Contraining the large-scale dynamics and structure of the lower mantle using observations of the geoid, dynamic topography and plate tectonics

对地球内部地震波的分析表明，地幔主要是长波（5000公里）结构。特别是环太平洋地区的下地幔地震波速度相对较快，而非洲和太平洋地区的下地幔地震波速度相对较慢。非洲和太平洋下面这两个地震缓慢的区域被认为是异常的，并且很可能是地幔内的热区，因为地震波在热物质中传播较慢。这些异常在靠近地核和地幔边界（core - mantle boundary或CMB）区域尤为强烈，通常被称为大低剪切速度省（large low shear-velocity province）或llsvp。在过去的2亿年里，llsvp包含了大部分的火山活动和大型火成岩省。这种地幔地震结构模式也与地表重力异常高度相关。结合地震研究，地球化学研究也表明，llsvp的成分可能与地幔的大部分不同，富含重的、不相容的元素，如放射性元素U和Th。考虑到火山脱气对地球气候的重要影响以及地幔黏度对冰期后反弹和海平面变化的重要影响，了解地幔长波长对流动力学、llsvp的产生以及llsvp与其地表表现（包括火山作用和重力异常）的关系具有重要意义。这个为期3年的项目的目标是通过分析和模拟地幔对流的流体动力学过程来寻求对这些基本问题的理解。该项目还将通过与澳大利亚地球动力学家的国际合作，在模拟地幔对流的计算技术方面取得重大进展。该项目包括培训一名研究生。该项目旨在解决以下三个具体问题：1)llsvp是纯热特性还是热化学特性？2)大地水准面和动力地形的观测能否与稳定的热化学llsvp相一致？3)长波长地幔结构和对流（例如-2度）能否与动态板块和现实岩石圈流变（包括低温塑性）动态自一致地生成？该项目为期3年，主要任务包括：1)通过建立包含浮力的地幔流动模型，并考虑llsvp和钙钛矿后相变的成分效应，从地震模型中推导出浮力模型，寻求解释llsvp的大地水准面、CMB超椭圆性和化学性质不同的llsvp的稳定性的浮力和粘度结构；2)利用具有强加板块运动历史的时变对流模型，验证了llsvp是纯热异常或热化学异常的假设；3)通过建立具有真实岩石圈流变学的地幔对流全动力学模型，了解包括llsvp在内的长波长地幔结构的起源；4)提高CitcomS的求解器的效率和鲁棒性。该项目将大大提高对llsvp起源、长波长地幔对流和板块构造的认识。这些结果将对地球动力学以外的其他领域的研究产生直接影响，包括地幔地球化学、地震学、超大陆旋回、重力异常和火山作用。

Analyzing seismic waves traveling through the inside of the Earth has revealed that the Earth's mantle has a predominately long-wavelength (5000 km) structure. Particularly, the lower mantle beneath the circum-Pacific regions has seismic wave speeds that are relatively fast, while seismic wave speeds in the lower mantle beneath Africa and Pacific are relatively slow. These two seismically slow areas beneath Africa and the Pacific are are considered anomalies, and are most likely hot regions within the mantle because seismic waves travel slower in hot materials. These anomalies are particularly strong near the boundary between the core and mantle (Core-Mantle Boundary or CMB) regions, and are often termed as large low shear-velocity provinces or LLSVPs. The LLSVPs encompass most volcanism and large igneous provinces for the last 200 million years. This pattern in mantle seismic structure also highly correlates with the gravity anomalies at the Earth's surface. Together with seismic studies, geochemical studies also suggest that the LLSVPs may be compositionally different from the bulk of the mantle, being enriched in heavy, incompatible elements such as radioactive elements U and Th. Considering the significant influence of volcanic degassing on Earth's climate and the important effect of mantle viscosity on post-glacial rebound and sea-level change, it is therefore important to understand the dynamics of the long-wavelength mantle convection, the generation of the LLSVPs, and the relationship between the LLSVPs and their surface expression including volcanism and gravity anomalies. The goal of this 3-year project is to seek understanding of these fundamental questions, by analyzing and modeling fluid dynamical processes of mantle convection. This project will also lead to significant improvement in computational techniques in modeling mantle convection via international collaboration with an Australian geodynamicist. The project includes training of a graduate student. The project seeks to address the following three specific questions: 1) Are the LLSVPs purely thermal or thermochemical features? 2) Can the observations of the geoid and dynamic topography be made consistent with stable, thermochemical LLSVPs? 3) Can the long-wavelength mantle structure and convection (e.g., degree-2) be generated dynamically self-consistently with dynamic plates and realistic lithospheric rheology including the low-temperature plasticity? Four tasks for the proposed three-year project include: 1) to seek buoyancy and viscosity structures that explain the geoid, CMB excess ellipticity, and stability of the chemically distinct LLSVPs, by formulating mantle flow models with buoyancy derived from seismic models with compositional effects from the LLSVPs and post-perovskite phase change; 2) to test the hypothesis that the LLSVPs are purely thermal or thermochemical anomalies, using time-dependent convection models with imposed plate motion history; 3) to understand the origin of long-wavelength mantle structures including the LLSVPs, by formulating fully dynamic models of mantle convection with realistic lithospheric rheology; 4) to improve solvers in CitcomS for efficiency and robustness. The project should significantly improve understanding on the origins of the LLSVPs, long-wavelength mantle convection and plate tectonics. The results should have direct implications for studies in other areas beyond geodynamics, including mantle geochemistry, seismology, supercontinent cycles, gravity anomalies, and volcanism.