A comparison of two methods for earthquake source inversion using strong motion seismograms
A comparison of two methods for earthquake source inversion using strong motion seismograms
复制标题
两种强震震源反演方法的比较
DOI：
10.4401/ag-4151
复制
发表时间：
1994
影响因子：
1
通讯作者：
G. Beroza
中科院分区：
文献类型：
--
作者：
B. P. Cohee;G. Beroza
在本文中，我们比较了两种广泛应用于利用强震地震图模拟地震破裂问题的两种时域反演方法。在多窗口方法中，断层上的每个点都允许破裂多次。这使得破裂时间以及破裂速度具有灵活性。滑移速度函数的变化通过每个时间窗口中滑移幅度的变化来调节。单窗口方法假设当破裂前沿经过时，断层上的每个点仅破裂一次。允许滑移幅度的变化，并且通过允许破裂时间变化来适应破裂速度的变化。由于多窗口方法具有更大的灵活性，因此有可能描述更广泛的故障行为；然而，随着灵活性的增加，自由度也随之增加，并且解的稳定性相对较差。我们使用 1992 年加利福尼亚州兰德斯 7.3 级地震的测试模型的合成数据证明了这种效应，然后将两种反演方法应用于实际记录。这两种方法对具有不同地震矩的强震动数据产生相似的拟合，表明该力矩并不能很好地仅受强震动数据的约束。使用这两种方法时，滑移幅度分布相似，但破裂传播模型存在重要差异。单窗口方法可以更好地恢复真实地震矩和平均破裂速度。当上升时间变化很大时，多窗口方法是优选的，但往往会高估地震矩。当上升时间与建模周期相比恒定或较短时，两种方法都效果良好。除非滑移幅度的分布受到独立数据的约束，否则这两种方法都无法恢复破裂传播的时间细节。
In this paper we compare two time-domain inversion methods that have been widely applied to the problem of modeling earthquake rupture using strong-motion seismograms. In the multi-window method, each point on the fault is allowed to rupture multiple times. This allows flexibility in the rupture time and hence the rupture velocity. Variations in the slip-velocity function are accommodated by variations in the slip amplitude in each time-window. The single-window method assumes that each point on the fault ruptures only once, when the rupture front passes. Variations in slip amplitude are allowed and variations in rupture velocity are accommodated by allowing the rupture time to vary. Because the multi-window method allows greater flexibility, it has the potential to describe a wider range of faulting behavior; however, with this increased flexibility comes an increase in the degrees of freedom and the solutions are comparatively less stable. We demonstrate this effect using synthetic data for a test model of the Mw 7.3 1992 Landers, California earthquake, and then apply both inversion methods to the actual recordings. The two approaches yield similar fits to the strong-motion data with different seismic moments indicating that the moment is not well constrained by strong-motion data alone. The slip amplitude distribution is similar using either approach, but important differences exist in the rupture propagation models. The single-window method does a better job of recovering the true seismic moment and the average rupture velocity. The multi-window method is preferable when rise time is strongly variable, but tends to overestimate the seismic moment. Both methods work well when the rise time is constant or short compared to the periods modeled. Neither approach can recover the temporal details of rupture propagation unless the distribution of slip amplitude is constrained by independent data.