Tecumseh's prophecy: Preparing for the next New Madrid earthquake: A plan for an intensified study of the New Madrid sei

U.S. GEOL()GICAL SURVBV CIRCULAR 1066 COVER

Woodcut of ~~scene of the Great Earthquake in the West II (Our First

Century: One Hundred Great and Memorable Events, 1877, p. 220).

ABOUT THE TITLE

According to legend, (Penick, 1981, p. 123), the Shawnee Indian chief Tecumseh traveled south in 1811 from his village of Tippecanoe on the Wabash River in an attempt to recruit supporters. At a town of Creek Indians near the site of Montgomery, Ala., Tecumseh declared that they would know he was sent by the Great Spirit because, upon leaving, he would go to Detroit, stamp his foot on the ground, and shake down all their houses. Tecumseh left. The Creeks counted the days; on the morning they had estimated for his arrival in Detroit, the earth began to shake, and houses fell down. It was the first of the New Madrid, Mo., earthquakes, which were centered about 250 mi away in the Mississippi Valley. TECUMSEH ·s PROPHECY: PREPARING FOR THE NEXT NEW MADRID EARTHQUAKE Edited by ROBERT M. HAMILTON and ARCH C. JOHNSTON

With contributions by the participants in the New Madrid Seismic Zone Workshop, November 15-16, 1989, Memphis, Tennessee

Federal Center, Box 25425 Denver, CO 80225

Summary 1 Introduction 2 The Danger: Seismic Risk Assessment 5 Introduction 5 Seismicity 8 Regional Tectonic Setting 9 Earthquake Effects 11 Potential Losses 12 The Response: Preparedness and Mitigation 15 The Plan: Preparing for the Next New Madrid Earthquake 17 Introduction 17 Purpose 18 Goals 18 Activities 18 Management and Coordination 27 Conclusions 27 References Cited 28 Appendix: New Madrid Seismic Zone Workshop Participants 30

1-5. Maps showing:

6. Photograph showing trees tilted by a landslide along the Chickasaw Bluffs
7. Aerial photograph taken in 1959 near Portageville, Mo. 14
8. Diagram illustrating the goals of community actions in implementing loss

9-11. Maps showing:

1. Damaging earthquakes in the New Madrid region 3
2. Earthquake probability estimates for the New Madrid seismic zone 8
3. Seismicity of the Central and Eastern United States and adjacent parts of Canada showing earthquakes of magnitude 3 or greater from 1568 to 1987 4
4. Areas of intensities VI and VII for two great earthquakes of about magnitude 8 and two major damaging earthquakes 6
5. Regional distribution of estimated Modified Mercalli intensities that would result from a recurrence of the 1811-12 earthquake series in the New Madrid seismic zone 7
6. Seismicity and major structural features of the New Madrid seismic zone 10
7. Seismicity and major geologic structures in the most active part of the New Madrid seismic zone 11

near Reelfoot Lake in northeastern Tennessee 13

9. Current regional seismic network in the New Madrid seismic zone 19
10. U.S. National Seismic Network stations in the Central United States 20
11. Current strong-ground-motion stations in the New Madrid seismic

Summary

earthquakes. Because such information is currently incomplete for the New Madrid region, some of the major goals of an intensified study should be to improve the information that would be used as the basis for mitigation.

An intensified study of the New Madrid seismic zone should focus on five goals:

1. Implementing earthquake-hazard mitigation meas-
2. Improving preparedness for earthquakes of magni-
3. Establishing a modern seismic network in the New
4. Locating faults that could generate destructive
5. Improving seismic-risk assessments.

Depending on the options selected and the pace of the study, these recommendations could be implemented at a funding level of $5 to $10 million annually for 5 years, at which time progress should be reviewed and a new plan prepared. An intensified study would be conducted under the auspices of the National Earthquake Hazards Reduction Program (NEHRP), which involves the Federal Emergency Management Agency (FEMA), the U.S. Geological Survey (USGS), the National Science Foundation (NSF), and the National Institute of Standards and Technology (NIST). An intensified study should include the full scope of NEHRP activities. Generally, mitigation measures are the responsibility of FEMA, NSF, and NIST, and research is the responsibility of USGS and NSF. Coordination of the study would be facilitated by establishing a coordinating committee and by holding an annual workshop to review recent developments and progress and to recommend priorities.

INTRODUCTION

The New Madrid seismic zone in the central Mississippi Valley poses the greatest earthquake danger in the United States east of the Rocky Mountains. Four earthquakes of about magnitude 8 (table 1) occurred in the seismic zone in the winter of 1811-12, and the area continues to experience the highest level of seismicity in the central and eastern parts of the Nation (fig. 1). Concern about the probable effects of future New Madrid earthquakes convinced Congress to direct the U.S. Geological Survey to prepare a plan for an intensified study of the New Madrid seismic zone. The first step in preparing the plan was to convene a workshop of about 70 academic, private, and State and Federal Government experts (appendix) in ures.

tude 6 or larger.

Madrid seismic zone to monitor the sizes, locations, and characteristics of the earthquakes and to determine the nature of the ground motions that they generate.

earthquakes, determining the recurrence rates of earthquakes, and delineating areas of potential damage. Memphis, Tenn., on November 15-16, 1989. to prepare recommendations for such an intensified study. This publication summarizes the results of that worksho'J.

Earthquake studies in the United States an" conducted mainly under the auspices of the National Earthquake Hazards Reduction Program (NEHRP), which was authorized by the Earthquake Hazards Reduction Act of 1977 and its subsequent amendments and reauthorizations. The Federal agencies that participate in NEHRP and th~ir areas of responsibility are as follows:

• Federal
• U.S. Geological Survey (USGS): Earthquake
• National Science Foundation (NSF): Earthquake
• National Institute of Standards and Technology

Representatives from the four NEHRP Federal agencies attended the workshop. The recommendations ir this report cover the full scope of responsibilities under NEHRP.

NEHRP not only involves a partnership among four Federal agencies but also encompasses participation by scientists, engineers, and others from universitie~. State and local government agencies, private groups, and other Federal agencies. University scientists, in particular, have played an important role in NEHRP, a role that developed from their pioneering studies of U.S. earthquakes. Federal Government agencies that are not formally part of NEHRP also have contributed significantly to earthquake studies in the New Madrid region. Most importantly. the U.S. Nuclear Regulatory Commission (NRC) has supported a comprehensive scientific research effort that included expansion of regional seismic network coverage during the late 1970's and early 1980's. Recently, the NRC has been gradually withdrawing financial support for tl'~ regional seismic networks; support is scheduled to end in 1992. In Emergency Management Agency (FEMA): Lead agency of NEHRP; planning, coordination, and program review; annual report to Congress; opportunities for participatior by States, localities, private organizations, and individuals; assistance to State and local governments to implement comprehensive earthquake-hazarc reduction programs; improved seismic design and construction techniques and standards for application; public education and awareness programs; and coordination of Federal response to catastrophic earthquakes.

potential; earthquake prediction; earthq'Iake information and data services; earthquake hazards and risk assessments; strong-ground-motion data and estimates; and technical assistance in fostering implementation of loss-reduction measures.

engineering research; earthquake preparedness and emergency response research; social, economic. and political impacts; earthquake research information; and fundamental studies on the sources and mechanisms of earthquakes and on earth struc+ure.

(NIST): Seismic design and constructio'1 standards and Federal construction practices. 1986, the NRC and the USGS signed an agreement to establish the U.S. National Seismic Network to satisfy NRC's seismologic data needs. The U.S. Army Corps of Engineers has conducted a variety of important studies of the fluvial deposits of the Mississippi Valley and maintains instruments at Corps facilities to record strong ground motion from earthquakes. The Department of Veterans' Affairs has installed instruments to record strong ground motion in some of its hospitals in the region. The NSF established the National Center for Earthquake Engineering Research, which conducts various studies and operates instruments in the the New Madrid seismic zone.

