Fluid oscillations in conduit-reservoir systems, very long period seismic signals at Kilauea volcano, and the phenomenology of unsteady magma ascent

火山爆发和与之相关的人类危害很难预测，因为它们涉及隐藏在地球表面以下的岩浆的高度不稳定运动。尽管对推断地下岩浆运动（如地震活动和次声）至关重要的观测数据越来越多地被收集起来，但我们缺乏一个理论框架来理解产生这些信号的物理过程。这项工作将是在短时间尺度上研究火山导管内的岩浆流动，以了解引起各种喷发现象的物理过程。第一个目标是研究夏威夷基拉韦厄火山的火山活动，该火山与岩石落入活跃的熔岩湖有关，这会引起管道内岩浆的共振振荡，有时会在地表发生小而危险的爆炸。这些充分记录的自然实验为浅层管道和岩浆储层几何模型以及多相岩浆流体特性提供了独特的测试。第二个目标是开发一个高性能的计算建模框架，以预测火山管道中岩浆在一系列流动条件下的瞬态运动，包括在爆炸性喷发期间。这项工作将有助于弥合地球物理观测、火山物理学和先进计算之间的差距。该项目将支持研究生进行跨学科的科学研究，并为“火山聆听项目”的持续发展做出贡献，该项目致力于将火山数据以声音和动画的形式呈现出来。本文描述了火山导管中小振幅振荡岩浆流的研究。第一个目标是研究夏威夷基拉韦厄火山的短期（数十分钟）动荡事件，这些事件与岩石落入活跃的熔岩湖和上升的气体鼻涕虫有关。这些扰动引起管道内多相岩浆的“非常长周期”（VLP, 5−40 s）振荡、爆炸、不稳定的地表气体通量和熔岩湖高度变化，这些都记录在附近的地球物理仪器网络上。这些记录良好的自然实验为非定常管道流动模型提供了一种独特的测试方法，该模型将用于反演地下管道和储层几何形状以及岩浆流变学和初级挥发物含量。以前nsf资助的工作开发了一个初步框架，用于根据耦合管道-水库系统的共振特征模态建模和反演VLP地震数据，其中流体压力变化导致周围固体地球的弹性变形，这些变形被仪器记录下来。这一框架将应用于基拉韦厄火山Halema 'uma 'u喷发十年期间的数千次事件。贝叶斯马尔可夫链蒙特卡罗反演将结合地震活动、地面变形、连续重力、岩石学决定的熔体粘度和挥发性含量以及熔岩湖几何形状的限制。第二个目标是生成一个正演模型框架来预测岩浆流复杂状态下的波动。模拟气体和液体之间的相对运动将预测瞬态动荡事件期间的地表气体通量数据，以及由岩崩引发的瞬态爆炸。波浪状扰动也将在爆炸性喷发的背景下进行研究，其中这种振荡流体运动可能在爆发期间的状态转移中发挥关键作用，例如碎裂和爆炸行为的开始。火山系统中的流动通常没有在这里提出的短时间尺度上进行研究。要实现多个地球物理数据集之间的一致性，必须明确考虑非平衡气泡生长和吸收、包括分支裂缝在内的复杂管道几何形状，以及具有强界面（如气泡析出或岩浆破碎）的分层多相流体流动。这种方法也允许对准稳定管道流动模型进行严格检查，这些模型可以被证明对扰动是有条件的线性不稳定的。这为研究喷发样式的转变提供了一种新的方法。最后，对强分层多相系统中不稳定波动的研究将有助于数值方法的发展，其应用将超越火山学。利用可证明稳定的高阶有限差分方法的最新发展在数值上解决流动不稳定性提供了一个独特的机会来推进火山管道流动模型，发现新的喷发现象，并与火山监测工作联系起来。该项目将涉及俄勒冈大学地球科学和计算机科学的两名博士生。软件将是开源的，并可供社区使用。项目组将访问并与美国地质勘探局（夏威夷火山观测站）合作研究基拉韦厄火山。正在进行的建模和地震数据分析的结果将被纳入公共演讲和“火山聆听项目”的推广工作，致力于将火山数据以声音和动画的形式呈现出来。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。

Volcanic eruptions and the human hazards associated with them are challenging to predict because they involve highly unsteady motions of magma hidden below Earth’s surface. Although observations critical for inferring subsurface magma motions, such as seismicity and infrasound, are being collected more and more often, we lack a theoretical framework for understanding the physical processes that give rise to such signals. This work will be a study of magma flow within volcanic conduits on short timescales, in order to understand the physical processes that give rise to a variety of eruptive phenomena. The first goal is to study volcanic activity at Kilauea volcano, Hawai’i’, associated with rocks falling onto an active lava lake, which cause resonant oscillations of magma within the conduit and sometimes small but hazardous explosions at the surface. These well-documented natural experiments provide a unique test for models of the shallow conduit and magma reservoir geometry as well as multiphase magma fluid properties. The second goal is to develop a high performance computing-enabled modeling framework to predict transient motions of magma in volcanic conduits under a range of flow conditions, including during explosive eruptions. This work will help bridge the gap between geophysical observations, volcano physics, and advanced computing. The project will support graduate students in interdisciplinary scientific research, and contribute to ongoing development of “The Volcano Listening Project”, an outreach effort dedicated to representing volcano data as sound and animations.This proposal describes a study of small amplitude oscillatory magma flow in volcanic conduits. The first goal is to study short term (tens of minutes) unrest