Three-dimensional spontaneous dynamic rupture models on geometrically complex faults with state-of-the-art frictional parameterization

动态自发地震破裂模型已被证明是研究地震物理和帮助预测地面运动的有价值的工具。这些数值模型从物质结构、摩擦行为和断层几何形状的基本假设出发，计算断层滑动的时空演变（通常是由此产生的近源地面运动）。这种动态模型通常要么使用实验室推导的摩擦定律，要么使用实际复杂的断层几何形状，但不能两者兼而有之。研究人员建议将这两个独立的轨道结合起来，使动态地震建模向前迈出重要的一步：他们将使用实验室导出的摩擦定律来模拟具有实际复杂几何形状的断层上的自发破裂传播和滑动。他们期望在断裂传播、滑动和地面运动中获得一阶效应，这将不同于以前的建模工作，从而更好地理解地震过程，更好地预测断层行为和地面运动。动态地震模型历来遵循两条轨道：1)研究摩擦参数化和应力模式对简单平面断层的影响；2)研究断层几何形状对地震过程的影响，使用简单的摩擦参数化。pi将把这两种轨迹结合起来，产生新一代的动态地震模型。高滑移率下的实验室实验数据和理论模型表明，在地震期间观察到的高滑移率下，必须修改典型的速率-状态摩擦公式，以在更大的长度尺度上纳入更大程度的弱化。此外，对复杂、不对称几何断层的研究表明，非平面断层不可避免的正应力随时间变化对破裂动力学有重要影响。为了正确地模拟断层行为的这两个方面，他们将开发一种现代摩擦参数化方法，并用它来模拟几何上复杂的断层行为，例如具有台阶和分支的系统。新的三维有限元方法将结合一种新的、现实的断层应力松弛方法，这对于避免此类断层系统上的病理性应力积累是必要的。地震物理学的这些重要成分以前从未在单一的建模方法中组合过，而这种组合的结果将是最先进的地震物理建模工具。研究人员将解决有关台阶和分支断层行为的重要问题，包括确定是否存在如何预测分支破裂路径的一般规则，以及破裂跨越台阶的能力。拟议的研究将对地震科学以及更广泛的科学和教育界产生重要影响。建模方法的一个关键用途将是深入了解几何复杂断层系统上地震的潜在规模，例如洛杉矶地区的断层系统。许多断层系统是由断层间隙和断层走向变化等几何特征所限定的；所提出的数值模型将有助于确定地震破裂可能跨越这些分段边界的情况，并产生更大的地震和更大的地面运动。此外，所提出的研究将有助于更好地估计滑移分布和破裂锋演化，这也强烈影响地面运动。由此得到的改进震源模型有助于地震大小和地面运动模式的概率评估，从而对地震危险性和建筑规范和设计产生潜在影响。

Dynamic spontaneous earthquake rupture models have proven themselves to be valuable tools to investigate the physics of earthquakes and to help predict ground motion. These numerical models start from basic assumptions about material structure, frictional behavior, and fault geometry, and calculate the spatiotemporal evolution of fault slip (and often the resultant near-source ground motion). Such dynamic models typically use either laboratory-derived friction laws or realistically complex fault geometry, but not both. The researchers propose to bring dynamic earthquake modeling an important step forward by combining these two separate tracks: they will use laboratory-derived friction laws to model spontaneous rupture propagation and slip on faults with realistically complex geometry. They expect to obtain first-order effects in rupture propagation, slip, and ground motion that will differ from previous modeling efforts, leading to both a better understanding of the earthquake process and better predictions of faulting behavior and ground motion. Dynamic earthquake models have historically followed two tracks: 1) investigations of the effect of frictional parameterization and stress pattern on simple planar faults, and 2) investigations of the effects of fault geometry on the earthquake process, using simple frictional parameterizations. the PIs will combine these two tracks to produce a new generation of dynamic earthquake models. Data from laboratory experiments at high slip rates and theoretical models imply that at the high slip rates observed during earthquakes, the typical rate-and-state frictional formulation must be modified to incorporate a greater degree of weakening over a larger length scale. Additionally, research on faults with complex, asymmetrical geometry shows that temporal variation of normal stress, which is inevitable on non-planar faults, can have a significant effect on rupture dynamics. To correctly model both these aspects of faulting behavior, they will develop a modern frictional parameterization and use it to model the behavior of geometrically complex faults, such as systems with stepovers and branches. The new 3D finite element method that will incorporate a new, realistic method for off-fault stress relaxation, which is necessary to avoid pathological stress buildup on such fault systems. These important ingredients of earthquake physics have never before been combined in a single modeling method, and the result of such a combination will be a state-of-the art tool to model the physics of earthquakes. The researchers will address important questions about the behavior of faults at stepovers and branches, including determining if there are general rules for how to predict rupture path at branches, and the ability of rupture to span stepovers.The proposed research will have important implications for both earthquake science and the broader scientific and educational community. A key use of the modeling method will be to gain insight into the potential size of earthquakes on geometrically complex fault systems, such as those in the Los Angeles region. Many fault systems are bounded by geometrical features such as fault gaps and changes in segment orientation; the proposed numerical models will help determine the circumstances under which earthquake rupture may propagate across these segment boundaries, and generate larger earthquakes with larger ground motion. In addition, the proposed research will lead to better estimates of the slip distribution and rupture front evolution, which also strongly affect ground motion. The resulting improved earthquake source models can help in the probabilistic assessment of earthquake size and ground motion pattern, with subsequent potential impacts on seismic hazard and building code and design.