Assessing submarine great-earthquake statistics fails

Geologists who study turbidites assume that the distinctive graded beds from which they are constructed and a range of other textures represent flows of slurry down unstable steep slopes when submarine sediment deposits are displaced. Such turbidity currents were famously recorded by the severing of 12 transatlantic telecommunication cables off Newfoundland in 1929. This happened soon after an earthquake triggered 100 km hr-1 flows down the continental slope, which swept some 600 km eastwards.

Load structures on turbidite sandstones, Crook...

Typical structures in Upper Carboniferous turbidites near Bude, Cornwall, UK (credit: Flickr, Earthwatcher)

Sea beds at destructive margins provide the right conditions for repeated turbidity currents and it is reasonable to suppose that patterns should emerge from the resulting turbidite beds that in some way record the seismic history of the area. British and Indonesian geoscientists set out to test that hypothesis at the now infamous plate margin off Sumatra that hosted the great Acheh Earthquake and tsunamis of 26 December 2004 to kill 250 thousand people around the rim of the Indian Ocean (Sumner, E.J. et al. 2013. Can turbidites be used to reconstruct a paleoearthquake record for the central Sumatra margin? Geology, v. 41, p.763-766).

Animation of 2004 Indonesia tsunami

Animation of Indonesian tsunami of 26 December 2004 (credit: Wikipedia)

Cores through turbidite sequences along a 500 km stretch of the margin formed the basis for this important attempt to test the possibility of recording long-term seismic statistics. To avoid false signals from turbidity currents stirred up by storms, floods and slope failure from rapid sediment build-up 17 sites were cored in deep water away from major terrestrial sediment supplies, which only flows triggered by major earthquakes would be likely to reach. To calibrate core depth to time involved a variety of radiometric  and stratigraphic methods

Disappointingly, few of the sites on the submarine slopes recorded turbidites that match events during the 150-year period of seismic records in the area, none being correlatable with the 2004 and 2005 great earthquakes. Indeed very little correlation of distinctive textures from site to site emerged from the study. Some sites on slopes revealed no turbidites at all from the last 150 years, whereas turbidites in others that could be accurately dated occurred when there were no large earthquakes. Only cores from the deep submarine trench consistently preserved near-surface turbidites that might record the 2004 and 2005 great earthquakes.

These are surprising as well as depressing results, but perhaps further coring will discover what kind of bathymetric features might yield useful and consistent seismic records from sediments.


3 responses to “Assessing submarine great-earthquake statistics fails

  1. Just came across this blog and thought I’d comment on it. I’m the lead PI on a marine paleoseismic project in Sumatra, and have been working on this since our cruise in 2007. We collected 109 piston, gravity and multicores cores along the margin in both trench and basin sites, and in short, our results are just the opposite of Sumner et al. The problems with the Sumner paper are too long to list, but here are a few of them: For turbidite paleoseismology, understanding of flow paths, source regions for failure, and depositional potential of the sites are critical. Without knowledge of these factors, results may be completely determined by site selection, which appears to be the case in this study. Cores 5MC, 8MC, and 12MC are in flat floored wide basins that have proven to be of limited use for paleoseismology in slope settings. Core 5MC is located >30 km from the mouth of a local canyon, and 20 km from subdued local slopes. These wall failures are not channelized, and thus the turbidity currents rapidly weaken as they spread across the basins (Patton et al., 2013). For example, in Hydrate Ridge Basin West (Cascadia) we found that a good record can fade in <2 km across the proximal basin floor (Johnson et al., 2004; Goldfinger et al., 2012). Core 4MC is located in a small basin in an area of complex topography, with subdued nearby slopes and no clear pathway for turbidity currents. Cores 7MC and 14MC are located in a slope basin ~1 km from the steep local slope and do have turbidite records.
    The authors cored at several segment boundaries (Simeulue and Batu areas; 7 out of 9 coring locations), which typically are characterized by low or no slip during earthquakes, plus complex structural and slip transfer mechanisms from one segment to another, and many small earthquakes occurring in between large megathrust earthquakes, which might disturb the record. Paleoseismologists avoid segment boundaries whenever possible. The authors assume that because they don’t see the 2004 turbidite, that great earthquakes do not always generate turbidity currents. Given that Patton et al., (2013) do see the 2004 turbidite, and that none of the Sumner cores were actually in the 2004 rupture zone, a more likely explanation is poor experimental design.

    So the bottom line is location, location, location. You can't just go anywhere, take a core, and assume that you can interpret a small number of poorly placed cores to mean much of anything. The small number of very short cores in the Sumner paper, with poorly located sites, tells a tale not of Sumatra earthquakes, but rather of poorly conceived planning and execution. In our work (one paper out, two more in the mill) we find abundant evidence of both the 2005 and 2004 earthquakes in numerous cores.

    Cheers, Chris

    Dr. Chris Goldfinger
    Director, Active Tectonics and Seafloor Mapping Laboratory
    College of Earth, Ocean and Atmospheric Sciences
    Oregon State University
    1+ 541 737 5214 fax 1+ 541 737 2064
    Earthquake Blog:
    mail: Ocean Admin Bldg 104, Corvallis OR 97331, USA

  2. Referring to my photo above, showing small-scale load structures in the Bude Formation, please note that you have got the age wrong. The Bude Formation is upper Carboniferous in age, not (as your caption states) Devonian. The age is clearly given in the description of the photo on my Flickr page.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s