trenched sackungen have recorded all of the earthquakes that have occurred since the bluff sediment was deposited, it means there was a long time period with no earthquakes between 11,000 years ago and 4,400 years ago, followed by an active period with four major earthquakes in the last 4400 years. Importantly, this investigation represents the longest continuous earthquake record for the Reelfoot fault.

The sackungen analysis has enabled scientists to look further back in time and better constrain the earthquake timeline of the region, while informing seismic hazard. All indications are that these sackung in the New Madrid area were created by earthquakes on the Reelfoot fault. Collectively, studies indicate that earthquakes do not occur on a regular schedule in this region. Additionally the study has documented the potential role of sackungen in studies of past earthquakes in other areas where the clues have been buried.

- written by Lisa Wald, USGS, April 2019

The field team - Ryan Gold, Randall Jibson, Jaime Delano, and Chris DuRoss standing on the trench ledge.
Randall Jibson and Jaime Delano gathering samples from the trench wall.
Chris DuRoss taking notes in the trench.

For More Information

The Scientists Behind the Science

Rich Briggs.

Rich Briggs is a Research Geologist with the USGS in Golden, Colorado. He studies the geology of earthquakes. When not at work, he has a soft spot for vintage hi-fi, the theory and practice of pilsner, and freestyle home improvement. He mostly loves hanging out with his family.


Jaime Delano.

Jaime Delano is a geologist with the USGS in Golden, CO who studies past earthquakes by looking for evidence in the landscape with high-resolution imagery and digging into the sediment record. She also enjoys podcasts, playing in the mountains, and experimenting with sourdough bread.


Chris DuRoss with samples from a trench in the Tetons near Jackson Hole, Wyoming.

Chris DuRoss is a Research Geologist with the USGS in Golden, Colorado. As a paleoseismologist, he studies geologic evidence of large, prehistoric earthquakes in areas of active faulting such as the Wasatch Front. When not working, he can be found (or not) with his family in the remote parts of the West.


Ryan Gold mapping the Teton fault in Wyoming in a paleoseismic trench at the Buffalo Bowl site.

Ryan Gold is a Research Geologist with the USGS. His work focuses on documenting the timing and magnitude of past (prehistoric) earthquakes using clues from the modern landscape and trench exposures. When he’s not working, he can usually be found trail running or skiing in the Colorado Rocky Mountains.


Randy Jibson investigating the stability of Tsho Rholpa, a glacial lake in the Himalayas, following the 2015 Gorkha, Nepal earthquake.

Randy Jibson is a Research Geologist with the USGS working in Golden, Colorado. With backgrounds in both geology and geotechnical engineering, he has studied landslides that have been triggered by earthquakes throughout the world for more than 40 years. When not chasing earthquakes, he enjoys being with his family for back-road travel, hiking, biking, and all that other outdoor stuff.