episodes at Kilauea volcano, Hawai’i’, associated with rock falls onto an active lava lake and from rising gas slugs. These disturbances cause ‘very long period’ (VLP, 5−40 s) oscillations of the multi-phase magma within the conduit, explosions, unsteady surface gas flux, and lava lake height variations, recorded on a nearby network of geophysical instruments. These well-documented natural experiments provide a unique test for unsteady conduit flow models, which will be used to invert for subsurface conduit and reservoir geometry as well as magma rheology and primary volatile content. Previous NSF-funded work developed a preliminary framework for modeling and inverting VLP seismic data in terms of the resonant eigenmodes of coupled conduit-reservoir systems, where fluid pressure changes cause elastic deformations of the surrounding solid Earth that are recorded instrumentally. This framework will be applied to thousands of events spanning the ten-year lifespan of the Halema’uma’u vent on Kilauea. Bayesian Markov-Chain Monte Carlo inversions will incorporate constraints from seismicity, ground deformation, continuous gravity, petrologically-determined melt viscosity and volatile content, and lava lake geometry. The second goal is to generate a forward modeling framework to predict wave motion in complex states of magma flow. Modeling relative motion between gas and liquid will predict surface gas flux data during transient unrest events, as well as transient explosions triggered by rockfalls. Wave-like disturbances will also be studied in the context of explosive eruptions, where such oscillatory fluid motions may play a key role in state shifts during eruptions such as the onset of fragmentation and explosive behavior. Flow in volcanic systems is typically not studied at the short timescales proposed here. Explicit consideration of non-equilibrium bubble growth and resorption, complex conduit geometry that includes branching cracks, and stratified, multiphase fluid flow with strong interfaces (such as bubble exsolution or magma fragmentation) is necessary to achieve consistency between multiple geophysical datasets. This approach also permits a critical examination of quasi-steady conduit flow models, which can be shown to be conditionally linearly unstable to perturbations. This suggests a new approach to studying transitions in eruption style. Finally, the study of unstable wave motions in strongly stratified multiphase systems will contribute to numerical method developments with applications beyond volcanology. Numerically resolving flow instabilities using recent developments in provably stable high-order finite difference methods provides a unique opportunity to advance models of volcanic conduit flow, discover new eruptive phenomenology, and connect with volcano monitoring efforts. The project will involve two PhD students at the University of Oregon across Earth Science and Computer Science. Software will be open-source and available to the community. The project team will visit and collaborate with the USGS (Hawaiian Volcano Observatory) to study Kilauea. Ongoing results of modeling and seismic data analysis will be incorporated into public presentations and to “Volcano Listening Project” outreach effort dedicated to representing volcano data as sound and animations.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.