Scaling laws for large earthquakes: Consequences for physical models

Scaling laws for large earthquakes: Consequences for physical models
复制标题

DOI:
--
复制
发表时间:
1982-02
影响因子:
3
通讯作者:
C. Scholz
C. Scholz
中科院分区:
地球科学3区
文献类型:
--
作者:
C. Scholz

文献摘要

被引文献

相似文献

观测到大地震的平均滑动与断层长度成线性关系,而与断层宽度无关。根据两种可能的大地震模型解释了这一观察结果:W模型和L模型,在W模型中,应力降和滑动由断层宽度决定,在该模型中,这些参数基本上由断层长度决定。在W模型的解释中,应力降随L/W、长宽比以及相应的地震矩而系统地增加。滑移与长度的相关性意味着破裂长度由动应力降决定。这与以下观点相矛盾:大地震的持续时间通常受以前地震的相邻破裂带或构造障碍的控制。这也与对小地震的观测结果相矛盾,即应力降几乎是恒定的,并且在很大范围内与震源半径无关。在L模型的解释中,滑动与长度的相关性意味着应力降是恒定的,即板间走滑、逆冲和日本板内地震的应力降分别约为7.5bar、12bar和60bar。L模型要求断层在底部不受机械约束。W模型预测平均质点速度随断层长度增加而增加,但上升时间不变。L的模型做出了相反的预测。
abstract It is observed that the mean slip in large earthquakes is linearly proportional to fault length and does not correlate with fault width. This observation is interpreted in the light of the two possible classes of models for large earthquakes: W models, in which stress drop and slip are determined by fault width, and L models, in which these parameters are fundamentally determined by fault length. In the W model interpretation, stress drop systematically increases with L/W , the aspect ratio, and, as a consequence, seismic moment. The correlation of slip with length means that the rupture length is determined by the dynamic stress drop. This conflicts with the observation that the length of large earthquakes is often controlled by adjacent rupture zones of previous earthquakes or by tectonic obstacles. It also conflicts with the observations for small earthquakes that stress drop is nearly constant and does not correlate with source radius over a broad range. In the L model interpretation, the correlation between slip and length means that stress drop is constant, namely about 7.5, 12, and 60 bars for interplate strike-slip, thrust, and Japanese intraplate earthquakes, respectively. L models require that the fault be mechanically unconstrained at the base. W models predict that mean particle velocity increases with fault length, but rise time is constant. L models predict the opposite.