State organizations and universities have taken some very important initiatives in addressing earthquake hazards. Monitoring of the New Madrid seismic zone was begun by seismologists at Saint Louis University when they installed instruments at Cape Girardeau, Mo., and Little Rock, Ark., in 1929. The university continues to monitor seismicity through networks funded by the USGS and the NRC. Faculty members, particularly the late Otto Nuttli, have spent considerable time informing the public of the hazard posed by the New Madrid seismic zone. Professor Nuttli, beginning in the late 1960's, struggled virtually alone for several years to alert State and Federal agencies, as well as the public, to that hazard. The State of Tennessee recognized the earthquake threat in 1977, when it formed the Center for Earthquake Research and Information (CERI) as both a State agency and an independent research unit of Memphis State University. The State has con+inued its strong support for CERI to the present time and provided resources for major programmatic expansions in 1984 and 1986. The State of Kentucky is one of only a few States in the Nation that provide funding for a State seismic network. Kentucky also adopted building code provisions for seismic loading in 1981 and upgraded them in 1988. The State of Missouri recently established an earthqualre-hazard

N = 817

• High-rise buildings in San Francisco (most of which are located on very firm soil or rock foundation) were not damaged significantly by the earthquake; however, because the ground motion in San Francisco on sites underlain by rock or stiff soil was only about 10 percent of gravity and the duration of strong shaking was only about 10 seconds-generally below the level that the buildings were designed to withstand- the Lorna Prieta earthquake did not provide an extreme test of their design. Schools, hospitals, and other critical facilities suffered little damage. In contrast, numerous structures built on filled land experienced ground motion of about 25 to 30 percent of gravity and sustained heavy damage. Earthquake-resistant design and construction practices apparently were effective. Many cities in the New Madrid region have not implemented earthquake-resistant design practices, and most old and new buildings in the region were not designed to withstand earthquakes.
• Government agencies and the general public in northern California responded well to the effects of the Lorna Prieta earthquake. Preparedness plans and exercises had helped lay the groundwork. Until recently, little effort had been made to prepare for earthquakes in cities in the Central United States, and the public is largely complacent about the earthquake danger. Regional and local centers-for example, the Central U.S. Earthquake Consortium (CUSEC), CERI of Memphis State University, Saint Louis University, the Governor's Earthquake Hazards and Safety Technical and Advisory Panel of Kentucky, and the Center for Earthquake Studies at Southeast Missouri State University in Cape Girardeau-have developed a good knowledge base for preparedness measures, but the scope of implementation is limited by the meager resources available. The National Center for Earthquake Engineering Research has implemented an active program addressing preparedness and mitigation issues in the Eastern United States in which the New Madrid region has a prominent role.
• The location of the Lorna Prieta earthquake had been forecast accurately (U.S. Geological Survey, 1988). It occurred on a section of the San Andreas fault system in the Santa Cruz Mountains that was identified as the fault segment most likely to produce a strong earthquake in the bay area. This successful forecast shows that understanding of the mechanics of the San Andreas fault zone has improved to the point that mitigation actions can be directed to the most hazardous parts of a fault system. In contrast, the little that is known about the relative probabilities of earthquakes occurring along The appropriate level of support for New Madrid

studies in relation to studies of other U.S. seismic zones, as well as the balance of effort among the various NEHRP Federal agencies, can be determined only through evaluation of the whole NEHRP, a task that is beyond the scope of this report. Also, funding for NEHRP agencies to carry out the recommendations in this plan is, of course, subject to established budget procedures. This plan is intended primarily to provide a basis for budget proposals from each NEHRP agency. It can also be used by other organizations, such as State geological surveys and emergency agencies, to prepare related budget initiatives. Nevertheless, it is necessary to base this plan on an approximate level of funding to indicate the types of products that are needed and the types of projects that are foreseen in the study. The work described here could be carried out with total funding among all four NEHRP agencies of $5 to $10 million annually for 5 years. The exact amount would control the pace of the study and the choice of options. After 5 years, progress should be reviewed and the plan modified.

In recent years, New Madrid research studies have been funded at under $1 million per year. In fiscal year (FY) 1990, $3 million was allocated for New Madrid studies from funds appropriated to the USGS in Public Law 101-130 (FY 1990 Dire Emergency Supplemental to Meet the Needs of Natural Disasters of National Significance). This funding will permit some of the studies described in this plan to be initiated; however, most types of research require several years to complete, and some, such as seismic network operation, are of an ongoing nature and require a stable budget. Thus, the FY 1990 funds will be used mainly for capital expenditures (equipment and data acquisition) and short-term studies that meet urgent needs.

Introduction

The earthquake dangers in California and Alaska are well known to most Americans. Small but sharply felt earthquakes occur frequently in those States, and destructive shocks are common enough to maintain public awareness. The relative infrequency of destructive earthquakes east of the Rocky Mountains leads to a perception that the earthquake danger there is insignificant in comparison with the dangers from other natural hazards, such as tornadoes and hurricanes. But the low frequency of occurrence of eastern earthquakes is offset by the vast area that they affect (fig. 2). A recurrence of the great New Madrid earthquakes of 1811-12 would be felt from Denver to New York City, topple chimneys in Chicago, Knoxville, Dallas, and Kansas the numerous major fault zones of the New Madrid region emphasizes the need for additional concentrated studies.

"' "c a"' ::"0

"'z "'~

c:m

Figure 2. Areas of intensity VI and VII for two great earthquakes of about magnitude 8-New Madrid, Mo., in 1811 and San Francisco, Calif., in 1906-and two major damaging earthquakes-Charleston, S.C., in 1886 and San Fernando, Calif., in 1971 (Rankin, 1977). At intensity VI, minor damage occurs; at intensity VII, principally architectural damage occurs. Areas affected in the East are much larger than those affected in the West owing to lower seismic-wave attenuation in the East.

VII

about how often. where. and when such events are likely to occur and about whether other areas in the Eastern United States that have not yet experienced destructive earthquakes of that magnitude might experience one in the future. One goal of earthquake research is to reduce this uncertainty about the locations and frequency of damaging earthquakes in the eastern part of the country. Such information, which is currently inadequate. is essential for deciding where to focus efforts to prepare for and cope with earthquake hazards and for allocating resources accordingly for preparedness and mitigation measures.

Information and knowledge that would enable earthquake-hazards assessment in any area are:

• Geologic (prehistoric)~ historical, and instrumen-
• Geologic setting: Is the area situated along a bound-
• Current rates and past history of land (crustal

tal records of past seismicity: Where, how often, and how regularly have strong earthquakes occurred in the past? When will the next one occur? How big have they been? How big can they be?

ary between two tectonic plates (like the San Andreas fault) or in the middle of an old, quasistable plate (like the New Madrid seismic zone)? What is the location, length. depth, distribution, and direction of movement on the faults that cause the earthquakes? What is their structure? What properties make some faults more susceptible to movement than others?

and surface) deformation: What is the orientation of stresses that cause earthquakes? What is the cause

• Efficiency and nature of seismic-wave propaga-
• Vulnerability of existing structures: How vulner-
• Effects of strong ground motion on the land
• Effects of strong ground motion on ground fail-

For the New Madrid region, the data currently aYailable for assessing the earthquake danger are deficient in many respects. This deficiency can be reduced only through further research on the topics described in this report.

Seismicity

The history of earthquake activity in the New Madrid seismic zone (table 1) is dominated by the earthquake sequence that struck the area in the winter of" 1811-12 (Fuller, 1912; Nuttli. 1973). The sequence began on December 16. 1811. with a tremendous earthauake, followed by another large shock 6 hours later. ~11bsequent great earthquakes occurred on January 23 and February 7, 1812. In terms of magnitude, these earthquakes are believed to be the largest shocks known to have occurred in a so-called stable continental interior; in term<;' of area affected, they may be the strongest historical shocks in the world (Johnston and Kanter, 1990). In contrast to the typical pattern of a single principal shock followed by a series of aftershocks, the 1811-12 sequence consisted of four very large shocks. each of which was followed by aftershocks, many of which were themselves significant earthquakes. Six aftershocks had magnitudes of 6 to 7, and more than 1. 800 aftershocks large enough to be recorded as far away as Louisville, Ky., occurred in the firs" 5 months following the December 16, 1811 , event. Aftershocks continued until at least 1817. About as many felt earthquakes occurred in the Mississippi Valley in 5 months as occurred in southern California in the 40-year p~riod from 1932 through 1972. The most intense earthquake activity in the Eastern United States continues to be in the New Madrid seismic zone (fig. 1).

Other strong earthquakes have occurred in the New Madrid region since the 1811-12 sequence (ta'Jle 1). Of particular note were a magnitude 6.4 earthqua1-e in 1843 near Marked Tree, Ark., and a magnitude 6.8 earthquake in of the stress? How fast is stress accumulating, and what level does it have to reach before faults slip?

tion: How large an area would be affected by destructive earthquake shaking? What a~e the predominant vibration frequencies?

able are existing structures of various mzterials and construction types to a strong earthquak~?

surface, buildings, and lifeline systems: What will be the duration of shaking? Will the ground diminish or amplify the vibrations? How will the structures and lifeline systems respond?

ure: Where will liquefaction and landslides take place? What properties and physical setti"'gs lead to ground failure of unconsolidated sediments? 1895 near Charleston, Mo. These events occurred at the southwestern and northeastern ends, respectively, of the area most affected by the 1811-12 sequence; perhaps the 1811-12 fault movement modified stresses and stimulated fault movement in those areas. The 1843 earthquake cracked walls. felled chimneys, and broke windows in Memphis, Tenn. The 1895 earthquake damaged many buildings in Charleston and caused sand to liquefy and erupt onto the land surface. At Cairo, Ill., buildings swayed, chimneys toppled, and church steeples twisted.

