1
A 7,500 year earthquake history in the region of the 2004 Sumatra-Andaman EQ
2
Jason R. Patton1, Chris Goldfinger1, Ann Morey1, Ken Ikehara2, Chris Romsos1, Yusuf
3
Djadjadihardja3, and Udrekh Udrekh3
4
1College
5
National Institute of Advanced Industrial Science and Technology. 3Bandan Penghajian Dan Penerapan
6
Teknologi BPPT 2nd Building, 19th Floor, Jl.MH. Thamrin 8, Jakarta, 10340 Indonesia.
7
Abstract
8
144 sediment cores were collected along the entire length of the Sumatra margin in order to interpret the strata as
9
they relate to seismogenic triggering of submarine landslides. Cores were collected in the trench and in lower
of Oceanic and Atmospheric Sciences, Oregon State University, Corvallis, OR 97331 USA. 2
10
slope piggyback basins of the accretionary prism. Based on correlation of at least 19 turbidites, the 14C age based
11
Recurrence Interval (RI) estimate for large to great earthquakes in the 2004 rupture region for the last 7.5±0.06
12
ka is 330±60 years, consistent with the nascent terrestrial paleoseismic record. 210Pb age control is used to
13
evaluate the timing of the most recent sedimentary deposits, most likely generated by the 2004 earthquake. We
14
find that the seismic shaking trigger interpretation may be related to seismic moment release and relative
15
amplitude.
16
17
Introduction
18
Earthquakes and tsunamis are some of the most deadly natural disasters, with the 26 December 2004 Sumatra-
19
Andaman earthquake and tsunami responsible for the deaths of nearly a quarter of a million people. Knowledge
20
of the earthquake cycle, through many cycles, is fundamental to a deeper understanding of the seismogenic
21
process and seismic hazard. GPS data yield detailed short-term strain information during one or two strain
22
cycles, yet in many regions even a single cycle may span centuries. To evaluate earthquake/tsunami risk and
23
better understand the fundamental tectonic processes involved, a long record of earthquake cycles in concert with
24
the latest technology and modern monitoring efforts is necessary. Paleoseismology is one method that garners
25
information about faults as they behave through many cycles.
26
Historic Sumatra-Andaman subduction zone (SASZ) earthquakes in this region 1 (1847, 1881, 1941) were much
27
smaller (magnitude < 8.0) than the 2004 earthquake (Fig. 1). However, recent investigation of secondary
28
evidence left behind by tsunami as sand sheets in northern Sumatra 2, western Thailand3, and the Andaman-
29
Nicobar Islands4, along with coral microatoll evidence5 suggests the penultimate large earthquake most likely
30
occurred 500-700 years ago, and an ante-penultimate earthquake earthquake/tsunami in Sumatra2,5, the A-N
31
Isles4, and India6 likely occurred ~ 900-1,200 years ago.
32
Strong ground shaking from rupture of earthquakes has been inferred to trigger turbidity currents that leave a
33
potentially very long record of past earthquakes7,8,9,10,11,12,13,14,15,16,17. While there are not unequivocal global,
34
regional, or local criteria to distinguish between turbidite generation processes, the combined evidence from
35
sedimentology, tests of synchroneity, stratigraphic correlation, and analysis of non-earthquake triggers can be
36
used to develop a reliable earthquake record for submarine fault zones in some cases 5,6,10.
37
This paper describes initial results from a paleoseismic investigation conducted offshore Sumatra, in the region
38
of the 26 December 2004 Mw 9.1-9.3 earthquake18,19,20,21. Marine turbidite stratigraphy is first examined and
39
tested for suitability for paleoseismology, then interpreted to estimate earthquake recurrence for the last ~7,500
40
years.
41
Seismogenic Trigger Rationale and Physiography
42
Adams7 and Goldfinger8,9,10,12 suggest eight plausible triggering mechanisms for turbidity currents: 1) storm wave
43
loads, 2) earthquake, 3) tsunami wave loads (local or distant), 4) sediment loads, 5) hyperpycnal flows, 6)
44
volcanic explosions, 7) submarine landslides, and 8) bolide impacts. Other mechanisms may reduce slope
45
stability, but are likely random and not regional nor synchronous11,22. The basis for determining the origin of
46
turbidity currents is that regional and synchronous deposition is unlikely to have been generated by a trigger
47
other than an earthquake9,10,23.
48
We test the plausibility of a seismogenic source in northern Sumatra by 1) using tests for synchronous triggering
49
of sedimentologically isolated turbidite systems and 2) using secondary constraints that consider sedimentologic
50
characteristics of the turbidites. When turbidites can be correlated between sites separated by a large distance or
51
between sites isolated from land sources and from each other, synchronous triggering can be inferred and most
52
other triggering mechanisms can be eliminated7,8,9,10,12,24,25,26.
