Collaborative Research: Multi-scale models of subduction zone earthquake cycle observations

地球的构造板块在俯冲带被循环回到地幔。最大的地震发生在那里，这些巨型逆冲环境还产生了一系列其他地质灾害，包括海啸和火山，这使得他们的研究具有重大的社会意义。最近，地震学和大地测量揭示了一系列与巨型逆冲行为有关的现象，这些现象是缓慢加载和灾难性破裂的简单粘滑地震周期模型所不能捕捉到的。这些新发现的现象包括在十年尺度上的瞬时、蠕动的断层滑动事件，这可能表明断层系统的预备行为，可能与主要地震事件有系统地联系。力学模型没有完全跟上这些新的发现，我们对这些现象背后的物理过程的理解是不完整的。因此，需要一个新的综合框架来理解复杂变形和地震活动模式背后的物理机制和断层本构关系。该项目寻求开发这样一个力学模型，最初是针对数据丰富的日本背景，以了解区域巨型逆冲动力学和断层相互作用模式，以及改进地震危险估计。以后，这种机械模型可能会被部署在卡斯卡迪亚等其他俯冲带，并协助解释现有的和计划中的监测数据流，用于地震预报和预警。为了捕捉这一复杂问题所涉及的空间和时间尺度，德克萨斯大学奥斯汀分校(UTIG)、普渡大学和东京大学的研究人员将合作开展三个子项目：1)基于日本粘弹性、地震间、地震前、地震前和地震后断层加载的数值模型，开发一套多地震周期情景。2)全球地幔流动和区域性、随时间演化的地幔对流模式的发展，以了解该地区板块边界和背弧的长期俯冲诱导的强迫。3)发展了三维跨时间尺度的粘弹塑性模型，在动态一致的热力对流框架中考虑了速率和状态摩擦以及其他断层本构关系。例如，这将使研究地震前缓慢滑动现象对断裂带演化的作用成为可能，最终在流体流动的情况下。所有项目部分将被整合，并对不同方法的结果进行交叉检查。此外，所有分项目都是相辅相成的，将指导建立一般俯冲带模型，这是理解日本和其他地方的地震周期观测和地震活动所必需的。

Earth's tectonic plates get recycled back into the mantle at subduction zones. The largest earthquakes happen there, and these megathrust environments also generate a range of other geohazards including tsunamis and volcanoes, making their study of great societal relevance. Recently, seismological and geodetic measurements have revealed a range of phenomena associated with megathrust behavior that are not captured by a simple, stick-slip earthquake cycle model of slow loading and catastrophic rupture. These newly discovered phenomena include transient, creeping events of fault slip on decadal scales which may indicate preparatory behavior of the fault system, perhaps systematically linked to the main seismic event. Mechanical models have not quite kept up with these new discoveries and our understanding of the physical processes behind these phenomena is incomplete. A new, integrative framework is therefore needed to understand the physical mechanisms and fault constitutive laws behind complex deformation and seismicity patterns. This project seeks to develop such a mechanical model, initially for the data rich Japan setting, in order to understand regional megathrust dynamics and fault interaction patterns, as well as improve seismic hazard estimates. Later, such a mechanical model may potentially be deployed at other subduction zones such as Cascadia and assist in interpreting existing and planned monitoring data streams for earthquake forecasting and early warning. To capture the spatial and temporal scales involved in this complex problem, three sub-projects are to be pursued in collaboration between researchers at University of Texas Austin (UTIG), Purdue University, and the University of Tokyo: 1) The development of sets of multi earthquake-cycle scenarios based on numerical models of visco-elastic, inter-, pre-, co- and post-seismic fault loading in Japan. 2) The development of global mantle flow and regional, time-evolving mantle convection models to understand long-term, subduction induced forcing of plate boundaries and backarcs in the region. 3) The development of cross-timescale, visco-elasto-plastic models in 3-D, incorporating rate and state friction as well as other fault constitutive laws in a dynamically consistent, thermo-mechanical convection framework. This will enable, for example, studying the role of pre-seismic, slow slip phenomena for fault zone evolution, eventually in the presence of fluid flow. All project parts will be integrated, and results from different approaches cross-checked. Moreover, all sub-projects are complementary and will guide the establishment of the general subduction zone model that is needed to understand earthquake cycle observations and seismicity in Japan and elsewhere.