Barbara Luke

Improved parameterization to invert Rayleigh-wave data for shallow profiles containing stiff inclusions

Inversion of shear-wave velocity profiles from phase-velocity measurements of Rayleigh-wave energy for sites containing stiff layers can be erroneous if such layers are not characterized in the starting or reference model. Incorporation of a priori knowledge then is key for converging upon a realistic or meaningful solution. Resolving soil profiles in desert regions where stiff layers cemented with calcium carbonate are intermixed with softer, uncemented media is an application for which locating shallow stiff inclusions has important implications. Identification of the stiff layers is critical for foundation design and cost estimating of excavations. A parameterization that seems adequate for this problem is to solve for anticipated high-stiffness layers embedded in a coarser (background) profile that captures the general shear-wave velocity trend of the study area. The optimization is accomplished by using simulated annealing. Uncertainty measures resulting from the inversion are helpful for describing the influence of the parameterization on final model estimates.

Shallow shear velocity and seismic microzonation of the urban Las Vegas, Nevada, basin

Las Vegas Valley has a rapidly growing population exceeding 1.5 million, subject to significant seismic risk. Surveys of shallow shear velocity performed in the Las Vegas urban area included a 13-km-long transect parallel to Las Vegas Blvd. (The Strip), and borehole and surface-wave measurements of 30 additional sites. The transect was completed quickly and economically using the refraction microtremor method, providing shear velocity versus depth profiles at 49 locations. The lowest velocities in the transect, nehrp d class, are near intrabasin faults found near Interstate 15 and Lake Mead Blvd. Calcite cementation of alluvium (a.k.a. caliche) along the Las Vegas Strip elevates Vs30 values to 500–600 m/sec, nehrp c class. Our transect measurements correlate poorly against geologic map units, which do not predict the conditions of any individual site with accuracy sufficient for engineering application. Some usda soil map units do correlate, and Vs30 predictions based on measurements of soil units match transect measurements in the transect area. Extending soil-map predictions away from the area of dense measurement coverage generally failed to predict new measurements. Further, for several test sites the predictions were not conservative, in that the soil model predicted higher Vs30 than was later measured (predicting lesser potential ground motion). Subsurface information is needed to build a Vs30 model extending predictions throughout Las Vegas Valley. A detailed stratigraphic model built by correlating >1100 deep well logs in Las Vegas predicts Vs30 better than surface maps, but again only in parts of the Valley well-measured for velocity. The stratigraphic model yields good predictions of our transect Vs30 measurements. It is less accurate, although at least conservative, when extended to sites away from the transect.

Interpreting Surface-Wave Data for a Site with Shallow Bedrock

The inversion of dispersive Rayleigh-wave data has been shown to be successful in providing reliable estimated shear-wave velocities within unconsolidated materials in the near surface. However, in a case where the multi-channel analysis of surface waves method was applied to a site consisting of clay residuum overlying basalt bedrock, inversion for the fundamental-mode Rayleigh wave resulted in shear-wave velocities within the rock that are less than half of expected values. Forward modeling reveals that the fundamental-mode dispersion curve is hardly sensitive to bedrock velocity perturbations over a practical range of wavelengths, leading to poorly constrained solutions. Standard surface-wave methods can fail because of a shortage of phase-velocity estimates at the low frequencies that are necessary to properly constrain shear-wave velocities at depth. The commonly used guideline that maximum investigation depth is roughly half of the largest recorded wavelength can be misleading. Data at much lower fr...

Comparison of Rayleigh Wave Dispersion Relations from Three Surface Wave Measurements in a Complex-Layered System

The study presented herein documents procedures to assess uncertainty in shear wave velocity determined via SASW, and demonstrates use of these uncertainty estimates to guide the inversion process to provide a quality, realistic shear wave velocity profile. Based upon the analyses performed, the inversion protocol guided by uncertainty estimates is shown to successfully invert dispersion data. For the site example presented, the trend in shear wave velocities determined via the protocol is quite consistent with the material classification and standard penetration test results available for the site. It is shown that uncertainty in shear wave velocity is a function of the number of soil layers used in the inversion. As the number of layers increases, uncertainty in shear wave velocity also increases. The results also suggest that the layer resolution of SASW inversion on soil sites is approximately 0.3048 m (1 ft).

Inversion of Rayleigh Wave Data For Shallow Profiles Containing Stiff Layers

Summary A non-linear inversion scheme for interpreting complex shear wave velocity profiles from the fundamental mode of Rayleigh wave data is presented. The scheme incorporates a priori information in the form of expected ranges of layer depth, thickness and shear wave velocity to constrain stiff layers that otherwise might not be resolved without this information. The scheme, based on the simulated annealing optimization algorithm, inverts for a shear wave velocity profile with a fixed number of layers in which velocity primarily increases with depth, and for velocities and depths of a given number of stiff inclusions. The proposed scheme is tested to resolve a shallow soil profile in a desert site where stiff cemented layers are intermixed with softer, uncemented soil.

