Laurie G. Baise

Estimating Unknown Input Parameters when Implementing the NGA Ground-Motion Prediction Equations in Engineering Practice

The ground-motion prediction equations (GMPEs) developed as part of the Next Generation Attenuation of Ground Motions (NGA-West) project in 2008 are becoming widely used in seismic hazard analyses. However, these new models are considerably more complicated than previous GMPEs, and they require several more input parameters. When employing the NGA models, users routinely face situations in which some of the required input parameters are unknown. In this paper, we present a framework for estimating the unknown source, path, and site parameters when implementing the NGA models in engineering practice, and we derive geometrically-based equations relating the three distance measures found in the NGA models. Our intent is for the content of this paper not only to make the NGA models more accessible, but also to help with the implementation of other present or future GMPEs.

Impediments to predicting site response: Seismic property estimation and modeling simplifications

Abstract We compare estimates of the empirical transfer function (ETF) to the plane SH -wave theoretical transfer function (TTF) within a laterally constant medium for invasive and noninvasive estimates of the seismic shear-wave slownesses at 13 Kiban-Kyoshin network stations throughout Japan. The difference between the ETF and either of the TTFs is substantially larger than the difference between the two TTFs computed from different estimates of the seismic properties. We show that the plane SH -wave TTF through a laterally homogeneous medium at vertical incidence inadequately models observed amplifications at most sites for both slowness estimates, obtained via downhole measurements and the spectral analysis of surface waves. Strategies to improve the predictions can be separated into two broad categories: improving the measurement of soil properties and improving the theory that maps the 1D soil profile onto spectral amplification. Using an example site where the 1D plane SH -wave formulation poorly predicts the ETF, we find a more satisfactory fit to the ETF by modeling the full wavefield and incorporating spatially correlated variability of the seismic properties. We conclude that our ability to model the observed site response transfer function is limited largely by the assumptions of the theoretical formulation rather than the uncertainty of the soil property estimates.

Impediments to predicting site response

Abstract We compare estimates of the empirical transfer function (ETF) to the plane SH -wave theoretical transfer function (TTF) within a laterally constant medium for invasive and noninvasive estimates of the seismic shear-wave slownesses at 13 Kiban-Kyoshin network stations throughout Japan. The difference between the ETF and either of the TTFs is substantially larger than the difference between the two TTFs computed from different estimates of the seismic properties. We show that the plane SH -wave TTF through a laterally homogeneous medium at vertical incidence inadequately models observed amplifications at most sites for both slowness estimates, obtained via downhole measurements and the spectral analysis of surface waves. Strategies to improve the predictions can be separated into two broad categories: improving the measurement of soil properties and improving the theory that maps the 1D soil profile onto spectral amplification. Using an example site where the 1D plane SH -wave formulation poorly predicts the ETF, we find a more satisfactory fit to the ETF by modeling the full wavefield and incorporating spatially correlated variability of the seismic properties. We conclude that our ability to model the observed site response transfer function is limited largely by the assumptions of the theoretical formulation rather than the uncertainty of the soil property estimates.

Validation and Application of Empirical Liquefaction Models

Empirical liquefaction models (ELMs) are the standard approach for predicting the occurrence of soil liquefaction. These models are typically based on in situ index tests, such as the standard penetration test (SPT) and cone penetration test (CPT), and are broadly classified as deterministic and probabilistic models. No objective and quantitative comparison of these models have been published. Similarly, no rigorous procedure has been published for choosing the threshold required for probabilistic models. This paper provides (1) a quantitative comparison of the predictive performance of ELMs; (2) a reproducible method for choosing the threshold that is needed to apply the probabilistic ELMs; and (3) an alternative deterministic and probabilistic ELM based on the machine learning algorithm, known as support vector machine (SVM). Deterministic and probabilistic ELMs have been developed for SPT and CPT data. For deterministic ELMs, we compare the “simplified procedure,” the Bayesian updating method, and the ...

