Slow slip at elevated pore pressures on brittle faults in a compliant subduction channel

大约十年前，利用全球定位卫星（GPS）系统的观测发现了断层最深处的“慢滑”现象。这些短暂的事件现在已经在地球的大部分俯冲带被记录下来，包括太平洋西北部的卡斯卡迪亚边缘，俄勒冈大学已经在那里对它们的行为进行了分类，以及在其他地质环境中，如圣安德烈亚斯和夏威夷。顾名思义，慢滑事件发生的滑动速度比地震慢得多。然而，它们能够释放在间隔时间内（通常是几个月到几年）积累的大量应变能，它们在确定不常见的大地震发生前的断层状况方面发挥着重要作用，大地震对这些地区的人口和基础设施构成了一些最大的风险。我们的工作重点是揭示导致慢滑发生的地质条件。我们正在对沿俯冲界面的孔隙流体（即水）的压力如何由沉积物的压实和矿物颗粒释放水的脱水反应控制进行预测性理解。我们还在开发数学模型，以确定在“俯冲通道”中嵌入更柔顺矩阵的脆性断层上滑动是否会导致与观测结果一致的滑动行为。这项研究工作是培养一名有前途的新研究生的一部分，它将使一名本科生参与我们更广泛的研究工作，它将使我们能够更有效地向更广泛的科学界传播俄勒冈大学慢滑源参数目录。为了进一步了解慢滑，需要在设计和实施预测慢滑模型时考虑到更好的地质约束条件。特别值得注意的是，高孔隙压力似乎是缓慢滑动发生所必需的，而在发掘出的俯冲带岩石中，结构特征表明脆性和韧性变形机制共存。为了探索高孔隙压力是如何发展的，我们考虑了多孔逆冲下沉积的固结和流体流动，并考虑了化学脱水反应的影响，从而对从海沟到慢滑区域的俯冲体之后的孔隙压力演化进行了综合量化。在高压和高温条件下的半深海泥岩单轴变形实验中，通过测量沉积物渗透率和孔隙比的变化来约束本构行为，可能包括蒙脱石-伊利石转变的影响。一项补充的建模工作考察了嵌入韧性剪切带的断层上的摩擦滑动所产生的变形。初步的工作集中在一个固定宽度的俯冲通道的二维处理上，该通道在一个柔性矩阵中包含一个单一的断层；随后的建模阶段将面向更现实的描述，包括在柔顺矩阵中刚性、有效材料的变化，以及沿断层滑动方向的粗糙度的影响，与挖掘出的大型结构的观察结果一致。我们的新建模框架受到缓慢滑动事件的扩展目录和缩放关系的限制，这些关系是从沿着卡斯卡迪亚边缘的大地测量观测推断出来的。通过理论、观测和实验相结合的方法，这项有针对性的研究工作将提高我们对慢滑动力学和俯冲边缘构造行为和危害的理解。

About a decade ago, observations using the global-positioning satellite (GPS) system led to the discovery of "slow-slip" behavior along the deepest extent of faults. These transient events have now been documented along most of the Earth's subduction zones, including the Cascadia margin of the Pacific Northwest, where the University of Oregon has been cataloging their behavior, and in other geologic settings such as the San Andreas and Hawaii. As their name suggests, slow-slip events occur at sliding rates that are much slower than those during earthquakes. They nevertheless are capable of releasing large amounts of the strain energy that accumulates during the time intervals (typically months to years) that separate them and they play an important role in determining the conditions along faults just prior to the infrequent, large earthquakes that pose some of the greatest risks to populations and infrastructure in these settings. Our efforts focus on unraveling the geological conditions that enable slow-slip to occur. We are developing a predictive understanding for how the pressure of pore fluids (i.e. water) along the subduction interface is controlled by the compaction of sediments and dehydration reactions that release water from mineral particles. We are also developing mathematical models to determine whether sliding on brittle faults embedded in a more compliant matrix in a "subduction channel" can cause slip behavior that is consistent with observations. This research effort is part of the training for a promising new graduate student, it will involve an undergraduate in our broader research effort, and it will enable us to more effectively disseminate the University of Oregon catalogue of slow-slip source parameters to the broader scientific community.Further progress in understanding slow slip requires that improved geologic constraints be included in the design and implementation of predictive slow-slip models. Of particular note are the high pore pressures that seem to be required for slow-slip to occur and the structural features in exhumed subduction zone rocks that suggest coexistence of brittle and ductile deformation mechanisms. To explore how high pore pressures develop, we consider consolidation and fluid flow in porous underthrust sediments and incorporate the effects of chemical dehydration reactions to develop a synoptic quantification of pore pressure evolution that follows a subducted volume from the trench to the region of slow slip. Constitutive behavior will be constrained using measured changes in sediment permeability and void ratio during uniaxial deformation experiments on hemipelagic mudstones under elevated pressure and temperature conditions, potentially including the influence of the smectite-illite transition. A complementary modeling effort examines the deformation produced by frictional sliding on faults embedded within a ductile shear zone. Preliminary efforts focus on a two-dimensional treatment of a subduction channel of fixed width that contains a single fault in a compliant matrix; subsequent modeling phases will be geared towards more realistic descriptions that incorporate variations in the fraction of rigid, competent material within the compliant matrix and include the effects of roughness along the direction of fault slip, consistent with observations of exhumed m elange structures. Our new modeling framework is constrained by an expanding catalog of slow-slip events and scaling relationships that have been inferred from geodetic observations along the Cascadia margin. With a combined theoretical, observational, and experimental approach, this targeted research effort will improve our understanding of the dynamics of slow slip and the tectonic behavior and hazards along subduction margins.