Quaternary Fault and Fold Database of the United States

The New Madrid seismic zone is east of Crowleys Ridge, in and east of the Eastern Lowlands and the Mississippi River (for example, Wheeler and Rhea, 1994 #754). Therefore, the seismic zone is generally 30 km or more east of the area summarized here. The New Madrid seismic zone is spatially associated with the Reelfoot rift. The main upper crustal expression of the rift is the Mississippi Valley graben. All of the Western Lowlands paleoseismological sites studied by Vaughn (1994 #3924), Tuttle and others (1998 #3817), and Tuttle (1999 #3921) lie several tens of kilometers northwest of the floor of the graben, and also northwest of the wide, step-faulted margin of the graben (Rhea and Wheeler, 1994 #3919). However, several of the sites lie near or on an aeromagnetic lineament that trends northeast (Vaughn, 1991 #3922; Tuttle, 1999 #3921). The lineament is part of the Commerce geophysical lineament, which, after magnetic and gravity modeling, Langenheim and Hildenbrand (1997 #3917) attributed to a mafic dike swarm that they suggested may have formed parallel to and coevally with the Reelfoot rift. If this part of the lineament, the inferred dike swarm, and the rift formed coevally, then likely the mafic magmas intruded fractures that formed or reopened with an extensional component of movement. Thus, the formation of the source of the Commerce geophysical lineament may have involved at least a component of normal faulting. This reasoning indicates that the paleoseismological sites in the Western Lowlands may lie near one of the northwesternmost normal or normal-oblique faults of the rift, even if the sites lie outside the main part of the step-faulted rift margin.

Paleoliquefaction sites 1, 2, 6, and 7 of Vaughn (1994 #3924) surround Dudley ridge. Two high-resolution seismic-reflection profiles across the western slope of the ridge are interpreted as showing three faults that extend from Paleozoic and Cretaceous bedrock upward into Quaternary sediments (Shoemaker and others, 1997 #3920; Shoemaker and others, 1998 #3955). Another high-resolution reflection profile across the Commerce geophysical lineament approximately 30 km southwest of Dudley ridge shows about 20 m of dip separation on a reflector within the Quaternary sequence (Stephenson and others, 1999 #3954). Thus, the profile interpretations support the liquefaction evidence of Quaternary faulting.

Vaughn (1991 #3922) and Boyd and Schumm (1995 #3916) identified several geomorphic anomalies, including diverted rivers, in the Western Lowlands, and speculated about possible Quaternary tectonism. However, any such tectonism need not occur coseismically, or even involve faulting. Accordingly, this assessment of the Western Lowlands focuses on paleoseismological evidence for prehistoric earthquakes.

The first two events are best represented by crosscutting dikes at site 1. A dike of event 1 fed a crater and a sand blow. Three dikes of event 2 cut the older dike, crater, and sand blow. Both events are bracketed by dates of 22,750 radiocarbon yr B.P., from wood 35 cm below the event 1 sand blow, and of 3,570 yr B.P., from wood at the base of a paleochannel that post-dates both events. Elsewhere at site 1, a late Wisconsinan channel deposit contained a bone dated at 12,570 yr B.P. The channel deposit is cut by a dike, which is itself cut by a younger dike. The younger dike is truncated by a second, late Holocene channel deposit with wood at its base that yielded a date of 590 yr B.P. Therefore, the two dikes represent two liquefaction episodes between 12,570 yr B.P. and 590 yr B.P. These episodes might or might not correspond to events 1 and 2. At site 6, deformed beds beneath a sand blow contain wood and bone; the bone gave a date of 17,990±150 yr B.P. Vaughn (1994 #3924) assumed that the sand blow was coeval with event 1. Thus, events 1 and 2 are younger than 22,750 yr B.P. Note that, depending on the validity of the correlations elsewhere in site 1 and to site 6, both events may be younger than 17,990 yr B.P. Alternatively, the dikes elsewhere in site 1 might have formed in unrecognized events between events 2 and 3. Regardless, events 1 and 2 occurred before 3,570 yr B.P.