53
As in all paleoseismic investigations, selecting the correct sites is critical. We use continental margin
54
physiography to narrow the selection of sites to those most likely to preserve earthquake generated turbidites at
55
the expense of other possible sources. Sumatra presents some physiographic difficulties in this respect, and some
56
advantages, in comparison to Cascadia where significant work has been conducted10. Continental margin
57
morphology in western Sumatra is dominated by the upper plate structure of a Tertiary and Quaternary
58
accretionary prism with structural highs and forearc basins. Fold and thrust belt topography forms longitudinally
59
linked basins that can be either isolated, or drain to the trench. Canyon systems tend to be short and drainage
60
catchments are relatively small, limiting the areal extent of source areas for turbidity currents 27. This is a
61
significant disadvantage as many sites are more proximal to small earthquakes and local random landslides.
62
However, the outer forearc is also isolated from northern Sumatran sediment sources by the broad, unfilled Aceh
63
forearc basin, though limited input from the offshore islands of Simeulue, Nias, and Siberut Island is possible for
64
some nearby basins. Similarly, the southward-deepening trench is likely blocked from input from northern
65
sources by the subducting Ninety-East Ridge, and a large landslide at 14º N28. The trench deepens from 4.5 km.
66
to 6.5 km from north to south, and is filled with sediment several km thick in the north from the Nicobar fan,
67
partially burying lower plate bending moment normal faults and fracture zones that trend across the trench. This
68
morphology controls turbidite channel flow within the trench.
69
The Sumatran slope and trench physiography results in trench segments that are isolated from each other, and
70
from isolated slope basins in part, and therefore offers numerous opportunities to compare the stratigraphic
71
sequences in sites that have unique sediment sources. The basins and trench segments are fed by small canyons
72
that have limited drainage areas on the slope and form small fan/aprons. Slope sites used in this study include
73
cores 108, 104, 103, and 96, which all have isolated sediment sources. In the trench, cores 107, 105, 98, and 94
74
receive sediment both from upstream in the trench and downslope transport from the continental slope. Fig. 2
75
shows the source and flow relations for these cores. Correlation of strata is supported by 14C ages of individual
76
turbidites and hemipelagic intervals between earthquakes (radiocarbon methods are in Table S1).
77
Stratigraphic Correlation
78
We use integrated stratigraphic correlation techniques, including lithostratigraphic description (color, texture,
79
and structure, etc.), visual analysis, Computed Tomography (CT) image analysis, and core log “wiggle
80
matching” of Multi Sensor Core Logger (MSCL) geophysical data. The most sensitive criteria for correlating
81
fine grained turbidites (which may not be visible to the naked eye) is the density profile, which we augment with
82
very high resolution CT density profiles (Inouchi11) and 3D imagery.
83
The primary goal of correlating turbidite strata is to establish spatial extent of individual turbidites that are
84
carefully “fingerprinted” with geophysical data. Geophysical wiggle matching (fingerprinting23) of turbidites is
85
based on the correlation of identifiable unique stratigraphic characteristics using MSCL core log data: gamma
86
density and magnetic susceptibility (ms). These “fingerprints” represent the time-history of deposition of the
87
turbidite, and have been shown to correlate between independent sites separated by large distances and
88
depositional settings8,9,10,29. The turbidite itself is commonly composed of multiple coarse fraction pulses,
89
probably deposited over minutes to hours which form much of the basis of the turbidite “fingerprint.” The fact
90
that these cores, in sedimentologically isolated and hydrodynamically unique systems, share turbidites with
91
depositional histories that match in considerable detail, suggests that they also share a common trigger
92
mechanism. Goldfinger and others9,10,23 have proposed this shared characteristic is, in part, the time history of
93
ground shaking during the earthquake.
94
Northern Sumatran Core Stratigraphy
95
Lithostratigraphy in northern Sumatran slope cores is dominated by turbidites interbedded with massive
96
hemipelagic mud and less common tephras. Bioturbation is common and core-induced deformation is observed
97
in some cores. Turbidites are composed of coarse silt to coarse sand bases, with fining upward sand and silt to
98
clay sub-units. The coarse fraction is composed of mica and quartz grains with rare mafics. Some basal turbidite
99
sub-units are foraminiferal hash. Sand sub-units commonly range in thickness from 0.5 to ~20 cm and are
100
laminated and cross bedded, commonly underlying massive sand units. Finer material is composed of silt to clay
101
sized particles. 0.5- to 10.5-cm thick primary tephras are rare and can be correlated between sites using electron
102
microprobe and laser ablation ICPMS data31.
103
Trench turbidite lithology is dominated by fine mature quartz sand and mica, consistent with the well-known
104
Himalayan source of the accreting Bengal and Nicobar fans32. At 2- to 4-km depth, slope basins contained
105
similar lithology, which we interpret to have been accreted and recycled into the slope basin stratigraphy, with
106
the addition of abundant forams.
107
Sumatra turbidite bases are sharp and generally fine upwards, with some deposits composed of multiple upwards
108
fining sub-units, similar to those observed in Cascadia10. The coarser Sumatra turbidites commonly are
109
amalgamated and this results in incomplete or irregular structural sequences, suggesting multiple turbidity
110
current pulses and a longitudinally heterogeneous source turbidity current.
