Geodesy and Gravity/Tectonophysics
The Sumatra subduction zone: A case for a locked fault zone extending into the mantle
Article first published online: 2 OCT 2004
Copyright 2004 by the American Geophysical Union.
Journal of Geophysical Research: Solid Earth (1978–2012)
Volume 109, Issue B10, October 2004
How to Cite
2004), The Sumatra subduction zone: A case for a locked fault zone extending into the mantle, J. Geophys. Res., 109, B10402, doi:10.1029/2003JB002958., , , and (
- Issue published online: 2 OCT 2004
- Article first published online: 2 OCT 2004
- Manuscript Accepted: 26 JUL 2004
- Manuscript Revised: 28 JUN 2004
- Manuscript Received: 23 DEC 2003
- locked fault zone;
- interseismic deformation;
 A current view is that the portion of the subduction interface that remains locked in the time interval between large interplate earthquakes, hereinafter referred to as the locked fault zone (LFZ), does not extend into the mantle because serpentinization of the mantle wedge would favor stable aseismic sliding. Here, we test this view in the case of the Sumatra subduction zone where the downdip end of the LFZ can be well constrained from the pattern and rate of uplift deduced from coral growth and from GPS measurements of horizontal deformation. These geodetic data are modeled from a creeping dislocation embedded in an elastic half-space and indicate that the LFZ extends 132 ± 10/7 km from the trench, to a depth between 35 and 57 km. By combining this information with the geometry of the plate interface as constrained from two-dimensional gravimetric modeling and seismicity, we show that the LFZ extends below the forearc Moho, which is estimated to lie at a depth of ∼30 km, at a horizontal distance of 110 km from the trench. So, in this particular island arc setting, the LFZ most probably extends into the mantle, implying that either the mantle is not serpentinized, or that the presence of serpentine does not necessarily imply stable sliding. From thermal modeling, the temperature at the downdip end of the LFZ is estimated to be 260 ± 100°C. This temperature seems too low for thermally activated ductile flow, so that aseismic slip is most probably due to pressure and/or temperature induced steady state brittle sliding, possibly favored by fluids released from the subducting slab.