SCEC Award Number 14020 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Methodology for Validating Median Values and Aleatory Variability of Broadband Platform Ground Motion Simulations
Name Organization
Jeff Bayless URS Corporation Andreas Skarlatoudis URS Corporation Paul Somerville URS Corporation
Other Participants
SCEC Priorities 6e SCEC Groups EEII, WGCEP, SIV
Report Due Date 03/15/2015 Date Report Submitted N/A
Project Abstract
In Part (1) we evaluate the simulation residuals (relative to data) for seven crustal earthquake simulations from the Phase 1 SCEC Broadband Platform Validation project, utilizing the 50 source realizations of each validation event, and 5 different simulation methods. We compare the variability of the residuals with the 2013 NGA-West2 empirical models, both for individual events, and for the entire validation catalog as a whole. The variability of the empirically-derived GMPEs, while not perfect for comparisons, serve as guidelines for the variability of the simulation residuals. Additionally, the variability of the 5 simulation methods are compared with each other. We find the variability of event terms to be smaller than those from the GMPEs, as expected, for all methods but one. And we find that generally, the total standard deviations behavior are in line with those from the NGA-West2 GMPEs, especially for the California events and for the panel-recommended simulation methods. In Part (2) we implement new median validation metrics (Anderson, 2004) in the time and frequency domains for use with future validation type simulations, and apply them to the existing validation events. These new GOF metrics utilize various intensity measures (IMs) to supplement the currently utilized measures – Pseudo Spectral Acceleration (PSA) GOF, and the GOF criteria of Olsen and Mayhew (2010). Finally, working in close collaboration with SCEC IT, we integrate the new validation metrics within the BBP.
Intellectual Merit The research contributes directly towards the long-term project of the SCEC Broadband Platform development, including a better understanding of the existing simulation methods, as well as the goal for gaining credibility and acceptance in the results coming from the simulations.
Broader Impacts The project contributes to the SCEC GMSV TAG as well as the ongoing BBP Validation group, and is a direct result of the strong sense of teamwork and collaboration within both of these groups. The project itself also embodies the collaborative spirit of these projects, since the products of the activity are useful to the group as a whole. One long term potential benefit is the acceptance by the engineering community of the kinematic simulations for use in developing/updating infrastructure.
Exemplary Figure Figure 3. Separation of residuals using Method 1, showing Graves & Pitarka standard deviation of residuals for the Loma Prieta validation simulation (solid lines) and the Abrahamson et al. (2013b) NGA-West2 GMPE standard deviation (dashes lines).