Aftershock study of the 2007 Sumatran earthquakes: aftershocks and stress diffusion

On September 12, 2007, 11:10 a great earthquake with MW=8.4 occurred off the West Sumatran coast near the city of Bengkulu and the Mentawai islands, followed by another major earthquake with MW = 7.9 just over 12 hours later to the north of the first rupture, and a significant aftershock with MW = 7.0 near the Mentawai island of Sipora at 2.2S. The distribution of aftershocks and first estimates of slip distribution (Fig. 1a) suggest that the slip patch of the 1833 earthquake was re-ruptured by these earthquakes, but with a much smaller maximum slip. Changes to the Coulomb stress have been recognised as an important factor determining the time-dependent seismic hazard. It is therefore of great importance to understand how the Coulomb stress evolves in response not only to the coseismic rupture but also to the postseismic relaxation and the aftershock sequence. The aftershock seismicity rate is generally observed to fall of as a function of time after the event according to the modified Omori law. The Omori law can arise naturally from the evolution of the Coulomb stress around a rupture in a system of rate-state friction controlled faults. However, a similar decay law is expected for postseismic relaxation governed by a power law rheology. So, an alternative possibility is that aftershocks arise primarily as side effects of largely aseismic stress diffusion. According to this view, most aftershocks would occur on small brittle (velocity-weakening) patches within a ductile or creeping (velocity-strengthening) background. A closely related question is whether the postseismic deformation is primarily caused by afterslip (i.e., aseismic slip on a fault surface), or by viscoelastic relaxation in the overriding plate or along a ductile zone at the plate interface. Although afterslip appears to dominate in the case of strike-slip faults it is harder to separate the mechanisms for subduction-zone megathrusts, where available postseismic deformation data can often be fit successfully by different relaxation mechanisms. In most subduction zones, it is very difficult to obtain any constraints at all on relaxation processes updip of the coseismic rupture due to the fact that it is usually under water and far from the coast. The Sumatran subduction zone is unusual in that it has a large number of forearc islands, and thereby better constraints on co- and postseismic deformation updip of the rupture can be obtained. At the time of the Mentawai earthquakes described above, a US-Indonesian network of continuous GPS stations along the mainland coast and on the Mentawai islands was in operation. Fortuitously, a survey team from GEOTEK-LIPI was in the area at the time of the earthquakes, allowing rapid measurements of coseismic uplift. Therefore, good geodetic constraints are available for the northern part of the MW = 8.4 rupture, as well as for the MW = 7.9 and MW = 7.0 earthquakes. We propose to install a temporary network of seismic stations to record aftershocks of these events in order to create a dataset with which the link between aftershocks and postseismic deformation can be explored. Good geodetic constraints also exist for the southern limit of the 2004 Aceh-Andaman earthquake and the 2005 Nias earthquake. However, temporary local seismic networks deployed after the Andaman-Aceh earthquake were either too far away from the geodetic constraints, or were installed too late to truly test the link between postseismic deformation and aftershocks. The recent events present a unique opportunity to record aftershocks above the slip patch of a great subduction zone earthquake (and which is geodetically well constrained) made possible by the rare occurrence of forearc islands.