USGS Peak Ground Acceleration 2% 50 years
USGS Ground 1.0 sec PSA 2% 50 years
Modified Deep Soil Peak Ground Acceleration 2% 50 years
Modified Deep Soil Ground 1.0 sec PSA 2% 50 years
Specific geographic area: Mississippi Embayment
Scientific question being addressed: Are there seismic site effects that will amplify or deamplify ground motion in the frequency band of engineering interest at sites in the Mississippi Embayment?
General project plan: Analyze three-component data from regional earthquakes recorded at four sites in the Embayment: Meeman-Shelby State Forest, TN; Ridgely, MO; Cape Girardeau, MO; and Memphis, TN. Compute spectral ratios from floodplain/bluff sites to identify site effects. Model 1-D site effects using near-surface velocity models. Add site effects to seismic hazard maps.
Project duration: continuing
Most significant uncertainties: Are weak-motion-derived site effects appropriate for strong shaking? We need a velocity model for generic Embayment sediments below ~30-50 m and above the Paleozoic rocks.
return to CUS Project Summaries Index
Specific geographic area covered: entire U.S. and territories.
Scientific question being addressed: We produce national probabilstic seismic hazard maps and related products for use in building codes, loss estimation, land-use planning, etc. These maps involve the integration of information on earthquake sources (estimated locations, magnitudes, and recurrence times) with ground-motion attenuation relations.
General project plan: We are planning to produce updated national seismic hazard maps by the end of the year 2000. These maps will incorporate recent findings on earthquake sources and ground-motion attenuation.
Project duration: continuing
Most significant uncertainties: Recurrence times, locations, and estimated magnitudes of large earthquakes in the New Madrid region. Characterization of other potential sources of damaging earthquakes in the Central U.S. Regionally-specific ground-motion attenuation relations which include finite-fault effects.
return to CUS Project Summaries Index
Specific geographic area covered: Illinois, Indiana, Kentucky (Wabash Valley Seismic Zone)
Scientific question being addressed: Crustal deformation associated with intraplate tectonism in the southern Illinois Basin
General project plan: GPS measurement of crustal deformation at 55 sites distributed about southern Indiana, southern Illinois, and western Kentucky, encompassing southern Illinois Basin, Rough Creek graben, Wabash Valley Fault System. Measurements were made during GPS campaigns in 1997 and 1998. Ongoing data analysis provides coordinate estimates with precision of several mm (in horizontal); any systematic strain should be reflected in temporal change in position estimates. We seek systematic regional patterns of strain that may reflect major tectonic processes responsible for intraplate deformation in the Wabash Valley seismic zone and surrounding regions.
Project duration: March 1997 - February 1999
Most significant uncertainties in analyses and interpretations: Systematic errors in site position estimates associated with instrument setup and unmodelled tropospheric delays in the GPS signal. Very small signal associated with low intraplate strain rates suggests that observable signal in a single year of GPS measurements will be very close to (or below) the level of noise.
return to CUS Project Summaries Index
Specific geographic area covered: 1x2 degree Memphis quadrangle
Scientific question being addressed: Estimation of soil amplification potential
General project plan: Our base map is the 1x2 degree Memphis quadrangle map and will focus on soil amplification potential during an earthquake. ArcView GIS is used to digitize the map to delineate regions based on the standard CUSEC categories: A-Hard Rock; B-Rock; C-Stiff Soils/Most Gravels; D-Sands, Silts, Stiff Clays; E-Stiff Clay (10-50 feet thick); E2-Stiff Clay (50-120 feet thick); F1-Soils Vulnerable to Liquefaction; F2-Peats/Organic Clays (>10 feet thick); F3-High Plasticity Clays (>25 feet thick); F4-Soft/Medium Clays (>120 feet thick). These categories are also based on thickness and type of surficial material, depth to ground water and top of bedrock, standard penetration tests (blow counts), and shear wave velocity of surface material. Saucier's maps on the geology of the Lower Mississippi River Valley are used to trace the geology onto mylar and then traced it onto a copy of the 1x2 degree Memphis map. The Memphis map is being digitized and input into ArcView using a CD-ROM. Although we have not decided on how to formally classify the different regions yet, our map will be a very general one. Crowley's Ridge (loess) will have its own classification and the eastern part of the Ozarks (western end of our map) will have a separate classification (hard rock). The rest of the map is all alluvium. Data from our Highway Department in Little Rock and as much bridge construction data (stratigraphy and blow counts) going across several rivers have been collect. Saucier's data have been modified in some places. Work is collaborative with Tom Hart of Tennessee, as part of his data overlaps onto our map. Once his data is digitized onto our map, agreement is reached on how to classify each delineated region, and a complete staff review of the map is completed, then we will be able to move onto the next phase of the project.
return to CUS Project Summaries Index
Specific geographic area covered: Reelfoot Scarp, western Tennessee
Specific Question being answered: What are the rates, timing, and style of folding related to slip on the blind Reelfoot fault
General Project Plan and Preliminary Results: The timing of folding is being determined using geomorphic analysis of aerial photographs to determine the relative age of abandoned Mississippi River meanders that are deformed by the Reelfoot scarp. We have also developed 3D anaglyphs from digitized 7.5 minute USGS Quadrangleswith 5' contour intervals and Reelfoot Lake Bathymetry with 1 foot contours to characterize folding styles over much of the Lake County Uplift. Dating of deformed river meanders is being accomplished by study of extensive archeological sites whose ages are tightly constrained by diagnostic artifacts and by radiocarbon ages from other regions where these artifacts are not present in point bar, abandoned meander, and overbank deposits. Samples for radiocarbon analysis were also collected from two 60 meter long trenches, (8) 3m deep backhoe pits and (2) 200 m long core profiles or transects. Preliminary results based on archeology indicate that folding of the southern part of the scarp (east of Tiptonville, TN) is entirely younger than A.D. 800.
