Wave-equation helioseismology

太阳是一颗磁星。它的磁场渗透到它的上层、大气层和太阳系的大部分地区。这种磁场以及太阳内部浅层相关物质流中的变化和复杂性被认为是导致一系列现象的原因。范围从太阳黑子的形成和演化，到太阳耀斑和物质抛射，再到太阳风和近地环境的性质变化，这些变化对电信产生了直接影响，并促进了地球上的全球变化。尽管它们很重要，但太阳变化的原因，磁场的作用，以及它们与太阳内部物质流动的关系，还没有被很好地理解。日震学家通过测量太阳表面发出的光的多普勒频移来记录太阳上的低频声波。当地的日震学使用这些声波，通过太阳上方几十兆米的区域传播，来拍摄太阳内部的图像。这些图像揭示了物理性质的复杂变化和复杂的流动模式。通过使用声波直接观测太阳内部，日震学家正试图了解太阳变异性的起源及其与磁场的关系，并最终能够预测太阳上那些直接影响地球和近地环境的事件的发生。该项目旨在为日震学带来一种新的技术方法，使产生的图像在空间和时间上都比使用现有技术可能获得更好的分辨率。这种新的方法，波动方程层析成像，是由对以高分辨率成像地球感兴趣的地震学家开发的，主要是为了帮助石油工业寻找石油和天然气。该行业已经在这些技术上进行了巨大的投资，这些投资的一些技术成果现在可以应用于一个新的科学领域。与地震或地球上的人造源不同，太阳声源不是离散的事件。相反，太阳源既分布在空间上，也分布在时间上。为了使用这种波来成像太阳内部，现有的技术必须首先对太阳相当大的区域和相当长的一段时间内的数据进行平均。这种平均化必然会限制这种方法所能达到的空间和时间分辨率。这些方法的成像还基于简化的射线物理理论，由于这不能完全解释声波的有限波长，因此它可能会在最精细的尺度上折衷它们的分辨率和精度。在波动方程层析成像中，波传播的全部物理过程是在计算机中模拟的，并作为成像的基础。这种方法既避免了其他方法所需的平均，也避免了与几何光学相关的近似。在一项初步研究中，我们已经证明了这项新技术适用于分布式来源的合成太阳数据。这种方法对计算要求很高，但即使在3D模式下，它也可以在现代计算机硬件上实现。在这个项目中，我们将开发这项技术，以便它可以常规地应用于当地的日震数据。我们将测试我们方法在合成数据上的准确性和分辨率，并将其应用于来自欧空局/NASA SOHO卫星的现有数据，以及将在定于2008年发射的NASA SDO任务中收集的新的更高分辨率的数据。我们特别感兴趣的是将该方法应用于太阳黑子和太阳上其他活动磁区的数据。在项目结束时，我们将发布计算机代码，以便其他正在研究日震数据的人可以将新方法应用到他们自己的问题上。

Summary The Sun is a magnetic star. Its magnetic field permeates its upper layers, its atmosphere, and much of the solar system. Variations and complexities within this magnetic field, and within related mass flows in the shallow interior of the Sun, are thought to be responsible for a wide range of phenomena. These range from the formation and evolution of sunspots, through solar flares and mass ejections, to changes in the properties of the solar wind and the near-Earth environment that have a direct effect on telecommunications and contribute to global change on Earth. Despite their importance, the causes of solar variability, the role of the magnetic field, and their relationship to the flow of material within the interior of the Sun, are not well understood. Helioseismologists record low-frequency sound waves on the Sun by measuring the Doppler shift in light emitted from the surface. Local helioseismology uses these sound waves, propagating through the upper few tens of mega-metres of the Sun, to make images of the solar interior. These images reveal complicated changes in physical properties, and complicated patterns of flow. By using sound waves to observe the interior of the Sun directly, helioseismologists are attempting to understand the origin of solar variability, its relationship to the magnetic field, and ultimately to be able to predict the occurrence of those events on the Sun that directly influence the Earth and the near-Earth environment. This project aims to bring a new technological approach to helioseismology that will allow images to be generated that are better resolved in both space and time than is possible using existing techniques. This new approach, wave-equation tomography, has been developed by seismologists interested in imaging the Earth at high resolution, principally to aid in the search for oil and gas for the petroleum industry. A huge investment in those techniques has been made by that industry, and some of the technical fruits of that investment can now be applied to a new area of science. In contrast to earthquakes or man-made sources on the Earth, sources of solar sound are not discrete events. Instead the solar source is distributed in both space and time. In order to use such waves to image the interior, existing techniques must first average data over a significant area of the Sun and over a significant time. This averaging necessarily limits the resolution in space and time that such methods can achieve. These methods also base their imaging on the simplified physics of ray theory, and since this does not fully account for the finite wavelength of sound waves, it can compromise both their resolution and accuracy at the finest scales. In wave-equation tomography, the full physics of wave propagation is simulated in the computer and used as the basis for imaging. This approach avoids both the averaging required by other methods, and the approximations associated with geometric optics. In a pilot study, we have shown that the new technique works on synthetic solar data with a distributed source. The approach is computationally demanding, but even in 3D it is achievable on modern computer hardware. In this project, we will develop this technique so that it can be applied routinely to local helioseismic data. We will test the accuracy and resolution of our method on synthetic data, and will apply it to existing data from the ESA/NASA SOHO satellite and to new higher-resolution data to be collected on the NASA SDO mission scheduled for launch in 2008. We are interested particularly in applying the method to data from sunspots and other active magnetic regions on the Sun. At the end of the project, we will release the computer codes so that others who are working on helioseismic data can apply the new methods to their own problems.