How do earthquake ruptures propagate through clay-rich fault zones?

在大的构造断层上，运动可以稳定地发生，产生断层蠕变，或者在地震发生的地方产生不稳定的滑动。断层蠕变通常与富含粘土的断层沟槽有关，这些断层沟槽可以容纳断层上的滑动。它们通常被认为比周期性地震发生的封闭性断层造成的地震危险要小。最近的研究表明，地震可以通过断层的爬行部分传播，带来毁灭性的后果。这个项目将联合领先的实验学家和建模者来研究地震破裂在什么条件下可以通过“爬行”断层传播。这项工作将利用一种独特的新型高压旋转剪切变形装置来复制和理解地震破裂时的物理反应，并利用破裂模型预测大规模反应。实验和建模的结果将用于为蠕变断层开发新的地震危险性评估，包括如何看待它们的地震活动性，以及破裂的性质如何影响辐射波场——这将影响地震的破坏性。我们从缓慢滑动的实验室实验中得知，地震不会在富含粘土的断层上形成核，因为随着滑动开始加速，它们会加强，从而阻止任何潜在的破裂。这一点很好地说明了在俯冲带富含粘土的增生弧前部分缺乏地震活动。然而，最近的事件表明，在断裂带中粘土含量较少的区域发生的大地震破裂可以穿透粘土含量丰富的区域，甚至大大增强滑动，例如2011年的Mw9.0 tohokuo - oki地震，其中有记录以来最大的同震滑动（约50米）发生在富含粘土的增生前弧中，造成了海底的大量偏移，导致了毁灭性的海啸。地震通过爬行断层传播的其他例子有1999年台湾的Mw7.6级池池地震，那里的破裂性质被明显改变（破裂速度和滑动速度增加），以及1944年的Mw7.4级北安那托利亚断层地震。这项研究将使用利物浦最近开发的独特实验室设备，该设备可以复制地震期间的条件，并允许我们测量断层的摩擦强度如何发展，这将决定地震破裂是否会在富含粘土的断层中传播或停止。它将允许在大约15公里深的完全受限条件下模拟地震破裂的接近。实验将在不同的施加位移和位移速率下监测实验断裂带的强度和性质。峰值加速度和应力减小将模拟不同大小的地震，并研究促进不稳定滑动所需的能量屏障。在另一种不同类型的实验中，当粘滑不稳定性（实验室地震）传播到实验室断层带的富含粘土的区域时，将对其进行监测。来自实验室的约束地震滑动物理反应的结果将被添加到大尺度模型中，以帮助我们理解(a)破裂传播，这将决定一个小地震是否会发展成大地震；(b)破裂传播的性质，例如它将以多快的速度传播，释放多少应力，用于概率地震危害评估。

On large tectonic faults movement can occur stably, producing fault creep, or by unstable slip where earthquakes occur. Fault creep has typically been associated with clay-rich fault gouges that accommodate slip across a fault. They have typically been thought to pose less seismic hazard than locked faults where earthquakes occur periodically. Recent studies have demonstrated that earthquakes can propagate through creeping sections of faults, with devastating consequences. This project will combine leading experimentalists and modellers to investigate under what conditions earthquake ruptures can propagate through 'creeping' faults. The work will utilize a unique new high-pressure rotary shear deformation apparatus to replicate and understand the physical response as an earthquake rupture passes and rupture models predict the large-scale response. Results from experiments and modelling will be used to develop new seismic hazard assessment for creeping faults, both in terms of how their potential for seismicity is viewed, and how the nature of a rupture would affect the radiated wavefield - which influences how destructive an earthquake will be.We know from slow-slip laboratory experiments that earthquakes are not expected to nucleate on clay-rich faults as they strengthen as slip starts to accelerate, thereby arresting any potential rupture. This is nicely illustrated by a lack of seismicity seen in the accretionary forearc clay-rich parts of subduction zones. However, recent events have suggested that large earthquake rupture, nucleated on a less clay-rich region of a fault zone can punch through clay-rich regions, and even greatly enhance slip, such as was seen in the Mw9.0 Tohoku-Oki earthquake in 2011, where the largest co-seismic slip ever recorded (~50m) occurred in the clay-rich accretionary forearc that produced a large offset of the seafloor leading to a devastating tsunami. Other examples of where earthquakes have propagated through creeping faults are The 1999 Mw7.6 Chi Chi earthquake in Taiwan, there the properties of the rupture were clearly modified (increase in the rupture velocity and slip speed), and the 1944 Mw7.4 North Anatolian Fault earthquake.This research will use unique laboratory equipment recently developed at Liverpool that can replicate the conditions during earthquakes and allow us to measure how the frictional strength of the fault develops, which will dictate whether or not an earthquake rupture will propagate or arrest in clay-rich faults. It will allow the approach of an earthquake ruptures to be simulated under fully confined conditions approximating to 15km depth. Experiments will be conducted where the strength and properties of the experimental fault zone is monitored under different imposed displacements and displacement rates. The peak acceleration and stress reduction will mimic earthquakes of different size and investigate the energy barrier required to promote unstable slip. In a different type of experiment, a stick-slip instability (laboratory earthquake) will be monitored as it propagates into clay-rich region of a laboratory fault zone. Results constraining the physical response of earthquake slip from the laboratory will be added into large-scale models to aid our understanding of (a) rupture propagation, which will dictate if a small earthquake will grow into large event and (b) what the properties will be, such as how fast it will travel and how much stress will be released, for use in probabilistic seismic hazard assessment.