Deformation of Late Holocene units is being characterized with the core transects, trenches and backhoe pits described above and by structural modeling. Our preliminary results indicate the Reelfoot Scarp grows primarily by multibend fault-bend folding above a convex upward (e.g. flattening from west to east) 22 degree bend in the Reelfoot Blind thrust. Our structural model indicates the faulting events that produced the scarp were associated with an average of 5.8 m of dip slip. Other regional aspects of folding illuminated by the analysis of the anaglyphs indicate the presence of broader wavelength folding located immediately above the top of the thrust ramp defined in the structural model. This style of deformation is interpreted to be produced by fault propagation folding, however total strain produced by this process (which we relate to fault slip) is less than 10% of that produced by kink-band migration (i.e. that which produced the Reelfoot Scarp). The implications of our structural analysis indicate significant mode switching where some small proportion of slip on the steeper segment of the blind Reelfoot Thrust is clearly terminated at the top of the thrust ramp at the point where the fault begins to flatten.
Ongoing aspects of our work include testing of new geomorphic models of fold scarp degradation in an attempt to define the amount of limb widening (which we relate to fault slip) which occurs on an event by event basis. We are also beginning to identify significant lateral variations in height along the Reelfoot Scarp along its length that may be related to the lateral termination of slip on past blind thrust earthquakes which produced uplift of Tiptonville Dome.
return to CUS Project Summaries Index
Specific geographic area covered: New Madrid
Specific Question being answered: Recurrence intervals and their uncertainties for major New Madrid earthquakes
General Project Plan and Preliminary Results: A procedure has been developed which allows us to invert observations of modern seismicity along with historic and prehistoric information on large earthquakes to obtain recurrence intervals and their uncertainties for large events. We assume that our event catalog contains three types of information: prehistoric (paleoseismic) events that occurred over a period of thousands of years, historic macroseismic events that occurred over a period of a few hundred years, and recent instrumental data. The prehistoric and historic parts of the catalog contain only the strongest events, whereas the instrumental data can be divided into several subcatalogs, each assumed to be complete above a specified magnitude threshold. The technique takes into account uncertainties in determinations of earthquake magnitudes.
Preliminary applications of the method to New Madrid earthquakes show that results depend strongly on the occurrence times and assumed magnitudes of the largest events and the uncertainties of those estimates. Instrumental data are taken from SLU quarterly reports between 1974 and 1997. Prehistoric events used in the inversions are those reported by Tuttle, Schweig, Saucier, and others. For our first inversions we assume that prehistoric events of magnitude of 7.0 or larger occurred in 1530 AD, 900 AD, and an event of 6.4 or larger occurred in 490 AD.
Project duration: November 1998 - ?
Most significant uncertainties in analyses and interpretations: I expect that our results will be greatly influenced by uncertainties in times of occurrence and magnitudes of prehistoric earthquakes.
return to CUS Project Summaries Index
Specific geographic area covered: Wabash Valley region
Scientific question being addressed: Seismity of Wabash Valley region
General project plan: Twenty, three-component digital seismographs borrowed from the IRIS-PASSCAL facility were deployed along the Wabash Valley Seismic Zone of southern Indiana and Illinois over a seven month interval in 1995-96. Sites were chosen for low noise conditions and shallow bedrock. We deployed ten, 2-Hz sensors that recorded continuously at 40 sps in a phased array with a one-km aperture. One site in the array also had a broadband sensor continuously sampled at 40 sps that operated for 3 months in 1996. Ten of the sites were located at 20-30 km intervals and constituted a conventional triggered monitoring network with recording at 250 sps. Three of the sites located on thick unconsolidated deposits utilized 12-element strings of 4.5 Hz triaxial geophone spread out in 55-meter long linear arrays. Noise tests demonstrated a reduction in noise by a factor of three at frequencies over 10 Hz through the use of these linear geophone arrays. The raw data set from this experiment exceeds 140 Gbytes in size. The data are dominated by mining explosions. Approximately 10 times as many explosions were detected as earthquakes making unambiguous discrimination of explosions a critical issue. In addition, because this experiment occurred during a period with the poorest regional network coverage in twenty years, we have a serious problem calibrating our magnitude estimates to a standard. We have adopted a Mw method using spectra computed by the multitaper method. Current results are suggesting seismicity rates consistent with estimates derived from historical data and instrumental data for the region. Some of our earlier results appear to have been biased by a coherent noise source related to a mine near the phased array. This source created periodic transients that were initially identified as low magnitude earthquakes near the detection limit.
Project duration: June 1995 - present.
Most significant uncertainties in analyses and interpretations: 1) Discrimination of mining explosions from earthquakes is normally clear, but ambiguous events exist. 2) Low level of seismicity and finite duration of the experiment (approximately 200 days) make the results prone to the statistics of small numbers. To maximize this we have focused on variations in detection limit with time, but statistical uncertainties remain.
return to CUS Project Summaries Index
Specific geographic area covered: Memphis and Shelby County, TN
Scientific question being addressed: How does the P- and S-wave seismic velocity structure in the upper 50 m vary in the Memphis area? Can we use these data to determine important seismic impedance boundaries for earthquake site response in the Memphis area?
General project plan: We have taken measurements of near-surface P- and S-wave reflection/refraction velocity profiles at 12 sites in Memphis using surface seismic reflection/refraction methods. Measurements have been made at sites in the Mississippi River flood plain, the Wolf River flood plain, and on loess within the city of Memphis. We are comparing results with R. Street and J. Harris, P. Bodin and S. Horton, and C. Mueller, M. Meremonte, and E. Cranswick.
Using P- and S-wave seismic-reflection traveltimes in Seattle, WA, Sherman Oaks, CA, and Cape Girardeau, MO, we have directly detected important site resonant frequencies that were also independently verified by data from earthquake seismograms. We will use this resonance frequency mapping technique on the Memphis data set. However, independent verification of site resonances from earthquake data will likely be lacking because of infrequent earthquakes. We will continue to check with Bodin and Horton who are conducting site response studies in the Memphis area using microtremors and have data on resonances. We will also analyze data for correlations between seismic velocity structures and mapped surficial geology and available borehole data. Sites will be classified based on S-wave velocity according to NEHRP soil categories A,B,C,D, and E.
