Preliminary ground motion prediction equations for the Central and Eastern United States

Wednesday, 17 December 2014
Vladimir Graizer, U.S. Nuclear Regulatory Commission, Rockville, MD, United States
At the current stage I used the database created under the Next Generation Attenuations (NGA-East) project by Cramer et al. (2013). In contrast to the active tectonic environment in the Western US (WUS) the strong motion record database for the stable continental environment in the Central and Eastern US (CEUS) is not sufficient to create purely empirical ground motion prediction equations (GMPE) covering required for the PSHA magnitude (4.5<Mw<8.5) and distance ranges (0<R<1000 km). Recorded data are sparse and cover mostly ranges of Mw<6.0 with only few data points for larger earthquakes. There is also only small number of recordings at close fault distances (R<10 km). Absences of large magnitude and near fault recordings significantly complicate constraining GMPE. I used previously developed GMPE models developed by Electric Power Research Institute (EPRI) in 2013 and Pezeshk et al. (2011) as main constraints especially for large magnitudes.

Current GMPE model is based on the same modular approach developed by Graizer and Kalkan for active tectonic environment (2007, 2009, and 2011). In this approach, a GMPE is expressed as a series of filters with each filter representing a certain physical phenomenon affecting the radiation of seismic waves from the source. Graizer-Kalkan GMPE for peak ground acceleration (PGA) was composed of five different filters. For consistency I keep same number of filters for the CEUS GMPE. The main filter modeling the distance attenuation of ground-motion is similar to the frequency response function of a damped single-degree-of-freedom oscillator with frequency replaced by distance to the fault. Following the approach developed in Graizer and Kalkan (2009) for WUS spectral acceleration (SA) prediction model explicitly integrates PGA as a scaling factor for the spectral shape, which is a continuous function of spectral period. Geometrical spreading of R-1 combined with anelastic attenuation Q0 is used allowing adjusting the model for regional anelastic attenuation. The model covers the range of 4.5<Mw<8.5, distances of 0<R<1000 km, S-wave velocities of 500<VS30<2800 m/sec and period range of 0.01 to 10 s. Any opinions, findings and conclusions expressed in this abstract are those of the author and do not necessarily reflect the views of the U.S. Nuclear Regulatory Commission.