Model Development and Validation for Intelligent Data Collection for Lateral Spread Displacements

The geotechnical earthquake engineering community often adopts empirically derived models. Unfortunately, the community has not embraced the value of model validation, leaving practitioners with little information on the uncertainties present in a given model and the model’s predictive capability. In this study, we present a machine learning technique known as support vector regression (SVR) together with rigorous validation for modeling lateral spread displacements and outline how this information can be used for identifying gaps in the data set. We demonstrate the approach using the free face lateral displacement data. The results illustrate that the SVR has relatively better predictive capability than the commonly used empirical relationship derived using multilinear regression. Moreover, the analysis of the SVR model and its support vectors helps in identifying gaps in the data and defining the scope for future data collection.

The Effect of Shallow San Francisco Bay Sediments on Waveforms Recorded during the MW 4.6 Bolinas, California, Earthquake

To investigate the effect of the shallow, low-velocity sediments on the seismic wave field in the northern San Francisco Bay, we modeled tangential component displacement seismograms recorded during the 18 August 1999 M W 4.6 Bolinas, California, earthquake. The modeling indicates that the velocity structure of Pleistocene horizons in the San Francisco Bay is important for simulations of weak ground motions for Bay Area earthquakes. Models including the Pleistocene sediments generate the 1-sec-period surface waves observed at several stations. Modeling of Treasure and Yerba Buena Island records requires structures approximately an order of magnitude higher in spatial resolution than the current 3D velocity models for the region. This pair of sites, located only 2 km apart in the bay, records a sixfold difference in peak ground acceleration during the Bolinas earthquake. Three transects are forward modeled using 1D frequency-wavenumber integration and 2D finite-difference methods. Generally the ground motions are characterized by a direct shear wave ( S ), a midcrustal reflection ( S 1), a near-receiver multiple ( S 2), and surface waves. The direct S arrival at all six stations requires a faster model than GIL7, the model routinely used to estimate earthquake source parameters using the Berkeley Digital Seismic Network. In addition, the timing of S 1 indicates the possibility of a dipping midcrustal interface. S 2 can be matched with a single strong impedance contrast at 3 km depth. A thin (200-m) surface layer of weathered rock and sediments simulates the surface waves that follow S 2 at the Richmond Field Station site. However, the surface waves at Treasure Island and the Berkeley sites are longer in duration and higher amplitude than at Richmond and require 2D structure. A simple shallow uniform basin model for the San Francisco Bay consisting of stiff sediments (shear-wave velocity, β = 400 m/sec; thickness ∼100 m) over weathered rock ( β = 1.5 km/sec) of the Franciscan assemblage produces surface waves in the 0.02-2 Hz passband at Treasure Island and the Berkeley sites. Manuscript received 27 July 2001.

Consistency of Ground-Motion Estimates Made using System Identification

This article isolates the systematic effects of the soil profile on earthquake ground motions recorded at vertical seismic arrays. An empirical Greens function is calculated for a soil interval between two sensors in a vertical array. The estimation technique used here falls into the category of site-response estimates but differs from the standard spectral ratio methods in that an extended Weiner filter (ARMA model) becomes the impulse-response function from one point to another in the soil profile. This method allows quantification of the accuracy of predicted ground motions at a given site with normalized mean square prediction errors generally under 10%, indicating effective and consistent estimates. The models are shown to reproduce site ground motions for inputs that differ over a wide range of peak ground acceleration (PGA), hypocentral locations, and over numerous occurrences. We demonstrate consistency of site ground-motion estimates at four sites that have undergone multiple strong shakings: Chiba, Japan; Garner Valley, California; Lotung, Taiwan; and Port Island, Japan. Consistency of site response over a wide range of shaking levels indicates an “effective” range of linear soil behavior for PGA up to 0.33 g at Chiba, 0.1 g at GVDA, and 0.21 g at Lotung. At Port Island, consistency was evident below 16 m at a surficial PGA of 0.54 g , but the surface layer liquefied. Results are presented as predicted time series and in the frequency domain with calculated variance. Results from Garner Valley and Chiba indicate that the location and magnitude of significant impedance contrasts within the soil profile control response. This article concentrates on levels of shaking that do not excite the soils into phase transitions. Our results indicate that strong nonlinear behavior (liquefaction) controls site response, but more minor and localized softening of the soil may not alter the site response enough to deter prediction within 10%.