Though specific sites are yet to be determined, about 10-15 more sites in Memphis-Shelby County will be measured by the USGS group for P- and S-wave velocity in 1999. Some of the preliminary goals for 1999 field work are: 1) image deeper, especially by S-waves, 2) examine the velocity variation between Wolf River flood plain deposits and sites located on loess, 3) measure the difference in seismic velocity between the 10-m-high man-made berm on President's Island and the adjacent unmodified land, 4) at the locations where they crop out east of Memphis (also the locations of microtremor recordings of Bodin and Horton), measure the velocities of formations that underlie Memphis, 5) fill in areas with data gaps.
Project duration:Started October, 1997. Ending 2000-2001?
Most significant uncertainties in analyses and interpretations: Based on similar studies in California and Washington, seismic velocities and layer depths determined from refraction/reflection profiles are accurate to within about 5 percent. Because of the calculation 1/T used to calculate quarter-wavelength resonant frequencies, where T is two-wave traveltime of the reflection, the error in picking resonant frequencies from the arrival times of seismic reflections increases as reflection times decrease; that is, a slight error in pick time is magnified for short traveltimes.
return to CUS Project Summaries Index
Specific geographic area covered: Data come from the world's SCRs. Application will be to the North American SCR, and particularly to the CEUS.
Scientific question being addressed: What are the large-magnitude parts of the recurrence curves for the North American SCR? Large-earthquake recurrence is poorly constrained for the North American SCR because the historical record is short and the recurrence intervals are likely to be long and variable. Better constrained North American SCR recurrence curves can come from two complementary approaches. (1) Paleoseismology adds prehistoric North American SCR earthquakes.
has the advantage of being source-specific, but the disadvantages of being time consuming, short of trained personnel east of the Rockies, and, at present and in most places, unable to distinguish the geologic record of one large earthquake from that of several smaller earthquakes that might have been separated by decades to centuries. (2) Compilation of global analogs would add historical earthquakes from other SCR's. Analog compilation complements paleoseismology because global analogs, while not source-specific, can be compiled comparatively quickly, require only a single geologist with appropriate tectonic expertise and seismologist collaborator, and utilize individual historical earthquakes. The paleoseismological approach is well under way. This project will add the global analogs.
General project plan: The use of global analogs is feasible now because EPRI funded Arch Johnston and colleagues to compile a worldwide database of large SCR earthquakes and their seismological characteristics and geological settings. However, the database in its present form cannot produce the best estimates of North American SCR recurrences, for three reasons. (1) The database is more than a decade old, and new information shows that some of the inferred geological settings are wrong. Wheeler is familiar with the pertinent new information for North America, and can correct the data from the North American SCR. The North American earthquakes comprise about 1/4 of the global total, so correcting their geological settings should indicate which earthquakes in the other SCR's merit critical reexamination. (2) More than a decade's new earthquakes should be added to the database. Johnston has been collecting results from these earthquakes and has offered to share his files, collection of foreign reports and maps, and advice. Other needed references will be obtained from the USGS libraries or through interlibrary loans. (3) The published analyses of the existing database grouped earthquakes according to geological settings that applied globally, e.g. Mesozoic, Paleozoic, or Precambrian extended or compressional terranes. To sharpen results, earthquakes should be regrouped according to the specific seismotectonic settings found in the US and southern Canadian parts of the North American SCR: Mesozoic and latest Precambrian-Cambrian extended terranes, Mississippian-Permian compressional terranes, and Precambrian cratons.
Project duration: start early calendar 1999, end calendar 2001.
Most significant uncertainties in analyses and interpretations: Numbers and types of errors that will be found in decade-old intrepretations of geologic settings of earthquakes.
return to CUS Project Summaries Index
Specific geographic area covered: New Madrid
Scientific question being addressed: What are the characteristics of contemporary seismicity in New Madrid?
General project plan: The New Madrid Rapid Earthquake Information System (REIS) is a multi-focused project designed to provide seismic information to a variety of end-users within sufficient time-frames to maximize the utility of the information for the particular customer. Products include rapid and reliable notification, timely parametric information, and detailed, high quality data archives. Station coverage includes a dense network of high-frequency sensors (72 3-component stations) designed to provide rapid epicentral estimates for larger events, and a low detection threshold permanent archive of parametric data for smaller events. The short-period network is complemented by a sparse network of broadband sensors (11 3-component stations) designed to provide high quality waveforms for each event. A sophisticated near-real-time data processing system is employed (earthworm). This system provides rapid automatic earthquake locations (and associated waveforms) and has been tuned to a relatively high detection threshold (>2.5) to prevent false alarms. It also provides a traditional sta/lta trigger algorithm to record waveforms for subsequent review and archive. A continuous, revolving data buffer facilitates interactive waveform acquisition for time blocks of particular interest.
Near term directions include paging and email for dissemination of the rapid automatic bulletins. These bulletins could be suitable for active mitigation purposes. Beginning in January, 1999 we will install an Oracle-based front-end (a prototype is in operation at Golden, CO). This will increase the efficiency of routine review and analysis as well as provide user friendly access to the archive (including a web-based search engine). Compliance with y2k is currently being tested and will be complete within 90 to 120 days. The system also provides the opportunity to link independent regional networks in near-real-time thereby expanding coveraging, and increasing reliability through redundancy. Initially, we will link CERI networks operating in the New Madrid and Southern Appalachian Seismic Zones. We will also be linking CERI with the USGS systems in Menlo Park, CA and Golden, CO. The flexibility of the system lends itself to including a variety of information (e.g. strong motion data) provided minimum standards (particulary with respect to timeliness), protocols and formats are adhered to. This flexibility is born out by the application of earthworm to the Tsunami Warning System.
Project duration: ongoing
Most significant uncertainties in analyses and interpretations: The speed of the bulletins is currently limited by magnitude estimation. A rapid body wave estimator is needed. The system is also hampered by communications bandwidth, and the high ongoing costs of long distance ISDN telephone. Reliance on the internet has been deliberately minimized due to the lack of guaranteed bandwidth, particularly in critical situations. Better digital communications, with lower ongoing costs is needed.
return to CUS Project Summaries Index
Specific geographic area covered: New Madrid
Scientific question being addressed: Pre-1811 earthquakes
Project Summary: We propose a systematic investigation of prehistoric earthquake liquefaction in the New Madrid seismic zone by introducing ground penetrating radar. As has been demonstrated in previous experiments, ground penetrating radar is capable of the high resolution needed to detect near surface discontunities such as sand dikes and to obtain very shallow GPR profile. Therefore, it has great potential for use in mapping liquefaction deposits and their feeding structures, or sand dikes. Because earthquake liquefaction tends to recur at the same site, by applying this technology to a wide distribution of sites along the edge of the 1811-1812 earthquake liquefaction distribution, we will be able to quickly identify prehistoric liquefaction distribution. Combining with intensive trenching and radiocarbon dating at the beginning, we will be able to interpret the time of prehistoric earthquakes and expand to the entire region thereafter.
