Collaborative Research: Mechanisms of Transient Deformation Following Great 2006-2007 Kuril Earthquakes: Frictional Afterslip or Viscoelastic Relaxation?

DESCRIPTION: The earthquake cycle in subduction zones includes strain accumulation caused by subduction, sudden strain release by the earthquake rupture, and postseismic time-dependent deformation. GPS observations of postseismic transients after large subduction earthquakes provide a natural experiment to study rheology of the megathrust and of the sublithospheric mantle. Scientists still disagree on what mechanism prevails in the postseismic transients: frictional afterslip on the coseismic rupture, viscoelastic relaxation in the asthenosphere, or poroelastic rebound. In 2006-2007, a doublet of great earthquakes (Mw > 8) struck in the center of the Kuril subduction zone at the northwest Pacific Ocean: a thrust event at the subduction interface followed by an extensional event beneath the oceanward flank of the Kuril trench. It is probably the greatest doublet that has been observed in the era of satellite geodesy and broadband seismology. In 2006, the Kuril GPS Array was installed, which allowed to measure all components of the earthquake cycle. The data collected in 2006-2009 before and after the 2006-2007 earthquakes outlined a broad zone of postseismic deformation with initial horizontal velocities as fast as 100 mm/a, and a regional uplift. Most of the postseismic signal after the great Kuril doublet is caused by the viscoelastic relaxation of shear stresses in the weak asthenosphere. Viscoelastic relaxation was discriminated from other candidate mechanisms by the pattern of horizontal and vertical motions. We predict that the postseismic deformation will die out in about a decade after the earthquake doublet, so the preferred mechanism can be tested with observations continued for several years. Our initial results suggest large variations among subduction zones in the upper mantle viscosity, one of the most important parameters that governs the stress distribution.