Physics-Based Volcano Geodesy with Application to Effusive Eruptions at Mount St Helens

非技术综述在火山喷发之前，岩浆聚集在地球-S地壳的浅层储集层中。因此，这些岩浆室中的压力增加，从而变形或膨胀？S地表；相反，在喷发期间，岩浆离开这些储集层，降低压力，导致地球？S地表？？放空。更好地了解这些信号可能有助于改善社会对火山喷发的反应，例如可能的疏散和火山附近航线的改变，比如2010年冰岛火山喷发之后。该项目正在开发新的以物理学为基础的火山变形模型，该模型可以与地球望远镜板块边界的变形测量相结合，以改进对喷发持续时间和可能喷发的物质数量的预测。该项目正在研究一种数据同化方法，利用现有数据对描述岩浆系统状态的参数进行概率估计。然后，这些被用来初始化预测未来行为的正向模型集合，包括喷发持续时间和总喷发量。考虑到改进的正演模型，这种方法具有与现有数据和基于物理的真实喷发模型相一致的优势。基于物理学的火山喷发预报在概念上类似于数值天气预报，后者将卫星和其他数据吸收到复杂的天气模型中。该项目采用马尔可夫链蒙特卡罗(MCMC)方法，利用基于物理的喷发正演模型，对连续的GPS位置、岩浆喷发和其他数据进行反演。包括物理上一致的喷发模型，使估算程序能够限制传统方法无法解决的有关参数，包括地壳岩浆室的体积和岩浆的初始含水率。这些参数影响喷发的大小和潜在的爆炸潜力。正在进行的工作增加了正演模型的真实性，包括：1)岩浆上升和压力下降时的平衡结晶；2)基于宾汉流体模型，显式考虑从分布的粘性流动到边界断层上滑动的固体推流的流变学转变；3)显式考虑横向和垂直扩散的气体损失(H2O和CO2)。其他目标包括更好的喷发开始和停止模型。该方法正在应用于2004-2008年圣海伦斯火山(MSH)形成穹顶的喷发，包括来自板块边界观测站(PBO)的GPS数据，并可应用于其他火山，包括阿拉斯加的奥古斯丁、日本的温岑和蒙特塞拉特的SouFriere Hills。

Non-technical summaryPrior to volcanic eruptions magma accumulates in shallow reservoirs in Earth?s crust. As a result, pressure increases in these magma chambers, which deforms or ?inflates? the Earth?s surface; in contrast, during eruptions, magma leaves these reservoirs, decreasing pressure and causing the Earth?s surface to ?deflate?. Better understanding of these signals could help improve societal responses to volcanic eruptions, such as possible evacuations and changes to airline routes near volcanoes like following the 2010 Icelandic eruption. This project is developing new physics-based models of volcano deformation, which can be coupled with deformation measurements from the EarthScope Plate Boundary to improve forecasts of the duration of an eruption and the volume of material that may be erupted. The project is investigating a data assimilation approach in which available data are used to develop probabilistic estimates for parameters that describe the state of the magmatic system. These are then used to initialize an ensemble of forward models that predict future behavior, including eruption duration and total erupted volume. Given improved forward models, this approach has the advantage of being consistent with both available data and realistic physics-based eruption models. Physics-based volcano eruption forecasts are similar in concept to numerical weather forecasts that assimilate satellite and other data into sophisticated weather models. Technical summaryThis project employs Markov Chain Monte Carlo (MCMC) inversion of continuous GPS positions, magma efflux, and other data using a physics-based forward model of an effusive eruption. Including a physically consistent eruption model allows the estimation procedure to constrain parameters of interest that are not resolved by traditional approaches, including the volume of the crustal magma chamber and the initial water content of the magma. These parameters influence the size and potential explosive potential of eruptions. Ongoing work is increasing the realism of the forward model by including: 1) equilibrium crystallization of the magma as it ascends and pressure decreases; 2) explicit consideration of the rheological transition from distributed viscous flow to solid plug flow with slip on bounding faults, based on a Bingham fluid model and 3) explicit consideration of gas loss (both H2O and CO2) through both lateral and vertical diffusion. Other goals include better models of the eruption onset and cessation. The method is being applied to the 2004-2008 dome forming eruption of Mount St. Helens (MSH), including GPS data from the Plate Boundary Observatory (PBO) and could be applied to other volcanoes, including Augustine in Alaska, Unzen in Japan, and the Soufriere Hills on Montserrat.