Collaborative research: Unraveling coseismic and postseismic deformation: A prerequisite for analyses of stress-coupling and tsunami genesis.

2004年12月26日苏门答腊-安达曼9.2级地震（SAE）是有史以来第三大地震，它破坏了印度-澳大利亚板块与缅甸板块之间的边界，引发了一场毁灭性的海啸，严重影响了印度洋沿岸的9个国家。在震中附近，海啸在苏门答腊北方的一些沿海地区造成了30多米的上升，几乎立即造成20多万人死亡。根据相关的地震数据和全球定位系统对海底和地面变形的测量，SAE的破裂包括20多米的断层滑动。这种变形的幅度提供了一个难得的机会，进行区域规模的现场流变实验，其中同震断层滑动的冲动和随后的变形是响应。对这些测得的扰动进行建模可以检验同震（包括海啸成因）和震后（包括地震耦合和海啸上升）行为的假设。具体而言，有限元模型（FEM）的震中附近的俯冲带允许流变分区和过程的作用，对应力，应变和孔隙压力，管理同震和震后行为的定量评价。然后，海啸传播模型可以使用FEM生成的海底变形来预测海岸的上升。SAE的建模和解释研究将解决以下科学问题：（1）材料特性的分布（即，结构，密度，孔隙度，和俯冲带岩层的刚度）影响断层滑动的估计？(2)这种分布如何影响海底变形、海啸成因和上升预测？(3)地震后孔隙弹性和粘弹性变形的时间和分布是什么？(4)需要什么后滑？(5)库仑应力和孔隙压力瞬变与余震发生相关吗？这些问题的基础是一个更基本的问题：我们如何构建和约束同震和震后行为的模型作为一个综合的过程，所有这些都有助于变形系统？因此，拟议研究的主要目标是确定描述SAE同震和震后行为的流变特性的分布和标定。更具体地说，有限元模拟将解决余震发生在空间和时间（包括SAE和2005年3月25日发生在距SAE震中350公里的尼亚斯M8.7地震之间的应力耦合）。有限元生成的海底变形预测将推动海啸传播模拟。由于我们预计材料特性的变化（可能还有次级展布断层）将导致长尺度和短尺度的海底变形，因此将使用色散长波模型FUNWAVE进行海啸生成和传播模拟。海啸危险将以印度洋受影响最严重地区的模拟上升和淹没表示（例如，北方苏门答腊），并与2004年12月26日海啸的观测影响进行比较。这种模拟同震和震后变形系统的天气学方法可能会显着提高SAE的构造和海啸沿海灾害评估能力，并影响未来对其他人口密度高的俯冲带（如美国西海岸（卡斯卡迪亚）和日本）的类似大型逆冲地震的评估。设计和实施有限元模型的技术将在本项目后期阶段的一次研讨会上向科学界传播。学生将使用Abaqus软件构建模拟断层滑动的有限元模型，该模型可用于变形的正演和反演模型以及震后过程的驱动，包括孔隙弹性和粘弹性变形。

The December 26, 2004 M9.2 Sumatra-Andaman Earthquake (SAE), the third largest earthquake ever recorded, ruptured the boundary separating the subducting Indo-Australian Plate from the overriding Burma Plate and triggered a devastating tsunami that significantly impacted 9 countries bordering the Indian Ocean. Near the epicenter, the tsunami caused over 30 meters of run-up in some coastal areas of Northern Sumatra, almost instantly killing over 200,000 people. The rupture of the SAE included more than 20 meters of fault-slip, based on associated seismologic data and GPS measurements of seafloor and ground deformation. This magnitude of deformation provides a rare opportunity to conduct a regional-scale in-situ rheological experiment, in which the coseismic fault-slip is the impulse and the subsequent deformation is the response. Modeling these measured perturbations can test hypotheses of coseismic (including tsunami-genesis) and postseismic (including earthquake-coupling and tsunami run-up) behavior. Specifically, Finite Element Models (FEMs) of the subduction zone near the epicenter allow for a quantitative evaluation of the role of rheologic partitioning and processes, on the stress, strain, and pore pressure that govern coseismic and postseismic behavior. Tsunami propagation models can then use FEM-generated seafloor deformations to predict coastal run-up. This modeling and interpretive study of the SAE will address the following scientific questions: (1) How does the distribution of material properties (i.e., structure, density, porosity, and stiffness of rock formations in the subduction zone) affect fault-slip estimations ? (2) How does this distribution influence seafloor deformation, tsunami genesis, and run-up predictions ? (3) What is the timing and distribution of poroelastic and viscoelastic postseismic deformation ? (4) What afterslip is required ? (5) Do Coulomb stress and pore pressure transients correlate to aftershock occurrence ? These questions are underpinned by a more fundamental question: How do we construct and constrain models of coseismic and postseismic behavior as a synthesis of processes, all of which contribute to the deformational system ? Accordingly, the primary goal of the proposed research is to determine the distribution and calibration of rheologic properties that describe coseismic and postseismic behavior of the SAE. More specifically, FEM simulations will address aftershock occurrence in both space and time (including stress-coupling between the SAE and the March 25, 2005 M8.7 Nias earthquake that occurred 350 km away from the SAE epicenter). FEM-generated seafloor deformation predictions will drive tsunami propagation simulations. Because we expect that variations in material properties (and possibly secondary splay faulting) will cause both long and shorter scale seafloor deformations, tsunami generation and propagation simulations will be performed with the dispersive long wave model FUNWAVE. Tsunami hazards will be expressed in terms of simulated run-up and inundation for the most affected areas of the Indian Ocean (e.g., Northern Sumatra), and compared to the observed impact of the 12/26/04 tsunami. This synoptic approach to simulating coseismic and postseismic deformational systems may significantly advance tectonic and tsunami coastal hazard assessment capabilities for the SAE and impact future assessments of similar mega-thrust earthquakes for other subduction zones hosting high population densities, such as the upper US West Coast (Cascadia) and Japan. Techniques for designing and implementing FEMs will be disseminated to the scientific community during a workshop in the latter stages of this project. Students will use Abaqus software to construct FEMs that simulate fault-slip, which can be used in forward and inverse models of deformation and drive of postseismic processes, including poroelastic and viscoelastic deformation.