111
Fig. 2 shows core data for cores 108, 105, 104, 103, and 96 (all cores are in Fig. S2). The light-grey sand bases of
112
turbidites are easily identified and MSCL maxima correlate well with the CT density maxima. CT data permit a
113
refined view of the strata and the effect of core disturbance, while gamma and magnetic data reflect signals that
114
average these effects. Similar turbidite stratigraphy is found in all slope basins (cores 108, 104, 103, and 96) and
115
trench sites (cores 107, 105, 98, and 94), spanning 350 km along strike.
116
We interpret that the uppermost turbidite is most likely the 2004 turbidite (See Fig. S3). The lack of hemipelagic
117
sediment overlying this turbidite, the lack of consolidation when compared to older strata, its great thickness
118
(over 3m) at one site, and excess 210Pb activity of the uppermost material are consistent with this interpretation.
119
Where multiple 14C ages exist for correlated turbidites, we test whether they agree with our stratigraphic
120
correlation model by using the ‘combine’ function in OxCal software 41. Combined ages for turbidites 5, 8, 9, 11,
121
and 14, are computed with the combination of either two or three ages (Table S4). Combinations are
122
substantiated by three measures of success: chi-squared, agreement index "Acomb" (>90), and convergence
123
integral "C" (>95)41(Table S4). Because the combined ages satisfy these measures they support our stratigraphic
124
correlations. Lithostratigraphy, radiocarbon ages and geophysical fingerprinting suggest a good correlation of
125
turbidites T1 through T19 between cores 108, 105, 104, 103, through T12 in core 96, and through T7 in 107;
126
spanning the 2004 rupture zone (Figs. 2 & 3). Less well correlated are turbidites 20 through 24, in cores 108 and
127
103. While most ages are consistent with our correlations, turbidite 10 in core 108PC is older than in other cores
128
most likely due to the basal erosion during deposition of the turbidite (evident from CT data). Likewise, turbidite
129
19 in core 108 is 600 years too young, which may be due to variations in marine reservoir corrections that are not
130
yet resolvable.
131
The stratigraphic sequences with the most unique “fingerprints” carry the strongest correlative weight and act as
132
"anchor points" for our correlations (Fig. S5). We note that some turbidites are absent across some intervals in
133
the 2004 rupture region. For example, T2 to T6 are not present in core 108 (Figs. 2 & 3). Likewise, T1 to T4
134
appear to be absent in core 104. We interpret this as the result of our site physiography, which is more proximal
135
than ideal in most cases, and would be expected to have greater variability as has been observed in Cascadia9,10.
136
As the sequence correlates well overall, local variability due to 1) basal erosion, 2) heterogeneous source areas
137
within the region, and 3) site geomorphology would be expected, regardless of the triggering mechanism.
138
We observe some sites with persistent local depositional variations that may be due to local geomorphologic
139
conditions. Core 104 records turbidites (T6, T10, T13, T14, and T17) that we correlate to other sites, but which
140
also have many thin coarse pulses as part of their structure. The core location for 104 is at the base of a steep
141
slope and canyon mouth with 1.5 km of relief (Fig. S6), suggesting possible local retrogressive failures. Another
142
example is Core 96, which has finer grained turbidites overall. Core 96 is located in a closed slope-basin fed by
143
very low relief terrain that does not form large channels, possibly explaining the fine-grained multi-pulse
144
structures. Core 103 is in a location that has a low gradient, wide floodplain channel directly upslope, explaining
145
the large variability of turbidite structure. Comparing two trench cores that have somewhat different turbidite
146
records, core 98 is located in a location insulated from high relief upslope sources due to a landward vergent fold
147
that deflects higher energy channelized flow further to the southeast (Fig. S6). This isolation may explain the
148
distal nature of the turbidite structures in core 98. In contrast, core 94 is directly downslope from an active
149
canyon system and has more structurally varying turbidites.
150
Discussion
151
The correlated framework shown in Figs. 2 & 3 represents a depositional history of turbidites spanning 7,500
152
years, with potential for a longer record with further work. But what triggered them? Detailed depositional
153
structure of turbidites is likely a combination of trigger source (e.g. hyperpycnal vs. seismogenic), source
154
proximity (e.g. proximal vs. distal), and flow dynamics (e.g. site or flow-path geomorphology). These factors
155
may compete depending upon core location and local physiography. Slope cores are proximal to the source and
156
this proximity likely dominates the structure and contributes to high variability in sedimentary structures. Trench
157
cores are located in varying proximity to the source, thus some trench cores appear to reflect a more distal setting
158
(98PC Fig. S2). Trench cores in this region also may record turbidity currents from sources further north along
159
the margin in Indian waters which were inaccessible to this project.
160
While the core sites were more proximal than ideal, the coherence of the turbidite “fingerprints” and radiocarbon
161
ages between isolated basin and trench sites over ~400 km along strike suggests that many or most of the
162
correlated turbidites have a common trigger.