If prehistoric earthquakes in the New Madrid seismic zone are of the same magnitude range as the 1811-1812 earthquakes, as inferred from the size of isolated liquefaction deposits, we have the potential to establish a regional correlation among different liquefaction deposits or even to build a liquefaction stratigraphy. This will help us to eliminate the temporal and spatial ambiguity in interpretation of pre-1811 earthquake events.
This technology will also have great potential of being applied to other earthquake regions such as Charleston earthquake area in the east or even in sedimentary basins in the California.
Project duration: 2/1999-1/2000
return to CUS Project Summaries Index
Specific geographic area covered: National
Scientific question being addressed: Develop site specific ground motions
General project plan: Software system of integrated tools and procedures necessary to develop design earthquake ground motions for seismic analysis of Corp's facilities. This is an integrated system of guidance, earthquake data, and procedures to develop and evaluate design earthquakes. Specific tools include: coupled strong motion catalog and time history database to search for accelerograms that match site parameters, attenuation relationships, 1-D site response model calculations, module to query USGS National Hazard maps and data. Future plans are to develop a deterministic seismic hazard module to identify faults and source zones within specified ranges of site, synthetic accelerogram development module, and site specific probabilistic hazard calculations (SEISRISK III model). This system is being developed for use by the Corps to support seismic stability studies for the Dam Safety program for all levels of analysis. This system will provide a new capability, procedures and tools to efficiently, accurately, following CE policy and guidance, develop economically and technically defendable estimates of site seismic hazard. This system will enable expedient seismic reviews and appropriate ground motions that will enable more efficient analysis and design. This system will also provide the Corp with a technology transfer platform for our corporate knowledge to sustain our capability to assess the seismic hazard at project sites.
Project duration: 1996-2000
return to CUS Project Summaries Index
Specific geographic area covered: Mid-America
Scientific question being addressed: Response spectra and ground motion simulation
General project plan: Construction of elastic and inelastic response spectra, develop ground motion simulation method
Project duration: Jan. 1, 1998 Dec. 31, 1999
Most significant uncertainties in analyses and interpretations: Effect of large events and local (site) conditions.
return to CUS Project Summaries Index
Specific geographic area covered: Central Mississippi Embayment, MO, IL, TN
Scientific question being addressed: S wave ground motion in sub-regional distances
General project plan: Contribution of Moho reflection to ground motion
Project duration: (Jan. 1998 Sept. 1998)
Most significant uncertainties in analyses and interpretations: Amplitude of S wave how it varies with distance near critical distance (about twice the Moho depth, or 80 km)
return to CUS Project Summaries Index
Specific geographic area covered: All of southern IN and IL
Scientific question being addressed: Paleoliquefaction studies have been conducted by many researchers in the region since 1990, in order to assess the paleoseismic record. Seismically induced liquefaction features that have been discovered through much of the area. And, ages of specific earthquakes, and their epicenters and magnitudes, have been evaluated for many paleo-earthquakes. Results will be presented, current as of the fall of 1997.
General project plan: Search for paleoliquefaction features in banks of streams, date them, perform geotechnical testing and analysis at the sites of liquefaction.
Project duration: (Ongoing)
Most significant uncertainties in analyses and interpretations: All the large and very large earthquakes in the region struck in mid-Holocene time or earlier. No large earthquakes have been documented the past 4,000 years. Is it possible that the region is no longer prone to hosting large earthquakes?
return to CUS Project Summaries Index
Specific geographic area covered: CUSEC States
Scientific question being addressed: Soil amplification
General project plan: Generate maps at 1:250,000 and 1:24,000 scales.
Project duration: 1:250,000 maps completed Spring '99, start 1:24,000 maps Spring '99
return to CUS Project Summaries Index
Collaborators: Sharon Diehl, Joan Gomberg, and John Sims, USGS; Robert Lafferty, MCRA; Brian Noonan, Texas A&M University; Julie Morrow, Arkansas Archeological Survey; Lorraine Wolf, Auburn University; Paul Bodin, University of Memphis; Paul Mayne and Tim Stark, MAEC
Specific geographic area covered: New Madrid region including parts of Arkansas, Kentucky, Missouri, and Tennessee
Scientific question being addressed: What is the earthquake potential of the New Madrid seismic zone (NMSZ)? Is the NMSZ likely to generate 1811-1812-type earthquakes in the future and if so how soon? To date, results indicate that the NMSZ generated two significant earthquake sequences circa A.D. 900 +/- 100 yr and A.D. 1535 +/- 135 yr. These events induced severe and widespread liquefaction similar to the 1811-1812 earthquake sequence. The A.D. 900 and A.D. 1530 events are estimated to be of M > 7.4 and M > 7.2, respectively, based on established relationships between earthquake magnitude and distance of liquefaction (Ambraseys, 1988; Obermeier et al., 1993; and Pond and Martin, 1997). Comparison of the size distribution of sedimentological units comprising historic and prehistoric sand blows suggests that the prehistoric earthquakes had similar sources and magnitudes to the three major earthquakes of 1811-1812. An average recurrence interval of 456 years is derived for 1811-1812-type earthquake sequences. The recurrence interval for these events may vary from 141 to 870 years.
General project plan: We are in the process of mapping the age and size distribution of liquefaction features in the New Madrid region by conducting regional reconnaissance and detailed investigations of earthquake-induced liquefaction features. It is important to conduct regional reconnaissance to identify the best sites for dating paleoliquefaction features as well as to establish the extent of liquefaction induced by prehistoric earthquakes. Detailed investigations are also important to constrain the ages and measure the sizes of liquefaction features. Archeological sites, rich in organic material as well as artifacts, have been extremely helpful in dating liquefaction features.
