Why Do Slow Earthquakes Occur Favorably in Hot Subduction Zones? : 2D Numerical Analysis

Abstract:

It is puzzling why slow earthquakes occur in hot subduction zones only. We study numerically how antigorite dehydration coupled with slip-induced dilatancy and thermal pressurization affects rupture behavior to solve the above puzzle. Recent related studies actually suggest importance of antigorite dehydration. We assume faulting in a 2D thermoporoelastic medium. The mineral reaction is assumed using a first order Arrhenius law. Nondimensional parameters important in the modeling are Su, Su’ and X’ according to our former study; Su’ and Su are proportional to permeability and increase rate of slip-induced porosity. X’ denotes volume change induced by the reaction. Our calculation shows that moment release rate and fault tip growth rate are smaller for larger values of Su, smaller values of X’ or smaller values of Su’. These two rates are found to be negligibly small compared with the solutions for the dynamic elasticity analysis when Su>1 and X’<0 are satisfied. Slow sustained fault growth occurs for such values of Su and X’. This suggests that Su>1 and X’<0 are satisfied at hot subduction zones; the condition X’<0 is consistent with the reaction expected at hot subduction zones. In cold subduction zones however, antigorite dehydration will occur at depth greater than 60km, with -0.1<X’<0. Our calculation shows that fault growth rate accelerates to the shear wave speed soon after the rupture nucleation for any values of X’ when Su<1 is assumed. However, the moment release rates for negative values of X’ are found to be a few orders smaller than expected by the dynamic elasticity analysis. This occurs because the slip velocity evolution is more pulse-like for larger values of |X’|. In such way, antigorite dehydration could also be the source of a non-negligible fraction of intermediate depth seismicity. The condition Su<1 will be due to the fact that slip-dilatancy is more prohibited at larger depth.