163
geologically synchronous would be very large storms. Hyperpycnites are reported to initially coarsen upwards
164
and then fine upwards, representing the waxing and then waning of the hyperpycnal flow 15,33. Amalgamation is
165
an indicator that the turbidites were deposited from multiple turbidity currents, merging and forming a
166
longitudinal structure reflected in the deposit28, possibly from multiple slides on slopes. In Lake Biwa, Nakajima
167
and Kanai28 conclude these multiple pulses are the results of synchronous triggering of multiple parts of the
168
canyon system. Cascadia margin cores contain deposits that pass multiple tests of synchronous earthquake origin,
169
and
170
turbidites9,10. Goldfinger et al.10 correlate the number of coarse units to the rupture length of the causative
171
earthquakes. Sumatra turbidite bases are sharp and generally fine upwards, with some deposits composed of
172
multiple upwards fining sub-units, similar to Cascadia10. Coarser Sumatra turbidites commonly are amalgamated
173
and this results in incomplete or irregular structure sequences, suggesting multiple turbidity current pulses and a
174
longitudinally heterogeneous turbidity current. Thus we find no support for a hyperpycnal origin for the
175
turbidites investigated in this study.
176
The good stratigraphic correlation between sites isolated from each other, land sediment sources, and from other
177
triggering mechanisms, coupled with compatible radiocarbon ages suggest that the most likely triggering
178
mechanism is regional earthquakes. Uncorrelated events present at some sites may be random sediment failures
179
or smaller local earthquakes. We note that the lack of cyclones in equatorial waters all but rules out massive
180
regional storms as a sediment source. The isolation of our sites from Sumatra also prevents terrestrial input to
generally
contain
multiple
coarse
One of the few potential triggers that may be both regional and
fining
upward
sub-units,
consistent
with
other
seismo-
181
the outer forearc slope. Were cyclones present in the past, such storms would not likely trigger turbidity currents
182
on the forearc slope with minimum depths relevant to this study of ~1,000 m.
183
A 7500 year earthquake record
184
Based on 14C ages and the oldest turbidite in core 108PC, the average recurrence interval for earthquakes in the
185
region of the 2004 earthquake is 330 ± 60 years BP. Given a fault normal convergence rate of 34-37 mm a-1 in
186
this region21, earthquakes like the 2004 earthquake (slip of ~20 m19) would yield an average recurrence of 530 -
187
590 years. This estimate is close to our estimate, particularly given that we know little of the magnitudes of past
188
earthquakes, nor of the coupling ratio along the SASZ (e.g. for earthquake slip < 20m, recurrence would
189
decrease). This is consistent with the RI estimate from Chlieh et al. 19 of 140-420 yrs. Given a RI of ~330 years
190
for events large enough to generate observable turbidites, and that smaller earthquakes may be unrecorded, our
191
estimate is probably a maximum. If turbidite thickness is consistent along strike and relates to earthquake
192
magnitude10, thickness may be used to distinguish between earthquakes of different magnitude. Given the
193
variation and frequency of larger turbidites, recurrence of earthquakes similar to the 2004 events may be
194
approximately 3ka.
195
Based on paleotsunami evidence in Thailand, the last tsunamigenic earthquake recorded onshore in Phra Thong
196
was 500 - to 700 - years ago3,34 (Fig. 3). Based on uplifted marine abrasion platforms, penultimate and ante-
197
penultimate earthquakes occurred ~600 years ago and ~1,000 years ago respectively in the Andaman-Nicobar
198
Islands4. Similarly, paleotsunami evidence in Aceh, Sumatra, suggests the penultimate large tsunami = 600 - 700
199
years ago and the ante-penultimate EQ = 1,000 – 1,200 years ago2. Based on coral microatoll paleoseismology5
200
earthquake deformation occurred around 510 and 560 years ago. These estimates are all consistent with our
201
turbidite based evidence which suggests the ages of the previous events were ~600 and 1,000 years ago.
202
Paleotsunami, microatoll, and uplifted abrasion platform evidence also may not record all earthquakes and thus
203
may represent maximum intervals for recurrence of great earthquakes sensitive to the recording thresholds of the
204
different methods.
205
The temporal distribution of these earthquakes shows apparent clustering, with four gaps of 500- to 700-years;
206
ending at ~ 6.5, 3.8, 2.7, and 1.4 ka. We also note that cluster timing based on the paleoearthquake record of the
207
Padang segment35 is consistent with the four youngest turbidite ages from this study, but inconsistent with
208
earthquakes on Simuelue5. Thus while long earthquake sequences such as the 2004-2010 sequence along
209
Sumatra may have occurred in the past, they are not necessarily the rule over the 7,500 year record. Further
210
analysis of the turbidites in the 2004-2005 rupture areas may determine whether the 2004-2005 stress triggering
211
relations9 are a persistent feature along the Sumatra margin. Cores in the 2005 region contain turbidites that are
212
less frequent and stratigraphy is inconsistent between cores. Historic earthquakes in this region are smaller and
213
further downdip36 from our outer forearc high sites, possibly explaining this observation.
