Temporal features of aftershock activities of four large earthquakes in the continental part of China were studied quantitatively. The anomalous change of aftershock activity has been found before the occurrence of large aftershock.
Aftershock activity shows an obvious decrease from the level expected from the modified Omori formula before the occurrence of a large aftershock in the same sequence. In some cases, the aftershock activity recovered after the decrease to the normal level or even exceeded the normal level just before the occurrence of a large aftershock. This recovery before large aftershocks is more appreciable in the case of the single main shock-aftershock type sequence than that in the case of the multiple event type sequence. The seismic activity after the largest foreshock of the Haicheng earthquake also decreased before the occurrence of the main shock.
The attenuation coefficients of aftershock activity (p-values in the modified Omori formula) in the continental part of China are smaller than those in Japan.

View full abstract

We observed the waveforms from a small earthquake (M=4.8), which corresponds to the largest aftershock of the 1983 Tottori earthquake, just above its hypocenter. The records have significant information on complicated source features. The P-waveforms recorded by a velocity seismograph are decomposed into a longer- (4 Hz) and a shorter-period (10-20 Hz) component. The source of the longer-period P-waves and S-waves can be theoretically explained by a uniform dislocation over a rectangular fault plane (1, 200 m × 800 m). Aftershocks accompanying this earthquake are distributed around the periphery of this fault plane. The source of the shorter-period P-waves corresponds to a multiple shock composed of two subevents whose source dimensions are each about 150m. Stress drops of the two subevents are about 10 to 100 times as large as that of the longer-period process.
The distribution of aftershocks has a clustering structure. Events in each cluster appear to take place on a common fault plane appropriate to each cluster with a dimension of 100-200 m. The clustering structure shows heterogeneous distribution of fracture strength in the aftershock area. We may suppose such heterogeneities composed of pre-existing small weak zones also in the source area of this earthquake (M=4.8), in order to explain its complex source process. Namely, the rupture is considered to have extended over the rectangular fault including several weak zones, and the shorter-period waves with high stress drop mentioned above are expected to have been radiated from the fractures of strong patches among weak zones.
The source process of micro to small earthquakes in a heterogeneous crust may be systematically explained as follows: The growth of rupture in the case of a microearthquake is confined within one or a few neighboring pre-existing weak zones (NISHIGAMI, 1987). On the other hand, the rupture in a small earthquake grows more extensively by fracturing not only several weak zones but also strong patches among them.

View full abstract

We have investigated spatial and temporal variation of seismicity before and after four moderate-size earthquakes (M=4.3-6.2) that occurred in active seismic zones in the Inner Zone of Southwest Japan.
The results of the present study mainly from space-time distribution diagrams show that the typical variation in seismicity consists of the following four stages. In the first stage of the variation, precursory earthquake swarms occur. Secondly, a seismic gap is formed until the occurrence of a main shock. Thirdly, the area of aftershocks expands in two ways, that is gradual growing and leaping. In the fourth stage, seismic activity spreads out from the aftershock area. We propose calling the phenomenon in the fourth stage "diffusion of earthquakes." The velocity of the diffusion has been estimated at 100 to 200 km/year.

View full abstract