The Great Sumatran Earthquake took place on 26th December 2004. One month into the
aftershock sequence, a dense swarm of earthquakes took place beneath the Andaman
Sea, northeast of the Nicobar Islands. The swarm continued for ∼11 days, rapidly
decreasing in intensity towards the end of that period. Unlike most earthquake swarms,
the Nicobar cluster was characterised by a large number of shocks with moment
magnitude exceeding five. This meant that centroid moment tensor data could be
determined, and this data in turn allows geometric analysis of inferred fault plane
motions.
The classification obtained using program eQuakes shows aftershocks falling into distinct
spatial groups. Thrusts dominate in the south (in the Sumatran domain), and normal faults
dominate in the north (in the Andaman domain). Strike-slip faults are more evenly spread.
They occur on the Sumatran wrench system, for example, but also on the Indian plate itself.
Orientation groups readily emerge from such an analysis. Temporal variation in behaviour is
immediately evident, changing after ∼12 months. Orientation groups in the first twelve
months are consistent with margin perpendicular extension beneath the Andaman Sea
(i.e. mode II megathrust behaviour) whereas afterward the pattern of deformation appears to
have reverted to that expected in consequence of relative plate motion. In the first
twelve months, strike-slip motion appears to have taken place on faults that are
sub-parallel to spreading segments in the Andaman Sea. By early 2006 however normal
fault clusters formed that showed ∼N-S extension across these spreading segments
had resumed, while the overall density of aftershocks in the Andaman segment
had considerably diminished. Throughout this entire period the Sumatran segment
exhibited aftershock sequences consistent with ongoing Mode I megathrust behaviour.
The Nicobar Swarm marks the transition from one sort of slab dynamics to the
other.
The earthquake swarm may have been facilitated by hydrothermal activity related to a
seamount, or by magma intrusion. However, the swarm is located where the transpressional
regime of the Sumatran strike-slip fault system changes to that of the ‘microplate-bounding’
transtensional wrench involved in the Andaman Sea spreading centre. The swarm thus may
be the result of the confluence of two tectonic modes of afterslip on the main rupture, with
arc-normal compression to the south, and arc-normal extension to the north. The orientations
of the controlling faults can be related to the right-lateral Sumatran strike-slip system, and to
oceanic transforms in the spreading system. Faults parallel to the Andaman Sea spreading
system axis reactivated as left-lateral strike-slip faults during the period of afterslip.
Analysis of the orientation groups shows that the swarm involved synchronous but
geometrically incompatible movements on opposing but conjugate fault plane sets with
trends that are consistent with Mohr-Coulomb failure, even though the orientation
groups delineated require slip in many different directions on these planes. The
fault planes allow inference of regional deviatoric stress axes with the principal
compressive stress parallel to the prior distortion inferred using satellite geodesy. |