Constraining interseismic deformation processes in subduction zones through simple mechanical models

Inter-seismic slip deficit and the implied seismic hazard are typically determined by inverting surface deformation for a distribution of back-slip on the megathrust plate interface, often using analytical Green's functions computed for faults embedded in an elastic half-space. In this framework, regions with high back-slip are interpreted to have little relative motion and regions of low back-slip are interpreted as freely sliding. These contrary assumptions are fundamental to the back-slip approach and introduce artifacts in the quantitative analysis and/or interpretations of the results. Using a suite of subduction zone finite element models (FEMs) with boundary conditions chosen to more closely represent the frictional state of the interface, we assess the validity of back-slip techniques and reexamine interpretations of frictional characteristics, strain accumulation, and seismic hazard. These FEMs consist of a simple homogeneous elastic subduction zone to isolate the effects of the interface boundary conditions and compare with analytical solutions. On the megathrust is a locked (100% coupling) asperity spanning the seismogenic zone (10-50 km depth), surrounded by a region that can slide freely (0% coupling). Back-slip inversions of the surface deformation in these models using analytical Green's functions appear to be consistent with the distribution of slip deficit in the FEM; however, the slip deficit does not have a 1:1 correspondence with the asperity. The asperity is surrounded by a rim of significant slip deficit (on freely sliding sections of the interface) that decays into the full plate motion slip deficit of the surrounding region. This slip deficit transitional rim is a consequence of the continuity of the plates and their elasticity. In particular, although the region up-dip of the 10-km deep upper limit of the asperity may be frictionally free to slide, this section still accumulates 75-100% slip deficit because of its proximity to the coupled section. The degree to which inversions based on analytical solutions can identify these regions of slip deficit may vary depending on the state of the subduction zone, so we also explore variations in the asperity configuration (size, location, and number of locked patches) to test the sensitivity of the system to geometrical parameters.