Exploring the location of large earthquakes using fault stress accumulation -A case study upon Mw9.0 Tohoku-Oki earthquake in Japan

Where a large earthquake will occur is the first problem to be solved in earthquake prediction. Using the fault stress changes obtained from inversion for the GNSS observation data in the 7 years (2004 2010) before the 2011 M9.0 Tohoku-Oki earthquake, we have found the seismogenic zone of the earthquake. In order to further study its evolution process, we investigate the fault stress changes from 1997 to 2003 in this paper. Through the two phases of inversion work, we see that the main characteristics of the annual change pattern of the fault stress changes are basically stable during the 14 years, the stress in the seismogenic zone increases year by year, and there exist obvious stress increase and decrease areas. The former is consistent with the rupture zone of the earthquake, and the latter is consistent with the foreshocks, small repeated earthquakes and aseismic sliding zones. The areas of significant shear stress increase are not only consistent with the main shock, but also with the rupture areas of the large aftershocks. We found that the contours of zero-shear and zero-normal stress change on the fault plane do not coincide, the former has a greater depth than the latter, which means that there exists a normal stress reduction zone or a shear strength reduction zone in the shear stress increase zone because the shear strength is proportional to the normal stress. The initial rupture point of the fault seems to be more preferred the location close to the contour of zero-normal stress change, because it is not only close to the area of shear strength reduction, but also located in the most significant area of shear stress accumulation. The results obtained show that normal stress changes have an impact on the initial rupture of a large earthquake, and the inversion method of fault stress change can be used to predict the location of large earthquakes.