论文部分内容阅读
利用地震矩解目录外推估算了全球长期地震概率。该预测基于地震矩张量解,预测结果用一幅表示预测的地震发生率密度和震源机制方位图来表示。震源机制首先用于平滑地震活动图以获得期望的危险性图,然后用于预测未来地震的机制。使用了几种类型的平滑影响函数核:在空间域,使用1/距离影响函数核来计算任一震中周围的地震活动性分布。该影响函数核使用两个可调参数:即最大距离和方向性因子(任一震中周围考虑了一次地震震源机制的地震活动性分布)。对于时间预报,使用了地震时间行为的泊松假设。我们用下述方法检验这些预测,利用一个目录的前半部分平滑地震活动水平,用该目录的后半部分进行证实及优化预测。为了说明该方法,我们利用哈佛大学地震矩解目录中的可用资料评估了几个地震区的地震活动性图。该方法可用于类似的目录,分析方法完全公式化,不需要人为干预,因而对其预測结果可进行客观检验。此外,对照具有相同样本空间(相同区域、时间窗、可接受的判据)的任何其他预测模式,该方法的预测图可用作似然方法的泊松零假设检验。
The global long-term seismic probability is estimated by extrapolation of the catalog of moment solutions. The prediction is based on moment tensor solutions and the prediction is represented by a seismic intensity density and focal mechanism map representing the prediction. Focal mechanisms are first used to smooth seismic activity maps to obtain the desired hazard map and then to predict the mechanism of future earthquakes. Several types of smoothing kernels are used: in the spatial domain, the 1 / distance kernels are used to calculate the seismicity distribution around any epicenter. The kernel of the influence function uses two tunable parameters: the maximum distance and the directivity factor (the seismicity distribution of an earthquake focal mechanism is considered around any epicenter). For time prediction, the Poisson assumption of seismic time behavior is used. We test these predictions using the following method, using the first half of a catalog to smooth seismic activity levels, validating and optimizing forecasts with the second half of the catalog. To illustrate this method, we used seismograms from the Harvard Earthquake Momentum to assess the seismicity of several seismic zones. The method can be used in a similar catalog, the analytical method is completely formulated, and does not require human intervention, so its prediction can be objectively tested. In addition, the predictive graph of this method can be used as a Poisson's null hypothesis test for likelihood methods against any other prediction mode with the same sample space (same region, time window, acceptable criterion).