Damaging earthquakes also have occurred in the region surrounding the area of greatest 1811-12 activity, including southern Illinois and Indiana, western Kentucky, and eastern Missouri. The strongest earthquake in the Central United States since 1895 occurred in 1968 in south-central Illinois on the margin of the New Madrid seismic zone. It had a magnitude of 5 .4 and was felt over all or portions of 23 States, from Minnesota to Florida and from North Carolina to Kansas. Damage consisted primarily of bricks thrown from chimneys, broken windows, toppled television antennas, and cracked or fallen plaster. A magnitude 5.2 earthquake in 1980 centered in north-central Kentucky was felt over 15 States. Property damage exceeded $1 million at Maysville. Ky., where 37 commercial structures and 269 residences sustained some damage. More recently, in 1987, a magnitude 5.0 earthquake in southeastern Illinois was felt in 21 States and southern Canada and caused one injury and minor damage in southern Illinois and Indiana.

Information on the underlying cause of New Madrid seismicity is derived from earthquake focal mechanisms and other stress indicators, which show that the New Madrid seismic zone is being compressed in an east-northeast-westsouthwest direction (Zoback and Zoback, 1981). This orientation is consistent with the direction of drift of the North American plate with respect to the mantle and with the direction of stress caused by pressure on the plate as it moves away from the mid-Atlantic ridge. Although these explanations for the origin of the stresses in the crust of the North American plate are plausible, they do not explain why earthquakes are concentrated in the New Madrid seismic zone.

Regional Tectonic Setting

The New Madrid seismic zone, which geologically is situated in the upper Mississippi embayment (fig. 4). is an area of abundant seismicity that extends generally from near Charleston, Mo., on the northeast to Marked Tree, Ark., on the southwest (fig. 5). The zone of intense seismicity lies within a larger region, including parts of Arkansas, Missouri, Illinois. Indiana, Kentucky, and Tennessee, in which earthquakes are more dispersed and less common. Major structural and tectonic elements in the New Madrid region include the Ozark uplift. the southern part of the Illinois basin, the Cottage Grove-Rough Creek-Shawnertown fault system, the Ste. Genevieve fault. the Wabash Valley fault system, and the Reelfoot rift.

The most seismically active part of the New Madrid seismic zone lies within the Reelfoot rift (fig. 5), a northeast-striking graben about 40 mi wide and 200 mi long (Hildenbrand, 1985). Limited geologic evidence indicates that the graben formed about 500 million years ago and was filled with thousands of feet of sandstone, shale, limestone, and dolomite. Geophysical surveys, particularly aeromagnetic and seismic-reflection data, show that the top of crystalline basement rocks is generally about 1.2 mi deep outside the rift but is 2.4 to 4. 2 mi deep along the axis of the rift. A very thick section of sedimentary rocks originally filled an ancient basin along the rift axis. The basin was subsequently uplifted to form the Blytheville arch (Hamilton and McKeown, 1988). The well-defined, northeasttrending zone of earthquake hypocenters between Marked Tree and Caruthersville, Mo., coincides with the Blytheville arch. Near Caruthersville. the pattern of intense seismicity changes to a northwesterly trend, w;.ich coincides with the intersection of the Blytheville arch and the northwest-striking Pascola arch (Grohskopf. 1955). The areal extent of the Pascola arch cannot be ma'1ped adequately from the available subsurface data, but seismic-reflection and sparse drill-hole data show that tl'~ Pascola arch also was formerly part of a deep sedimentary basin. Like the Blytheville arch, this deep basin is now 2 structural high but underwent severe erosion that removed several thousand feet of sedimentary rocks.

Seismic-reflection data show that the boundaries of the Reelfoot rift are marked by fault zones several miles wide that displace basement rocks by as mud' as 1 mi vertically and, in some places, also displace younger rocks several tens of feet. A few earthquakes are spatially associated with the rift-bounding faults, especially along the southeastern margin (fig. 5). The po~bility that the southeastern margin of the rift could produce a strong earthquake is of major concern. because such an event could be within 20 mi of Memphis.

The dispersed seismicity outside the area of intense seismicity in the New Madrid seismic zone has net yet been linked definitely to any major fault zones. In so·~theastem Missouri, the topographically and structurally highest part of the Ozark uplift-the St. Francois Mountains-is encircled by seismicity. This seismicity and geomo':phic evidence suggest that minor uplift may be occurrjng in the Ozarks.

Similarly, a crude spatial correlation between seismicity and the Ste. Genevieve fault raises the pmsibility of active deformation on that fault. Reconnaissance studies of the geomorphology of stream terraces across the fault, however, show no evidence of current deformation. Along the Cottage Grove-Rough Creek-Shawneetown fault system, there is no obvious correlation between modem

seismicity and the major faults. Along the Wabash Valley fault system, a crude spatial association between the fault system and modem seismicity may be evidence that some of the faults are active. Seismologic data from magnitude 4 to 5 earthquakes in the area show that the style of faulting is predominately strike-slip on northeast- or northwesttrending faults in response to a nearly east-west maximum horizontal stress orientation.

')---MISSISSIPPI -- ~-- -.---\ ALABAMA

so-

v -~..·

In summary, the tectonic framework in the New Madrid region beyond the New Madrid seismic zone is only poorly known. Modem seismicity cannot be related to specific geologic features except in a general way. Until a cause-and-effect relationship is developed, the earthquake potential of the entire region cannot be reliably assessed. Preparedness and mitigation measures cannot t·:- effected fully until such assessments are made.

,_J

k

Lake miles; the most notable area of subsidence was at Reelfoot Lake in northwestern Tennessee. Several islands in the Mississippi River completely disappeared because of flowage of liquefied sediments. Along the bluffs on the eastern side of the valley, landslides were commonplace from near Cairo, Ill., to south of Memphis (fig. 6). Throughout the region, fissures formed at the ground surface, river banks caved hundreds of miles from the epicentral area, and huge quantities of sand and water erupted onto the surface (fig. 7). The eruption of sand and water, which resulted from liquefaction of subsurface sand layers, flooded approximately 4,000 mi with as much as 3 ft of sand and water. More than 250 mi of timber were destroyed by flooding, violent ground shaking, or landsliding. Eyewitness accounts describe the entire land surface as being disrupted and, in many places, uninhabitable.

Accounts of the shaking caused by the 1811-12 earthquakes (Fuller, 1912; Penick, 1981) describe almost unbelievable ground motion. Godfrey Le Sieur, a young boy and resident of Little Prairie (near Caruthersville, Mo.) at the time of the earthquakes, later wrote,

James Audubon, the famous naturalist, experienced one of the shocks while riding in Kentucky:

L. Bringier, who lived with Indians near New Madrid and was later a surveyor in New Orleans, wrote that the water forced its way through the surface deposits,

The few structures that were present in the area were heavily damaged. Masonry and stone structures were damaged as much as 150 mi away. Chimneys were destroyed in Louisville, Ky., about 250 mi away, and less extensive chimney damage was reported at distances of over 400 mi. The earthquakes were felt southward to the Gulf Coast, southeastward to the Atlantic shore, and northeastward at least to Quebec. No reliable reports to the west are documented.

If estimated isoseismal areas to the west a·e included, the earthquakes of December 16, 1811, and February 7, 1812, had the greatest potential damage and thr. largest felt areas known in the earthquake history of the United States and possibly of the world (Johnston and Kanter, 1990) (fig. 2). The area of potential damage (the area shaken at intensity VII or greater) was about 250,000 mi• For comparison, a reasonable extrapolation of the area of intensity VII or greater for the 1964 Alaska earthquake covers an area of about 80,000 to 100,000 mi. The 1906 San Francisco earthquake affected an area of only about 12,000 mi at intensity VII (or 24,000 mi, if isoseismal symmetry to the west in the Pacific Ocean is assumed). Thus, the area of strong shaking associatei with the 1811-12 earthquakes is two to three times larg~r than that associated with the 1964 Alaska earthquake and 10 times larger than that associated with the 1906 San Francisco earthquake.