A Geospatial Liquefaction Model for Rapid Response and Loss Estimation

We describe an approach to model liquefaction extent that focuses on identifying broadly available geospatial variables (e.g., derived from digital elevation models) and earthquake-specific parameters (e.g., peak ground acceleration, PGA). A key step is database development: We focus on the 1995 Kobe and 2010–2011 Christchurch earthquakes because the presence/absence of liquefaction has been mapped so that the database is unbiased with respect to the areal extent of liquefaction. We derive two liquefaction models with explanatory variables that include PGA, shear-wave velocity, compound topographic index, and a newly defined normalized distance parameter (distance to coast divided by the sum of distance to coast and distance to the basin inland edge). To check the portability/reliability of these models, we apply them to the 2010 Haiti earthquake. We conclude that these models provide first-order approximations of the extent of liquefaction, appropriate for use in rapid response, loss estimation, and simulations.

Regional Wave Propagation in New England and New York

We validate and improve 1D velocity models of the two main crustal provinces in the northeastern United States (NEUS), using seismograms from the 20 April 2002 M 5 Au Sable Forks earthquake, which is the largest earthquake in the region to be recorded by multiple, recently deployed, good-quality, regional broad- band stations. To predict and mitigate the effects of future earthquakes in the north- eastern United States, more information is needed regarding both the local earthquake sources and how seismic waves travel through the region. We investigate the source and regional wave propagation for the Au Sable Forks earthquake. The earthquake epicenter is located near the boundary of two distinct geological provinces, the Appalachian (New England) and Grenville (New York). We use a forward-modeling approach to study the waveforms recorded at 16 stations located within 400 km of the epicenter. We generate synthetic seismograms using the frequency-wavenumber method, testing several published models for the two provinces. Several models per- form well at low frequencies (<0:1 Hz). We refine these models and generate two alternative 1D crustal models for intermediate frequencies (<1 Hz) of engineering interest. Our new Grenville model performs better than previously published models for all six source-station paths modeled in that province according to goodness of fit statistics: variance reduction and correlation coefficient. Our alternative Appalachian model improves the fit of synthetics to data for five of the ten paths modeled in that province. From the results, we identify two specific sources of wave-field complex- ities that should be investigated in future studies of earthquake ground motions in NEUS: 3% azimuthal anisotropy in the Appalachian Province and complex wave paths along the boundary between the two provinces.

Ground-Motion Suite Selection for Eastern North America

Ground-motion suite selection for Eastern North America (ENA) is distinguished from suite selection for high seismic regions by uncertainty related to earthquake intensity, spectral shape, and the wide range of relevant periods experienced by low-ductility structures. Whereas trends in high seismic regions point toward developing smaller, more efficient suites for use in practice based on reliable intensity parameters, current research on moderate seismic regions requires the development of ground-motion suites capable of exciting the widest range of structural periods while accounting for uncertainty related to ground-motion intensity. This paper discusses uncertainty related to ENA ground motions in terms of the logic tree in the probabilistic seismic hazard analysis (epistemic uncertainty) and the deaggregation of hazard into magnitude and distance bins (aleatory uncertainty), recommends a suite selection process for addressing this uncertainty without amplitude scaling, and evaluates the effectiveness of a specific suite in the context of reliability-based performance assessment procedures. DOI: 10.1061/(ASCE)ST.1943-541X.0000258.