Dikes and buried sand blows at site 2 were attributed to event 2. A carbonaceous silt 58–62 cm below a sand blow was dated at 13,430±170 yr B.P. Early Archaic archeological artifacts are on the terrace surface that overlies the sand blow, indicating that the event predated 9,000–10,000 yr B.P. (Vaughn, 1994 #3924).

Events 3 and 4 are known only from site 3 (Vaughn, 1994 #3924). Two surficial sand blows about 100 m apart bury dated soils, and Vaughn (1994 #3924) suggested that the ages of the soils approximate the ages of the events. Organic material in the soil horizon immediately beneath the larger sand blow was dated as 1660±70 and 1110±65 yr. B.P., which give calibrated ages of AD 240–560 and 770–1020, respectively (Vaughn and others, 1993 #3923; Vaughn, 1994 #3924). Thus, event 3 was estimated to have occurred during AD 240–1020. Similarly, organic material from the soil horizon immediately beneath the smaller sand blow was dated as AD 1440–1540 (calibrated), which is the estimated age of event 4.

Tuttle and others (1998 #3817), and Tuttle (1999 #3921) sought to define the ages and size distribution of liquefaction features in and near the New Madrid seismic zone, including the Western Lowlands. She summarized results from 116 liquefaction and archeological sites, of which 25 are in the Western Lowlands. Of these 25 sites, 15 contained prehistoric liquefaction features attributed to earthquakes. Most of the features are sand dikes. Only three sandblows were found and two of these were dated at Current River 8 and at Current River 2 of Tuttle (1999 #3921). Current River 8 is at or near site 4 of Vaughn (1994 #3924), and Current River 2 is newly discovered.

At the Current River 2 site, a sand blow buried a cypress knee. Later, a tree root, now burned, penetrated a paleosol above the sand blow. Calibrated dates on the cypress knee and the burned root constrained the sand blow to have formed during AD 1310–1450 (Tuttle, 1999 #3921).

The Current River 8 site records two to four liquefaction episodes (Tuttle, 1999 #3921). Sand dikes and a sand blow (episode I) were intruded by later sand sills (episode II). Other sills were intruded (episode III) into sediment that had been deposited on top of the sand blow. The sills of episode III were themselves cut by younger dikes (episode IV). No structural or stratigraphic evidence precludes episodes II–IV from having formed during the same earthquake or earthquake sequence. The only age control is from wood found 15 cm below the episode I sand blow. The wood yielded a calibrated date of 3490–3470 BC and 3380–3090 BC. Thus, episodes I–IV postdate 3490 BC. (Tuttle, 1999 #3921).

Tuttle (1999 #3921) combined her results with those of several other paleoliquefaction studies, including that of Vaughn (1994 #3924), to deduce the occurrence of very large to great earthquakes or earthquake sequences in the New Madrid seismic zone in AD 900±100 yr, AD 1530±130 yr, and, of course, 1811–12. Additional liquefaction events occurred before AD 900, but their ages and geographic distributions remain poorly constrained. Because no sand blows were found in the area that could be attributed to the 1811–12 New Madrid earthquakes, Tuttle (1999 #3921) suggested that the sand blow at Current River 2, which formed in AD 1380±70 yr, may be too large to have been caused by an earthquake in the New Madrid seismic zone 80 km or more distant. Instead, Tuttle (1999 #3921) suggested that the sand blow may be related to a separate source in the Western Lowlands. Earlier, Vaughn (1994 #3924) had attributed his event 2 to a possible source in a different part of the Western Lowlands. The two to four liquefaction episodes at Current River 8 (site V4 of Vaughn, 1994 #3924) are known only to postdate 3490 BC. One or more of liquefaction episodes II–IV at Current River 8 might have been coeval with the single episode at Current River 2 (Tuttle, 1999 #3921). Events 3 and 4 at site 3 of Vaughn (1994 3924; site V3 of Tuttle, 1999 #3921) are attributed to the AD 900±100 yr and AD 1530±130 yr earthquakes or earthquake sequences in the New Madrid seismic zone (Tuttle, 1999 #3921, figure 47). The zone is only approximately 50 km away from site 3, at which Vaughn (1994 #3924) suggested that a smaller surficial sand blow may be attributed to the 1811–12 New Madrid earthquakes.