Potential Losses

Preliminary estimates of potential earthquake losses for dwellings in the Central United States are very large. Analysis of the intensity patterns of the 1811-12 earthquakes and other earthquakes in the Central United States in conjunction with the present-day distribution cf dwellings gives an upper-bound loss estimate of about $:0 billion to dwellings (in 1980 dollars) for a recurrence of the 1811-12 sequence (Algermissen, 1990). The expected maximum loss to dwellings in a 50-year period (at a 10-percent chance of being exceeded) is about $5.6 billion (1980 c.ollars), and a magnitude 6 to 7 earthquake would cause ap:'Jroximately $3.6 billion in dwelling losses (S.T. Algermissen, unpublished data, 1990; 1990).

The central Mississippi Valley is a majo~ communication and transportation corridor. Communication facilities, such as radio and microwave towers and telephone trunk lines, are vulnerable. Barge traffic on the river, natural gas and crude oil pipelines, interstate highways, and power lines all provide essential services, the loss of which would have a heavy impact on the entire Central and Eastern United States These transportation and lifeline facilities are highly vulnerable to earthquake damage, particularly because their structural integrity depends, to a great extent, on the stability of the ground. De~truction of pipelines would severely disrupt the delivery of ~nergy fuels to the northern and northeastern parts of the Urited States.

The Mississippi River would be strongly affected by a large earthquake. After the February 7, 1812, earthquake, two waterfalls or heavy rapids formed in the rive· and halted traffic for several days. Extensive changes in tte course of the river channel made navigation difficult and hazardous. Eliza Bryan, a resident of New Madrid, de:"cribed the effects of the 1811-12 earthquakes on the river (Fuller, 1912):

j

l ______ _

At first the Mississippi seemed to recede from its banks, its waters gathered up like mountains, leaving boats high upon the sands. The waters then moved inward with a front wall15 to 20 feet perpendicular and tore boats from their moorings and carried them up a creek closely packed for a quarter of a mile. The river fell as rapidly as it had risen and receded within its banks with such violence that it took with it a grove of cottonwood trees. A great many fish were left upon the banks. The river was literally covered with wrecks of boats.

Caving of steep banks along rivers and streams was the most widespread type of landsliding triggered by the 1811-12 earthquakes. According to witnesses (Penick, 1981),

In many places, the banks of the river ... sunk hundreds of acres together, leaving the tops of the trees to be seen above the water. [Banks] fell in "large columns"; in some places "five, ten, and fifteen acres ... sunk down in a body .... "

In a future large earthquake, barge traffic would almost certainly be interrupted, perhaps for weeks, if substantial dredging were needed to clear river channels. Foundations of structures and levees could fail in numerous places. Facilities along the river bank, particularly large grain silos and fuel and fertilizer tanks, could collapse and cause fires and spills that would pollute the river. Damage to landfills, wastewater-treatment facilities, pipelines, and chemical storage facilities could cause widespread contamination of ground water and disruption of water-supply systems.

Damage to the interstate and other highway systems would be extensive in a repetition of a great earthquake. The various forms of ground failure (liquefaction, landslides, and so on) would weaken or destroy highway foundations. Ground failure combined with strong ground shaking would severely impact bridges and overpasses, which generally have not been designed with earthquakes in mind. The bridges across the Mississippi River and its numerous tributaries are located on ground that is most susceptible to failure and amplified shaking.

The central Mississippi Valley is one of the most productive agricultural areas in the United States. Not only would farming be disrupted by general earthquake damage, but the extensive canal system that drains water from lowlands also could be damaged both by liquefaction and landslides and by uplift and subsidence of the land surface associated with a great earthquake. Large areas could be inundated by several feet of water, and fields could be covered by liquefied sand vented to the ground surface. Electric power failures could also affect pumping and other equipment systems.

An earthquake in the magnitude 6 to 7 range, such as the one that occurred near Charleston, Mo., in 1895, would also cause severe shaking, but the area affected would be smaller than the area that would be affected by repr.tition of one of the great shocks of the 1811-12 series. Hurdreds of square miles near the epicenter would experience ground motions strong enough to jeopardize many buildings and other facilities; within an additional tens of thou~::t.nds of square miles, shaking would be strong enough to cause damage to structures located on unfavorable geologic foundation material. An example of the effect of a modem earthquake of similar magnitude in a different geologic setting is provided by the San Fernando, Calif., ea'thquake of 1971 (magnitude 6.6) (intensity pattern in fig. 2). That shock killed 58 people, caused a near-catastrophic dam failure and collapses of freeway overpasses, and produced $500 million (1971 dollars) in damages. We have not had a modem magnitude 6. 8 earthquake in a region similar to the New Madrid seismic zone. The 1895 shock occurrr.d before construction of the numerous facilities that are nov' present in the area.

A magnitude 6 earthquake beneath a met~opolitan center would cause substantial damage. Recent examples of such a shock outside of the New Madrid seismic zone are the Whittier Narrows, Calif., earthquake of 1987 (magnitude 5. 7), which killed 8 people and caused propetty losses of over $300 million (1987 dollars), and the magnitude 5.5 Newcastle, Australia, earthquake of December 27, 1989, which killed 12 people and caused $1 billion in property losses. Unreinforced masonry buildings, of which there are many in the metropolitan areas of the New Madrid seismic zone, were particularly hard hit in the Whittier Narrows shock. The region of intense shaking in a magnitude 6 earthquake is, however, small enough that there are areas of the New Madrid seismic zone in which such an errthquake could occur without strong shaking extending to heavily populated areas.

THE RESPONSE: PREPAREDNESS AN[' MITIGATION

Preparedness activities must be initiated long before an earthquake to prepare emergency managers for the response and recovery periods. Such activities encompass the following:

• Establishing a process to develop emergency
• Utilizing information gained from scientific and
• Developing effective programs to disseminate infor-

Mitigation activities, like preparedness activities, also are carried out before the event, their goal being the

LOSS REDUCTION MEASURES

Figure 8. Goals of community actions in implementing loss reduction measures. (Figure preparec by W.W. Hays, U.S. Geological Survey, Reston, Va.)

response and recovery plans that are based on the best available earth-science and engineering data.

engineering studies and from hazard and risk assessments.

mation and educate the public on the potential vulnerability of the community in such a way as to stimulate real action and accomplish actual changes in structural and nonstructural mitigation as well as actual public preparedness on both an individual and a community basis. reduction or prevention of damage and societal disruption. They include the following:

• Developing realistic estimates of pot~ntial losses
• Reducing vulnerability in the community.
• Utilizing seismic zonation; that is, identifying the
• Adopting and implementing codes and Handards for
• Adopting and implementing criteria fo':" the siting,

Earth scientists, engineers, and social scientists build the knowledge base that practitioners use in preparedness and mitigation activities. Earth scientists and engineers evaluate the physical nature of the earthquake hazards of ground shaking, earthquake-induced ground failure, surface-fault rupture, regional tectonic deformation, flooding from dam failure, fire following an earthquake, and the aftershock sequence. Their goal is to understand the physand societal impacts for one or more possible events.

spatial distribution and nature of specific types of hazards and enacting land-use restrictions in hazardous areas.

the siting, design, and construction of nr.w buildings and lifelines (for example, energy, water, transportation, and communication systems).

design, and construction of essential and critical facilities (for example, schools, hosp:tals, emergency command centers, and conve"'tional and nuclear powerplants) that are vital to th~ life of the community and must remain functioral after an event. ical system for each type of earthquake effect, the· parameters that control the cause-and-effect relationships of the physical system, the central tendency and variability in space and time of each parameter, and the sensitivity to extrapolation of parameters beyond the limits of the data.

Social scientists study the social components of earthquake hazards. They analyze societal systems and focus their research on the behavior of individuals, households, organizations, and communities. Within this framework of individual and organizational behavior, social scientists analyze how people and institutions respond to, prepare for, and mitigate earthquake and related hazards. Social scientists are, for example, interested in how perception of risk is communicated to vulnerable communities. In addition, some social scientists concentrate on questions of policy development or policy analysis. For example, they might analyze how technical and nontechnical information is transferred from researchers to practitioners or how social, political. and (or) cultural factors constrain or promote the utilization of hazard-reduction information.

Experience has shown that knowledge alone makes no contribution to earthquake preparedness and mitigation measures if the knowledge base is unknown, misunderstood, inappropriate, unintelligible, misdirected, or ignored. Consequently, a vigorous awareness and education program would be needed to support earthquake-hazard mitigation and preparedness activities in the New Madrid region.

Introduction

Knowledge of the New Madrid seismic zone lags behind that of most seismic zones in the Western United States, particularly the San Andreas fault zone, for several reasons:

• The seismically active faults in the main part of the
• Earthquakes occur more frequently in California.
• The San Andreas fault zone is the boundary between

New Madrid seismic zone are concealed by as much as 3,000 ft of poorly consolidated sedimentary rocks, whereas most of the San Andreas fault zone is exposed at the surface.

