Co- and post-seismic deformation of the M7.3 foreshock triggering the 2011 M9.0 Tohoku Earthquake

A large foreshock with M7.3 occurred on March 9, 2011 at the subducting Pacific plate interface followed by the M 9.0 Tohoku earthquake 51 hours later. We propose a simple rectangular fault model of the foreshock based on a dense GPS network and Ocean Bottom Pressure gauge (OBP) sites. The multiple OBP gauges were installed before the M7.3 earthquake in and around the focal area. In this time, we succeed to collect the OBP gauge data in 4 sites. Inazu et al. [this meeting] will be made more detail discussion for the OBP data and it's processing. The horizontal coseismic displacements are estimated based on baseline analyses to show the broad area of displacement field up to ~30mm directing to the focal area. In contrast, there is no coherent signal in the vertical components. The several OBP sites, for example, TJT1 site located the eastward (offshore) from the hypocenter of the foreshock, clearly detected the coseismic displacement. After the earthquake occurred at 2:45 (UTC), some oceanographically disturbance appeared at TJT1 that might be caused by the environmental change due to the ground shaking. The small negative coseismic offset, however, is clearly seen as a difference in pressure before and after the earthquake. The negative change in the pressure corresponds to the uplift of the site amounting to -2.1 hPa (~21 mm) at TJT1. Based on the land GPS and OBP data, we constructed a simple rectangular fault model using a non-linear inversion approach. The fault model can explain observations including the vertical displacement obtained at the OBP sites. The amount of moment release is equivalent to Mw 7.15. The spatio-temporal aftershock distribution of the foreshock shows a southward migration from our estimated fault model. We suggest that aseismic slip occurred after the M7.3 earthquake. The onshore GPS data also supports the occurrence of the afterslip in the southwestward area of the coseismic fault. We estimated the sub-daily coordinates every three hours at the several coastal GPS sites to reveal the time evolutional sequences suggesting the postseismic deformation, especially in the horizontal components. We also examine volumetric strain data at Kinka-san Island, which is situated at the closest distance from the hypocenter. The time series also clearly show the postseismic signal after the M7.3 earthquake. The both of strain meter and GPS time series did not show any acceleration expected as a nucleation process of the M9.0 event; rather, both time series show deceleration of the postseismic deformation before the M9.0 event, even for only two days after the M7.3 earthquake. We suggest that aftershocks of the March 9 event may have been caused by stress concentrations along the edge of the afterslip. One of these aftershocks may trigger the huge M 9.0 Tohoku earthquake, although more investigations are required to confirm this scenario.