214
We note that the 2004 turbidite has three fining upward coarse pulses in our cores, and also that the 2004 SASZ
215
earthquake has three primary slip sub-events37. The success of turbidite correlation in Cascadia, the northern San
216
Andreas, and Sumatra suggests that some first order structure of the turbidity current maintains integrity despite
217
the fluid dynamic complexity of turbidity currents. We suggest that the longitudinal heterogeneity of the current,
218
imparted by the heterogeneous earthquake rupture itself
219
be recorded in the deposits. Our model predicts that these sub-events, minutes apart, may be recorded as
220
discernable coarse pulses within the turbidite that can be correlated over large distances, (Fig. S3). A similar
221
conclusion was drawn for Cascadia earthquake turbidites10 the mechanism has been tested in flume studies 39 and
222
is predicted by theory and analog models39 and references therein.
223
Conclusions
224
Turbidite stratigraphy is ubiquitously present in isolated slope basin and trench sites along the northern Sumatra
225
margin. Physiography isolates these sites from terrestrial sediment input and from large storms, providing good
226
localities to investigate the potential for earthquake paleoseismology. Stratigraphic correlation and radiocarbon
227
ages support serial deposition of synchronous turbidites over the past ~ 7500 years. Detailed correlation between
228
sites with no sedimentologic communication supports a regional earthquake origin most of these turbidites. The
229
youngest turbidite most likely correlates with the 26 December 2004 great Sumatra-Andaman Mw 9.1-3
230
earthquake. This event triggered turbidity currents in multiple submarine drainage systems that left stratigraphic
231
evidence in the form of multi-pulse turbidites in isolated slope basin and trench depocenters. Previous great
232
earthquakes in the same region have shaken sufficiently to trigger at least 19 (and as many as 24) turbidity
9,10,38
(the source-time function of the earthquake) may
233
currents and deposit corresponding turbidites during the past ~ 7.5 ka, with an average repeat time of ~ 330
234
years. The offshore turbidite record is consistent with both plate motion and land paleoseismic ages for eight
235
previous earthquakes approximately 400, 600, 800, 1,000, 1,500, 2,300, 6,000, and 7,100 years ago. This long
236
record includes earthquake clustering with gaps of 500- to 700-years between events large enough to trigger
237
recordable turbidity currents. The timing of the four most recent earthquakes is similar to tight clusters of
238
earthquakes recorded by live corals35 along the Padang segment of the Sumatra subduction zone, 400 km to the
239
south, suggesting along-strike sequences of earthquakes over relatively short time intervals.
240
References
241
242
243
244
245
246
247
248
1.
249
5.
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
R. Bilham, R. Engdahl, N. Feldl, and S. P. Satyabala, Partial and Complete Rupture of the Indo-Andaman Plate Boundary 1847-2004.
Seis. Res. Let. 76, 299-311 ( 2005).
2.
K. Monecke, Finger, W., Klarer, D., Kongo, W., McAdoo, BG., Moore, A., and Sudrajat, S., A 1,000-year sediment record of tsunami
recurrence in northern Sumatra. Nature 455, 1232-1234 (2008).
3.
K. Jankaew, Atwater, B. F., Sawai, Y., Charoentitirat, T., Martin, M. E., & Prendergast, A., Medieval forewarning of the 2004 Indian
Ocean tsunami in Thailand. Nature 455, 1228-1231 (2008).
4.
K. Rajendran, et al., Age estimates of coastal terraces in the Andaman and Nicobar Islands and their tectonic implications.
Tectonophysics 455, 53-60 (2008).
A. J. Meltzner et al., Coral evidence for earthquake recurrence and an A.D. 1390–1455 cluster at the south end of the 2004 Aceh–
Andaman rupture. Jrnl. of Geophys. Res. 115, B10402 (2010).
6.
J. Adams, Paleoseismicity of the Cascadia Subduction Zone: Evidence from Turbidites off the Oregon-Washington Margin. Tectonics
9, 569-583 (1990).
7.
C. P. Rajendran, et al., Crustal Deformation and Seismic History Associated with the 2004 Indian Ocean Earthquake: A Perspective
from the Andaman–Nicobar Islands. BSSA 97, S174-191 (2007).
8.
C. Goldfinger, C. H. Nelson, J. E. Johnson, Holocene Earthquake Records From The Cascadia Subduction Zone And Northern San
Andreas Fault Based On Precise Dating Of Offshore Turbidites. Annu. Rev. Earth Planet. Sci. 31, 555-577 (2003).
9.
C. Goldfinger et al., Late Holocene Rupture of the Northern San Andreas Fault and Possible Stress Linkage to the Cascadia
Subduction Zone. BSSA 98, 861-889 (2008).
10.
C. Goldfinger, et al., Turbidite Event History: Methods and Implications for Holocene Paleoseismicity of the Cascadia Subduction
Zone” (U.S. Geological Survey Professional Paper 1661-F, 2011).
11.
Y. Inouchi et al., Turbidites as records of intense paleoearthquakes in Lake Biwa, Japan. Sed. Geol. 104, 117-125 (1996).
12.
T. Shiki et al., Sedimentary features of seismites, seismo-turbidite, Lake Biwa, Japan. Sed. Geol. 135, 37-50 (2000).
13.
