The Mw 9 March 11, 2011 Tohoku earthquake, resulting in 25,000 fatalities, was the largest earthquake to strike Japan in modern times. The failure of long-term forecasts to anticipate an event of this magnitude is leading to a reevaluation of our scienti c understanding of subduction zone fault dynamics. At the same time, the enormous Japanese investment in seismic and geodetic monitoring provides an unprecedented opportunity to study processes leading to giant earthquakes. The investigators will focus here on two aspects of the quake that raise central issues for our ability to forecast future hazards globally: 1) the constraints that geodetic strain accumulation place on the seismic moment budget, and 2) how the distribution of mechanical properties on the plate interface control seismic versus aseismic slip.
Forecasts of seismic hazard rely primarily on the historical record, paleoseismology, and geodetic strain accumulation. While a strict time-predictable model has proven inconsistent with geodetic data in some areas, the rate of strain accumulation must strongly in uence the recurrence rate and/or size of large earthquakes. Previous analyses of the extensive Japanese GPS data set regularize the underdetermined inverse problem; the inferred plate coupling is thus conditional on the choice of regularizing functional. In addition, most assumed no coupling at the trench, an area that effectively lies in the model null-space. The researchers propose rather to estimate rigorous bounds on the moment deficit rate, independent of any regularizing functional. They also propose to systematically explore the effects of time-dependent viscoelastic effects on estimates of slip deficit.
The Tohoku earthquake also raises profound questions about the frictional properties of the subducting plate interface. Mw > 7 events along the Japan-Kurile trench have repeated in nearly the same locations. Substantial afterslip following these events suggested that the fault surrounding these seismic `asperities' have different frictional properties, e.g., velocity weakening `asperities' surrounded by velocity strengthening regions. The fact that the March 11 earthquake seems to have ruptured through both areas raises important questions. Is the asperity paradigm fundamentally flawed? Or did the M 9 activate other weakening mechanisms (thermal pressurization?) that allowed it to rupture through velocity strengthening regions?
To distinguish between competing hypotheses the researchers will use a combination of GPS data based hypothesis testing and physics based fault modeling. They will test whether or not post-seismic GPS data can be fit with afterslip only in regions that have not experienced seismic slip in the last several decades. They will test alternative physical models that could account for the observed behavior, by extending our numerical codes that couple rate-state friction, thermal pressurization, and dilatant strengthening to three dimensions. This will allow them to test for example whether dilatancy could stabilize slip in otherwise velocity weakening regions, or thermal pressurization could allow ruptures to extend into velocity strengthening areas.
