Jack K. Odum

Simulation of Broadband Ground Motion Including Nonlinear Soil Effects for a Magnitude 6.5 Earthquake on the Seattle Fault, Seattle, Washington

The Seattle fault poses a significant seismic hazard to the city of Seattle, Washington. A hybrid, low-frequency, high-frequency method is used to calculate broadband (0–20 Hz) ground-motion time histories for a M 6.5 earthquake on the Seattle fault. Low frequencies ( 1 Hz) are calculated by a stochastic method that uses a fractal subevent size distribution to give an ω -2 displacement spectrum. Time histories are calculated for a grid of stations and then corrected for the local site response using a classification scheme based on the surficial geology. Average shear-wave velocity profiles are developed for six surficial geologic units: artificial fill, modified land, Esperance sand, Lawton clay, till, and Tertiary sandstone. These profiles together with other soil parameters are used to compare linear, equivalent-linear, and nonlinear predictions of ground motion in the frequency band 0–15 Hz. Linear site-response corrections are found to yield unreasonably large ground motions. Equivalent-linear and nonlinear calculations give peak values similar to the 1994 Northridge, California, earthquake and those predicted by regression relationships. Ground-motion variance is estimated for (1) randomization of the velocity profiles, (2) variation in source parameters, and (3) choice of nonlinear model. Within the limits of the models tested, the results are found to be most sensitive to the nonlinear model and soil parameters, notably the overconsolidation ratio.

Near-surface structural model for deformation associated with the February 7, 1812, New Madrid, Missouri, earthquake

The February 7, 1812, New Madrid, Missouri, earthquake (M [moment magnitude] 8) was the third and final large-magnitude event to rock the northern Mississippi Embayment during the winter of 1811–1812. Although ground shaking was so strong that it rang church bells, stopped clocks, buckled pavement, and rocked buildings up and down the eastern seaboard, little coseismic surface deformation exists today in the New Madrid area. The fault(s) that ruptured during this event have remained enigmatic. We have integrated geomorphic data documenting differential surficial deformation (supplemented by historical accounts of surficial deformation and earthquake-induced Mississippi River waterfalls and rapids) with the interpretation of existing and recently acquired seismic reflection data, to develop a tectonic model of the near-surface structures in the New Madrid, Missouri, area. This model consists of two primary components: a north-northwest–trending thrust fault and a series of northeast-trending, strike-slip, tear faults. We conclude that the Reelfoot fault is a thrust fault that is at least 30 km long. We also infer that tear faults in the near surface partitioned the hanging wall into subparallel blocks that have undergone differential displacement during episodes of faulting. The northeast-trending tear faults bound an area documented to have been uplifted at least 0.5 m during the February 7, 1812, earthquake. These faults also appear to bound changes in the surface density of epicenters that are within the modern seismicity, which is occurring in the stepover zone of the left-stepping right-lateral strike-slip fault system of the modern New Madrid seismic zone.

Surface Seismic Measurements of Near‐Surface P‐ and S‐Wave Seismic Velocities at Earthquake Recording Stations, Seattle, Washington

We measured P- and S-wave seismic velocities to about 40-m depth using seismic-refraction/reflection data on the ground surface at 13 sites in the Seattle, Washington, urban area, where portable digital seismographs recently recorded earthquakes. Sites with the lowest measured V s correlate with highest ground motion amplification. These sites, such as at Harbor Island and in the Duwamish River industrial area (DRIA) south of the Kingdome, have an average V s in the upper 30 m (V¯ s30 ) of 150 to 170 m/s. These values of V¯ s30 place these sites in soil profile type E (V¯ s30 < 180 m/s). A “rock” site, located at Seward Park on Tertiary sedimentary deposits, has a V¯ s30 of 433 m/s, which is soil type C (V¯ s30 : 360 to 760 m/s). The Seward Park site V¯ s30 is about equal to, or up to 200 m/s slower than sites that were located on till or glacial outwash. High-amplitude P- and S-wave seismic reflections at several locations appear to correspond to strong resonances observed in earthquake spectra. An S-wave reflector at the Kingdome at about 17 to 22 m depth probably causes strong 2-Hz resonance that is observed in the earthquake data near the Kingdome.