Most significant uncertainties in analyses and interpretations: The current methodology of dating organic material in horizons that bound sand blows leads to age estimates of liquefaction features that have ranges of two hundred years or more, even under the best of circumstances. This, in turn, leads to uncertainties in estimating dates of prehistoric earthquakes, making regional correlations of similar-age features, and therefore, in estimating source areas, magnitudes, and recurrence intervals of large earthquakes.
Project duration: April 1, 1998 to March 31, 2000
Collaborators: K. Dyer-Williams, MTA; Robert Lafferty, MCRA; Joan Gomberg and Eugene Schweig, USGS; Bob Bauer, John McBride and Wen-June Su, Illinois State Geological Survey; Dave Hoffman, Missouri Department of Natural Resources; Tim Stark, University of Illinois
Specific geographic area covered: St Louis region including parts of Missouri and Illinois
Scientific question being addressed: During an NRC-funded paleoseismology study in southwestern Illinois and southeastern Missouri, we documented liquefaction features, including sand dikes and sand blows, along the Big Muddy, Marys, Meramec, and Kaskaskia Rivers and Mud, Shoal, and Silver Creeks (Tuttle et al., 1996 and 1998). Our preliminary conclusion was that a large to very large earthquake occurred east of St. Louis about 6,500 yr B.P. (or 5,670 +/- 80 radiocarbon years B.P). This was similar to the result of McNulty and Obermeier (1997) and Obermeier (1997) who also found liquefaction evidence along the Kaskaskia River and Shoal Creek for a large earthquake about 6,500 to 7,000 radiocarbon yr B.P. We proposed three possible earthquake scenarios to account for the observed pattern of liquefaction in the St. Louis region: (1) a M 7.2 event centered near Germantown, Illinois; (2) a M 7.7 event near Centralia, Illinois; or (3) a M 6.3 near Germantown, a M 5.3 near St. Louis, and a M 7.7 near New Madrid, Missouri. During this USGS-funded project, we will further resolve the timing, source areas, and magnitudes of prehistoric earthquakes that generated liquefaction in southwestern Illinois and southeastern Missouri. A better understanding of prehistoric earthquakes will contribute to more realistic estimates of earthquake recurrence in this region.
General project plan: We plan to conduct additional reconnaissance and detailed investigations of earthquake-induced liquefaction features in the St. Louis region and to cooperate with others in testing geotechnical properties at liquefaction sites and in modeling the dynamic deformation field associated with the proposed earthquake scenarios.
Most significant uncertainties in analyses and interpretations: The age and size distributions of liquefaction features are poorly understood in the St. Louis region. This leads to significant uncertainties in estimating source areas and magnitudes of causative earthquakes. Liquefaction features whose ages can be well-constrained are not as prevalent here as they are in the New Madrid region. In addition, only a small percentage of river and stream cutbanks have been searched so far. Therefore, much more reconnaissance is needed to find the best features for dating prehistoric earthquakes and to assure a representative sampling of liquefaction features across the region.
Project duration: December 1, 1998 to November 30, 1999
Specific geographic area covered: Memphis/Shelby Co/TN, West Memphis/AR, Blytheville/AR, 155 Bridge & Steele/MO
Scientific question being addressed: Site-specific soil liquefaction assessment throughout the NMSZ embayment by in-situ testing. The seismic piezocone penetration test (SCPTU) is an efficient and economical exploration tool for evaluating the in-situ liquefaction potential of soils, postcyclic undrained residual strength, and site amplification properties (Gmax). Using a light cone truck outfitted with special earth-anchoring system, field tests over 30+ meters deep have been conducted in these seismic regions to obtain site-specific data and four independent readings within a single sounding: tip resistance (qt), sleeve friction (fs), penetration porewater pressure (ub), and shear wave velocity (Vs). The Vs is obtained in a downhole manner using a interval-velocity procedure with horizontally-polarized vertically-propagating shear waves from a surface source.
General project plan: The SCPTU tests have been performed at prior geologic-mapped paleo-liquefaction sites having sand dikes, subsidence features, carbon-dating techniques, and other liquefaction evidence. Field sites have been organized and coordinated with Prof. Roy Van Arsdale/Univ. Memphis, Prof. Martitia Tuttle/Univ. MD, USGS (Buddy Schweig and Joan Gomberg), and MoDOT. Representative soundings in Memphis and Shelby County/Tennessee, West Memphis and Blytheville/Arkansas, Steele and I-155 Bridge/Missouri are presented and evaluated in terms of interpreted assessments of soil liquefaction response. The SCPTUs provide very detailed and continuous stratigraphic information, as well as reliable means of interpreting soil properties, such as strength and stiffness. Both the qt and Vs readings can be used independently to assess inplace liquefaction potential (e.g. Robertson & Wride, 1998), thus providing redundancy in evaluation. Recent methods proposed by the Japanese (i.e., Suzuki et al., 1995, use both the qt and fs measurements together to accommodate variations in soils fines content. Moreover, the small-strain stiffness (denoted by Gmax) is needed in amplification analyses and obtained directly from the shear wave velocity measurements.
The SCPTU is much less disruptive than conventional soil borings since only a 36 to 44 mm hole is hydraulically-pushed during the sounding and no auger cuttings or spoil is produced. The method is considerably quicker too in that the test is conducted at 20 mm/sec for the cone penetration portion and left-right strikes are conducted each 1-meter interval for the shear wave portion. A complete 30-meter sounding may take only 1.5 to 2 hours, versus 8 to 10 hours for conventional drilling. The measured qt indicates the friction angle and relative density of sands, whilst in clays gives information on the undrained strength and degree of overconsolidation. If desired, dissipation readings of the porewater pressures can be taken to provide evaluations of permeability and time rate parameters (coef. of consolidation).
Project duration: MAE Center Project 1GT-3A (Jan. 1998 - Oct. 1999).
