Collaborative Research: Probing the frictional behavior of the Tohoku megathrust using GPS, seismicity, and physics-based models

世界上最大的地震发生在俯冲带，例如美国西北部的卡斯卡迪亚俯冲带，那里的海洋板块潜入地球地幔。2011年在日本北部发生的毁灭性的9级大地震就是一个明显的例子，也是到目前为止历史上仪器最好的大地震。这次地震的数据为了解是什么控制了俯冲带地震的频率和规模提供了一个独特的机会。在2011年之前，人们知道日本北部俯冲带既有缓慢滑动(没有地震的断层蠕动)松弛应力，也有地震时的快速滑动。它曾错误地认为，爬行带将限制在大地震中可能发生滑动的区域，因此不可能发生9级地震。研究人员过去的工作表明，与预期相反，断层的同一部分可以同时表现出蠕变和地震。这个新项目开发了对控制故障蠕变发生方式和时间的机械特性的严格的基于物理的理解。模型将与精确的GPS测量(量化应力积累的速度)以及跟踪断层蠕变的小地震进行测试。之前对日本记录的GPS数据的分析表明，在9级地震之前的几十年里，断层蠕动加速；值得注意的是，来自小规模重复地震的独立数据证实了这一点。我们将检验这一假设，即当板块边界上的应力积累时，蠕变区域会随着时间的推移而扩大。P.I.Segall将参与面向中学和高中理科教师的Geoscape Bay Area专业发展计划。他将开发一个模块，将日本东北地震的科学转化为美国的地震危险，这与喀斯喀特俯冲带特别相关。该项目探索模型来解释日本东北部独特的震间和震后观测，这些观测受到重复地震和GPS数据的限制，包括2011年9兆瓦地震前后的数据。我们关于加速震间蠕变和后滑重叠历史破裂的假设是，当地震在速度弱化区孕育时，由于强烈的动力减弱而破裂为速度强化区。在地震间期，蠕变进入速度强化带，随时间侵蚀锁定的凹凸体。这些模拟包括速率状态摩擦和热压，以及针对GPS和重复地震数据的测试预测。与时间相关的粗糙度侵蚀似乎是解释长时间应变瞬变的一种很有前途的机制。仅限于孕震深度的锁定凹凸体的弹性模型不能解释上个世纪记录的海岸下沉；这些数据要么需要深耦合(60？100公里)，要么需要地幔松弛。全球定位系统和海底大地测量，在MW9之后，需要一些后滑移和深部地幔流动的组合。东北海岸经历了全新世的隆升，但目前尚不清楚仅靠震后过程是否能恢复观测到的震间和同震下沉。我们将使用耦合的粘弹性和基于物理的余滑模型来检验地幔流动和界面耦合对地震间变形的相对贡献，并进一步检验侵蚀粗糙度假说。这些基于物理的断层摩擦和地震周期变形模型能够整合不同的数据集，并提供这一独特板块边界的力学行为的综合模型。该项目指导一名暑期本科生研究地球科学与工程专业的学生分析东北地区的GPS垂直时间序列，并利用这些结果来限制板块边界上的地震间滑移率。学生将参加关于准备GRE、申请研究生院和了解地球科学职业的研讨会。

The largest earthquakes in the world occur in subduction zones, such as the Cascadia Subduction Zone in the northwestern US, where oceanic plates dive into Earth's mantle. The devastating magnitude 9, 2011 Tohoku-oki earthquake in northern Japan is a notable example, and by far the best instrumented giant earthquake in history. The data from this quake presents a unique opportunity to understand what controls the frequency and size of subduction zone earthquakes. Prior to 2011 it was known that the northern Japan subduction zone exhibited both slow sliding (fault creep without earthquakes) which relaxes stress, as well as rapid slip in earthquakes. It had been, erroneously, thought that the creeping zones would limit the area that could slip in large earthquakes such that magnitude 9 events were not possible. The researchers past work has shown that, contrary to expectation, the same part of the fault can exhibit both creep and earthquakes. This new project develops a rigorous physics-based understanding of the mechanical properties that control how and when fault creep occurs. Models will be tested against precise GPS measurements, which quantify how rapidly stress builds up, as well as small earthquakes which track fault creep. Previous analysis of GPS data recorded in Japan showed that fault creep accelerated in the decades leading up to the magnitude 9; remarkably this was confirmed by independent data from small repeating earthquakes. We will test the hypothesis that creeping areas expand with time as stress builds up on the plate boundary. P.I. Segall will participate in the Geoscape Bay Area Professional Development program for middle and high school science teachers. He will develop a module to translate science from the Tohoku earthquake in Japan to earthquake hazards in the U.S., which has particular relevance to the Cascade subduction zone. This project explores models to explain unique interseismic and post-seismic observations from northeast Japan, constrained by repeating earthquakes and GPS data, both before and after the 2011 Mw 9 earthquake. Our hypothesis for accelerating interseismic creep as well as afterslip overlapping historical ruptures is that while earthquakes nucleate in velocity weakening regions, they rupture into velocity strengthening regions due to strong dynamic weakening. In the interseismic period, creep penetrates into velocity strengthening zones, eroding locked asperities with time. The simulations include rate-state friction and thermal pressurization, and test predictions against GPS and repeating-earthquake data. Time dependent asperity erosion appears to be a promising mechanism to explain the long-duration strain transient. Elastic models with locked asperities restricted to seismogenic depths cannot explain coastal subsidence documented over the last century; these data require either deep coupling (60? 100 km) or mantle relaxation. GPS and seafloor geodetic measurements, post-Mw 9, require some combination of afterslip and deep mantle flow. The Tohoku coast underwent Holocene uplift, yet at present it is unknown whether postseismic processes alone will even recover the observed inter- and co-seismic subsidence. We will use coupled viscoelastic and physics-based afterslip models to examine relative contributions of mantle flow and interface coupling to interseismic deformation and further test the eroding asperity hypothesis. These physics-based models of fault friction and earthquake-cycle deformation are able to integrate diverse datasets and provide a comprehensive model of the mechanical behavior of this unique plate boundary. This project mentors a SURGE (Summer Undergraduate Research in Geoscience and Engineering) student to analyze GPS vertical time series from Tohoku and use these results to constrain interseismic slip rate on the plate boundary. The student will participate in workshops on preparing for the GRE, applying to graduate school, and understanding geoscience careers.