SCEC Award Number 12113 View PDF
Proposal Category Collaborative Proposal (Integration and Theory)
Proposal Title Enabling the SCEC Modeling Community to Conduct Physics-Based, Deterministic Broadband Earthquake Simulations using a High-Performance Portable Parallel Software
Investigator(s)
Name Organization
Jacobo Bielak Carnegie Mellon University Ricardo Taborda Carnegie Mellon University
Other Participants It is expected that we will leverage additional help from Carnegie Mellon Ph.D. students Haydar Karaoglu, Yigit Isbiliroglu, and Dorian Restrepo.
SCEC Priorities 6c, 6e, 6a SCEC Groups CS, GMP, CME
Report Due Date 03/15/2013 Date Report Submitted N/A
Project Abstract
 The main goal of this project was to prepare a tool-framework that would enable SCEC’s modeling community and outside partners to conduct physics-based, deterministic earthquake simulations using Hercules—the octree-based finite-element portable parallel software for seismic wave-propagation problems developed by the Quake Group at Carnegie Mellon University (CMU) (Tu et al., 2006; Taborda et al., 2010). To that end, during the last year we have taken a series of steps to ensure future use of Hercules beyond its group of developers at CMU. These activities include: (1) moving the Quake Group Web site and computer codes archive from a private to a public space; (2) porting Hercules source code from a private CVS-based version control system, with a CMU-access only repository, to a Web-based hosting service using the Git revision control system, with a membership access repository; (3) providing an issue tracker and a Wiki documentation infrastructure for supporting the code; and (4) creating a light version of Hercules (Hercules-LT) for non-developers. These steps constitute a first stage in preparation of Hercules as a community code that will enable the SCEC modeling community to conduct physics-based, deterministic simulations at increasingly higher frequencies (~5 Hz). Other impacts of bringing Hercules up to a more open framework for community development is its potential future use in interdisciplinary projects such as CyberShake and the Broadband Simulation Platform, or its use as a reference tool for other groups working on simulation, tomography, and verification and validation of the various source and velocity models.
Intellectual Merit Hercules has been thoroughly verified in multiple large-scale simulations in regional basins under realistic conditions, with strong heterogeneities and complex seismic faulting. Examples of this are the simulations performed with Hercules for the TeraShake and the ShakeOut scenario earthquakes and the latest validation efforts for simulations of the 2008 Chino Hills earthquake. Having prepared Hercules to become a community code will allow members of the SCEC modeling community to continue advance deterministic physics-based earthquake simulations as a complementary alternative for seismic hazard analysis.
Broader Impacts Preparing Hercules to become a community code will facilitate its use in interdisciplinary projects such as CyberShake or the Broadband Simulation Platform. We expect Hercules to be used as a reference tool for groups working on tomography problems, or in the validation of community velocity models such as CVM-S and CVM-H. In turn, this will also help SCEC in building collaborations with outside partners interested in employing deterministic simulations for seismic hazard analysis. Examples of such potential outside users are USGS personnel, the Pacific Gas and Electric Company, and Southern California Edison.
Exemplary Figure Figure 7. Current branching of Hercules in development (for developers) and Hercules LT version (for general users).