Collaborative Research: Forward and inverse models of global plate motions and plate interactions

合作研究：全球板块运动和板块相互作用的正向和逆向模型

• 批准号: 1646337
• 负责人: Georg Stadler
• 金额: $ 9.8万
• 依托单位: New York University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2017
• 资助国家: 美国
• 起止时间: 2017-01-01 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1646337&HistoricalAwards=false
• 关键词: Collaborative; Research; Forward; inverse; models

Part 1: This is a project in which geoscientists at Caltech and mathematicians at New York University will collaborate to better understand the forces responsible for the motion of continents and those responsible for great earthquakes. This is a mathematical and modeling project that will use existing observations to infer the forces which push and pull the tectonic plates around and the forces which cause the plates to slow down-essentially the resistance or friction inside the earth. The driving forces arise from the lower temperatures associated with the cold tectonic plates when they return to the inside of the earth. The resistance comes from several sources including the strength of the tectonic plates and the friction between the tectonic plates. Beyond those simple statements, scientists do not know precisely how those forces are balanced and this gap in their understanding means that society cannot precisely forecast where great earthquakes will occur. Great earthquakes, like the one that caused the Fukushima nuclear disaster in Japan in 2011, are the most destructive of all earthquakes. By harnessing new computational methods developed by the team and using supercomputers supported by NSF and the Department of the Energy, the collaborators will bring a new level of understanding to this problem. Essentially, the team will develop what are called inverse models in which they will go from the data (like the motion of the tectonic plates) to maps showing the relative strength of different tectonic boundaries. Inverse models are powerful methods, especially in the area of big data, but they require considerable investment in new mathematical techniques, algorithms and computer software. Key for this project is that for the first time, the key aspects of the physics will be incorporated at high resolution. Besides the scientific output, a key outcome of this project will be the training of graduate students and postdoctoral scholars. They will gain enormous experience with sophisticated numerical methods, inversion & optimization methods, supercomputing technologies, and the linkage of data with numerical models. The graduate students in the respective fields of geophysics and applied mathematics are highly sought after by the research community (academia and government research laboratories) and private industry (including the hydrocarbon industry). A key outcome will also be sophisticated new computer software that will efficiently utilize the largest supercomputers supported by the government and industry and this software will be donated to the NSF-supported software center for geophysics (the Computational Infrastructure for Geodynamics, CIG) for broader distribution throughout the scientific community.-Part 2: The team will use global forward and inverse models to determine the variations in mechanical coupling in subduction zones using present day plate motions and the along strike variation in the depth of oceanic trenches. Using plate motions, they will then apply the methods to the geological past and more fully establish the dynamic link between changing plate motions and how they translate into and result from changes in subduction including plate coupling, with specific linkages to geophysical and geological observations. The models will be high resolution (e.g. locally 1 km or less where needed) and fully incorporate the extreme variations and non-linearity in viscosity between the slab and mantle, hinge zone and interface between subducting and over-riding plates. They will use mathematically rigorous and computationally robust PDE-constrained optimization and uncertainty quantification approaches to infer mechanical parameters and coupling coefficients between plates. Global plate motions are primarily driven by the negative buoyancy within subducted slabs, but beyond that statement there is much less consensus on either the relative importance of other driving forces or the strength and nature of resisting forces. The failure to reach consensus on the complete force balance of plate tectonics and mantle flow manifests itself in fundamental unanswered questions, including the causes of changes in plate motions over millions of years, the variability in occurrence of great earthquakes at subduction zones, and the thermal and tectonic evolution of the earth, for example. These processes are likely related to how the forces within a subducting slab are transmitted and resisted through the bending (hinge) zone from the slab and into the subducting plate. In this project, the team will take a new generation of plate-mantle models in which the details of plate boundaries are resolved while cast as a mathematically rigorous and computationally robust parameter estimation problem linked to key observational data sets. They will achieve a much deeper understanding on the forces driving and resisting plate motions, the formal trade-off between parameters, the variation of plate coupling between subduction zones, and the evolution of the force balance on million-year time scales. The most important broader impact will be the training of several Ph.D. students in geophysics and applied mathematics and a postdoctoral scholar. The students and post-docs will gain enormous experience with sophisticated numerical methods, inversion & optimization methods, supercomputing technologies, and the linkage of data with numerical models. The project will allow a young investigator in applied mathematics to more closely collaborate with earth scientists on critical scientific interest. At the end of the project the team will release the adjoint Rhea-II code to the community. Scientific impacts will be with the GPlates-CitcomS-Rhea linkage. The team will reach a new level of integration between dynamic flow models, plate kinematics and seismic tomography, that could have broad application in the geosciences. Their work could impact several large programs. The geodynamic hypotheses could result in predictions that could be tested with deep sea drilling under the auspices of IODP ocean drilling. The inference of the mechanical coupling along strike in subduction zones will be of significant use to better evaluate the role of geodynamic factors influencing the occurrence of great earthquakes and indirectly assist in the evaluation of seismic hazards.

项目成果

Advanced Newton Methods for Geodynamical Models of Stokes Flow With Viscoplastic Rheologies

具有粘塑性流变学的斯托克斯流地球动力学模型的高级牛顿方法

• DOI: 10.1029/2020gc009059
• 发表时间: 2020
• 期刊: Geosystems
• 作者: Rudi, Johann;Shih, Yu‐hsuan;Stadler, Georg
• 通讯作者: Stadler, Georg

Scalable simulation of realistic volume fraction red blood cell flows through vascular networks

• DOI: 10.1145/3295500.3356203
• 发表时间: 2019-09
• 期刊: Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis
• 作者: Libin Lu;Matthew J. Morse;Abtin Rahimian;G. Stadler;D. Zorin