Toward an Improved Understanding of Subduction Zone Mechanics Finite Element Modeling Results From the Jalisco GPS Project, 1993-2004

1
Toward an Improved Understandingof Subduction
Zone MechanicsFinite Element Modeling Results
Fromthe Jalisco GPS Project, 1993-2004
S. Schmitt, C. DeMets, J. Stock, O. Sanchez, T.
Masterlark, B. Marquez-Azua Past contributions
from W. Hutton, T. Melbourne, I.S.E. Carmichael,
K. Hudnut
2

• Postseismic subduction zone processes
• Relocking of rupture area
• Downdip propagation of slipafterslip
• Viscoelastic relaxation

3
Jalisco GPS network setting
4

• Modeling
• Finite element modeling with ABAQUS code
• Allows for realistic geometry and material
  properties
• Can model static and transient deformation

5
9 October 1995MW 8.0 6.5 - 7.5 m maximum
slip 22 January 2003MW 7.4 2.5 m maximum
slip
Major earthquakes
Geodetic and seismic solutions agree
6
COLI time series
7

• Interseismic motion
• Elastic model
• Completely locked interface
• at full plate convergence rate

8
Interseismic
Coseismic
Coseismic displacement Release of elastic strain
9

• Near term postseismic
• What are the transient processes?
• Afterslip and/or viscoelastic flow
• Viscoelastic model fails to predict
• direction and vertical motion
• Log-linear model fits better
• than log-only model
• Relocking

10
Very near-term postseismic Coseismic subsidence
subsequent uplift Migration of slip down-dip
11
Long-term postseismic Assume interseismic
velocity from pre-1995 Extrapolate decaying
afterslip rate from near-term postseismic Need
another process for westward biasviscoelasticity
12
Long-term postseismic
Viscoelastic prediction agrees well with
subsequent observation
13

• Summary
• Afterslip is needed to fit postseismic vertical
  motion, site direction, and decaying motion.
• Viscoelastic flow is necessary to predict the
  directions in the velocity field.
• Linear component implies relocking of the
  seismogenic zone soon after the earthquake.