Constraints from multiple low frequency data on the long wavelength density structure in the deep mantle

为了在全球范围内了解地球的动力学（例如，构造板块如何移动，导致破坏性地震和火山爆发），我们必须了解地幔的性质，地幔是位于地球表面下3000公里的固体岩石。特别是，两个神秘的圆顶状结构被称为“大横波速度省”（或llsvp），从地幔底部上升到大约1000公里以上，一个在非洲下面，另一个在太平洋下面。地震产生的地震波有时会穿过这些llsvp，当它们进入这些结构时速度会减慢。目前还不清楚为什么会出现这种情况：llsvp可能比周围环境更热（因此浮力更大），或者它们的化学性质不同（可能密度更大）。两者都表明，在数百万年的时间里，地幔是如何流动的，因此它如何影响构造板块运动和相关的自然灾害。pi计划首次结合传统地震数据集的信息，以及特殊类型地震波和潮汐的测量结果，以揭示这些llsvp的浮力。特殊类型的地震波被称为斯通利模式，它是沿地核-地幔边界（CMB）捕获的振动，而地球潮汐是固体地球在太阳和月球的引力作用下每天两次的变形。该奖项将支持加州大学伯克利分校的两名研究生和两名女pi的培训，其中一人是加州大学伯克利分校的早期职业科学家和新教员。在这个项目中开发的模型将作为专门为这个项目开发的网页的一部分，以及通过地震学联合研究所（IRIS）设施提供。它们将用于加州大学伯克利分校的本科生研讨会，以及面向高年级研究生和早期职业科学家的跨学科CIDER研讨会。在过去的两年中，独立的研究对地球深部地幔地震层析成像的大低剪切速度省（llsvp）的密度结构提出了看似矛盾的结果。特别是，一项研究使用了多种地震学数据（正模分裂、地震走时和波形数据），另一项研究使用了半昼夜的潮汐数据，得出结论认为，LLSVP的很大一部分代表了密度过大的区域，表明LLSVP异常的来源主要是化学非均质性。相比之下，一项以测量地球的斯通利模式为特征的研究得出了相反的结论，这意味着LLSVP异常的来源在很大程度上是热的。这两种结论都描绘了一幅非常不同的地幔对流图景：前者暗示了llsvp是潜在的缓慢上升流或稳定的堆，而后者暗示了一种更有活力的地幔循环模式。在这个项目中，研究人员将结合这些不同的大地测量和地震数据集，这些数据集对弹性和密度结构的深度敏感性不同，以确定在假设llsvp的过量密度可能局限于CMB上方的薄层（~200公里或更少）下，先前的结果是否可以调和。在此过程中，他们将进一步研究横波和纵波速度结构，各向异性，衰减，以及CMB地形和最外层核心可能的结构。首先，他们将考虑现有的全球地震波形、潮汐和正常模式分裂数据（包括Stoneley模式数据）的组合，并使用经典的一阶模式扰动形式对分裂数据进行建模，对最深地幔的结构进行反演。他们将为此目的实现优化和跨维蒙特卡罗反演方法。在很大程度上，这将作为指导，为下一阶段进行的更高级的反转做准备，同时产生有趣的中间结果。在第二阶段，在完成自己的正态谱数据集后，他们将考虑更严格的理论方法和直接反演模态谱，并结合长周期波形和潮汐数据（辅以新获得的数据）来研究CMB附近的密度和速度结构。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

To understand the dynamics of the Earth on a global scale (e.g., how tectonic plates move, resulting in destructive earthquakes and volcanic eruptions), we must understand the nature of the mantle, 3000 km of solid rock that lies beneath the Earth's surface. In particular, two enigmatic dome-shaped structures called the "Large Shear Wave Velocity Provinces" (or LLSVPs) rise from the base of the mantle to around 1000 km above, one beneath Africa and the other beneath the Pacific Ocean. Seismic waves that emanate from earthquakes sometimes travel through these LLSVPs and are slowed down when they enter these structures. It is still unclear why this might be the case: LLSVPs could be hotter (and thus more buoyant) than their surroundings, or they could be chemically distinct (and likely denser) than their surroundings. Both suggest fundamental differences in how the mantle flows over millions of years and therefore how it influences tectonic plate motions and the related natural disasters. The PIs plan to, for the first time, combine information from traditional seismic data sets but also measurements from special types of seismic waves and earth tides to shed new light on the buoyancy of these LLSVPs. The special type of seismic waves are called Stoneley modes and are vibrations trapped along the core-mantle boundary (CMB) and the earth tides are the twice-daily deformation of the solid Earth under gravitational forces from the Sun and Moon. This award will support the training of two graduate students at UC Berkeley and two female PIs, one of whom is an early career scientist and new faculty member at UC Berkeley. The models developed in this project will be made available as part of a webpage specifically developed for this project, as well as through the Incorporated Research Institution for Seismology (IRIS) facility. They will be used in undergraduate workshops at UC Berkeley and at the interdisciplinary CIDER workshop geared towards senior graduate students and early career scientists.Within the last two years, independent studies have presented seemingly contradictory results on the density structure of the large low shear velocity provinces (LLSVPs) imaged by seismic tomography in the earth's deep mantle. In particular, one study used a variety of seismological data (normal mode splitting, seismic travel time and waveform data), and another, semi-diurnal earth tide data, concluding that a significant part of the LLSVPs represent regions of excess density pointing to the source of the LLSVP anomalies being dominated by chemical heterogeneity. In contrast, a study featuring measurements of Earth's Stoneley modes concluded the opposite, implying that the source of LLSVP anomalies are in large part thermal. Both conclusions paint a very different picture of mantle convection: the former implies that the LLSVPs are potentially sluggish upwellings or stabilized piles, while the latter implies a more energetic mode of mantle circulation. In this project, the investigators will combine these different geodetic and seismic datasets - which have different depth sensitivities to elastic and density structure - to establish whether previous results can be reconciled under the hypothesis that the excess density in LLSVPs may be confined to a thin layer (~200 km or less) above the CMB. In doing so, they will investigate further trade-offs with shear and compressional wave speed structure, anisotropy, attenuation, as well as CMB topography and possible structure in the outermost core. First, they will consider a combination of existing global seismic waveform, tide and normal mode splitting data, including Stoneley mode data, and perform inversions for structure in the deepest mantle, using the classical first order mode perturbation formalism for modeling splitting data. They will implement both an optimization and a trans-dimensional Monte Carlo inversion method for this purpose. This, in large part, will act as a guide to prepare for the more advanced inversions to be undertaken in the next stage, while yielding interesting intermediate results. In the second stage, after completing their own dataset of normal mode spectra, they will consider a more rigorous theoretical approach and directly invert mode spectra, together with long period waveform and tide data (complemented by newly acquired data) for density and velocity structure in the vicinity of the CMB.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.