Collaborative Research: Frameworks: Seismic COmputational Platform for Empowering Discovery (SCOPED)

合作研究：框架：增强发现能力的地震计算平台（SCOPED）

• 批准号: 2103621
• 负责人: Hatice Bozdag
• 金额: $ 61.34万
• 依托单位: Colorado School of Mines
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-09-01 至 2025-08-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2103621&HistoricalAwards=false
• 关键词: Collaborative; Research; Frameworks; Seismic; COmputational

Seismology is the most powerful tool for investigating the interior structure of Earth—from its surface down to the inner core—and its wide range of processes, including earthquakes, volcanic activity, glacial processes, oceanic and environmental processes, and human-caused processes such as nuclear explosions or hydraulic fracturing in oil and gas exploration. Seismology cannot achieve its greatest potential without harnessing state-of-the-art computing capabilities for the dual purpose of scientific modeling and analysis of rapidly increasing data sets. The SCOPED (Seismic COmputational Platform for Empowering Discovery) project establishes a computing platform that delivers data, computation, and service to the seismological community in a way that promotes education, innovation, and discovery, and enables efficient solutions to outstanding scientific problems in geophysics. By focusing on openly available data, openly available software, and virtual training, SCOPED opens seismological research to a broad range of users. Four research components emphasize openly available software for the purpose of characterizing Earth's subsurface structure and the wide range of natural and man-made events that are recorded by seismometers every day. Training of seismologists is a central focus of the project. SCOPED training workshops (seismoHackweeks) are open to the community. Emphasis on virtual research and training diversifies strategies to engage minority groups entering computational geosciences. The project trains a new generation of seismologists to harness the latest capabilities for processing and modeling large data sets. The SCOPED project establishes cyberinfrastructure that provides fast access to large seismic archives from a suite of containerized open-source computational tools for big data analysis, machine learning, and high-performance simulations. The implementation focuses on four interconnected, compute- and data-intensive research components: seismic imaging of Earth’s interior, waveform modeling of earthquakes and Earth structure, monitoring of Earth structure using ambient noise, and precision monitoring of earthquakes and faults. Each research component is enabled by open-source codes that meet, or aspire to meet, best practices for software development. The project contains several transformative components. First, it offers compute performance for both model- and data-driven seismological problems. Hundreds of terabytes of waveform data are directly accessible both to modelers—for data assimilation problems—and to data scientists for processing, analysis, and exploration. Second, it establishes a direct collaborative link among four teams of seismologists at four institutions and a team of computational scientists at Texas Advanced Computing Center. This unity reflects the necessity of both groups to achieve research-ready codes that can exploit high-performance computing (HPC) and Cloud systems. Third, it establishes a gateway with ready-to-run (or adapt) container images and data as a service for the seismological community. Fourth, it develops computational tools that promote the democratization of HPC/Cloud with cutting-edge data processing and modeling software through their scalability from laptops to HPC or Cloud systems and through their portability with containerization. Finally, although the development of cyberinfrastructure is the main priority, ancillary scientific results from advanced techniques are expected to offer insights into fundamental seismological problems. The project has the potential for discoveries across fields (seismology, Earth science, computer science, data science, material science), as well as societal relevance in the realms of seismic hazard assessment, environmental science, cryosphere, earthquake early warning, energy systems, and geophysical detection of nuclear proliferation.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

地震学是研究地球内部结构（从地表到内核）及其各种过程的最有力工具，包括地震、火山活动、冰川过程、海洋和环境过程以及人为过程，如核爆炸或油气勘探中的水力压裂。如果不利用最先进的计算能力来实现科学建模和分析快速增长的数据集的双重目的，地震学就无法发挥其最大的潜力。SCOPED（Seismic COmputational Platform for Empowering Discovery）项目建立了一个计算平台，以促进教育，创新和发现的方式向地震学界提供数据，计算和服务，并为地球物理学中的突出科学问题提供有效的解决方案。通过专注于公开可用的数据，公开可用的软件和虚拟培训，SCOPED向广泛的用户开放地震学研究。四个研究组成部分强调公开提供的软件，目的是描述地球地下结构和地震仪每天记录的各种自然和人为事件。地震学家的培训是该项目的中心重点。SCOPED培训研讨会（seismoHackweeks）向社区开放。对虚拟研究和培训的重视使吸引少数群体进入计算地球科学的策略多样化。该项目培训新一代地震学家，以利用最新的能力来处理和建模大型数据集。SCOPED项目建立了网络基础设施，可以从一套用于大数据分析、机器学习和高性能模拟的容器化开源计算工具中快速访问大型地震档案。实施重点是四个相互关联的，计算和数据密集型的研究组成部分：地球内部的地震成像，地震和地球结构的波形建模，使用环境噪声监测地球结构，以及地震和断层的精确监测。每个研究组件都由满足或渴望满足软件开发最佳实践的开源代码实现。该项目包含若干变革性组成部分。首先，它为模型和数据驱动的地震学问题提供了计算性能。数百TB的波形数据可直接供建模人员（用于数据同化问题）和数据科学家（用于处理、分析和探索）访问。其次，它在四个机构的四个地震学家团队和德克萨斯州高级计算中心的一个计算科学家团队之间建立了直接的协作联系。这种统一性反映了两个小组实现研究就绪代码的必要性，这些代码可以利用高性能计算（HPC）和云系统。第三，它建立了一个网关，其中包含准备运行（或调整）的容器图像和数据，作为地震社区的服务。第四，它开发了计算工具，通过从笔记本电脑到HPC或云系统的可扩展性以及通过容器化的可移植性，利用尖端的数据处理和建模软件促进HPC/云的民主化。最后，虽然网络基础设施的发展是主要的优先事项，但先进技术的辅助科学成果预计将为基本的地震学问题提供见解。该项目具有跨领域发现的潜力（地震学、地球科学、计算机科学、数据科学、材料科学），以及在地震灾害评估、环境科学、冰冻圈、地震预警、能源系统、该奖项反映了NSF的法定使命，并被认为值得通过使用基金会的学术价值和更广泛的影响审查标准。