Page 36 - Sumatra Geology, Resources and Tectonic Evolution

GRAVITY FIELD 23

part of the active margin of SE Asia lay in this area during the Late high seismic velocities, there is a strong circumstantial case for
Cretaceous and Palaeogene (e.g. Wilson & Moss 1999). From attributing high velocity in the lower mantle to lithospheric
southeastern Borneo the line of subduction then curved sharply material that has sunk to aseismic depths. The close correlation
to pass through western Java and on to Sumatra. Subducted litho- between high velocity in the lower mantle (Widiyantoro & van
sphere associated with this phase of convergence can be expected der Hilst 1996, 1997) and high gravity field provides additional
to have accumulated beneath Borneo and the Malacca Straits. support for this hypothesis.
Moreover, many theories of the evolution of Borneo require Tomography also provides an explanation for the absence of
there to have been subduction beneath its northwestern margin earthquake hypocentres at depths of more than 300 km beneath
during the Late Cretaceous and Palaeogene, leading to the com- Sumatra. There is no high-velocity material at these depths
plete destruction of a 'proto-South China Sea' and collision (Widiyantoro & van der Hilst 1996) and hence, presumably, no
between the Borneo block and attenuated continental crust rifted subducted slab. Taken together with the interpreted presence of
from the South China margin (e.g. Milsom et al. 1997). The a large volume of dense and fast material below 700 kin, this
extent of the long-wavelength gravity high suggests that it may observation supports hypotheses that involve the rupturing of
be recording effects from material subducted beneath Borneo slabs and the independent sinking of their detached lower portions
from the south, east and west (Milsom & Rocchi 1998). under gravity. Even stronger support comes from farther east,
In northwestern Sumatra, the margin of the long-wavelength north of Java, where the upper part of the detached slab protrudes
high curves to an almost northerly trend and peak values decrease above the 700 km limit and is both seismically 'fast' and seismo-
quite rapidly, suggesting that there is no significant deep sub- genic (Widiyantoro & van der Hilst 1996).
ducted material beneath the Andaman Sea. This seems reasonable The Sumatra region also conforms to the global pattern of lack
since, although the plate boundary west of the Andaman and of correlation between high gravity and subducted lithosphere
Nicobar islands is marked by a (rather poorly defined) trench, within the seismogenic zone, i.e. at relatively shallow depths.
the local convergence vector is almost parallel to the trench axis. Hagar (1984), amongst others, has used this global observation
Further light on the sources of the long wavelength gravity to support a model of dynamic flow that produces, at GEM-T3
anomalies has been provided by the improvements in, and wavelengths, close to perfect cancellation between the effects of
standardization of, seismic observatory instrumentation and the positive and negative density anomalies in the upper mantle.
dramatic increases in speed and memory of relatively cheap Some doubt has, however, been thrown on this model by
computers. Thanks to these two developments it is now possible Wheeler & White (2002), who used oil-industry borehole data to
to use observations of travel times for S and P waves from argue that, at least in offshore SE Asia, dynamic topography
remote earthquakes to model the variations of seismic wave amounts to no more than 300 m. Predictable improvements in
velocities in the mantle. This seismic tomography is providing data quality will undoubtedly lead to considerable refinements
ever stronger evidence for the penetration of subducted lithosphere in interpretation and resolution of this apparent discrepancy,
through the discontinuity between the upper and lower mantle at but it is sufficient to note that as far as the present review is con-
about 700 kin, below which it is not seismogenic. Because cerned, the GEM T-3 gravity field provides an excellent guide to
Wadati-Benioff seismic zones marking the sites of subducted the extent of Palaeogene, but not Neogene, subduction beneath
lithosphere in the upper mantle are invariably associated with Sundaland.