Investigation and Implications of Mechanically Stabilized Earth Wall Corrosion in Nevada

The Nevada Department of Transportation has more than 150 mechanically stabilized earth (MSE) walls at 39 locations. Wall reinforcement corrosion was found by accident during construction projects at two of these locations. The resulting investigations of these walls produced direct measurements of metal losses and electrochemical properties of the MSE-reinforced fill. One MSE wall was replaced with a cast-in-place concrete tie-back wall at great expense. A statistical analysis addresses the variability in measured corrosion and electrochemical data to predict corrosion behavior. The original MSE-reinforced fill approval electrochemical test results are significantly different from those measured in postconstruction investigations. A correlation has been developed between two distinctly different soil resistivity test methods: the Nevada T235B and AASHTO T-288 methods. Overprediction made by the Nevada T235B method has proven detrimental to the service lives of MSE walls. The internal stability analyses (using AASHTO 2007 load and resistance factor design specifications) of two remaining MSE walls at an intersection were also performed by using metal loss models developed from the statistical analysis. The findings of the study were subsequently extrapolated to other Nevada MSE walls. Through review of the reinforced fill approval data, suspect Nevada MSE walls have been identified relative to estimated reinforced fill aggressiveness.

Characterizing anomalous ground for engineering applications using surface-based seismic methods

Shallow seismics are in demand today in tectonically active regions to characterize and classify sites for earthquake response studies. The surface-based seismic methods are the most widely used for this purpose. In developed areas, the passive-source methods, also known as microtremor methods, are popular because of their efficiency and because the available frequency content is appropriate to determine an average shear-wave velocity for the upper 30 m. This information is required by the International Building Code, which is used by many municipalities in the US and elsewhere.

Application of seismic surface waves at a pre‐Columbian settlement in Honduras

A technique to locate culturally significant features by means of mechanical stiffness contrasts using seismic surface waves was applied at a Pre-Columbian settlement along the Talgua River near the town of Catacamas, Department of Olancho, Honduras, Central America. This technique was adapted from geotechnical engineering, where it is used to establish layering of soil and rock. With this surface-based, non-intrusive technique, a stress pulse is applied at the ground surface and the resulting ground motion is sensed and recorded using a pair of geophones placed on the ground surface a short distance away. Through spectral analysis, the dispersive characteristics of the earth between the two geophones are determined. Anomalies were mapped by comparing dispersion “signatures” along linear arrays. For corroboration, electrical resistivity measurements were also conducted. Ground truth was established through excavations, during which several significant features were unearthed. The most valuable result of the geophysical investigation was the discovery that a thick layer of fill had been brought in to level the site. This documented that the Talgua Village was constructed on top of a massive modification of the natural landscape and implies a higher level of political power and complexity than suggested by the modest size of the surface mounds. Copyright

Evaluating the Seismic Response of Deep Sandy Soil Deposits

We investigated the influence of depth to the half-space on evaluations of seismic response of a dry sandy soil deposit with thickness exceeding 100 m. Our investigation is based on the premise that the best estimate of surface response of such deposits is obtained using a statistically derived attenuation relationship, and the best estimate of other parameters required for geotechnical earthquake-engineering evaluations is obtained by calibrating the evaluated spectral characteristics of the site response using the best-estimate surface response. Results of our site-specific evaluations indicate that the preferred depth to the half-space does not coincide with the depth to bedrock, nor is the preferred depth uniquely defined in the seismic site-response analysis. We find that for a given accelerogram, the preferred depth to the half-space corresponds closely to the depth for which the peak horizontal ground acceleration (PHGA) predicted in the site-response analysis matches the median PHGA developed from the appropriate statistical attenuation relationship. Our results further indicate that the effect of the magnitude of the shear-wave velocity assigned to the half-space is minor in comparison with depth to the half-space.

Feasibility of Approximating Depth to Shallow Bedrock Directly from the Rayleigh Wave Dispersion Curve

A complementary pair of straightforward, data-driven techniques is presented for use in making an initial estimate of the depth to shallow bedrock for a simple profile consisting of homogeneous soil over bedrock, using only the high-frequency portion of the fundamental-mode Rayleigh wave dispersion curve. These techniques constitute tools that can be useful for developing high-quality starting models for the inversion process that is followed to resolve shear wave velocity profiles from Rayleigh wave dispersion curves and for conducting rapid checks on theoretical dispersion calculations. Synthetic studies addressed a suite of one-dimensional shear wave velocity profiles, each representing a homogeneous soft-sediment layer above the bedrock halfspace. The shapes of the dispersion curves and the wave velocities at the high-frequency limits of the curves were considered in order to develop an empirical relationship to estimate the depth to bedrock. A less subjective, but more time-intensive, estimate of the depth to bedrock is obtained by constructing a suite of master curves against which to compare the experimental data. Both tools were tested upon two experimental datasets collected at shallow bedrock sites. The tools accurately yielded the depth to bedrock in the presence of a homogeneous overburden. In the presence of a layered overburden, the tools were less effective.