An example of neotectonism in a continental interior - Thebes Gap, Midcontinent, United States

Abstract Some of the most intense neotectonic activity known in the continental interior of North America has been recently discovered on a fault zone in the Thebes Gap area, Missouri and Illinois. This faulting almost assuredly was accompanied by large earthquakes. The zone is located approximately 30 km north of the New Madrid seismic zone and consists of complex north-northeast- to northeast-striking, steeply dipping faults that have had a long-lived history of reactivation throughout most of the Phanerozoic. Geophysical studies by others suggest that the faults are rooted in the deeply buried Late Proterozoic and Early Cambrian Reelfoot rift system. Quaternary deposits are cut by at least four episodes of faulting, two of which occurred during the Holocene. The overall style of neotectonic deformation is interpreted as right-lateral strike-slip faulting. At many locations, however, near-surface displacements have stepped from one fault strand to another and produced normal and oblique-slip faults in areas of transtension and high-angle reverse faults, thrust faults, and folds in areas of transpression. There is evidence of reactivation of some near-surface fault segments during the great 1811–1812 New Madrid earthquakes. Quaternary faulting at Thebes Gap demonstrates that there are additional seismic-source zones in the Midcontinent, U.S., other than New Madrid, and that even in the absence of plate-margin orogenesis, intense neotectonic activity does occur over long time periods along crustal weaknesses in continental interiors.

Shallow P- and S-Wave Velocities and Site Resonances in the St. Louis Region, Missouri-Illinois

As part of the seismic hazard–mapping efforts in the St. Louis metropolitan area we determined the compressional and shear-wave velocities (Vp and Vs) to about a 40-m depth at 17 locations in this area. The Vs measurements were made using high-resolution seismic refraction and reflection methods. We find a clear difference in the Vs profiles between sites located on the river floodplains and those located in the upland urban areas of St. Louis. Vs30 (average Vs to 30-m depth) values in floodplain areas range from 200 to 290 m/s (NEHRP category D) and contrast with sites on the upland areas of St. Louis, which have Vs30 values ranging from 410 to 785 m/s (NEHRP categories C and B). The lower Vs30 values and earthquake recordings in the floodplains suggest a greater potential for stronger and more prolonged ground shaking in an earthquake. Spectral analysis of a M3.6 earthquake recorded on the St. Louis–area ANSS seismograph network indicates stronger shaking and potentially damaging S-wave resonant frequencies at NEHRP category D sites compared to ground motions at a rock site located on the Saint Louis University campus.

Surface seismic refraction/reflection measurement determinations of potential site resonances and the areal uniformity of NEHRP site class D in Memphis, Tennessee

We determined S-wave velocities (Vs) to about 40-m depth at 65 locations in the Memphis-Shelby County, Tennessee, area. The Vs measurements were made using high-resolution seismic refraction and reflection methods on the ground surface. We find a clear difference in the Vs profiles between sites located on the Mississippi River flood plain and those located to the east, mostly covered by loess, in the urban areas of Memphis. The average Vs to 30-m depth at 19 sites on the modern Mississippi River floodplain averages 197 m/s (±15 m/s) and places 17 of these sites at the low end of NEHRP soil profile category type D (average Vs 180-360 m/s). The two remaining sites are type E. Vs to 30-m depth at 46 sites in the urban areas east of the modern floodplain are more variable and generally higher than the floodplain sites, averaging about 262 m/s (±45 m/s), still within category D. We often observed the base of the loess as a prominent S-wave reflection and as an increase in Vs to about 500 m/s. Based on the two-way travel time of this reflection, during an earthquake the impedance boundary at the loess base may generate resonances in the 3- to 6-Hz range over many areas of Memphis. Amplitude spectra from four local earthquakes recorded at one site located on loess indicate consistent resonance peaks in the 4.5- to 6.5-Hz range.

Comparison of P- and S-wave velocity profiles obtained from surface seismic refraction/reflection and downhole data

High-resolution seismic-reflection/refraction data were acquired on the ground surface at six locations to compare with near-surface seismic-velocity downhole measurements. Measurement sites were in Seattle, WA, the San Francisco Bay Area, CA, and the San Fernando Valley, CA. We quantitatively compared the data in terms of the average shear-wave velocity to 30-m depth (Vs30), and by the ratio of the relative site amplification produced by the velocity profiles of each data type over a specified set of quarter-wavelength frequencies. In terms of Vs30, similar values were determined from the two methods. There is <15% difference at four of the six sites. The Vs30 values at the other two sites differ by 21% and 48%. The relative site amplification factors differ generally by less than 10% for both P- and S-wave velocities. We also found that S-wave reflections and first-arrival phase delays are essential for identifying velocity inversions. The results suggest that seismic reflection/refraction data are a fast, non-invasive, and less expensive alternative to downhole data for determining Vs30. In addition, we emphasize that some P- and S-wave reflection travel times can directly indicate the frequencies of potentially damaging earthquake site resonances. A strong correlation between the simple S-wave first-arrival travel time/apparent velocity on the ground surface at 100 m offset from the seismic source and the Vs30 value for that site is an additional unique feature of the reflection/refraction data that could greatly simplify Vs30 determinations.