Most significant uncertainties in analyses and interpretations: For the liquefaction assessment, the current State-of-Practice is to use Cyclic Stress Ratio (CSR) which is based on max. horiz. ground acceleration = ???). The CPT provides the measured penetration resistances, qc (or alternatively, shear wave velocity measurements) which is site-specific. The CSR depends upon amax which unfortunatelty is unknown, yet traditionally assessed using the computer program SHAKE which is a one-dimensional & nonlinear assessment of surface amplification effects. SHAKE was only calibrated for soil column depths of up to 150 meters. Other programs (2-d SASSY and 3-d DESRA) are likewise only applicable to those depths. Effects of nonlinearity on the G/GO degradation for use in deep soil strata are also unknown.
return to CUS Project Summaries Index
Specific geographic area covered: Eastern North America
Scientific question being addressed: There have been several relations proposed in the last few years to describe the amplitudes of ground-motion in eastern North America (ENA). These relations differ significantly in their assumptions concerning the amplitude and shape of the spectrum of energy radiated from the earthquake source. Collaborative work with D.M. Boore of the U.S. Geological Survey was undertaken to evaluate earthquake source models against the empirical database. This work, summarized below, was recently published in Atkinson, G. and D. Boore, 1998: Evaluation of models for earthquake source spectra in eastern North America. Bull. Seism. Soc. Am., 88, 917-934. The response spectra database is available to interested parties by writing to the authors.
General project plan: Ground motions predicted for alternative source models are compared against the sparse ENA ground-motion database. The source models evaluated include the two-corner models of Boatwright and Choy (1992), Atkinson (1993), Haddon (1996), and Joyner (1997a,b), and the one-corner model of Brune (as independently implemented by Frankel et al. (1996) and by Toro et al. (1997)). The database includes data from ENA mainshocks of M>4 and historical ENA earthquakes of M>5.5, for a total of 110 records from 11 events of 4<M<7.3, all recorded on rock. We also include 24 available rock records from 4 large earthquakes in other intra-plate regions; conclusions are checked to determine whether they are sensitive to the addition of these non-ENA data.
The Atkinson source model, as implemented in the ground-motion relations of Atkinson and Boore (1995), is the only model that provides unbiased ground motion predictions over the entire period band of interest, from 0.1 to 10 seconds. The source models of Frankel et al. (1996), Toro et. al. (1997) and Joyner (1997a,b) all provide unbiased ground motion estimates in the period range from 0.1 to 0.5 seconds, but overestimate motions at periods of 1 to 10 seconds. The Haddon (1996) source model overpredicts motions at all periods, by factors of 2 to 10. These conclusions do not change significantly if data from non-ENA intra-plate regions are excluded, although the tendency of all models towards overprediction of long-period amplitudes becomes more pronounced.
The tendency of most proposed ENA source models to overestimate long-period motions is further confirmed by an evaluation of the relationship between Ms, a measure of the spectrum at 20-second period, and moment magnitude. A world-wide catalogue of shallow continental earthquakes (Triep and Sykes, 1996) is compared to the Ms-M relations implied by each of the source models. The Atkinson source model is consistent with these data, while other proposed ENA models overpredict the average Ms for a given M.
The implications of MMI data from historical earthquakes are also addressed, by exploiting the correlation between felt area and high-frequency souce spectral level. High-frequency spectral amplitudes, as specified by the Atkinson and Boore (1995), Frankel et al. (1996), Toro et al. (1997) and Joyner (1997a,b) source models, equal or exceed the levels inferred from the felt areas of most of the large ENA events, with the noteable exception of the Saguenay earthquake. By contrast, high-frequency spectral amplitudes specified by the Haddon (1996) source model agree with the felt area of the Saguenay earthquake, but overpredict the the felt areas of nearly all other large events. In general, models which fit the Saguenay data - be it intensity data, strong-ground-motion data, regional seismographic data or teleseismic data - will not fit the data from the remaining earthquakes.
A source model derived from the California database, suitably modified for regional differences in crustal propertiies, is also evaluated. This model is not significantly different from the Atkinson model for ENA. There is an important practical application of this similarity, which we develop as an engineering tool: empirical ground-motion relations for California may be modified to predict ENA ground motions from future large earthquakes.
return to CUS Project Summaries Index
Specific geographic area covered: Eastern North America
Scientific question being addressed: Research was conducted into a finite-fault model, that explains the empirical characteristics of eastern ground motion, as documented in the study described in the previous section. This work, summarized below, will be published in Beresnev, I. and G. Atkinson, 1999: Generic finite-fault model of earthquake ground motions in eastern North America. Bull. Seism. Soc. Am., 89. The FORTRAN code FINSIM (Beresnev and Atkinson, 1998) used in the modeling, with copies of all input files, are available to interested parties by writing to the authors.
General project plan: Ground motion models based on the Brune point-source approximation have an underlying w2 spectrum, with a single corner frequency. These models overpredict observed spectral amplitudes at low-to-intermediate frequencies (~ 0.1 to 2 Hz), for earthquakes with moment magnitudes M of 4 or greater. The empirical spectra of moderate-to-large events tend to 'sag' at these frequencies, relative to the level suggested by the Brune point-source model. A model that accounts for the finite extent of the fault plane correctly describes the observed spectral shapes. The model represents seismic radiation as a sum of contributions from several subfaults. Each subfault may be represented as a point source, and each subevent has an w2 spectrum. When contributions to ground motion at an observation point are summed over all subfaults, the resulting spectral shape has two corner frequencies and more closely matches observed spectra. The more realistic spectral shape obtained through finite-fault modeling reflects the underlying reality that the radiation from real faults is formed by ruptures of their smaller parts, whose corner frequencies are higher than those implied by the full fault dimension. The two corners appear naturally as a result of subevent summation.
We use the stochastic finite-fault methodology to simulate the recorded ground-motion data from all significant earthquakes in eastern North America (ENA). These data include 8 events of M > 4 recorded on modern digital instruments (regional seismographs and strong-motion instruments), and 3 historical events of M 5.8 to 7.3 recorded on analog instruments. The goodness of fit of synthetics to the data is defined as simulation bias, which is indicated by the difference between the logarithms of the observed and the simulated spectrum, averaged over all recordings of an earthquake. The finite-fault simulations provide an unbiased fit to the observational database over a broad frequency range (0.1 to 50 Hz), for all events.
