Conveners: C.B. Crouse and Tom Jordan
Dates: May 12, 2014 (10:00 - 17:30)
Location: SCEC Boardroom, USC, Los Angeles

SUMMARY: The SCEC Committee for Utilization of Ground Motion Simulations (or "UGMS Committee") is tasked to develop long-period response spectral acceleration maps for Los Angeles region for inclusion in NEHRP and ASCE 7 Seismic Provisions and in Los Angeles City Building Code. The maps would be based on 3-D numerical ground-motion simulations, and ground motions computed using latest empirical ground-motion prediction equations from the PEER NGA project. This project is coordinated with (1) the SCEC Ground Motion Simulation Validation Technical Activity Group (GMSV-TAG), (2) other SCEC projects, such as CyberShake and UCERF, and (3) the USGS national seismic hazard mapping project.

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 They may contain unpublished or preliminary information and should only be used while viewing the talk.

Monday, May 12, 2014

PARTICIPANTS

*Remote Participants

NEW MEMBERS

Marshall Lew and Marty Hudson, geotechnical engineers in AMEC’s Los Angeles office, were added as new members and both attended the meeting.

PROJECT OVERVIEW

Crouse presented an overview of the pilot project, with the ultimate objective of producing long period response spectral acceleration maps of the Los Angeles Metropolitan region. The plan is to develop the maps in coordination with (1) the SCEC Ground Motion Simulation Validation Technical Activity Group (GMSV-TAG), (2) other SCEC projects, such as Cybershake and UCERF, and (3) the USGS national seismic hazard project. The maps would be submitted to the NEHRP Provisions Update Committee next code cycle. The tentative schedule was attached to the minutes of the 1st UGMS committee meeting in April 2013.

TECHNICAL PRESENTATIONS

Much of the meeting was devoted to a discussion of the CyberShake platform for generating ground-motion simulations throughout Southern California. T. Jordan first presented an overview of CyberShake and the SCEC ground motion simulation program. Jordan followed with presentations on the SCEC community velocity models, and comparisons of ground-motion predictions from CyberShake and the empirical NGA equations. Jordan also showed that Cybershake was able to reproduce long period ground motions from recent local earthquakes. Jordan’s final presentation was plans for future CyberShake development, which included extending the frequency band to 1.3 Hz to better define 1-sec period motions.

DISCUSSION

Deterministic/stochastic methods were mentioned as possible approaches to expand CyberShake’s ability to model higher frequency ground motions. This was not viewed as a necessary development for the successful production of the long period maps, because the maps at higher frequencies would continue to be developed using the traditional empirical ground- motion prediction equations. Jordan asked the committee whether long period maps for the vertical component would be useful to structural engineers, since CyberShake automatically generates this component in addition to the two orthogonal horizontal components. There was general agreement that it would be informative to see vertical ground motions as a check with provisions for the vertical component that were added to the 2009 NEHRP Seismic Provisions.

Nico Luco said (and Jordan confirmed) that unlike the USGS seismic source model, the CyberShake so far has only computed ground motions from ruptures on known active faults, whereas, the USGS also includes random earthquakes in its seismic hazard analysis. The question to be addressed is whether the random earthquakes increase the ground motion hazard at long periods by an insignificant amount. The USGS can quickly answer this question by examining the source deaggregations in its PSHA.

Another discussion topic was the performance measures that should be used in validating the simulations from CyberShake. Simulated ground motions will be compared to ground motions from the empirical NGA equations, as a benchmark, but that exercise does not necessarily represent a validation. Art Frankel questioned whether the absolute values from CyberShake could be trusted because of the uncertainties in the fault slip function. Art is using 3-D simulations to compute long period basin amplification factors for the Puget Sound region. The uncertainty in the source slip function needs to be considered. Rob Graves was not present to offer his views on the subject, but the topic will need to be addressed in the future.

RESOLUTIONS

The UGMS continued to affirm that the long period mapping project will proceed on two parallel tracks, in which PSHA/DSHA will be conducted from the 3-D numerical simulations using CyberShake and from the traditional empirical approach using the NGA-West equations. It was agreed that the results from the 3-D simulations could be used to refine the equations ultimately used in the empirical approach. The topic of how to use the results from the two approaches to construct the long period maps was discussed will be deferred until the results are generated.

Sensitivity studies will be conducted to evaluate the effect of the near surface velocity structure on the long period ground motions. Varying the mesh size in the near surface would address the issue, and depending on the results, some refinements the the near surface velocity structure may be made.

FOCUS OF NEXT MEETING

T. Jordan will present results of the 3-D Cybershake model at selected sites in Southern California. The results will consist of 5% damped, horizontal component response spectra, generated according to the requirements of Chapter 21 in the ASCE 7 standard. The approach will be the same as used by the USGS in creating the national risk-targeted Maximum Considered Earthquake (MCER) maps at 0.2-sec and 1.0-sec periods. The probabilistic and deterministic MCER response spectra would be computed for periods in the 2.0 to 10 sec range, and compared with similar results generated using the NGA West2 equations.

ACTION ITEM

Crouse will select sites for the CyberShake calculation of the MCER response spectra, and N. Luco will provide the Matlab routine to do calculate the probabilistic MCER from the hazard curves. Crouse will provide details on the calculation of the maximum direction shaking component.

DATE FOR NEXT UGMS COMMITTEE MEETING

The next UGMS committee meeting is tentatively scheduled for one day in the fall of this year in the time window, November 3 to November 21. A Doodle poll will be sent to members to arrange a date. 

CYBERSHAKE HAZARD MAPS
Map of CyberShake sites. Purple sites are along a 10 km x 10 km mesh, orange sites are near SCSN broadband stations, and yellow sites are locations of interest.
Site Name	NGA-1 3 second curve	NGA-1 5 second curve	NGA-2 3 second curve	NGA-2 5 second curve
LADT	LADT 3 sec	LADT 5 sec	LADT 3 sec	LADT 5 sec
CCP	CCP 3 sec	CCP 5 sec	CCP 3 sec	CCP 5 sec
P22	P22 3 sec	P22 5 sec	P22 3 sec	P22 5 sec
s429	s429 3 sec	s429 5 sec	s429 3 sec	s429 5 sec
STG	STG 3 sec	STG 5 sec	STG 3 sec	STG 5 sec
s684	s684 3 sec	s684 5 sec	s684 3 sec	s684 5 sec
s688	s688 3 sec	s688 5 sec	s688 3 sec	s688 5 sec
PAS	PAS 3 sec	PAS 5 sec	PAS 3 sec	PAS 5 sec
s758	s758 3 sec	s758 5 sec	s758 3 sec	s758 5 sec
SBSM	SBSM 3 sec	SBSM 5 sec	SBSM 3 sec	SBSM 5 sec
STNI	STNI 3 sec	STNI 5 sec	STNI 3 sec	STNI 5 sec
WNGC	WNGC 3 sec	WNGC 5 sec	WNGC 3 sec	WNGC 5 sec
s355	s355 3 sec	s355 5 sec	s355 3 sec	s355 5 sec

RUPTURE GENERATOR COMPARISONS
Hypocenter distribution for source 244, rupture 5 (M6.55, Puente Hills)	Slips for a single realization of source 244, rupture 5 (M6.55, Puente Hills)
Hypocenter distribution for source 128, rupture 1296 (M8.15, Southern San Andreas)	Slips for a single realization of source 128, rupture 1296 (M8.15, Southern San Andreas)