This frequency thereby yields more data that are critical to understanding their cause, poses a greater danger, and attracts greater scientific and political attention. Consequently, resources devoted to earthquake research are considerably greater in the West than they are in the East.

the Pacific and North American plates, a relatively well understood tectonic regime. In contrast, the tectonic framework for .. intraplate" earthquakes in the New Madrid region (and worldwide) is poorly understood. Thus, much research is needed in the New Madrid seismic zone to achieve a level of knowledge comparable to that achieved in California.

Although much remains to be done, significant progress in understanding the New Madrid seismic zone has already been achieved. As recently as the early 1970's, before the formation of NEHRP, the pattern of ea'ihquake epicenters in the region was poorly resolved; it looked like a buckshot pattern from a shotgun blast. No geologic explanation accounted for New Madrid seismicity. Following the installation of the seismic network, the dispersed epicenter pattern was resolved into several well-defined, linear trends that correspond to the locations of buried fault zones, and most of the earthquakes can now be attributed to the reactivation of faults in the Reelfoot rift. Many of the faults in the rift have been identified on seismic-reflection profiles, and the amount of vertical movement has been measured on some of them. Exploratory trenches have been dug in several places where faults extend close to the ground surface and, in one place, have yielded esf'llates of the rates of fault movement.

Some of the most important questions that cannot yet be satisfactorily answered are:

• How can New Madrid seismicity be exp~::tined in terms of global tectonic processes?
• What are the rates and modes of crustal deformation?
• Do the main seismogenic structures extend northeastward into southern Illinois and Indiara, or are the earthquakes in these areas caused by movement on structures unrelated to the Reelfoot rift?
• How far does the New Madrid seismic zone extend southwest into central Arkansas?
• How frequently do great (magnitude 8) anrl smaller destructive earthquakes occur? Is their o~currence periodic, quasi-periodic, or episodic? Ove~ the long term (centuries or millennia), does seismicity migrate to other intraplate areas?
• On which fault segments of the New Madr~d seismic zone and surrounding areas are the next d ~structive earthquakes likely to occur?
• What will be the effects of future earthquakes on the poorly consolidated ground in the Mississippi Valley?
• What is the potential for amplification of t"-e ground motion and for ground failure? The answers to these questions not only will improve

the basis for mitigating earthquake hazards in the New Madrid region but also will provide a better understanding of earthquakes elsewhere in the Central and Eastc'n United States. Moreover, results from the New Madrirl seismic zone will improve the general understanding of intraplate earthquakes and thereby help to reduce casualties and damage from earthquakes in similar environments worldwide.

Purpose

The overall purpose of an intensified study of the New Madrid seismic zone, which is consistent with the general mission of NEHRP, is to reduce casualties and damage from earthquakes through improved estimates of seismic risk and implementation of earthquake-hazard preparedness and mitigation measures.

Goals

Five goals have been identified for an intensified study of the New Madrid seismic zone:

1. Implementing earthquake-hazard mitigation
2. Improving preparedness for earthquakes of
3. Establishing a modern seismic network in the
4. Locating faults that could cause destructive
5. Improving seismic-risk assessments.

Activities

The activities that have been identified to achieve the goals of an intensified study of the New Madrid seismic zone are listed below. Activities associated with implementation of mitigation and preparedness and those associated with research should proceed concurrently. Each incremental improvement in implementing earthquake-hazard mitigation and preparedness measures would have a corresponding benefit in terms of loss reduction.

In this discussion of activities, some goals have been subdivided into objectives to show more clearly how certain activities relate to different aspects of a goal.

Goal1: Implementing earthquake-hazard mitigation measures

a. Facilitate adoption of structural design codes that measures.

magnitude 6 or larger.

New Madrid seismic zone to monitor the sizes, locations, and characteristics of the earthquakes and to determine the nature of the ground motions that they generate.

earthquakes, determining the recurrence rates of earthquakes, and delineating areas of potential damage.

include appropriate provisions for soil effects and earthquake resistance and of land-use policies that take into account earthquake hazards.

b. Develop effective and economical to.chniques to

c. Increase education and communicatio'l concerning

Four major research needs have been identified to implement earthquake-hazard mitigation mea"ures in the Central United States (Hanson, 1986): (1) assess the capability of existing buildings to withstand earthquakes, (2) verify the techniques for repairing, strengt'Iening, and retrofitting structures, (3) implement remedial measures, and (4) develop construction methods for enh2ncing earthquake resistance. Each need requires reevaluation of existing methods and development of new procedures drawing on experience in regions of high seismicity. Because resources for strengthening and repairing existing structures and for constructing new buildings are limited, earthquake mitigation practices in regions of moderate seismicity may vary from those developed for regions of high seismicity.

To meet these needs, research and development work is required in three general areas: (1) behavior of structures in response to earthquake shaking, (2) nondest~uctive testing of structures, and (3) building codes. Subs.. antial work needs to be done in evaluating the earthquake response of structures that may have little earthquake re"istance. A feasible, cost-effective solution for these structures in the New Madrid seismic zone entails more than simply modifying building codes that have been calibrated frimarily for highly seismic regions. New procedures for nondestructive testing and new design and evaluation rrethods are required, and they must be communicated to a professional community that has not dealt extensively with earthquake risk. For the Central United States, assessment of the earthquake hazard is a very critical issue. An overly conservative estimate of the hazard could result in an overestimation of the cost of solving the problem, which in tum could result in nothing being done to reduce the risk.

Goal 2: Improving preparedness for earthquakes of magnitude 6 or larger

a. Identify, assess, and retrofit, if neces"ary, those

b. Prepare loss estimates for large cities.

c. Prepare earthquake response and recovery plans.

d. Provide training tailored to the needs of those who facilitate repairing, strengthening, and retrofitting unreinforced masonry and other vulnerable buildings and structures.

earthquake risk and earthquake-hazards preparedness and mitigation measures.

structures and facilities having special importance to survival and recovery in the event of an earthquake.

respond to emergencies.

Although earthquake preparedness in the New Madrid seismic zone lags behind that in some of the western earthquake zones, some important progress has been made. The Central U.S. Earthquake Preparedness Project (CUSEPP) was initiated by FEMA in 1982. CUSEPP's

intensity of effects for selected earthquakes in the New Madrid seismic zone, (2) inventory structures, lifelines, and critical facilities in selected cities in the Central United States, and (3) assess the risk in those cities. Little Rock, Ark., Carbondale, Ill., Evansville, Ind., Paducah, Ky., Poplar Bluff, Mo., and Memphis, Tenn., were selected for these assessments (Federal Emergency Management Agency, 1985). The long-term goals ofthe project are to (1) increase the awareness of public officials and the private sector of earthquake hazards, (2) develop options for engineering solutions based on the balance of risk and cost, (3) accelerate the implementation of earthquake-hazard mitigation and preparedness strategies, and (4) improve earthquake response plans.

To facilitate its earthquake preparedness efforts, FEMA formed the Central U.S. Earthquake Consortium (CUSEC) in 1983. CUSEC consists of representatives of the seven States that are most vulnerable to damage from earthquakes in the New Madrid seismic zone: Arkansas, Illinois, Indiana, Kentucky, Mississippi, Missouri, and Tennessee. The goal of CUSEC is to ensure a coordinated program for achieving earthquake preparedness and mitigation goals common to all seven States. CUSEC developed and encouraged the adoption of a multistate-interstate compact, which allows neighboring States to share resources during an earthquake emergency in fields such as medicine, engineering, law enforcement, and fire fighting. The compact addressed liability protection for emergency personnel who render services in States where they are not licensed, registered, certified, or insured.

Goal 3: Establishing a modern seismic network in the New Madrid seismic zone to monitor the sizes, locations, and characteristics of earthquakes and to determine the nature of the ground motions that they generate

a. Upgrade the present regional seismic network (fig.

