The SCEC PetaSHA3 project will develop computational models of system-level earthquake processes and, running these models on NSF supercomputers, derive more accurate estimates of the strong ground motion expected from future earthquakes in California. SCEC?s CyberShake computational platform recently produced the first physics-based probabilistic seismic hazard analysis (PSHA) for the Los Angeles region?a major scientific innovation of the PetaSHA2 project and a breakthrough in workflow-managed, high-capacity computing (50-day continuous run on > 4000 processors of the TACC Ranger supercomputer). SCEC will implement CyberShake on the largest NSF supercomputers, including Blue Waters, with the goal of calculating high-resolution, high-frequency PSHA maps for the entire State of California in the next two years. This will require validation against very large data sets and a hundredfold increase in computational efficiency.
The project will use the CyberShake platform to represent time-dependent seismic hazard for operational earthquake forecasting (OEF). The goal of OEF is to provide authoritative information about the time dependence of seismic hazards to help communities prepare for potentially destructive earthquakes. The first OEF model for California (UCERF3) is now being developed by SCEC, USGS and CGS under a contract with the California Earthquake Authority. By coupling UCERF3 into the CyberShake system, the project will produce PSHA maps that show how seismic hazards are changing on a daily basis?i.e., it will extend CyberShake from the mapping of long-term hazard (seismic climate) to short-term hazard (seismic weather)?an entirely new capability not possible using conventional PSHA techniques. If successful, these developments are expected to have substantial impacts on seismic engineering, insurance rate-setting and reinsurance strategies, and earthquake preparedness activities, as well as basic research on earthquake predictability.
The SCEC PetaSHA3 project has developed computational models of system-level earthquake processes. By running these models on NSF supercomputers, we have derived more accurate estimates of the strong ground motion expected from future earthquakes in California. Among our achievements, we improved PetaSHA simulation capabilities by incorporating new codes that can model geologic complexities including topography, geologic discontinuities, and source complexities such as irregular, dipping, and offset faults; developed PetaSHA codes that achieve performance of sustained petaflops on petascale open-science computers; used dynamic rupture simulations to investigate the effects of realistic friction laws, geologic heterogeneities, and near-fault plasticity on seismic radiation, and improved pseudo-dynamic rupture models of hazardous earthquakes; employed the realistic rupture simulator, RSQsim, to evaluate static and dynamic stress transfer and assess their effects on strain accumulation, rupture nucleation, and stress release; increased the upper frequency limit of deterministic ground-motion predictions above 2 Hz and compared simulations with observed seismograms using goodness-of-fit measures of engineering relevance; improved the SCEC 3D community velocity models by iterated, full-3D inversion of large suites of observed waveforms from the Southern California Seismic Network, including earthquake phases and large ambient-noise datasets; and tested the CyberShake hazard model using existing seismic data and surveys of precariously balanced rocks in Southern California. We developed four new physics-based probabilistic seismic hazard models for the Los Angeles region using the CyberShake platform on the NSF Blue Waters and Stampede supercomputers. The CyberShake technique employs hundreds of thousands of deterministic wave propagation simulations to capture important physical effects at a given site, such as source directivity and basin excitations, and it does so more accurately than the conventional attenuation relationships based on data regression. We developed a new technique for the analysis and comparison of hazard models called Averaging-Based Factorization (ABF). ABF uses a hierarchical averaging scheme to separate the shaking intensities for large ensembles of earthquakes into relative (dimensionless) excitation fields representing site, path, directivity, and source-complexity effects, and it provides quantitative, map-based comparisons between models with completely different formulations. The CyberShake directivity effects are generally larger than predicted by the NGA directivity factor, but those calculated for the new models are smaller than those obtained previously, owing to the greater incoherence of the wavefields from the more complex rupture models. The CyberShake basin effects are generally larger than those from the three NGA models that provide basin-effect factors. However, the basin excitations calculated from one community velocity model (CVM-H) are smaller than those from another (CVM-S), and they show a stronger frequency dependence. Owing to this difference, the substitution of CVM-H for CVM-S reduces the CyberShake-NGA basin-effect discrepancy. Among the NGA models, that of Abrahamson & Silva (2008) is the most consistent with the CyberShake CVM-H calculations, with a basin-effect correlation factor greater than 0.9 across the frequency band 0.1-0.3 Hz. We used these comparisons to draw conclusions regarding strategies for reducing epistemic uncertainties in simulation-based hazard models. This project sponsored major efforts to build SCEC-sponsored collaborations between geoscientists and computer scientists that can apply petascale technology to socially relevant research in earthquake system science. In particular, we equipped a diverse scientific workforce with the tools to formulate and verify models, run simulations in a petascale environment, validate the predictions against observation, and assimilate data into model improvements. cross-trained diverse groups of undergraduate interns and early-career scientists in geoscience and computer science and teach them how to solve fundamental problems.