A. Noda, paper presented at the Workshop on Turbidites as Earthquake Recorders, Tsukuba, Japan (2004).
14.
Y. Okamura, paper presented at the Workshop on Turbidites as Earthquake Recorders, Tsukuba, Japan (2004).
15.
G. St-Onge, et al., Earthquake and flood-induced turbidites in the Saguenay Fjord (Québec): a Holocene paleoseismicity record.
Quat. Sci. Rev. 23, 283-294 (2004).
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
16.
A. Dallimore, R. E. Thomson, M. A. Bertram, Modern to Late Holocene deposition in an anoxic fjord on the west coast of Canada:
Implications for regional oceanography, climate and paleoseismic history. Marine Geology 219, 47-69 (2005).
17.
R. E. Karlin, M. Holmes, S. E. B. Abella, R. Sylwester, Holocene landslides and a 3500-year record of Pacific Northwest earthquakes
from sediments in Lake Washington. Geol. Soc. Am. Bull. 116, 94-108 ( 2004).
18.
M. Ishii, P. M. Shearer, H. Houston, J. E. Vidale, Extent, duration and speed of the 2004 Sumatra– Andaman earthquake imaged by
the Hi-Net array. Nature 435, 933-936 (2005).
19.
M. Chlieh et al., Coseismic Slip and Afterslip of the Great Mw 9.15 Sumatra–Andaman Earthquake of 2004. BSSA 97, S152-S173
(2007).
20.
S. Stein, E. Okal, Ultralong Period Seismic Study of the December 2004 Indian Ocean Earthquake and Implications for Regional
Tectonics and the Subduction Process. BSSA 97, S279-S295 ( 2007).
21.
C. Subarya, et al., Plate-boundary deformation associated with the great Sumatra–Andaman earthquake. Nature 440, 46-51 (2006).
22.
P. Bryna, et al., Explaining the Storegga Slide. Marine and Petroleum Geology 22, 11-19 (2005).
23.
C. Goldfinger, et al., Rupture lengths and temporal history of significant earthquakes on the offshore and north coast segments of the
Northern San Andreas Fault based on turbidite stratigraphy. Earth Plan. Sci. Ltrs 254, 9-27 (2007).
24.
T. Shiki, Reading of the trigger records of sedimentary events-a problem
for future studies. Sed. Geol. 104, 249-255 (1996).
25.
D. S. Gorsline, T. De Diego, E. H. Nava-Sanchez, Seismically triggered turbidites in small margin basins: Alfonso Basin, Western Gulf
of California and Santa Monica Basin, California Borderland. Sedimentary Geology 135, 21-35 (2000).
26.
T. Nakajima, Kanai, Y., Sedimentary features of seismoturbidites triggered by the 1983 and older historical earthquakes in the
eastern margin of the Japan Sea. Sedimentary Geology 51, 1-19 (2000).
27.
D. Graindorge et al., Impact of lower plate structure on upper plate deformation at the NW Sumatran convergent margin from
seafloor morphology. Earth and Planetary Science Letters 275, 201-210 (2008).
28.
D. G. Moore, J. R. Curray, F. J. Emmell, Large Submarine Slide (Olistotrome) Asscoiated with Sunda Arc Subduction Zone, Northeast
Indian Ocean. Marine Geology 21, 211-226 (1976).
29.
E. Gràcia, et al., Holocene earthquake record offshore Portugal (SW Iberia): testing turbidite paleoseismology in a slow-convergence
margin. Quaternary Science Reviews 29, 1156-1172 (2010).
30.
G. Biscontin, J. M. Pestana, F. Nadim, Seismic triggering of submarine slides in soft cohesive soil deposits. Marine Geology 203, 341354 (2004).
31.
M. Salisbury et al., Deep-sea ash layers reveal evidence of large Pleistocene and Holocene volcanic eruptions from Sumatra,
Indonesia, Abstract V11D-2330 presented at the 2010 Fall Meeting, AGU, San Francisco, Calif., 13-17 Dec.
32.
D.A. Stow., et al., Sediment facies and processes on the distal Bengal Fan, Leg 116. Proc, ODP, Scientific Results, 116 , 377-396
(1990).
33.
T. Mulder, et al., Marine hyperpycnal flows: initiation, behavior and related deposits. A review. Marine and Petroleum Geology 20,
861-882 (2003).
34.
S. Fujino, et al., Stratigraphic evidence for pre-2004 tsunamis in southwestern Thailand. Marine Geology 262, 25-28 (2009).
35.
K. Sieh, et al., Earthquake Supercycles Inferred from Sea-Level Changes Recorded in the Corals of West Sumatra. Science 322,
1674-1678 (2008).
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
36.
