Collaborative Research: High-Resolution Study of Subduction Zone Seismicity and Structure

俯冲带是地球板块构造系统的重要组成部分之一。这反映在两个主要关注俯冲带的国际科学计划的存在上：俯冲工厂（SubFac）和孕震带实验（SEIZE），它们都是边缘计划的一部分。SubFac倡议推动了大量矿物学和岩石学工作，研究候选板块成分（特别是水合相）的稳定性场及其与中深度地震活动性的关系。一种可行的假说认为，中深度地震与俯冲板块的变质脱水反应有关。这一假设的检验是不完善的，主要有四个原因——板热结构、板成分、板地震速度结构和地震位置的不确定性。后者影响地震位置，也影响板模型中地震速度与候选矿物学之间的关联。该项目涉及对世界各地俯冲带进行多方面的地震调查。该项目的主要目标是完善对下行板块内地震活动分布的认识，并改进其结构的地震模型。项目中应用的分析方法包括：(1)改进的波形对准技术；(2)高精度的远震和局地地震定位；(3)相关震源机制确定；(4)双差地震层析成像。方法1、2和3适用于一套广泛的俯冲带，特别是在存在或不存在双地震带的情况下。方法1和方法4应用于一组选定的俯冲带，那里有足够的当地地震资料-智利北部，阿拉斯加，哥斯达黎加和台湾。以往的研究表明，用双谱波形对准方法验证的波形互相关（WCC）时间延迟比用标准阈值准则选择的时间延迟具有更好的定位效果。预期这种改善有两个主要原因。首先，可以利用双谱检测去除不可靠的WCC时延数据。其次，双谱验证过程可以为两个距离近的地震提供更多的时间延迟测量，从而更好地控制它们的相对位置。根据以前对远震和局部地震事件的研究，可以预期相对位置散射减少近一个数量级。将精细波形对准技术和一种新的DD地震层析成像方法应用于几个俯冲带的局部数据集，以锐化地震速度结构图像。所有提议的目标区域都是以前一个或多个局部地震层析成像研究的主题，因此具有适合DD层析成像的数据集。基于将DD层析成像方法应用于本州北部的目录数据的初步结果，可以预期产生前所未有清晰的俯冲带地震图像。这个项目将从根本上有助于了解俯冲带和相关过程。俯冲带是推动地球演化的板块构造引擎的一个基本特征。构造、应力状态和下行板块的演化是已经获得大量科学关注的主题，但仍然相对知之甚少。该项目将提供关于下倾板块内地震活动分布的改进信息，以及板块和上覆地幔楔地震结构的详细模型，这些模型是地球动力学和岩石学模型的关键约束条件。特别是，关于中深度地震的性质有几个相互竞争的假设。通过对几个不同的俯冲带的详细研究，将获得支持或反对这些假设的地震证据。该项目将利用美国国家科学基金会支持的威斯康星大学麦迪逊分校的研究性教学与学习整合中心（CIRTL），根据研究结果准备基于网络的推广材料。软件和数据将与其他科学家共享，特别是通过提供使用DD断层扫描的培训。

Abstract Subduction zones are one of the most important components of the Earth's plate tectonic system. This is reflected by the existence of two major international science initiatives focused on subduction zones: the Subduction Factory (SubFac) and the Seismogenic Zone Experiment (SEIZE), both part of the MARGINS initiative. The SubFac initiative has motivated substantial mineralogical and petrologic work on stability fields of candidate slab constituents (especially hydrated phases) and their relation to intermediate-depth seismicity. One working hypothesis is that the intermediate-depth earthquakes are related to metamorphic dehydration reactions in the subducting slab. Tests of this hypothesis are imperfect for four main reasons - uncertainties in slab thermal structure, slab composition, slab seismic velocity structure, and earthquake locations. The latter impacts earthquake locations and also affects the association between seismic velocities in slab models and candidate mineralogies. This project involves a multi-faceted seismological investigation of subduction zones around the world. The principal project goals are refining knowledge of the distribution of seismic activity within down-going slabs and improving seismic models of their structure. Analysis methods that are being applied in the project include (1) improved waveform alignment techniques, (2) high-precision teleseismic and local earthquake location, (3) associated focal mechanism determination, and (4) double-difference (DD) seismic tomography. Methods 1, 2 and 3 are applied to a broad suite of subduction zones, especially where the presence or absence of a double seismic zone is in question. Methods 1 and 4 are applied to a selected set of subduction zones where adequate local earthquake data are available - Northern Chile, Alaska, Costa Rica, and Taiwan. Previous work indicates that waveform cross-correlation (WCC) time delays verified with the bispectrum waveform alignment method provide better relocation results than those selected with the standard threshold criterion. This improvement is expected for two main reasons. First, unreliable WCC time delay data can be removed using bispectrum checking. Second, the bispectrum verification process can provide more time delay measurements for two close earthquakes and thus more control over their relative positions. Based on previous work with both teleseismic and local events, close to an order of magnitude reduction in relative location scatter can be expected. Refined waveform alignment techniques and a new DD seismic tomography method are applied to local datasets from several subduction zones to sharpen images of seismic velocity structure. All of the proposed target areas have been the subject of one or more previous local earthquake tomography studies, and thus have datasets suitable for DD tomography. Based on initial results applying the DD tomography method to catalog data from northern Honshu, the production of subduction zone seismic images of unprecedented clarity can be anticipated. This project will contribute in a fundamental way to the understanding of subduction zones and related processes. Subduction zones are a fundamental feature of the plate tectonic engine that drives the evolution of the Earth. The structure, state of stress, and evolution of down-going slabs are topics that have received substantial scientific attention yet remain relatively poorly understood. This project will provide improved information on the distribution of seismic activity within down-going slabs along with detailed models of the seismic structure of the slab and overlying mantle wedge, which are critical constraints for geodynamic and petrologic models. In particular, there are several competing hypotheses about the nature of intermediate-depth earthquakes. Through a detailed study of several different subduction zones, seismic evidence for or against these hypotheses will be obtained. This project will take advantage of the new NSF-supported Center for the Integration of Research Teaching and Learning (CIRTL) at UW-Madison in preparing web-based outreach materials based on the research results. Software and data will be shared with other scientists, particularly by providing training on the use of DD tomography.