A Hybrid Empirical Ground Motion Prediction Model for Eastern North America

Shahram Pezeshk, The University of Memphis, Civil Engineering, Memphis, TN, United States, Arash Zandieh, Lettis Consultants International, Inc, Boulder, CO, United States, Kenneth W Campbell, CoreLogic EQECAT, Oakland, CA, United States and Behrooz Tavakoli, Bechtel Power Corporation, San Francisco, CA, United States

Contact First Author: Shahram Pezeshk; spezeshk@memphis.edu

Previously Published Material: Pacific Engineering Research Center NGA-East

Abstract ID#: 34430

 

English Abstract:
A hybrid empirical method (HEM) is utilized to develop two new ground-motion prediction equations (GMPEs) for Eastern North America (ENA) using five new NGA-West2 GMPEs developed by the Pacific Earthquake Engineering Research center (Bozorgnia et al., 2014). The two new GMPEs are derived for a moment magnitude (M) range of 4 to 8 and shortest distances to the fault rupture (RRUP) as far as 1000 km. The GMPEs are developed for the 5%-damped pseudo-acceleration response spectra and the peak ground acceleration (PGA) for hard-rock sites with VS30 = 3000 m/sec.

Seismological parameters for ENA are adopted from the most recent research and published information in ENA (Yenier and Atkinson, 2015a; Chapman, et al., 2014; Boore and Thompson, 2015; Hashash, et al., 2014). Seismological parameters for western North America (WNA) are adopted from a study by Zandieh and Pezeshk (2015) in which they performed a set of point-source inversions to match the median NGA-West2 GMPEs for M ≤ 6.0, RRUP ≤ 200 km, VS30 = 760 m/sec, strike-slip faulting, and sediment-depth parameters equal to the default values recommended by each of the NGA-West2 developers.

The two new sets of ENA GMPEs are based on two approaches: (1) using HEM to model magnitude scaling over the entire range of magnitudes; and (2) using HEM to model magnitude scaling for M ≤ 6.0 and using the magnitude-scaling predicted by the NGA-West2 GMPEs for M > 6.0.

The new GMPEs can be used as an alternative to those developed for ENA by other methods (e.g., stochastic and numerical simulation).