James Kaklamanos

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.

Modeling Dynamic Site Response Using the Overlay Concept

Finite element programs allow for an enhancement in the computational capabilities of earthquake site response models. However, many finite element programs involve complicated constitutive models that are difficult for the end user to implement. We present a methodology for modeling earthquake site response within a general finite element framework, using an overlay model to represent nonlinear soil behavior. Using parallel load-carrying elements with varying stiffness and yield stress, the behavior of any given backbone stress-strain relation can be replicated, along with hysteretic unloading-reloading (Masing) behavior. Our finite element modeling methodology makes use of existing conventional elastoplastic material models available in any general finite element program, without requiring the specification of any complicated constitutive models. To represent overlay elements in a finite element model, the user defines a number of finite elements and assigns each of them identical node numbers. The only necessary input parameters are density, elastic constants, and the backbone curve of a stress-strain relation. This paper focuses on the application of one-dimensional total-stress site response, but the framework could be easily extended to model cyclic hardening and softening, three- dimensional wave propagation, and soil-structure interaction.

A Case Study of Alternative Site Response Explanatory Variables in Parkfield, California

The combination of densely-spaced strong-motion stations in Parkfield, California, and spectral analysis of surface waves (SASW) profiles provides an ideal dataset for assessing the accuracy of different site response explanatory variables. We judge accuracy in terms of spatial coverage and correlation with observations. The performance of the alternative models is period-dependent, but generally we observe that: (1) where a profile is available, the square-root-of-impedance method outperforms VS30 (average S-wave velocity to 30 m depth), and (2) where a profile is unavailable, the topographic-slope method outperforms surficial geology. The fundamental site frequency is a valuable site response explanatory variable, though less valuable than VS30. However, given the expense and difficulty of obtaining reliable estimates of VS30 and the relative ease with which the fundamental site frequency can be computed, the fundamental site frequency may prove to be a valuable site response explanatory variable for many applications.