9) with modem three-component, broadband,

high-dynamic-range instruments. The network would cover approximately the area of intensity VII shaking shown in figure 2. b. Integrate the regional earthquake monitoring network with the U.S. National Seismic Network (fig. 10) and establish a common data transmission system. c. Upgrade and extend the spatial extent of the seismic coverage on the periphery of the New Madrid seismic zone.

a. Improve the current regional network of strongground-motion stations (fig. 11). b. Establish two field laboratories in areas of intense seismicity in the New Madrid seismic zone to acquire data on earthquake characteristics near the source and seismic-wave propagation, including strong ground motion. (Each array would consist WI CONSIN '-...,·-·"2S:':

\ _______ _ MISSOURI

telephone. Data would be relayed to research Figure 10. U.S. National Seismic Network (USNSN) stations in the Central United centers through the USNSN master station in Golden, Colo., by means of satellite antenStates. The area affected by shaking of intensity VII or greater during the 1811-12 New nas; this concept is illustrated by antenna symbols at Memphis, Tenn., and Saint Louis, Madrid earthquakes is shown by the solid line (dashed where there are no data). The area Mo., where the largest regional network covered by the regional seismic network is operators are located. (Figure prepared by R.E. Needham, U.S. Geological Survey, indicated by the dash-dot line. The regional Golden, Colo.) stations would be linked to satellite communication nodes of the USNSN by radio or

f-.------------, +

ARKANSAS

OPERATOR

U.S. Army Corps of Engineers Department of Energy National Center for Earthquake Engineering Research Center for Earthquake Research and Information University of Kentucky Tennessee Valley Authority

TENNESSEE

E8

r8J e. Develop a small, portable seismic network and

f. Deploy semiportable seismographs to study

a. Establish research centers to conduct and monitor

b. Upgrade USGS capabilities to support earthquake

Although the present seismic network is crucial to improving our knowledge of earthquakes in the New Madrid region, it now lacks a funding base and has major technical limitations that must be corrected to yield further advances. The current network was designed 15 to 20 years ago, when its goals were only to locate and determine the magnitudes of small earthquakes and to obtain a focal mechanism for a few larger ones. Today, more information must be obtained from the earthquake recordings to answer the complex questions related to earthquake hazards. The majority of seismometers in the network (fig. 9) do not sense the horizontal component of ground shaking, have limited frequency response and limited dynamic range, and are not installed in boreholes, where sensitivity would increase owing to the reduction of surface noise.

Horizontal-component seismometers would provide data to determine earthquake locations more accurately and would, for the first time, provide data on the transverse be stationed in the New Madrid seismic zone to implement this recommendation.)

provide for its rapid deployment to monitor aftershocks. (This network would consist of eightthreecomponent seismometers connected to a regional seismic-network satellite processor and would be used to define site-specific ground-motion characteristics.)

seismic-wave propagation to larger distances. (Selected studies could be conducted at St. Louis and Louisville, for example.)

seismic network and other instrument operations, locate earthquakes, and archive waveform and other earthquake data for future use. (These data centers, which would have separate functional responsibilities, would provide basic data to all research scientists for earthquake-hazard studies and would maintain the earthquake-monitoring networks.)

monitoring systems and to utilize the data in earthquake-hazard assessment projects (National Academy of Sciences, 1990). (The USGS projects would include support services to link the regional seismic network and the USNSN, acquisition of data for common use, and interpretation.) waves that are the main cause of earthqu"'ke damage. Without better locations, including accurate depth estimates, the geologic structures causing the earthquakes cannot be reliably identified. In addition, the~e data would be used for research on the shear modulus cf crustal and surficial materials, a property that characterizes the mineral composition and fluid content of rocks and can be used to understand the rupture process of earthquakes. The data would also permit studies of material anisotropy (for example, seismic waves traveling at different velocities in different directions) that could be used not only to map fault zones but also to monitor the stress buildup fo"" future large earthquakes. Finally, the horizontal data p-ovide better estimates of the attenuation of ground shaking with distance and are necessary for accurate determination of earthquake focal depths.

A major deficiency of the current seismic instruments is that they record earthquakes of about magnitude 3 and smaller on scale, but larger earthquakes caus~~ them to go off scale. The current instruments are set this way to sense the numerous small earthquakes that are used to delineate active faults. The tradeoff, however, sacrifices recording the complete waveform for the few larger (magnitude 4+) earthquakes. Thus, no usable near-source reccrd of ground motion exists for larger earthquakes in the New Madrid seismic zone that can be used to define the expected shaking in metropolitan areas during major earthquake". This deficiency can be resolved by using modern instruments having wide-dynamic-range digital telemetry and recording.

Several metropolitan areas and many critical facilities are located on poorly consolidated sediments in the New Madrid seismic zone. The prediction of ground motion at the surface is calculated from ground motion incident at the base of these sediments. Thus, for reliable estimates of ground motion at the surface, it is necessary to install seismometers at the base of the poorly consolidated sediments and also at the surface. The thickness of poorly consolidated sedimentary rocks in the most densely populated parts of the New Madrid region is as gr~at as that in any earthquake zone in the country (as much as 3,000 ft). Ground-motion prediction methods developed for California need to be calibrated for' this kind of geologic setting.

Goal 4: Locating faults that could cause d£~tructive earthquakes, determining the recurrence rC'tes of earthquakes, and delineating areas of potential damage

a. Expand geologic and geophysical studies; trace the areal extent of geophysical features at depth.

b. Conduct geologic mapping, including shallow

c. Acquire additional high-resolution and reconnais-

d. Identify deep faults that extend to the surface by

e. Drill one or more research holes to determine the

Accurately defining and characterizing seismic source zones require a clear understanding of the relationship between earthquakes and major geologic structures at the depths where earthquakes are generated. Subsurface geologic and geophysical studies conducted during the past 15 years in the New Madrid seismic zone have shown a strong correlation between seismicity and specific fault zones associated with the Reelfoot rift.

Geologic investigations of earthquakes in the New Madrid seismic zone have focused primarily on establishing the relationship between seismicity and specific structures within the Reelfoot rift. Subsurface geologic and geophysical data have been invaluable in identifying and characterizing the major seismic source zone along the axis of the rift. although a comprehensive, regional earthquake-hazard assessment must include information on the earthquake potential of other major faults both inside the rift and outside it, knowledge of which is now completely lacking. For example, in the past 100 years, five earthquakes of magnitude 4.5 to 5.5 have occurred in the Wabash Valley seismic zone of southern Illinois and southwestern Indiana. A magnitude 5.0 earthquake in southern Illinois in June 1987 was felt in 21 States and southern Canada. The historical record of the Central United States also shows a relatively high level of seismicity along the eastern and southeastern margins of the Ozark uplift. Thus, regional seismicity shows that the scope of deep structural studies should be expanded beyond the bounds of the Reelfoot rift to encompass important structural features such as the Wabash Valley fault zone, the Cottage Grove-Rough Creek fault zone, the Ste. Genevieve fault zone, and the Ozark uplift (fig. 3). Studies are needed to establish a geologic and geophysical basis for defining seismic source zones throughout the region and to determine the potential for damaging earthquakes in each zone.

Expanded regional studies will require that critical gaps in existing aeromagnetic and gravity data be filled. Aeromagnetic and gravity data acquisition was originally drilling to determine the physical properties of subsurface materials.

sance geophysical data (gravity. magnetic, seismic reflection) to fill critical gaps in existing data sets.

means of seismic-reflection and other geophysical data.

physical properties of geologic materials under realistic conditions, to calibrate geophysical data, and to permit borehole monitoring of seismogenic processes. designed to identify broad-scale structural features such as major tectonic blocks, shallow igneous intrusions, and anomalous zones of the crust in the northern Mifsissippi embayment. High-resolution aeromagnetic and graYity data sets that uniformly cover the rift and the surrounding region are now needed. These data are vital to correctly interpret the structural relationships of features at seismogeni':' depths and to resolve details about the structures in the upper crust in the area north of the Reelfoot rift. In addition, selected areas within the rift need specialized geophysical investigations. For example, magnetotelluric surveys in the Reelfoot rift could provide detailed information about s•Tuctural relationships in the middle and lower crust in the rift and help determine the physical properties of a low-velocity, high-attenuation zone along the center of the rift where earthquakes are concentrated. Similarly, detailed a~romag netic and gravity profiles could refine geophysical models and clarify the interrelationships among plutors, deep basins, fault zones, rift boundaries, and seismicity concentrations.

Detailed investigations of deep structure in the rift should continue as the new studies outside the Redfoot rift begin. Faults associated with the rift probably constitute the greatest seismic threat to the region. Therefore. detailed studies of structures that are apparently associated with seismicity, such as the Blytheville and Pascola arches. are imperative. Such studies would utilize a variety of geophysical techniques to define the extent and understand the origin of major structures in the rift that are associated with seismicity. Mapping these structures would re'luire the acquisition of additional seismic-reflection data. An extensive petroleum-exploration program in the Reelfoot rift in the early 1980's included the collection of almost 4,000 mi of seismic-reflection profiles. These data cover many of the most active parts of the New Madrid seismic zore and are available for purchase, but new data would have to be acquired over the Pascola arch and some other parts of the zone to improve understanding of the relationship between structure and seismicity.

Recent work in the New Madrid seismic zone has suggested that the crust in the most active part of the region is characterized by low seismic-wave velocities and high seismic-wave attenuation (Al-Shukri and Mitchell, 1987, 1988; Hamilton and Mooney, 1990). These properties may provide a criterion for evaluating whether a fault zone is capable of producing large earthquakes.