Shallow subsurface structure of the Wasatch fault, Provo segment, Utah, from integrated compressional and shear-wave seismic reflection profiles with implications for fault structure and development

Integrated vibroseis compressional and experimental hammer-source, shear-wave, seismic reflection profiles across the Provo segment of the Wasatch fault zone in Utah reveal near-surface and shallow bedrock structures caused by geologically recent deformation. Combining information from the seismic surveys, geologic mapping, terrain analysis, and previous seismic first-arrival modeling provides a well-constrained cross section of the upper ∼500 m of the subsurface. Faults are mapped from the surface, through shallow, poorly consolidated deltaic sediments, and cutting through a rigid bedrock surface. The new seismic data are used to test hypotheses on changing fault orientation with depth, the number of subsidiary faults within the fault zone and the width of the fault zone, and the utility of integrating separate elastic methods to provide information on a complex structural zone. Although previous surface mapping has indicated only a few faults, the seismic section shows a wider and more complex deformation zone with both synthetic and antithetic normal faults. Our study demonstrates the usefulness of a combined shallow and deeper penetrating geophysical survey, integrated with detailed geologic mapping to constrain subsurface fault structure. Due to the complexity of the fault zone, accurate seismic velocity information is essential and was obtained from a first-break tomography model. The new constraints on fault geometry can be used to refine estimates of vertical versus lateral tectonic movements and to improve seismic hazard assessment along the Wasatch fault through an urban area. We suggest that earthquake-hazard assessments made without seismic reflection imaging may be biased by the previous mapping of too few faults.

High-resolution seismic reflection surveys and modeling across an area of high damage from the 1994 Northridge earthquake, Sherman Oaks, California

Approximately 3.6 km of P -wave seismic-reflection data were acquired along two orthogonal profiles in Sherman Oaks, California to determine whether shallow (less than 1-km depth) geologic structures contributed to the dramatic localized damage resulting from the 1994 Northridge earthquake. Both lines, one along Matilija Avenue and one along Milbank Street, crossed areas of both high and low damage. We believe these data reveal a geologic structure in the upper 600 m that contributed to the increased earthquake ground shaking in the high-damage areas south of and along the Los Angeles River. Of interest in these data is a reflection interpreted to be from bedrock that can be traced to the north along the Matilija Avenue profile. This reflecting interface, dipping northward at 15°–22°, may be an important impedance boundary because it is the lower boundary of a wedge of overlying low-velocity sediments. The wedge thins and terminates in the area where we interpret down-warped reflections as evidence of a shallow subbasin. The low-velocity subbasin sediments ( V s of 200 m/sec V p of 500 m/sec) may be up to 150 m thick beneath the channelized Los Angeles River. The area across the subbasin experienced greater earthquake damage from possible geometric focusing effects. Three-dimensional basin effects may be responsible for the variable damage pattern, but from these seismic profiles it is not possible to determine the regional structural trends. Two-dimensional elastic and SH -mode finite-difference modeling of the imaged structural geometry along Matilija Avenue suggests that a peak horizontal-velocity amplification factor of two-and-over can be explained in the high-damage area above the shallow subbasin and sediment wedge. Amplification factors up to 5 were previously observed in aftershock data, at frequencies of 2 to 6 Hz. Amplification in the elastic simulation at the Santa Monica Mountains range-front on the southern end of the Matilija profile, with the geologic layering and geometry interpreted from the seismic data, is also consistent with aftershock observations.

VS30 and Spectral Response from Collocated Shallow, Active‐, and Passive‐Source VS Data at 27 Sites in Puerto Rico

Shear-wave velocity (VS) and time-averaged shear-wavevelocity to 30 m depth (VS30) are the key parameters used in seismic site response modeling and earth- quake engineering design. Where VS data are limited, available data are often used to develop and refine map-based proxy models of VS30 for predicting ground-motion intensities. In this paper, we present shallow VS data from 27 sites in Puerto Rico. These data were acquired using a multimethod acquisition approach consisting of noninvasive, collocated, active-source body-wave (refraction/reflection), active- source surface wave at nine sites, and passive-source surface-wave refraction micro- tremor (ReMi) techniques. VS-versus-depth models are constructed and used to calculate spectral response plots for each site. Factors affecting method reliability are analyzed with respect to site-specific differences in bedrock VS and spectral response. At many but not all sites, body- and surface-wave methods generally determine similar depths to bedrock, and it is the difference in bedrock VS that influences site ampli- fication. The predicted resonant frequencies for the majority of the sites are observed to be within a relatively narrow bandwidth of 1-3.5 Hz. For a first-order comparison of peak frequency position, predictive spectral response plots from eight sites are plot- ted along with seismograph instrument spectra derived from the time series of the 16 May 2010 Puerto Rico earthquake. We show how a multimethod acquisition approach using collocated arrays compliments and corroborates VS results, thus adding confi- dence that reliable site characterization information has been obtained.