A Petascale Cyberfacility for Physics-Based Seismic Hazard Analysis

点击翻译按钮获取中文摘要

This CME Collaboration proposes to transform probabilistic seismic hazard analysis (PSHA) into a physics-based science by deploying a new cyberfacility "PetaSHA" that can execute PSHA computational pathways and manage data volumes using the nation's petascale computing resources. The objectives of the project have been formulated in terms of three science thrusts: (1) Extend deterministic simulations of strong ground motions to 3 Hz for investigating the upper frequency limit of deterministic ground-motion prediction. (2) Improve the resolution of dynamic rupture simulationsby an order of magnitude for investigating the effects of realistic friction laws, geologic heterogeneity, and near-fault stress states on seismic radiation. (3) Compute physics-based PSHA maps and validate them using seismic and paleoseismic data. The investigators will assemble and operate a cyberfacility comprising four computational platforms, each designed to execute and manage the results from specific PSHA computational pathways. Two of these platforms, OpenSHA and TeraShake, have been developed by the CME Project. PetaShake will be an advanced research platform for petascale simulations of dynamic ruptures and ground motions with petascale capability computing and will be able to archive complete simulation data volumes (petascale data-intensive computing). CyberShake will be a new production platform that will employ advanced workflow management tools to compute and store the large suites of ground motion simulations needed for physics-based PSHA mapping. They will move these computational platforms forward in complexity and scale through a graduated series of calculations tied to a timeline with clear measures of success and will also develop a set of science gateways for the broader community, in particular researchers involved in the EarthScope and NEES Projects, to provide access to PSHA simulation capabilities and data products. The objectives include the cross-training of diverse groups of undergraduate interns and early-career scientists that will enable them to solve fundamental problems.