To date, very little effort has been made to evaluate the seismogenic potential of the major faults along the margins of the Reelfoot rift. If the rift-bounding faults are seismogenic, then Memphis would be only a few tens of miles from the epicenter of a potential large earthquake. Densely populated Shelby County, Tennessee, would lie even closer, and West Memphis, Ark., a community of 30,000, actually straddles the southeastern rift margin. Seismic-reflection data and detailed geophysical data are essential in evaluating the earthquake potential of major faults related to the rift margins.

To assess the seismogenic potential of deep faults that appear to extend close to the surface on seismic-reflection profiles, detailed studies using high-resolution seismic-reflection profiling and exploratory trenching, if appropriate. should be made. These studies would determine whether the faults deform young sediments near the surface where the resolution of conventional reflection data diminishes and thus constrain the likelihood of damaging earthquakes in specific source zones.

Because much of the New Madrid seismic zone is covered with soft sedimentary rocks, deep scientific drilling offers the only means of directly sampling and measuring the physical properties of the bedrock to depths that approach hypocentral depths. Better information on the properties of these rocks will contribute to improved interpretations of aeromagnetic, gravity, and seismic data. This information also will help determine the factors that affect the way seismic energy is attenuated as it propagates through these rocks in the uppermost crust. A deep drill hole would permit direct measurement of the orientation and possibly the magnitude of the modem stress field in the New Madrid region, which is the driving force for the seismicity. Much of the data and information derived from a deep drill hole are unique and cannot be obtained in any other way.

a. Continue and expand efforts to identify and determine the age of prehistoric earthquakes by studies of paleoliquefaction, surface faulting, landslides, b. Document land deformation produced by the

c. Identify deep faults that potentially extend to the

d. Determine the age of young sedimen+ary rocks on

e. Evaluate the status of geologic mapping and deter-

f. Conduct a comprehensive study of the fluvial

The accounts of the New Madrid earthquakes provide a fairly complete picture of the sequence of events in 1811-12 (Fuller, 1912), but the historical record offers no clues about the occurrence of strong earthquakes before 1811. Fuller (1912) briefly discussed five pre-1811 earthquakes, but his information is too sparse to determine even if they were Mississippi Valley events.

The historical record of earthquakes in nost parts of the world, including the United States, is too short to document the long-term behavior of seismogenic faults. Yet information on the locations and frequency of strong earthquakes in a region is vital to reliably assessing earthquake hazards. Because of the short historical reco:--d, geologic and archeological studies of prehistoric earthquake phenomena (paleoseismology) have played an increasingly important role in improving earthquake-hazard analyses and probabilistic hazard assessments. Thus, geolog::c studies on the locations, timing, and sizes of prehistoric earthquakes have successfully extended the historical seismicity record in many parts of the world.

Unlike many earthquake-prone places in the Western United States, the surface expression of young faults in the New Madrid seismic zone is very subtle to nonexistent. Geologic studies to determine the recurrenc~ of major earthquakes and the locations and nature of surface deformation in the region can contribute a great deal to hazard assessments, but these studies are difficult to conduct because of the subdued and concealed expressic n of tectonism.

The only fully reported geologic data on the recurrence of large earthquakes in the New Madric region are Russ's (1979) study of shallow faulting associated with the Reelfoot scarp, a 7-mi-long escarpment along the western margin of Reelfoot Lake in northwestern Tennessee. The 10- to 30-ft-high scarp is a prominent physiogra?hic feature in the Mississippi Valley, where the topographic relief is seismically induced disruption of lacustrine deposits, archeology, and dendrochronolo~y.

1811-1812 and earlier earthquakes.

surface by means of seismic-reflection and other geophysical data; target these faults for shallow, high-resolution, seismic-reflectio'l studies, ground-penetrating radar, and (or) trenching.

various alluvial terraces in the Missi~sippi Valley to improve understanding of rates of deformation.

mine if more or specialized mapping is needed.

geomorphology of the Mississippi Valley. otherwise very low. An exploratory trench across the scarp exposed numerous, small-displacement faults that are coincident with the scarp and revealed evidence of two strong earthquakes that predate the 1811-12 events (Russ, 1979). Stratigraphic relationships in the trench and radiocarbon dating show that both of the ancient earthquakes are less than 2,000 years old. Thus, the three strong earthquakes (1811-12 and the two older events) that have occurred at this site in the past 2,000 years imply an average recurrence interval of about 600 to 900 years for strong earthquakes but not necessarily for earthquakes of the size that occurred in 1811-12.

The geologic recurrence estimate of 600 to 900 years is in the range of estimates based on statistical analysis of historical and modem seismicity; these calculations yield recurrence estimates for maximum-magnitude events that range from about 200 years to more than 2,500 years but cluster around 600 to 1,200 years (Johnston and Nava, 1985, table 1). The wide range in the estimates from seismologic data and the paucity of constraints from geologic data demonstrate the need for greater efforts to determine the age of prehistoric earthquakes in the New Madrid seismic zone.

Geologic data on the timing of prehistoric earthquakes throughout the entire New Madrid region are needed not only to estimate the likelihood of future major earthquakes but also to understand the rate and processes of strain accumulation and release in the region. The most direct recurrence information is derived from studies of faults that have repeatedly ruptured the ground surface during successive events. However, other than the study of the Reelfoot scarp, efforts to identify tectonically significant surface faulting have been unsuccessful, in part, because it is difficult for displacements in competent bedrock to propagate to the surface through thousands of feet of soft sediment.

It is fortunate that geologic estimates of recurrence intervals need not rely solely on finding evidence of surface faulting. The 1811-12 events show that the strong ground motion and deformation from major earthquakes in the New Madrid region caused a variety of secondary geologic effects, including extensive liquefaction, earthquake-induced landslides, and uplift and subsidence. Ancient earthquakes surely produced similar effects that would be preserved in the geologic record. Identifying and dating these ancient effects would provide crucial information about the recurrence of strong ground motion- and thus large earthquakes-in the New Madrid seismic zone. Recurrence information is of such great importance that novel methods of dating prehistoric earthquakes need to be fully explored. For example, Reelfoot and Big Lakes are in the epicentral area of the seismic zone. Strong ground motion during the 1811-12 events may have disturbed the well-stratified sediments in these lakes, both of which probably were enlarged or even formed by the 1811-12 sequence. If sediment cores from these lakes show disruption from the 1811-12 events, deeper cores may reveal similar features caused by older earthquakes that C'Juld be dated.

Dendrochronology, the study of annual grow+h rings of trees, might also be useful in dating prehistoric earthquakes in the region. The strong ground shaking during earthquakes probably broke large branches, disturb~d root systems, tipped trees, and caused temporary changes in the water table, all of which stressed the trees. If this earthquake-induced stress were severe enough, it would have been recorded in tree growth rings, which could be dated. Very little old-growth forest remains; what is left should be systematically inventoried to determine the feasibility of dendrochronologic paleoseismic researd'.

Successfully obtaining recurrence information in the New Madrid seismic zone requires a strategy for locating the areas where evidence of prehistoric earthquakes is likely to be found. The areas that experienced substantial deformation in 1811-12 presumably suffered similar deformation during earlier events. Drainage patterns and the morphology of stream channels can be very sensitive indic"tors of tectonic deformation. Therefore, a first step in this strategy would be to analyze the fluvial geomorphology in the seismic zone to identify areas of significant deformation during 1811-12 and then to use this analysis as a basis for selecting sites for intensive investigations using techniques such as high-resolution seismic-reflection studies o.. exploratory trenching. An important related aspect of this geomorphic analysis would be to obtain better data on the age of alluvial and fluvial deposits in the Mississippi Valley to help determine the rates of deformation.

a. Identify areas of strain accumulation and quantify

b. Acquire information on the modem stress field

c. Drill one or more research drill holes.

Modem geodetic methods for determining rates of active crustal strain provide critical data about the occurrence of large intraplate earthquakes. The rate at which elastic strain energy is accumulating should be neasured; the relationship between such changes and earthquakes may provide a means for discriminating among regions where large intraplate earthquakes are likely to occur (and reoccur) and those where they are not. accumulated strain by using geodetic data.

that is independent of seismicity data (by using existing wells).

The opportunity to conduct this kind of research in intraplate seismic areas is facilitated by the presence of a comprehensive network of bench marks that have been accurately surveyed by triangulation techniques for many decades (in some cases, these networks extend back to the 19th century) and by the recent development of accurate, relatively inexpensive surveying methods using the Global Positioning System (GPS). By using GPS, it is now possible to measure crustal strain efficiently by resurveying old triangulation networks as well as by establishing new baseline data for future surveys.

