In the last decade there has been an increasing awareness that seismic hazard in the eastern United States may be greater in some areas than the historic earthquake record would suggest. Long recurrence intervals for major events, measured in centuries or millenia, far exceed the 200-year historic record. There is a tendency for many people to be lulled into a false sense of security in areas that may be prone to periodic large, damaging earthquakes because the area may have never experienced such an earthquake in historic times.
We need only to think of the series of great magnitude 8 earthquakes in 1811-1812 in New Madrid, Missouri, to realize that if they had occurred a century or two earlier our written record of these events would probably consist of a brief notice of a light shock felt in New England. There would be little realization that this area was capable of producing the largest earthquakes ever recorded in the continental United States.
Intensive studies of the New Madrid seismic zone have raised the inevitable question--how often do such large
earthquakes occur? In the absence of a written record, geologists turned to their book of the past, the record preserved in rocks and sediments. They soon began to realize that strong earthquakes cause some sediments to liquify into a fluidlike consistency and form dikes, sills, sand blows, and other ground-failure features. Thus was born the study of paleoseismicity and the search for earthquake-induced liquefaction features that could be dated by radiocarbon or archaeological associations and organized into a time sequence.
Generalized cross section of a stream bank showing two sets of vertical sand dikes and sand blows resulting from liquefaction of saturated sand by strong seismic shaking. Note that the dike on the right cuts through the sand blow generated by the dike on the left. This relationship indicates two separate seismic events. Modified from Obermeier and others (1993).
|
Typically, liquefaction is caused by upward propagation of shear waves from the bedrock into overlying unconsolidated sediments. Sand or gravelly sand that is saturated by a high water table and overlain by silt or clay is most susceptible to development of liquefaction features such as sand dikes. During an earthquake of sufficient intensity, the liquid sand-water mixture hydraulically fractures the overlying fine-grained materials. The sand-water mixture then typically protrudes up into the cap, forming a steeply dipping, tabular dike. In cross section, the dike may range from a few inches to a few feet in width. In plan view, the dike may extend for hundreds of feet.
Larger dikes tend to vent to the surface in the form of a sand blow, which may be a foot or two thick and more than 100 feet in diameter. In cross section the sand blows appear as horizontal layers of sand immediately overlying an ancient soil (paleosol). Later sedimentation may cover the surface sand deposit. Recurrent, strong earthquakes in an area may result in multiple sets of dikes and sand blows that exhibit a cross-cutting relationship. If each set of dikes can be dated, some prediction of recurrence intervals of large earthquakes can be made. In general, liquefaction features begin to appear during earthquakes of magnitude 5.5 or above.
However, in the eastern United States these features seem to be associated with larger earthquakes, generally magnitude 6.0 or larger.
This past summer, Ohio was fortunate to have the services of Stephen F. Obermeier of the U.S.
Geological Survey Branch of Earthquake and Landslide Hazards. He began a search for paleoliquefaction features that would indicate the occurrence of ancient great earthquakes in the state. Similar work by Obermeier and Patrick J. Munson of Indiana University in the Wabash Valley of Indiana and Illinois and other areas in this region indicated that at least seven strong earthquakes had occurred between about 20,000 years ago and 2,500 years ago. At least one of these events, about 6,100 years ago, is estimated to have had a magnitude on the order of 7.5.
Obermeier, accompanied by Ohio State University graduate student Erik Venteris, began his search in the western Ohio seismic zone, an area that has experienced at least 40 felt earthquakes since 1875 (see
Ohio Geology, Summer 1993). The largest of these, on March 9, 1937, is estimated to have had a magnitude of about 5.5. The limited exposures of sediments in this relatively flat area are confined to stream banks and sand and gravel pits. Obermeier and Venteris canoed more than 100 miles of streams and found more than 25 miles of freshly eroded stream banks that could be searched for seismically induced dikes and sand blows. In the western Ohio seismic zone, they canoed portions of the Auglaize, Great Miami, Stillwater, and
St. Mary's Rivers and Loramie Creek. Portions of the Scioto and Little Scioto Rivers in Marion County in north-central Ohio were examined, as were seven sand and gravel pits.
The good news, at least on a preliminary basis, is that Obermeier found no indisputable paleoliquefaction features in any of the outcrops he examined. He expresses some confidence that the western Ohio seismic zone has not experienced a very strong earthquake, above magnitude 7, in the last few thousand years. However, this evidence does not preclude the possibility that the area has had prehistoric earthquakes in the 6.0 to 6.5 range. If judgment can be drawn from the Charleston, Missouri, earthquake of 1895 (magnitude 6.5), liquefaction features only begin to appear at about this threshold magnitude and occur only in a very small epicentral area. Obermeier notes that large areas of western Ohio are unsuitable for development of liquefaction features such as dikes because of a lack of near-surface sand units and, therefore, would not exhibit evidence of strong prehistoric earthquakes, even if they did occur.
For the 1995 field season, Obermeier plans to examine stream exposures in northeastern Ohio in Lake and Geauga Counties. This area has experienced at least 20 felt earthquakes since 1836, including a magnitude 4.5 event in 1943 and a magnitude 5.0 event in 1986 (see Ohio Geology, Summer 1986). He also plans some additional investigations in western Ohio.
--Michael C. Hansen
FURTHER READING
Obermeier, S. F., Martin, J. R., Frankel, A. D., Youd, T. L., Munson, P. J., Munson, C. A., and Pond, E. C., 1993, Liquefaction evidence for one or more strong Holocene earthquakes in the Wabash Valley of southern Indiana and Illinois, with a preliminary estimate of magnitude: U.S. Geological Survey Professional Paper 1536, 27 p.