Testing Predictions from the Cascade and Pre-Seismic Slip Models for Foreshock Occurrence with the High Precision Catalog for Northern California

一个悬而未决的问题是，大地震是如何开始的，它们的成核过程能否与小地震区分开来？一些大地震前前震的存在是记录最好的主震前观测资料之一，因此可能为这些过程提供线索。一个相关的问题是，前震是否与任何其他地震不同？前震是一些大地震最明显、最有记录的前兆。理解它们发生的物理机制是至关重要的。如果单一的触发规律可以解释前震、主震和余震的观测行为，那么前震并不比任何其他地震具有更强的预测能力。然而，如果它们作为非地震成核阶段的副产品和间接证据出现，并与最终的主震震级成比例，它们本身和之间可能具有某种预测能力，或者证实和支持实验和理论研究预测的地震前滑动模型将发生在大地震之前。然后，可能会开发出检测这种成核相的其他方法。这项研究的影响将直接影响到地震的可预测性问题，从而对公众产生广泛的影响和兴趣，尤其是对灾害预测分析员、政府机构和决策者。这项提议将测试前震发生的两个相互竞争的模型。第一种是级联模型，在该模型中，前震是将主震推向崩溃的一系列连续的失败。第二个是震前滑动，前震是一个缓慢的、无地震的成核阶段的间接证据，据预测，这个阶段将随着主震强度的扩大而扩大。这两个预测都可以用我们为加州北部开发的高分辨率星表进行评估(最多提高三个数量级)。经过改进的震源参数和位置的静态应力变化计算可以确定这是否是增强或降低级联模型中主震失效压力的可行机制。相反，对前震带半径和破裂的测量可能显示出与地震前滑动模型中预测的主震震级的相关性。在以前的大多数研究中，由于位置误差大于成核区的预期大小，因此很难解释这些测量结果。我们在162个前震序列的较大数据集上的初步结果表明，前震的发生与正常应力呈负相关，这既表现在岩石静载荷随深度的变化上，也与主震滑动方向指示的区域构造应力呈负相关。我们提议的工作将检验一个单一的、统一的触发定律是否能在理论上解释我们数据中的观察结果。关于前震发生的两个相互竞争的模型的清晰明确的预测，现在可以用更高质量的数据和更多的数据进行测试，以确定有意义的统计数据和可靠的结论。由于位置误差，这些科学假设以前基本上是不可检验的，但现在有了加州北部30万个事件的完整高精度大目录，如果它们与观测数据一致，则可以单独检验级联模型的静态应力触发机制和地震前滑动模型的无震形核阶段。目前还不知道大地震是如何开始的，因此，了解和洞察近期前震的物理过程是很重要的。

An unanswered question is how do large earthquakes initiate and can their nucleation process be distinguished from small earthquakes? The existence of foreshocks before some large earthquakes is one of the best documented observables preceding mainshocks and therefore may offer clues into these processes. A related question is whether foreshocks are different than any other earthquakes? Foreshocks are the most obvious and well recorded precursors to some large earthquakes. It is essential to understand the physical mechanisms for their occurrence. If a single triggering law can explain the observed behavior for foreshocks, mainshocks, and aftershocks, then foreshocks have no more predictive power than any other earthquake. If, however, they occur as a byproduct and indirect evidence of an aseismic nucleation phase that scales with eventual mainshock magnitude they may have some forecasting ability, in and among themselves, or provide confirmation and support of the pre-seismic slip model predicted by experimental and theoretical research to occur before large earthquakes. Other methods of detecting this nucleation phase may then perhaps be developed. The implications of this research will bear directly on the question of earthquake predictability, thus having broad impact and interest to the general public, but especially among hazard forecasting analysts, governmental agencies, and policy makers. This proposal will test two competing models for the occurrence of foreshocks. The first is the cascade model where foreshocks are a successive series of failures that push the mainshock closer to failure. The second, pre-seismic slip, is one where foreshocks are indirect evidence of a slow, aseismic nucleation phase that is predicted to scale with mainshock magnitude. Both predictions can be evaluated with the high resolution catalog we have developed for northern California (up to three orders of magnitude improvement). Static stress change calculations with improved source parameters and locations can determine if this is a viable mechanism for enhancing or destressing mainshock failure in the cascade model. Conversely measurements of the foreshock zone radius and rupture may show a correlation with mainshock magnitude as predicted in the pre-seismic slip model. Interpretation of these measurements has been difficult before in most prior studies because the location errors were greater than the expected size of the nucleation region. Our preliminary results on a larger data set of 162 foreshock sequences show that there is a negative correlation of foreshock occurrence with normal stress manifested in both the lithostatic load with depth and with regional tectonic stress indicated by the orientation of mainshock slip. Our proposed work will examine if a single, unifying triggering law on theoretical grounds can explain the observations in our data. Clear well-defined predictions of two competing models for foreshock occurrence can now be tested with better quality data and much more of it to determine meaningful statistics and reliable conclusions. These scientific hypotheses were essentially untestable before due to location errors, but now with a large, 300,000 event complete high- precision catalog of northern California, both the static stress triggering mechanism of the cascade model and the aseismic nucleation phase of the pre-seismic slip model can independently be examined if they are consistent with the observed data. It is still not known how large earthquakes initiate and therefore it is important to gain knowledge and insight into the physical processes for immediate foreshocks.