Workshop on Implementation of GMSV Gauntlets on the Broadband Platform

Jump to: navigation, search
Organizers: Nico Luco, Christine Goulet, Andreas Skarlatoudis, Jeff Bayless, Phil Maechling, & Fabio Silva
Date: Monday, February 29, 2016 (10:00-16:00)
Location: SCEC Boardroom, University of Southern California, Los Angeles, CA
Participants: See below

Background & Objectives

One of the current projects of the SCEC Ground Motion Simulation Validation (GMSV) Technical Activity Group (TAG) is "Implementation of GMSV Gauntlets on the Broadband Platform." This multi-PI project is building on knowledge from previous SCEC GMSV projects to implement GMSV parameters beyond pseudo spectral acceleration (e.g., the ratio of maximum-direction to median spectral acceleration; see others in the first presentation posted at The objectives of this workshop are to demonstrate use of the newly implemented parameters (and corresponding empirical prediction models) and to obtain feedback from developers and users of the Broadband Platform.


Presentation slides may be downloaded by clicking the title of the presentation. PLEASE NOTE: Files are the author’s property. They may contain unpublished or preliminary information and should only be used while viewing the talk.

10:00 - 10:05 Welcome and introductions C. Goulet for T. Jordan
10:05 - 10:15 Overview of “Implementation …” Project N. Luco & F. Silva

10:15 - 10:30 Baker et al spectral ground motion parameters J. Baker
10:30 - 10:45 Example results from BBP C. Goulet
10:45 - 11:15 Discussion (may be intertwined with presentations) All

11:15 - 11:30 Rezaeian et al time-domain ground motion parameters S. Rezaeian
11:30 - 11:45 Example results from BBP J. Bayless
11:45 - 12:15 Discussion (may be intertwined with presentations) All

12:15 - 12:55 Lunch

12:55 - 13:10 Rezaeian et al scalar ground motion parameters S. Rezaeian
13:10 - 13:25 Example results from BBP N. Luco
13:25 - 13:55 Discussion (may be intertwined with presentations) All

13:55 – 14:10 Anderson et al goodness of fit J. Anderson
14:10 – 14:25 Example results from BBP A. Skarlatoudis
14:25 – 14:55 Discussion (may be intertwined with presentations) All

14:55 – 15:05 Break

15:05 – 15:15 Next steps, e.g., acceptance criteria N. Luco
15:15 – 15:30 Inter-period correlations for Fourier amplitude spectra J. Bayless
15:30 – 16:00 Discussion (may be intertwined with presentations) All


In-Person: 1) Kioumars Afshari, UCLA; 2) John Anderson, UNR; 3) Jacobo Bielak, CMU; 4) Nenad Bijelic, Stanford; 5) Greg Deierlein, Stanford; 6) Christine Goulet, SCEC; 7) Rob Graves, USGS; 8) Tran Huynh, USC/SCEC; 9) Ting Lin, Marquette; 10) Nicolas Luco, USGS; 11) Phil Maechling, USC/SCEC; 12) Kim Olsen, SDSU; 13) Fabio Silva, USC; 14) Andreas Skarlatoudis, AECOM; 15) Jonathan Stewart, UCLA; 16) Ricardo Taborda, UMemphis; 17) Farzin Zareian, UCI

Remote: 18) Norm Abrahamson, PG&E; 19) Ralph Archuleta, UCSB; 20) Jack Baker, Stanford; 21) Jeff Bayless, AECOM; 22) Brendon Bradley, UCanterbury; 23) Reagan Chandramohan, Stanford; 24) Jorge Crempien, UCSB; 25) Mayssa Dabaghi, AU-Beirut; 26) Steve Hartzell, USGS; 27) Taojun Liu, CU-Boulder/USGS; 28) Morgan Moschetti, USGS; 29) Sanaz Rezaeian, USGS; 30) Karim Tarbali, UCanterbury; 31) Peng Zhong, UCI

Unable to Participate: Domniki Asimaki (Caltech), Gail Atkinson (Western Ontario), Greg Beroza (Stanford), Douglas Dreger (UC-Berkeley), Tom Jordan (USC/SCEC), Paul Somerville (AECOM)

Action Items

For Baker et al spectral ground motion parameters ...

  • ASAP: Implement ratio of inelastic to elastic spectral displacement and its empirical ground-motion model (GMM) on the BBP.
  • ASAP: Implement empirical ground-motion model for RotD100/RotD50.
  • ASAP: Revise BPP bias plotting such that they illustrate RotD100/RotD50 for simulations, recordings, and GMMs, rather than (or in addition to) the ratios of these.
  • ASAP: Adapt existing BBP bias mapping (for pseudo spectral acceleration) to Baker et al spectral GMPs.
  • ASAP: Consider revising the period range of the BBP figures.
  • Later: After comparing inter-period correlations for Fourier amplitude vs. response spectra, implement one or both.
  • Later: Implement 2-D spectral displacement orbit plots (the basis for RotD100/RotD50).
  • Later: Implement a RotD100 ground motion parameter (GMP) and a corresponding GMM.

For Rezaeian et al time-domain ground motion parameters ...

  • ASAP: Resolve time-step issue so that time-domain GMPs for a recording do not vary when compared with different simulations.
  • ASAP: Plot time-domain GMPs for all realizations on each BBP figure. Consider plotting both horizontal components on each figure.
  • ASAP: Consider truncating time series in order to focus on levels of engineering interest.
  • Later: Consider generating BBP figures for the difference measures epsilon and nu.

For Rezaeian et al scalar ground motion parameters ...

  • ASAP: Add the already-implemented Afshari & Stewart (2016) GMM for duration to the BBP figures.
  • ASAP: Combine the GMPs for the two horizontal components, e.g. by plotting both on each BBP figure or by taking the geometric mean.
  • ASAP: Adapt existing BBP bias mapping (for pseudo spectral acceleration) to Rezaeian et al scalar GMPs.
  • ASAP: Consider adding a sixth panel for zeta to the BBP figure. The calculation of zeta is already implemented.
  • ASAP: Consider implementing source/magnitude-scaling figures. Correct for path/distance (and possibly focal mechanism) in such figures.
  • Later: Implement a more robust GMM for Arias intensity (e.g., Campbell & Bozorgnia, 2012).
  • Later: Consider replacing w_mid with mean period from Rathje et al (1998).

For Anderson et al goodness-of-fit (GOF) parameters ...

  • ASAP: QA/verify implementation.
  • ASAP: Adapt existing BBP bias mapping (for pseudo spectral acceleration) to Anderson et al GOF parameters.
  • Later: Consider changing the weights used in the current GOF parameters/scores and/or introducing new parameters.

For all parameters ..."

  • ASAP: Implement a means of distinguishing results from stations with Vs30 "close" to that of the simulations, i.e. 863 m/s.
  • ASAP: Post Python codes and documentation for parameters and corresponding empirical ground motion models on a SCEC GMSV TAG wiki.
  • Later: Prepare for post-event "blind" validations.