Collaborative Research: Exploring System-Wide Events on Complex Fault Networks using Fully-Dynamic 3D Earthquake Cycle Simulations

世界上最大的地震发生在嵌入地壳的相互连接的断层网络上。据信，大地震发生的条件取决于该地区数千年的构造运动历史，以及其他物理条件，如作用于断层本身的摩擦力。这项工作的目标是利用数学建模来扩展我们对决定地震发生地点和时间的物理学的理解。与模拟长周期地震活动相关的高计算成本限制了以前对简单情景的建模尝试（例如，只考虑单个断层，不产生破坏性地震波）。pi将通过使用他们最近开发的先进地震周期数值方案的高性能实现来克服这些限制。研究人员将探索的关键问题包括：为什么一些大地震发生在受到构造载荷时看似稳定的断层上？这些研究的结果将通过揭示如此大的地震是否以及如何在给定的断层网络上发生来改善地震危险性的估计。这些代码将在宽松的开源许可下公开提供，供其他人使用。除了支持一名初级教师（在她的领域中代表性不足）和一名职业生涯中期的教师外，该提案还支持指导两名研究生进行跨学科研究。这项工作的目标是通过使用大规模的、物理健壮的地震周期模型，扩大对系统范围地震可能发生的物理环境的理解。这项工作将开发一个大规模、高性能的框架，该框架可以解释复杂的断层几何形状、非断层材料特性和3D体积的完整动力学。该方法将震间载荷与同震破裂和波传播以自洽的方式耦合在一起。耦合方法将用于探索构造载荷，破裂历史，断层几何形状和其他物理特征在断层网络的全系统失效中所起的作用。这将包括诸如keystone fault hypothesis这样的研究，即在一个系统中，相对于区域应力场的最佳定向的断层被错误定向的keystone fault稳定，直到整个网络准备失败。与模拟震间期相关的高计算成本限制了以前的周期模型，使其仅限于简单的断层几何形状，抑制了动态影响，并且/或使用了较小的模拟体积。该项目将使用最近开发的地震间期混合方案，该方案非常适合解决过去计算成本过高的问题。这种混合方法将与一种新开发的动态破裂模拟技术相结合，以研究复杂断层网上的多次地震序列。通过利用最先进的算法和高性能计算，这项工作将导致复杂断层网络上震源过程的大规模，物理健壮的预测建模框架的发展。这个新的框架将允许探索地震科学中的基本问题，重点关注地震如何在定向错误的断层上形成核，这通常会引起巨大的、全网范围的事件。2010年的7.2 Mw El Mayor-Cucapah和2016年的7.8 Mw Kaikoura是两个众所周知的地震例子，它们的震级通过断层网络的多个部分级联而超出预期。这项工作将通过在物理健壮的建模环境中计算与复杂断层网络相关的事件概率，为地震危险性估计做出贡献，补充和改进目前用于地震预报的地震模拟器的更大努力。为了使更大的地震科学界受益，pi将在MIT许可下公开开发代码，并将采用开放的开发策略。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The world’s largest earthquakes occur on interconnected fault networks embedded in the Earth’s crust. The conditions under which large earthquakes occur is believed to be dependent on a millennia-long history of tectonic motion in the region, as well as other physical conditions such as frictional forces acting on the fault itself. The goal of this work is to use mathematical modeling to expand our understanding of the physics that dictate where and when an earthquake will occur. The high computational cost associated with simulating long periods of earthquake activity has limited previous modeling attempts to simple scenarios (for example, considering only a single fault, with no generation of damaging seismic waves). The PIs will overcome these limitations by using a high-performance implementation of their recently developed advanced numerical scheme for earthquake cycles. Key questions the researchers will explore include: why do some large earthquakes occur on faults that seem stable when subjected to tectonic loading? Results from these studies will improve seismic hazard estimates by shedding light on if and how such large earthquakes can occur on a given fault network. The codes will be made publicly available under a permissive open source license for use by others. In addition to supporting one junior faculty member (under-represented in her field) and one mid-career faculty member, the proposal supports the mentoring of two graduate students in interdisciplinary research.The goal of this work is to expand understanding of the physical settings in which system-wide earthquakes can occur through the use of large-scale, physically-robust earthquake cycle models. This work will develop a large-scale, high-performance framework that accounts for complex fault geometries, off-fault material properties, and full dynamics in 3D volumes. The method will couple interseismic loading with coseismic rupture and wave propagation in a self-consistent manner. The coupled approach will be used to explore the role that tectonic loading, rupture history, fault geometry and other physical features play in system-wide failure of fault networks. This will include such studies as the keystone fault hypothesis, namely, that faults in a system that are optimally oriented with respect to the regional stress field are stabilized by misoriented keystone faults until the entire network is primed to fail. The high computational cost associated with simulating the interseismic period has limited previous cycle models to simple fault geometries, suppressed dynamic effects, and/or use small simulation volumes. The project will use recently developed hybridized scheme for the interseismic period, which is well suited for problems that in the past have been too computationally expensive. The hybrid method will be coupled with a newly developed technique for dynamic rupture simulation to study sequences of multiple earthquakes on complex fault networks. By leveraging state-of-the-art algorithms and high-performance computing, this work will lead to the development of a large-scale, physically robust, predictive modeling framework of earthquake source processes on complex fault networks. This new framework will allow the exploration of fundamental questions in earthquake science, focusing on how earthquakes can nucleate on misoriented faults, often giving rise to huge, network-wide events. The 2010 Mw 7.2 El Mayor-Cucapah and the 2016 Mw 7.8 Kaikoura are two well-known examples of earthquakes whose magnitudes exceeded expectations by cascading through multiple segments of a fault network. This work will contribute to seismic hazard estimates by calculating event probabilities associated with complex fault networks in a physically robust modeling environment, complementing and improving the greater efforts of the earthquake simulators currently used to inform earthquake forecasting. In order to benefit the larger earthquake science community the PIs will make the developed codes publicly available under the MIT license, and will use an open development strategy.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.