The 2011 M9.0 Tohoku-oki Earthquake Provides Unprecedented Observations

The tragic M9.0 Tohoku-oki earthquake in Japan, which occurred on March 11, 2011, generated an extraordinary volume of observations that will be used to further our understanding of large seismic events. The devastating images, video, and reports showing widespread destruction serve as a catalyst to the scientists at Caltech to further strengthen our resolve to mitigate hazards associated with damaging earthquakes. Caltech scientists and collaborators at JPL continue to analyze the data, with the aim of describing the earthquake with the highest fidelity possible.

Animation showing back-projection of the teleseismic waves to their source. Note the north-south bilateral split of the fault rupture, with significant displacement of the fault continuing to the south, while mild to moderate displacement continues north.

We show a preliminary animation derived by back-projection of seismic waves, which models the initial 180 seconds of the earthquake rupture. Created by Lingsen Meng and Jean-Paul Ampuero of Caltech, the model uses data collected from USArray seismic stations located in the United States (in addition to European stations), and shows that ~100 seconds into the earthquake, the fault rupture undergoes a complex change, where the rupture front splits bilaterally north and south.

Ray count and analysis of high-frequency waveforms calculate 90% of the energy was released during the first 250 seconds of the fault rupture. This correlates to the slow speed of the bilateral fault rupture, shown in the previous animation, which models the speed of the rupture at ~2.8km/second. This is slightly faster than the 2004 Great Sumatra Earthquake, but still slower than typical fault rupture speeds.

High frequency analysis of the teleseismic records by Caltech Seismological Laboratory seismologists Risheng Chu and Don Helmberger further confirm the bilateral fault rupture as seen in the back-projection animation above, and the data demonstrate a slower than average rupture velocity, indicating this fault took longer to rupture than other faults of similar size and geometry.

Caltech scientists Shengji Wei and the members of the ARIA project (a collaboration between Caltech and JPL), along with Anthony Sladen of Geoazur, have created a  model of the distribution of slip on the subsurface fault.  The model is derived by combining observations from long distance teleseismic body waves and near source GPS observations. A  3D image of the fault slip can be viewed in Google Earth (link to KML file here: http://www.tectonics.caltech.edu/slip_history/2011_taiheiyo-oki/tohoku11-joint.kml ), and projects the fault rupture in three dimensions, which can be viewed from any point of reference. The analysis shows that the rupture originated ~24 km (15 miles) deep, and the maximum slip on the fault was 100 feet, ultimately thrusting the island of Japan as much as 5m (16 feet) east from its location before the earthquake.

The graphic above projects the fault geometry above the surface of the earth in three dimensions. Hot/cold colors correlate to maximum/minimum displacements of the fault, respectively. Click on the image above to download the Google Earth KML file.

Caltech scientists, in conjunction with Geoazur in France, have modeled the deadly tsunami by combining GPS and teleseismic data (http://www.tectonics.caltech.edu/slip_history/2011_taiheiyo-oki), creating an animation which is solved under the non-linear shallow water approximation using a finite difference code.

Maximum fault slip, shown in red, was nearly 30 m (90 feet). Surface projection of the fault plane slip distribution is shown above. Red dots are plate boundaries, whereas grey dots are aftershocks during the first day. The mainshock epicenter is denoted by the red star. This image was modeled by Shengji Wei, Anthony Sladen, and the ARIA group (Caltech-JPL).

The deadly Tohoku-oki earthquake occurred at the boundary between the Eurasian Plate and the Pacific Plate; this is a plate boundary where the plates are converging at a rate of ~8cm/year, respectively. No earthquake of this magnitude had occurred in this area within recorded history, and many in the scientific community did not think a M9 or larger was possible in this area. This event will change the way seismologists and geophysicists view subduction zone dynamics in the future.