A surprising conclusion of these simulations is that the subfault size that best fits the observed spectral shape increases linearly with moment magnitude, in an apparently deterministic manner. This strongly suggests that the subfault size can be unambiguously defined by the magnitude of the simulated earthquake. In this case, the radiation-strength factor (s), which is proportional to the square root of the high-frequency Fourier acceleration level, remains the only free parameter of the model. Its value is related to the maximum slip velocity on the fault. The strength-factors for all modeled ENA events are within the range of 1.0 to 1.6, with the exception of the Saguenay main shock (s = 2.2). This suggests a remarkable uniformity in earthquake slip processes.
Specific geographic area covered: Mid-America
Scientific question being addressed:What is the best estimate of hazard?
General project plan: Use the USGS 1996 NEHRP probabilistic hazard analysis as a starting point. Incorporate first order deep soil column effects for the Mississippi Embayment. The overriding constraint is to have a documented transition from the 1996 NEHRP maps to more realistic regional maps.
The figures present a comparison of expected peak accelerations and 1.0 sec pseudo-acceleration for 2% in 50 year hazard using the USGS ground motion model (e.g., not 50% USGS and 50% Toro-McGuire) to that of a prototype deep soil model. The anelastic attenuation in the deep soils overwhelms the site amplification at Memphis for peak acceleration, but at 1.0 sec, the deep soil site amplification increases the motion.
|
USGS Peak Ground Acceleration 2% 50 years
|
USGS Ground 1.0 sec PSA 2% 50 years
|
|
Modified Deep Soil Peak Ground Acceleration 2% 50 years
|
Modified Deep Soil Ground 1.0 sec PSA 2% 50 years
|
The prototype soil model Vs and Qs are show in the following figure for soil thicknesses from 1 to 2000 meters. This model was constrained by Paul Mayne's surface measurements for S-velocity and density, by a shear wave 'kappa' at New Madrid and an S-wave travel time in the New Madrid sediments. A power law relation from EPRI studies is used to estimate the shear velocity. The power law relation for Qs is ad-hoc'd. We use
Vs (m/sec) = 250 H**0.18 Qs = 6 H**0.25 Rho (gm/cc)= 0.8*alog10(Vs) - 0.1 (where H(m) is depth in sediment column) Beneath the sediments is a 1 km thick Paleozoic layer Vs (m/sec) = 3000 Qs = 500 rho (gm/cc)= 2.7 Followed by the basement: Vs (m/sec) = 3500 Qs = 500 rho (gm/cc)= 2.7
Project duration: Jan. 1, 1998 Sep. 30, 2001
Most significant uncertainties in analyses and interpretations: Everything! In order of magnitude of effect and likelihood of resolution: specification of Mid-America deep soil column velocity and Q model; refinement of ground motion scaling with distance; source spectrum scaling of mid-American earthquakes; re-examination of the method of including a finite fault for the characteristic New Madrid earthquakes; re-examination of size and recurrence of the characteristic New Madrid earthquake.
return to CUS Project Summaries Index
Specific geographic area covered: Mid-America
Scientific question being addressed:What is scaling of ground motion with distance?
General project plan: Use regional seismic network data of the Cooperative New Madrid Seismic NEtwork and the University of Memphis PANDA deployment to define an internally self-consistent parameterization of Fourier velocity spectra, signal duration and peak filtered velocities for vertical and horizontal component motions. Compare these to Atkinson and Boore (1995). The result is a specification of geometrical spreading, frequency depdendent Q and signal duration for small events.
The figures present a comparison of normalized peak ground velocities filtered at frequencies of 1.0 through 16.0 Hz. The figures have been corrected for a 1/R geometrical spreading to emphasize differences in the two models. Regional data indicate that higher motions are expected in the central U.S. than predicted by the Atkinson-Boore (1995) model for distances greater than 150 km. This will increase the hazard at St. Louis for New Madrid events. (The New Madrid results use data from 2500 seismograms)
Atkinson-Boore (1995)
|
Central US model
|
Project duration: Ongoing
Most significant uncertainties in analyses and interpretations: It is possible to estimate an effective "kappa" for the New Madrid nominal 600 meter deep soil site at 0.045 sec for horizontal motion (presumably S-waves in the sediment) and 0.025 sec for vertical motion (presumably P waves in the sediments).
The New Madrid data set consists of small earthquakes. The medium Green's function for propagation is constrained but nothing is learned about source spectrum scaling as a function of earthquake size.
return to CUS Project Summaries IndexSpecific geographic area covered: Southern Illinois Basin (IL and IN)
Specific Question being answered: Can coherent patterns of seismicity be recorded in the Illinois Basin and what is the relationship between seismicity and pre-existing geologic structure?
General Project Plan: Modern seismic monitoring in the central U.S. focuses mainly around the well-defined New Madrid seismic zone (NMSZ). A widened distribution of seismicity continues farther north into the southern Illinois Basin and southeastern Missouri. The accuracy in epicentral location for most events in southern Illinois is not sufficient to warrant associating earthquakes with a particular structure. Despite the high concentration of modern seismicity in the NMSZ, it is, however, notable that the southern Illinois region has had larger and deeper twentieth-century earthquakes than the NMSZ. It is unclear how the very different patterns of seismicity in the two areas could be kinematically related; we have few constraints on the structural control of earthquakes beneath the Illinois Basin. The relationship between subsurface structures and seismogenic zones should be one of the key pieces of information needed for an effective assessment of earthquake hazard in the southern Illinois Basin. Since currently available earthquake databases from regional seismic networks are inadequate for associating seismicity with structures, we propose two approaches, one to determine upper crustal structure from existing seismic reflection lines in the study region, and the other to deploy a PANDA seismic array in the southern Illinois Basin. Results from analysis of industrial seismic reflection lines will not only identify critical local structures but also will generate an important upper crustal P-wave velocity model for input to a 3-D tomographic velocity inversion, and will also provide critical constraints for the final tectonic interpretations. 45 three-component short-period PANDA II stations and five portable broadband stations will be distributed over the southern Illinois Basin for a period of 16 months. From the high-resolution PANDA II data, we expect to (1) determine a representative crustal velocity model for both P and S waves from 3-D tomographic inversion, (2) determine reliable hypocenter locations, (3) determine the regional stress pattern and modes of fault movement, and (4) correlate seismicity with known geological structures obtained from surface geology and seismic reflection lines.
Project duration: November 1999-November 2001 +?