One hypothesis frequently cited to account for intraplate earthquakes is that they occur in zones of extreme crustal weakness. To test this hypothesis, it is necessary to accurately map the orientation of the stresses in the Earth's crust. Although the ability to map the crustal stress field has improved markedly in the past decade, data coverage in areas such as the New Madrid seismic zone is quite sparse and completely inadequate to examine this hypothesis. It is now possible to obtain data on the orientation of the crustal stress field by making straightforward geophysical measurements in boreholes (for example, wellbore breakouts). These measurements could be made either in holes drilled for other purposes or in shallow holes drilled specifically for stress measurements. Such measurements would provide a greatly improved knowledge of the physical mechanisms responsible for intraplate seismicity. The advantage of wellbore breakout data is that they are independent of seismicity; the disadvantage is that they are confined to shallow crustal depths and generally do not sample the depths of principal seismogenic strain release.

Goal 5: Improving seismic risk assessments

a. Delineate seismic source zones.

b. Identify areas expected to experience strong

c. Identify areas having soil deposits that can amplify

d. Identify areas subject to ground failure.

e. Identify areas vulnerable to flooding from levee or

f. Inventory high-occupancy hazardous buildings.

g. Conduct analytical and experimental studies on

h. Identify procedures to reduce vulnerability

1. Evaluate the 1988 editions of the seismic design

ground shaking.

ground motion.

dam failure and eruption of water associated with liquefaction.

the earthquake response of engineered facilities.

through structural and nonstructural means.

provisions of the principal model building codes (that is, Standard Building Code, Uniform Building Code, National Building Code, and the j. Provide a data base for seismic zonation.

Assessment of seismic risk requires three elements: an earthquake-hazards model, an exposure model (inventory), and a vulnerability model. A generic earthquake-hazards model requires knowledge of ground shaking and its probability of nonexceedance, surface-fault rupture, earthquake-induced ground failure, and regional tectonic deformation. The exposure model, or inventory, involves the spatial distribution of population and various structures, including buildings, utility and transportatio.'l structures, hydraulic structures (dams, reservoirs, levees, and so on), and others. The vulnerability model is a means for predicting losses among the inventoried items. One element commonly used in a vulnerability model is a frr~ility curve, which shows the probability of damage versus the level of ground motion for a specific type of structur~. Predicting damage is very difficult for a variety of reasons: nonuniform design and construction, variability in material properties, changing design strength of aging structures, U'lcertainty in the level of ground shaking, uncertainty in structural response, and uncertainty in the response of seiiments that are susceptible to liquefaction, to name a few.

Improved techniques are needed for estimating seismic risk in the Central and Eastern United States because development of the current methodology was based primarily on experience in high-seismicity regions of the Western United States. More attention should be devoted to structural materials and systems that are no longer used in locations such as coastal California because of their perceived inherent low resistance to earthquake shaking but that are still widely used in the central and east~rn areas of the country. The response of these materials and systems to both strong and moderate levels of ground motion will greatly influence the pattern of losses in a future, strong New Madrid earthquake. Nondestructive testing would be necessary in evaluating the response of such stLictures.

Methods have been developed and apried in the Western United States to predict the probability and geographic extent of landsliding that would be t:iggered by earthquakes of given magnitudes and locations. Such methods could be applied to the New Madrid seismic zone but only after detailed investigation to (1) determin~ the distribution of susceptible slopes, (2) model the ground shaking that those slopes would experience, and (3) develop computer models to simulate how the slope mate:-ials would respond to the strong shaking. A predictiYe map of earthquake-induced landslide susceptibility developed in this way could be combined with an exposur~~ model to NEHRP Recommended Seismic Design Provisions) and recommend adoption of th ~ code that is most relevant.

CONCLUSIONS

4. Locating faults that could cause destructive earth-
5. Improving seismic-risk assessments.

Implementation of this five-point program would provide the information necessary to improve the reliability of seismic-risk assessments for the New Madrid region. The scope of activities envisioned under this program could be undertaken at a funding level of $5 to $10 million a year, for 5 years, after which time the status of earthquake-hazard information for the New Madrid region would be reexamined and the plan modified to reflect the improved status of each element. Coordination would be facilitated by establishin,a a New Madrid seismic zone coordinating committee, holding an annual workshop, and presenting results in publications and meetings.

REFERENCES CITED

tions, and characteristics of earthquakes and to determine the nature of the ground motions that they generate.

quakes, determining the recurrence rates of earthquakes, and delineating areas of potential damage.

APPENDIX:- NEW MADRID SEISMIC ZONE WORKSHOP PARTICIPANTS

Algermissen, S.T. (U.S. Geological Survey) Beavers, James (Martin Marietta Corporation) Bologna, Antonio (Memphis/Shelby County Building Code Enforcement Board) Braile, Lawrence (Purdue University) Branton, Fiona (U.S. Senate Committee on Commerce, Science, and Transportation)

Chang, T .S. (Memphis State University) Chin, Jer-Ming (Memphis State University) Chung, Wai-Ying (Memphis State University) Clarke, Timothy (Memphis State University) Crone, Anthony J. (U.S. Geological Survey)

Dorman, James (Memphis State University) Fanning, Harold (Reaves & Sweeney, Inc.) Guccione, Margaret T. (University of Arkansas) Hall, William (University of Illinois) Hamburger, Michael (Indiana University)

Hamilton, Robert M. (U.S. Geological Survey) Hanson, Robert (National Science Foundation) Hatcher, Robert D., Jr. (University of Tennessee) Hays, Walter W. (U.S. Geological Survey) Herrmann, Robert B. (St. Louis University)

Hester, Norman C. (Indiana Geological Survey) Heigold, Paul (Illinois State Geological Survey) Hildenbrand, Thomas G. (U.S. Geological Survey) Hill, William T. (Tennessee Department of Conservation) Howe, Warner (Gardner & Howe Structural Engineers)

Hughes, Terry (Memphis/Shelby Co. Code Enforcement) Hwang, Howard (Memphis State University) Jacob, Klaus H. (Lamont-Doherty Geological Observatory, Columbia University) Jibson, Randall W. (U.S. Geological Survey) Johnson, Gary D. (Federal Emergency Management Agency)

Johnston, Arch C. (Memphis State University) Kanter, Lisa R. (Memphis State University) Kernen, Michael (Tennessee State Representative) Kiefer, John (Kentucky Geological Survey) Krinitsky, Ellis (U.S. Waterways Experiment Station)

Lanelle, Kathleen Lew, H.S. (National Institute of Standards and Technology) Masse, Robert P. (U.S. Geological Survey) McCrary, Tom (Memphis State University) McFarland, John David (Arkansas Geological Commission)

McKeown, Frank A. (U.S. Geological Surv~y) McMullen, Richard (U.S. Nuclear Regulato"')'

Commission) Mitchell, Brian J. (St. Louis University) Murphy, Andrew J. (U.S. Nuclear Regulatcry Commission) Murphy, Loyal (P.E.)

Newton, Danny (Tennessee Emergency Management Agency) Nishenko, Stuart P. (U.S. Geological Survey) Obermeier, Stephen F. (U.S. Geological Su':'vey) Padovani, Elaine R. (U.S. Geological Surve:r) Palmer, James (Missouri Geological Survey)

Paul, James (U.S. House Committee on Science, Space, and Technology) Reiter, Leon (U.S. Nuclear Regulatory Commission) Russ, David P. (U.S. Geological Survey) Ryland, Harvey (Central U.S. Earthquake Consortium) Schweig, Eugene S. (Memphis State Univers~ty)

Sexton, John L. (Southern Illinois University) Sharrocks, Fred (Federal Emergency Manag~ment Agency) Shedlock, Kaye M. (U.S. Geological Survey) Stauder, William, S.J. (St. Louis University) Stevens, Jill (Memphis State University)

Stewart, David (Southeast Missouri State University) Street, Ron (University of Kentucky) Suiter, Lacy (Tennessee Emergency Managem~nt Agency) Sutter, John (U.S. Geological Survey) VanArsdale, Roy B. (University of Arkansaf) Vaughn, James (Missouri Geological Survey) Wesnousky, Stephen G. (University of Nevada, Reno) Wesson, Robert L. (U.S. Geological Survey) Whitcomb, James (National Science Foundation) Zoback, Mark (Stanford University)

Where this page came from

This page was imported from U.S. Geological Survey. Published by the U.S. Geological Survey and, as a work of the United States government, in the public domain.

Nobody has written it yet — it is the source material at a new address, which is why search engines are asked to skip it and why no one earns from it. It is up for grabs: take it on, and it is yours to rewrite and to earn from.