144 deep sea sediment cores were collected along the entire length of the Sumatra margin in order to interpret the strata as they relate to seismogenic triggering of submarine landslides. Cores were collected in the trench and in lower slope piggyback basins of the accretionary prism. Based on correlation of at least 19 Holocene turbidites, the ¹⁴C age based Recurrence Interval (RI) estimate for large to great earthquakes in the 2004 rupture region for the last 7.5±0.06 ka is 330±60 years, consistent with the nascent terrestrial paleoseismic record in Thailand, the Andaman Islands, India and northern Sumatra. The preliminary work from this project indicate that the technique of submarine paleoseismology, well developed in Cascadia, is applicable to the Sumatran continental margin.  A primary test of the methods is provided by the search for a turbidite related to the 2004 earthquake.  One particularly exciting result is that the radiocarbon age determination for the sediment underlying the uppermost turbidite in core RR0705-96PC is consistent with our interpretation of this deposit being the result of the 2004 Sumatra-Andaman subduction zone (SASZ) earthquake (30+-30 cal yr BP), a result consistent with the Pb210 analyses, and the lack of overlying hemipelagic sediment. Data from cores RR0705-90MC and RR0705-102MC have excess Pb210, and model curves suggest an essentially zero age for these samples at the seabed.  Although these result lack precision, they are consistent with a very young turbidite.  This event forms the groundtruth for the rest of the study.  Investigations of this type in Cascadia and along the Northern San Andreas Fault (NSAF) have demonstrated the techniques are capable of revealing long-term seismic histories. Recent work of others in other locations in New Zealand, Japan, Chile, and in European lakes is also proving successful in similar approaches.

The Sumatran margin presents numerous difficulties including extremely deep water, lower sedimentation rates, an unfilled trench with axial transport, and a more spatially variable earthquake record than present in Cascadia. Yet our early results suggest that it is possible to develop a good record of past earthquakes for the region. We find that it is possible to mitigate the shortcomings and take advantage of other aspects of the Sumatran margin to develop a method that is region specific for Sumatra.  Results from this project suggest that the return period of large earthquakes is variable, averaging from 250-400 years. 

In light of the Tohoku earthquake, models of earthquake recurrence based on short seismological records must now be considered to have failed. Long records from paleoseismology offer a new avenue to understanding of long term plate boundary behavior, as well as the simple case of understanding the maximum credible earthquake a particular fault is capable of producing. In the absence of global models, which may prove elusive, or perhaps may never exist, seismic hazards may be assessed reliably using long term recurrence histories.  The basic information from this project will also provide information valuable for tsunami modeling and for societal planning for future events.

We constructed estimates of static slope stability Factor of Safety (FOS). and used ground motion prediction equations to estimate ground motion intensity  and acceleration as possibly generated by fault rupture, to constrain seismic loads.  All slope and trench sites are statically stable and sensitive to ground motions generated by earthquakes of magnitude greater than 7. We conclude that for earthquakes of magnitude 6 to 9, PGA of 0.4-0.6 to 1.4-2.5 g would be expected from existing GMPE’s.  However, saturation of accelerations in the accretionary wedge may limit actual acc...

Please report errors in award information by writing to: awardsearch@nsf.gov.