328
Acknowledgements This research was funded by the Ocean Sciences and Earth Sciences Divisions
329
of the National Science Foundation. We thank Morgan Erhardt, Amy Monica Garrett, and Bran Black
330
for conducting lab analyses; Sarah Strano, Summer Praetorius, and Maureen Davies for their kind
331
assistance in preparing and running the
332
site development; NOC, IFREMER, and BGR for providing key bathymetry and sub-bottom data; UTM,
333
for providing science crew; NOC for providing Russ Wynn; BGR for providing Stefan Ladage; and
334
AIST/GSJ for providing Ken Ikehara. We also thank coring technicians from OSU including Chris
335
Moser, Bob Wilson, Paul Wolscak. Scripps Resident Technicians, the R/V/ Roger Revelle Captain
336
Tom Djardins and crew, and student volunteers and faculty from OSU including Bart DeBaere and
F. J. Tilmann, et al., The updip seismic/aseismic transition of the Sumatra megathrust illuminated by aftershocks of the 2004 AcehAndaman and 2005 Nias events. Geophys. Jrnl. Int. 181, 1261-1274 (2010).
37.
T. Lay, et al., The Great Sumatra-Andaman Earthquake of 26 December 2004. Science 308, 1127-1133 (May 20, 2005).
38.
C. Goldfinger, Morey, A., Physical Property Correlations from Cascadia Great Earthquakes: What Are They Telling Us About The
Triggering Events? Eos Trans. AGU 85(47), Fall Meeting Suppl. Abstract OS21E (2004).
39.
C. Goldfinger, U.S.G.S. Final Technical Report, N.H.R.P. Award 07HQGR0064 (2011).
40.
C. Grand Pre, et al., Application of Microfossils to Reconstruct a Paleoeseismic Record of the Sunda Subduction Megathrust,
Northern Sumatra. Eos Trans. AGU 89(53), Fall Meeting Suppl. Abstract U51A-0010 (2008).
41.
M. Stuiver, P. J. Reimer, T. F. Braziunas, High-Precision Radiocarbon Age Calibration for Terrestrial and Marine Samples.
Radiocarbon 40, 1127-1151 (1998).
42.
C. Bronk Ramsey, Bayesian Analysis of Radiocarbon Dates. Radiocarbon 51, 337-360 (2009).
43.
P. J. Reimer, et al., Intcal09 And Marine09 Radiocarbon Age Calibration Curves, 0–50,000 Years Cal Bp. Radiocarbon 51, 11111150 (2009).
44.
M. Ishii, P. M. Shearer, H. Houston, J. E. Vidale, Teleseismic P wave imaging of the 26 December 2004 Sumatra-Andaman and 28
March 2005 Sumatra earthquake ruptures using the Hi-net array. Jrnl. of Geophys. Res. 112, B11307 (2007).
45.
M. Tolstoy, D. R. Bohnenstiehl, Hydroacoustic contributions to understanding the December 26th 2004 great Sumatra–Andaman
Earthquake. Survey of Geophysics 27, 633-646 (2006).
46.
M. Stuiver, H. A. Polach, Discussion: Reporting of 14C data. Radiocarbon 19, pp. 355-363 (1977).
47.
J. S. Noller, Lead-210 Geochronology, in: Quaternary Geochronology. J.S.Noller, Sowers, UJ.M., Lettis, W.R., Eds. (AGU Reference
Shelf 4, Washington, DC, 2000), pp. 115-120.
48.
S. Ladage, et al., Bathymetric Survey Images Structure off Sumatra. Eos Trans. AGU 87, 165, 168 (2006).
49.
W. H. F. Smith, D. T. Sandwell, Global Sea Floor Topography from Satellite Altimetry and Ship Depth Soundings. Science 277, 19571962 (1997).
14
C age samples; Chris Romsos for mapping the seafloor for
337
Maureen Davies. Further details regarding the cruise and the core locations, please refer to the cruise
338
report here: http://www.activetectonics.coas.oregonstate.edu/sumatra/report/
339
Supplementary Information is linked to the outline version of the paper at http://www.sciencemag.org
340
Author Contributions J.R.P and C.G. led the discussion. J.R.P, C.G., A.M., K.I., and C.R. led
341
fieldwork and lab analyses; K.I. assisted in sedimentologic description; Y.D. and U.H. assisted in field
342
work.
343
Author Information Reprints and permissions information is available at http://www.sciencemag.org.
344
The authors declare no competing financial interests. Correspondence and requests for materials
345
should be addressed to J. R. P. (jpatton@coas.oregonstate.edu)
346
347
Figure Captions
348
Figure 1. India-Australia plate subducts northeastwardly beneath the Burma plate at modern rates (GPS velocity
349
based on Nuvel-1A21). Historic ruptures19 are plotted in grey. Paleotsunami and paleoearthquake sites are plotted
350
in grey and RR0705 cores are plotted in white.
