Markov chain Monte Carlo inversion of Rock Deformation Data: Applications to the Dynamics of Oceanic Mantle

地幔的动力学是导致各种地质活动的原因，包括地震、火山爆发、大陆漂移，甚至是调节地表温度的长期碳循环。地幔动力学最重要的参数是粘度，它决定了岩石的变形速度，但粘度也是地球物质中最不为人所知的特性之一。岩石变形实验提供了重要的粘度约束条件，但实验室条件与地幔条件有很大不同；地幔的变形速率比实验室实验慢10个数量级。通过将实验数据分析的最新进展与一种新的地球动力学建模相结合，拟议的项目将建立一个理论框架，有效地连接岩石变形实验和地球物理观测，并允许改进对地幔粘度的理解。由于其相对简单的构造环境，海洋地幔的动力学将作为这个多学科项目的理想试验台。该项目支持研究生的培养，并为进行高级研究项目的本科生提供主题。硅酸盐岩石的变形取决于许多因素，包括温度、压力、应力、晶粒尺寸、含水量、熔体分数和氧逸度。近年来，将这些因素纳入地球物理模拟已变得越来越普遍。与此同时，在实验岩石力学中有两条重要的发展线索，这两条线索都可以直接影响将岩石力学知识纳入地球物理建模的努力。首先，通常被认为控制上地幔动力学的橄榄石聚集体的流变学，长期以来被认为主要是由扩散蠕变和位错蠕变的组合控制的，但最近的实验研究表明，晶界滑动可能比以前认为的起更重要的作用。其次，我们已经看到了一个新的统计框架的发展，使用马尔可夫链蒙特卡罗（MCMC）采样，这有助于我们解决从变形数据估计流动规律的全部复杂性。MCMC反演技术的改进，使我们能够利用实验岩石力学和观测地震学的最新进展，对海底地幔动力学进行一种新的地球物理模拟。本项目有三个主要目标：(1)基于已发表的实验数据构建橄榄石聚集体的候选流律模型；(2)对海底地幔动力学进行概率模拟；(3)识别关键的流动规律不确定性，并设计可能的实验装置来解决这些不确定性。基于概率建模的方法不仅可以更好地理解海底地幔的动力学，而且可以更好地理解橄榄石集合体的流变学。

The dynamics of Earth's mantle are responsible for all kinds of geological activities including earthquakes, volcanic eruptions, continental drift, and even the long-term carbon cycle that regulates the surface temperature. The most important parameter of mantle dynamics is viscosity, which dictates how fast rocks can deform, but viscosity is also among the least understood properties of Earth materials. Rock deformation experiments provide important constraints on viscosity, but laboratory conditions are vastly different from mantle conditions; the rate of deformation for the mantle is 10 orders of magnitude slower than that for the laboratory experiments. By combining recent progress in the analysis of experimental data with a new kind of geodynamical modeling, the proposed project will build a theoretical framework that effectively bridges rock deformation experiments and geophysical observations and allows an improved understanding of mantle viscosity. Owing to its relatively simple tectonic setting, the dynamics of the oceanic mantle will serve as an ideal test bed for this multidisciplinary project. The project supports the training of a graduate student and provides topics for undergraduate students conducting senior research projects.The deformation of silicate rocks depends on a number of factors, including temperature, pressure, stress, grain size, water content, melt fraction, and oxygen fugacity. Incorporating such factors into geophysical modeling has become increasingly more common in recent years. At the same time, there have been two important threads of development in experimental rock mechanics, both of which can directly impact such efforts to incorporate the knowledge of rock mechanics into geophysical modeling. First, the rheology of olivine aggregates, which is usually thought to control the dynamics of the upper mantle, has long been considered to be governed mostly by the combination of diffusion creep and dislocation creep, but recent experimental studies suggest that grain boundary sliding may play a more important role than previously thought. Second, we have seen the development of a new statistical framework, using Markov Chain Monte Carlo (MCMC) sampling, that helps us to tackle the full complexity of estimating flow laws from deformation data. The improved command of MCMC inversion puts us into a unique position to conduct a new kind of geophysical modeling for the dynamics of suboceanic mantle by taking advantage of recent progress in both experimental rock mechanics and observational seismology. This project has three major objectives: (1) construct candidate flow-law models for olivine aggregates based on published experimental data; (2) conduct probabilistic modeling for the dynamics of suboceanic mantle; and (3) identify key flow-law uncertainties and design possible experimental setups that can resolve such uncertainties. The approach based on probabilistic modeling has potential to bring a better understanding of not only the dynamics of suboceanic mantle but also the rheology of olivine aggregates.