Most significant uncertainties: We do not know a priori if coherent patterns of earthquake foci really exist in the southern Illinois Basin.
Specific geographic area covered: Southern Illinois Basin and adjacent areas of the Midcontinent.
Specific Question being answered: What are the physical properties corresponding to (1) major Precambrian seismic sequences underlying the Illinois Basin and (2) major reflector boundaries in the middle and lower crust that correspond to localized seismicity? Can we then use this information to better map and characterize the seismogenic source north of the New Madrid seismic zone?
General Project Plan: Seismic reflection data donated or purchased from industry has allowed us to map Precambrian "basement" sequences beneath the Illinois Basin and to relate internal structures, where possible, to patterns in seismicity. Additionally, specific deep reflectors have in a few instances been directly associated with major hypocenters. We plan to use 2-D filtering of gravity and magnetic intensity data to investigate any possible correlations with reflectivity patterns in the crust. This will allow us to constrain the rock-type interpretation of such patterns for the first time beneath the Illinois Basin and allow a better constraint on issues such as whether basement reflection sequences are sedimentary or igneous in origin, the regional extent of crustal boundaries or suture zones (e.g., Commerce Geophysical Lineament), and the relation of contemporary seismicity to these and other geologic features of the region. The filtering work will be accompanied by 2-D and 3-D forward modeling for specific structures.
Project duration: ongoing.
Most significant uncertainties: Do reflectivity patterns necessarily correlate with contrasts in density and/or magnetization sufficient to cause discernible gravity and magnetic anomalies?
Specific geographic area covered: South-central Illinois.
Specific Question being answered: Is the Du Quoin Complex an important seismic source zone?
General Project Plan: Because much of the central U.S. Midcontinent is covered by unlithified glacial drift, loess, and alluvial deposits and some older, weakly lithified clay, silt, sand, and gravel, recognition of anything but large fault offsets is difficult. Surveying and analysis of paleoliquefaction evidence has been a successful approach to studying paleoseismicity in the region. In this study, we will address the key question of how paleoliquefaction site patterns in southern Illinois north of the NMSZ (New Madrid Seismic Zone) are related to a candidate seismogenic source recognized from seismic reflection data in south-central Illinois, the Du Quoin Complex. Two tasks are proposed. One is to study recently released seismic reflection profiles and to examine the possibility of a complex structure as a potential seismic source. The other task is to study paleoliquefaction features in the immediate vicinity of the potential seismic source. The study area is centered around the Dowell Fault Zone-Du Quoin Monocline-Centralia Fault Zone Complex (referred to as the Du Quoin Complex) in south-central Illinois. The area is selected for three main reasons: (1) The Du Quoin Complex has structural characteristics of a potential seismic source zone; (2) The proposed study area has not been adequately surveyed for paleoliquefaction sites; (3) The study area is covered by a network of seismic reflection profiles recently released to us. The project will develop a body of data and interpretations essential for a better understanding of the long-term seismicity in southern Illinois.
Project duration: 1999--2000.
Most significant uncertainties: Is there a coherent pattern of paleoliquefaction in south-central Illinois that will match a pattern of contemporary seismicity?
Specific geographic area covered: New Madrid region and adjacent areas of adjoining states, including southern Illinois.
Specific Question being answered: What is the detailed structure of the seismogenic source for New Madrid and nearby seismic activity; how is contemporary and historical seismicity manifested in observed geomorphic anomalies and other neotectonic features (e.g., paleoliquefaction); how can we improve the seismic velocity model for the region from teleseismic recording using the existing seismometer coverage?
General Project Plan: Initial work will and has included collection of high-resolution seismic reflection data integrated with drill hole data in the New Madrid area. This effort will be expanded to include conventional industry seismic data and reprocessing of available deep seismic data; we intend to expand this effort geographically to include the southern Illinois Basin region. Results from exploration geophysics will be used to constrain geodynamic modeling of contemporary deformation in the New Madrid zone and also in southern Illinois where data permit. A parallel program will be developed to record teleseismic arrivals using the current seismometer station coverage in the New Madrid zone in order to develop a velocity model, with which to improve focal mechanism inversions, using a receiver function approach.
Project duration: 1997--ongoing.
Most significant uncertainties: Is the fine structure of the New Madrid seismic zone simple and coherent enough to be detected geophysically and modeled?
Specific geographic area covered: Southern Illinois Basin and adjacent areas of the Midcontinent.
Specific Question being answered: Where are possible source zones for earthquakes in the Illinois Basin region based on mapping structure within the Paleozoic section that may be related to basement tectonics and based on mapping seismogenic structure directly in basement?
General Project Plan and Results: Although the Illinois basin is one of the world's most intensively studied interior cratonic basins, little is known of its deep Precambrian structure or seismotectonic framework. For the past three years, we have vigorously sought to acquire seismic reflection profiles previously surveyed over the Illinois basin, principally by the petroleum industry. This effort has begun to provide critical information on lower Paleozoic and Precambrian "basement" structure that can be integrated with data derived from regional earthquakes north of the New Madrid seismic zone. Reflection profiles over the major structures of the basin reveal reverse faults, with a possible strike-slip component, that penetrate the upper Precambrian crust, disrupt the surface of Precambrian basement, and facilitate folding of Paleozoic sedimentary rocks. Several types of associations have been recognized to date between earthquake source parameters and structures in reflection profiles. In the most interesting case, an earthquake hypocenter (1968.11.09 mb=5.5; depth=21.2±5.4 km) can be directly associated with dipping high-amplitude mid-crustal reflections, interpreted as compressional tectonic structures possibly of Grenville age. These reflectors, centered around a depth of 21 km, are part of a major zone of west-dipping reflectors within the upper and middle crust situated beneath the Wabash Valley Fault System and adjacent parts of southern Illinois, western Indiana, and western Kentucky. We have also begun a program of high-resolution seismic reflection profiling target the upper 300 m over areas of suspected Quaternary faulting that might be related to the modern New Madrid zone. The results from the shallow reflection program will be augmented by a drilling (250 m, maximum depths) campaign in 1999-2000.
Project duration: 1995--ongoing.
Most significant uncertainties: Do pre-existing faults and other deformation features actually govern contemporary stress release?