Physical state of the subducting Juan de Fuca plate and the blind deconvolution problem in global seismology

在接下来的五年里，我的研究计划将集中在全球地震学的两个关键问题上，一个是观测问题，另一个是理论问题。 第一个项目涉及从北方温哥华岛到北方加州的胡安·德富卡板块俯冲的物理状态及其对地震发生的影响。 该地区发生了许多不同类型的地震，每一种地震对人口中心的地震危害都有自己的意义。与2004年12月印度尼西亚地震的震级相当的大型逆冲断层地震平均每550年发生一次。深层地震震级较小，但发生在靠近主要人口中心的地方，如温哥华和维多利亚，因此构成重大威胁。在过去的世纪里，不列颠哥伦比亚省西南部发生的最具破坏性的地震是温哥华岛中部的地壳地震。最近才发现的第四类地震涉及板块边界深处非常缓慢的滑动。虽然没有直接的危险，但越来越多的证据表明，通过俯冲板块的结构和其中产生的流体的迁移/运输，慢滑地震和其他类型的地震之间存在联系。本研究旨在澄清这些关联。 第二个项目解决地震学中的一个典型问题，即分离震源（例如地震）和结构（例如反射、模式转换）对地震图的贡献。对于低频波（f<0.1 Hz），地球看起来更像一个径向对称的行星，地震可以被建模为具有简单时间依赖性的点源。地球结构和地震破裂的复杂性只有在更高的频率（f>0.5 Hz）才开始成为焦点。这项研究将开发一种计算效率高的方法，解决更高频率下的“盲解卷积”问题，从而提高对精细地球结构和地震破裂的理解。

Over the next five years my research program will focus on 2 key problems in global seismology, one that is observational and the other theoretical. The first project concerns the physical state of the subducting Juan de Fuca plate from northern Vancouver Island to northern California and its influence on earthquake occurrence. A number of different styles of earthquake, each with its own significance for seismic hazard to population centers, occur in this region. Megathrust earthquakes, comparable in magnitude to the Indonesian earthquake in December 2004, occur on average every 550 years. Deep slab earthquakes are smaller in magnitude but occur in closer vicinity to major population centers such as Vancouver and Victoria and so pose significant threat. The most damaging earthquakes to occur in southwestern B.C. in the past century have been crustal earthquakes on central Vancouver Island. A fourth class of earthquake that has only recently been discovered involves very slow slip deep on the plate boundary. Although not directly hazardous, there is growing evidence for associations between slow slip earthquakes and the other classes of earthquake through the structure of the subducting plate and the migration/transport of fluids that are generated therein. This research aims to clarify these associations. The second project addresses a canonical problem in seismology, namely the separation of source (e.g. earthquake) and structural (e.g. reflections, mode conversions) contributions to seismograms. For low frequency waves (f<0.1 Hz) the Earth appears much like a radially symmetric planet and earthquakes can be be modelled as point sources with simple time dependencies. The complexities of Earth structure and earthquake rupture only begin to come into focus at higher frequencies (f>0.5 Hz). This research will develop a computationally efficient method of solving the ``blind deconvolution'' problem at higher frequencies that will lead to improved understanding of both fine-scale Earth structure and earthquake rupture.