351
352
Figure 2. Regional stratigraphic correlations and source areas. A. Core locations (orange dots) on bathymetric
353
map27 also showing key channel flow paths (light blue) to eight core sites (scale bar units in km). Slope basin
354
source areas (orange) were determined by outlining drainage divides surrounding all submarine topography
355
contributing potential gravity flows to a given core site. While the region that drains to core 104PC then drains to
356
the trench, the three other slope basins are closed (they do not drain to the trench). B. Flow path profile depth (km)
357
vs. forearc distance (km) are plotted in brown for basin flow paths and blue for trench flow paths. C. Stratigraphic
358
correlations between key cores using lithology, CT, geophysical property, and
359
CT imagery displays lower density material in darker grey and higher density material in lighter grey. Slope cores
360
are labeled brown; trench cores are labeled blue. Trench core sites were deeper than the Carbonate
361
Compensation Depth (CCD), the depth below which foraminiferid CaCO 3 tests dissolve faster than they are
14
C data. MSCL data are plotted and
362
deposited. Foraminiferid abundance was nil in trench core sediments, so
363
cores. Grey rectangles refer to Fig. S7.
364
Figure 3. Space-Time relations for stratigraphy cored in the 2004 rupture region are plotted (vs. forearc distance)
365
as blue circles with error reported to 2 sigma shown by error bars. Green tie lines show stratigraphic correlations
366
(thicker = correlation more certain, dashed = less certain). Region-wide events are designated by a dashed grey
367
line and labeled with peak ages on the left margin, along with the probability density functions for these event
368
ages. Terrestrial paleoseismic and paleotsunami data are plotted by an dots/ diamonds and triangles respectively.
369
Data plotted to the left of 108PC are not plotted vs. forearc distance as they are further north than the figure limits.
370
The 2004 earthquake extends beyond the latitudinal extent of this figure 19. See Fig. S4 for more detailed
371
radiocarbon methods.
372
Supporting Online Material
14
C age control applies only to the slope
373
374
Table S1. Radiocarbon ages for cores in the region of the 2004 SASZ earthquake.
375
376
Figure S2. A. Regional stratigraphic correlations for all cores plotted in Figs. 2 A and 3. Symbology is the same as
377
used for Fig. 2 C. B. Core location map48,49 for cores plotted in A; cores are plotted as orange dots (scale bar units
378
in km). Source areas for basin cores are outlined in orange. The 2004 SASZ EQ rupture region 19 is outlined and
379
shaded in tan. Inset map shows cores as they relate to historic ruptures and the Island of Sumatera.
380
381
Figure S3. What is most likely the 2004 turbidite is displayed in this figure from cores RR0705-96PC and RR0705-
382
96TC. A. From left to right, gamma density (g/cc), CT density (greyscale digital number), lithologic log, CT imagery,
383
point and loop magnetic susceptibility (ms, SI
384
for each core. B. Moment release (vs. latitude) in red19 and relative amplitude (vs. time) in green44, purple20, and
385
orange45 are scaled to match peaks in the ms data from core RR0705-96PC. A light brown tie-line connects the
386
two data sets at a lithologic contact. Thick grey lines correlate seismic peaks with each other and with core
387
geophysical data. C. Particle size distribution data from sample locations found in A are plotted by volume (%) vs.
388
particle size (μm, log scale) with lines generally designating samples’ depth where the lighter lines have a larger
389
mean size and are generally lower in section. Differential volume displays the percent volume of each particle
390
size.
391
), and mean grain size (μm, log scale) are plotted vs. depth (m)
10-5
392
Table S4. OxCal combinations and single ages used to provide age control for timing of turbidite deposition (Fig.
393
3).
394
395
Figure S5. Event Flattening process displayed for correlated turbidites (c.turb) 8 and 9 in slope cores RR0705-
396
108PC, RR0705-104PC, and RR0705-103PC, each with unique sedimentary sources. Entire cores are plotted in
397
Figure S2. A. Geophysical traces as in Figure 2. Cores are vertically aligned to the basal contact for correlated
398
turbidite 9. Calibrated ages plotted at sample locations; the older age of event 9 in core 108PC is due to basal
399
erosion, clearly visible in the CT data. The age of event 9 in core 104PC has a large error due to the small sample
400
size. Core section breaks are in brown. B. Core data are plotted vs. depth, with vertical scales normalized based
401
on stratigraphic correlations in A. Green tie-lines show how data in A and B relate.
402
403
Figure S6 Core locations are plotted so that the geomorphology of the core sites can be evaluated with respect to
404
preserved stratigraphy. Depth contours are also plotted. Bathymetry was provided by our collaborators from the
405
UK, Japan, France, and Germany48. A. map showing relative locations of cores as orange dots (Fig. S2). White
406
scale bar is 50 km. B. through H. Larger scale, low angle oblique views of core sites (orange dots). Orientation
407
arrows (in white) point in the direction of North and Up (vertical). Vertical and horizontal units are in meters.
408
409
Figure S7. Core sites form RR0705 are displayed with symbols designating core type and symbol size
410
representing relative core length. Terrestrial paleotsunami and paleoearthquake sites, in the 2004 SASZ
411
earthquake region, are also displayed. The RR0705 cruise trackline is plotted as a blue line. SRTM topography 49
412
underlies bathymetry provided by our collaborators from the UK, Japan, France, and Germany48. Chlieh, et. al.19
413
slip estimates for the 2004 SASZ EQ, in cm, are plotted as colored dots. Historic ruptures19 are plotted in orange
414
outline. Paleotsunami and paleoearthquake sites are plotted in grey and RR0705 cores are plotted.
415