SCEC Award Number 14182 View PDF
Proposal Category Individual Proposal (Integration and Theory)
Proposal Title SEPARATING LOW- AND HIGH-FREQUENCY WAVE PROPAGATION IN THE UCSB BROADBAND MODELING METHOD
Investigator(s)
Name Organization
Ralph Archuleta University of California, Santa Barbara
Other Participants Jorge Crempien
SCEC Priorities 6e, 6c, 4e SCEC Groups GMP, EEII, SIV
Report Due Date 03/15/2015 Date Report Submitted N/A
Project Abstract
 In this report we illustrate the undesired effects of rapid amplitude decay with distance of high-frequency (HF) synthetic ground based on 1D crustal velocity structures. We show a simple solution to this problem by separating the wave propagation problem into a simplified single layer on top of a half-space for the HF portion of ground motion and a more realistic 1D multilayer model for the low-frequency (LF) portion of ground motion. After this modification, the UCSB method passed the SCEC validation exercise.
Intellectual Merit As a consequence of separating wave propagation into LF and HF, the UCSB method passed the SCEC validation exercise. The HF portion of the synthetic simulations is no longer under-predicting with distance. When compared to ground motion prediction equations and observed ground motions produced by real earthquakes, the bias shows no trend with distance. There is still over-prediction at LF due to trapped energy in shallow portions of the 1D velocity structure. This can be corrected with 3D Green’s functions.
Broader Impacts The ground motion prediction equations (GMPEs) are by default determined from acceleration response spectra of recorded ground motions. GMPEs are the basic tool used in earthquake engineering for the design of earthquake resistant structures. However, the data are generally limited in both magnitude and distance; in particular, there are few data within 20 km of large magnitude (M>7) earthquakes, where the largest amplitudes are expected. Moreover, the data are limited to particular regions of the earth and may not be generalized to regions with different velocity/attenuation structures. Synthetic seismograms generated by earthquake scenarios can fill in the data gaps. Different approaches to generating synthetic data provide epistemic uncertainty.
Exemplary Figure Figure 3.
Figure 3: (a) shows the mean bias of 50 realizations of a M 6.6 strike-slip and reverse earthquake at each station. (b) and (c) show the median RotD50 spectra in red dots at each period for all stations and 50 realizations for M 6.6 strike-slip and reverse earthquakes respectively. The blue boxes represent 90% confidence intervals of mean and the black whiskers are the extrema of 50 simulations for all stations and realizations. The black solid line are the averages between the four 2008 NGA-West1 GMPE models, as described in Goulet et al. (2015). In all plots, results are for a set of stations at 20 km closest distance to the rupture plane.