Investigating the Presence of Preseismic Velocity Changes

该奖项由2009年《美国复苏和再投资法案》(公法111-5)资助。地震毫无征兆地来袭，是造成生命损失的最具破坏性和破坏性的自然力量之一。2008年5月，中国汶川大地震造成8万人死亡，2万人失踪。2004年苏门答腊岛地震引发了一场巨大的海啸，造成22.5万人伤亡，影响到11个国家。1976年，唐山地震再次发生在中国，造成24万人死亡，是有记录以来死亡人数最多的一次。哥伦比亚大学编制的统计数据显示，在包括洪水、干旱、飓风、龙卷风和火灾在内的世界范围内的所有自然灾害中，地震的总伤亡人数最高。许多其他自然灾害提供了某种类型的警告，以便即使财产损失很大，至少可以拯救生命。可以理解的是，人们一直在努力寻找一些线索，以预测即将到来的大地震，希望能减轻一些损失和痛苦。但地震预报仍然是一个极具挑战性的问题。大多数前兆的证据都是轶事，没有在严格控制的环境中得到科学验证，结果是可重现的。专业地震学家一直在尽其所能为地震预测做出贡献。在大地震之前寻找地震波速度或其他波形属性的前兆变化已经有很长的历史了。对震前速度变化的研究被称为地震学的圣杯。该项目将继续朝着这一目标迈出一小步。已经取得了进展。现在，科学上已经很好地记录了几次大地震的同震和震后速度变化都表现出了一致的行为。然而，这些研究中的数据没有所需的时间样本来确认震前速度变化的存在或不存在。这些发现之所以没有定论，仅仅是因为缺乏数据。在这项提案中，我们提出了三种不同的策略，旨在极大地改善震前时间采样不足的状况。它们涉及新数据和新技术，这些都是最先进的，最近才出现。第一种方法是对新的重复事件序列进行全面分析，这些新的重复事件序列已成为同一位置可重复波的来源。重复事件的丰富的新数据来源大大改进了这种现在测量速度变化的标准方法的时间采样。第二种方法利用最近才被应用于速度变化监测的环境地震噪声相关这一热门的新领域。这项技术利用来自永久监测站的连续数据流，因此能够对大地震前几天的速度变化进行精细采样。我们从相关检测器的研究中提出了对该方法的改进，以更好地从噪声中提取信号。第三种策略是开发一个原型的依赖时间的双差层析成像程序，以反转变化在空间上发生的位置，并能够使用大多数微震。实验室实验和理论都得出结论，在大地震之前，地壳中应该有可测量的速度变化。它们还需要在现场进行一致和可靠的测量。原因可能是缺乏高质量的数据，因为我们不知道下一次地震将在何时何地发生。然而，最近在帕克菲尔德进行的一项控制极好的实验观察到了两次地震前的速度变化，这是迄今为止最令人信服的证据，证明这些预测的变化确实发生在现场。但进行更多这类实验在经济上是不可行的，因此需要替代策略来在其他大地震之前证实地震前速度变化的存在。地震学家最常被普通公众问到的问题是，我们还能预测地震吗？答案是目前不是。虽然这是一个遥远的目标，但广大人民显然关心这一问题，并希望他们提供的资源的一部分将分配给这一目标。工程师可以建造更安全的建筑来大幅减少生命损失，但在许多较贫穷的国家，这些国家碰巧也是伤亡人数最多的国家，这在经济上是令人望而却步的，因此他们将从更好的地震预报中受益最大。

This award is funded under the American Recovery and Reinvestment Act of 2009 (Public Law 111-5).Earthquakes strike without warning and are some of the most destructive and devastating forces of nature in terms of loss of life. In May, 2008 the Wenchuan earthquake in China killed 80,000 people with 20,000 still missing. The 2004 Sumatra earthquake generated a giant tsunami that left 225,000 casualties in its wake, affecting eleven countries. In 1976 the Tangshan earthquake in China again, killed 240,000 people, the deadliest in recorded history. Statistics compiled at Columbia University indicate that of all natural disasters world-wide including floods, droughts, hurricanes, tornadoes, and fires?earthquakes rank the highest for total number of casualties. Many of the other natural disasters provide some type of warning so that at least lives can be saved even if property damage is extensive. Understandably people have diligently sought for some clues to foretell of impending great earthquakes in hopes of alleviating some of this loss and suffering. But earthquake prediction remains an extremely challenging problem. Most of the evidence of precursors is anecdotal and has not been scientifically verified in a rigorously controlled setting that produces reproducible results.Professional seismologists have been doing their best to contribute what they could to the challenge of earthquake prediction. There has been a long history of looking for precursory variations in the velocity of seismic waves or other waveform attributes preceding large earthquakes. The search for preseismic velocity changes has been called the holy grail of seismology. This project will continue taking small steps in that pursuit. Progress has been made. It is now scientifically well-documented that coseismic and postseismic velocity changes have been measured by a variety of means for several large earthquakes all demonstrating consistent behavior. The data in these studies, however, do not have the temporal sampling required to confirm the existence or non-existence of a preseismic velocity change. The findings are inconclusive simply because of a lack of data. We present in this proposal three different strategies designed to greatly improve the insufficient temporal sampling in the preseismic period. They involve new data and new technologies that are state-of-the-art and have recently come on the scene. The first approach is to do a comprehensive analysis of new repeating event sequences that have become available as a source for repeatable waves in the same location. The abundant new data sources for the repeating events substantially improves the temporal sampling for this now standard method for measuring velocity variations. The second approach takes advantage of the hot, new field of ambient seismic noise correlations which just recently has been applied for velocity change monitoring. The technique utilizes continuous data streams from permanent stations and so is able to finely sample velocity variations in the days leading up to large earthquakes. We propose enhancements to this method from our correlation detector research to better be able to extract the signal from the noise. The third strategy is to develop a prototype time-dependent double-difference tomography code to invert for where the changes are occurring spatially and to be able to use the majority of the microseismicity.Both laboratory experiments and theories have concluded that there should be measureable velocity changes in the crust preceding large earthquakes. They have yet to be consistently and reliably measured in the field. The reason for this may be due to lack of high quality data since we don?t know where and when the next earthquake will strike. However, an extremely well controlled experiment at Parkfield has recently observed velocity changes preceding two earthquakes presenting the most convincing evidence to date that these predicted changes do occur in the field. But conducting many more of these types of experiments is economically infeasible and so alternative strategies are needed to confirm the existence of preseismic velocity changes before other large earthquakes.The question that seismologists are most often asked by the lay public is, "Can we predict earthquakes yet?" The answer is not currently. While it is a far off goal, the general population clearly cares about this issue and desires that a portion of the resources they provide would be allocated towards that end. Engineers can build safer buildings to reduce the loss of life substantially, but in many of the poorer countries which also happen to be the ones that suffer the greatest number of casualties, this is economically prohibitive and so they would benefit the most from better earthquake forecasting.