金曜セミナー（5月22日）Christopher H Sholz氏
Some Comments on the Great Tohoku-oki Earthquake
Prof. Christopher H Scholz (Lamont-Doherty Earth Observatory)
I will discuss two issues concerning the great Tohoku-oki earthquake of 2011. What is the role of such earthquakes in the seismic cycle of the Japan Trench, and what is the rupture mode of the large shallow slip surge in that earthquake?
Several inversions of GPS data prior to 2011 indicated that the long term seismic flux (PS=Mo/µ) accumulation rate for the Honshu sector of the Japan Trench subduction zone is 3.2-3.7 W (109m3/yr). Summation of seismic moments released in the 100 years prior to 2011 indicate a seismic flux release rate of 1.7 W, only half the accumulation rate. The seismic flux of the 2011 earthquake was 1.3-1.6x1012m3. Paleoseismic studies (Minoura et al 2001) indicate that such great events occur about once every 1000 years, the penultimate one being the Jogan earthquake of 869 AD. Using that recurrence time for the 2011 earthquake increases the seismic flux release rate to 3.0-3.3 W, about the same as the accumulation rate obtained from the GPS studies. It appears that the seismic coupling along that sector has several summits that are frequently ruptured by smaller Miyagi-oki and Fukushima-oki type earthquake, and that very infrequently these peaks come into phase, yielding a much greater Tohoku-oki type earthquake.
Similar behavior occurs in the Hokkaido sector, where the geodetically determined seismic flux accumulation rate is 2.9 W, whereas the historic release rate is only 1.85 W. Paleoseismic data there (Sawai et al, 2009) indicate that much greater earthquakes, with approximate Mw=8.8, occur with an average 400 year recurrence time. If these earthquake are included in the seismic release rate calculation, it then agrees with the accumulation rate determined by GPS inversions. The last such great earthquake was about 350 years ago, so another should be expected in the not too distant future.
An unexpected feature of the 2011 Tohoku-oki earthquake was the large shallow surge that ruptured to the trench with slip as large as 50 meters. This large shallow surge, which was depleted in high frequency radiation, was largely responsible for generating the great tsunami from that earthquake. It had previously been assumed that such shallow slip could not occur in subduction zone earthquakes because of the presence of clay-rich velocity strengthening sediments in the fault zone within the accretionary prism. It has been proposed that a thermal pressurization mechanism could overcome the velocity strengthening to produce a stress-drop that drove the large shallow surge. However, normal faulting in the outer wedge during and after the earthquake indicates that the resistive shear stress accompanying slip must have been near zero. This precludes thermal pressurization as the weakening mechanism because a significant resistive shear stress is required to produce thermal pressurization.
The shallow parts of subduction zones such as this one have a marked contrast in elastic constants between the upper plate (slow) and the lower plate (fast). It is known that Mode II rupture on a bimaterial interface can occur by a wrinkle pulse mode in which there develops an opening between the rupture surfaces that allows slip to occur in a traction-free manner. Theory and experiment shows that such a wrinkle pulse can produce an instability in the face of velocity strengthening but that it requires a substantial nucleation pulse to initiate it. Such a wrinkle pulse can explain the large shallow surge and its depletion of high frequency radiation. The nucleation pulse threshold explains why only great subduction earthquakes can generate a shallow slip surge and that smaller events such as the Miyagi-oki earthquakes, cannot. It now appears that such shallow surges also accompanied the Mw 9.2 2004 Ache-Andaman and Mw 8.8 Maule, Chile earthquakes, so it appears that this is a common feature of the greatest subduction earthquakes.