Global survey of aftershock area expansion patterns

We developed an objective method to define the aftershock areas of large earthquakes as a function of time after the main shock. The definition is based upon the amount of energy released by aftershocks, the spatial distribution of the energy release is first determined and is contoured. The 1-day aftershock area is defined by a contour line corresponding to the energy release level of 10^(15.6) ergs/(100 km^2 · day). The 10-day, 100-day and 1-y aftershock areas are similarly defined by contour lines corresponding to 10^(14.8), 10^(14.0), and 10^(13.5) ergs/(100 km^2 · day), respectively. We also define the expansion ratios at time t by the ratio of the aftershock area at t to that at 1 day. Using this method we study the aftershock area expansion patterns of 44 large (M_s ⩾ 7.5) and five moderate shallow earthquakes which occurred from 1963 to 1980. Each aftershock sequence is examined at four different times, i.e., 1 day, 10 days, 100 days, and 1 y after the main event. We define the aftershock area expansion ratios η and η_e by S(100)/S(1) and L(100)/L(1), respectively: here S(t) and L(t) are the area and the length of the aftershock area, respectively, at time t. Our study suggests that a distinct regional variation of aftershock area expansion patterns is present; it is strongly correlated with the tectonic environment. In general, the subduction zones of the “Mariana” type have large expansion ratios, and those of the “Chilean” type have small expansion ratios. Some earthquakes that occurred in the areas of complex bathymetry such as aseismic ridges tend to have large expansion ratios. These results can be explained in terms of an asperity model of fault zones in which a fault plane is represented by a distribution of strong spots, called the asperities, and weak zones surrounding the asperities. The rupture immediately after the main shock mostly involves asperities. After the main rupture is completed, the stress change caused by the main shock gradually propagates outward into the surrounding weak zones. This stress propagation manifests itself as expansion of aftershock activity. In this simple picture, if the fault zone is represented by relatively large asperities separated by small weak zones (“Chilean” type), then little expansion of aftershock activity would be expected. On the other hand, if relatively small asperities are sparsely distributed (“Mariana” type), significant expansion occurs. The actual distribution of asperities is likely to be more complex than the two cases described above. However, we would expect that the expansion ratio is in general proportional to the spatial ratio of the total asperity area to the fault area