You are on page 1of 1

Outer Rise Seismicity at the Tonga Trench Recorded by Ocean Bottom Seismographs

R. B. White and D. A. Wiens


Department of Earth and Planetary Sciences, Washington University, St. Louis, MO - USA

Abstract Study Area Discussion


We are investigating the locations and fault plane solutions of outer rise earthquakes in the Tonga
subduction zone to understand the distribution of strain in the bending oceanic plate. Intraplate outer rise -175˚ -174.5˚ -174˚ -173.5˚ -173˚ -172.5˚ -172˚
10/24/94
∑ Having a nearby OBS array allows for well-constrained relative locations of nearby events. Using a
earthquakes are due to tensional failure at shallow depths and compressional failure at deeper depths. -17˚ -17˚
11/4/94d
multiple-event hypocenter relocation technique [6], we can constrain the relative locations and
9/20/94e
-172.6˚ -172.4˚ -172.2˚ -172˚
The tension-to-compression transition depth is an important constraint for the study of downgoing plate -18.2˚ -18.2˚ depths very well. We used reflectivity synthetic seismograms and regional waveform inversion to
stresses, interplate coupling along the subduction interface, and the rheology of the lithosphere in the 176˚ 177˚ 178˚ 179˚ 180˚ -179˚ -178˚ -177˚ -176˚ -175˚ -174˚ -173˚ -172˚ -171˚ -170˚ -169˚
9/19/94e
obtain focal mechanisms for some of the outer rise events. Three events had CMT solutions that
outer rise region. In addition, the subduction of large bathymetric features may cause important -13˚ -13˚ -17.5˚ -17.5˚ 9/19/94d allowed for verification and calibration of a focal mechanism determination technique.
perturbations on the seismic coupling and plate rheology. -14˚ -14˚ 10/31/94
Samoa 11/4/94a
∑ Due to the varying crustal structure of the region [4, 12], four different velocity models were used to
9/20/94d
Using a 1994 three-month ocean bottom seismograph deployment across the Tonga subduction zone at -15˚ -15˚ -18˚ -18˚
11/4/94b
compute synthetic seismograms of the waveforms. These were used to invert for the best-fitting
-18.4˚ -18.4˚
about 19 S, more than 70 local earthquakes were relocated to less than 10 km error using a multiple- Fiji Islands fault plane solution.
-16˚ Lau Basin -16˚
event hypocenter determination algorithm. Regional waveform inversion for the largest events using 11/4/94c
9/19/94a

local broadband land stations yielded focal mechanisms for some earthquakes. An interesting feature in -17˚ -17˚ -18.5˚ -18.5˚ ∑ Perhaps the most interesting feature of this data set is a major cluster of earthquakes on the
the distribution of events is a cluster of 17 earthquakes with depths of 30-50 km below the Capricorn 9/21/94
Capricorn Seamount. It is a sequence of 17 events occurring mostly over two days (September 19-
Seamount, 13 of which span only two days. These earthquakes occur immediately beneath large normal -18˚ -18˚ 11/22/94
21, 1994). The close proximity of these events to each other suggests the possibility that they could
faults seen with high-resolution sea beam sidescan sonar bathymetry. Knowledge of the focal -19˚ -19˚ -19˚ -19˚ -18.6˚ -18.6˚ be an aftershock sequence. The focal mechanisms found for these earthquakes support this claim.
mechanisms for some of the events in this swarm offers the opportunity to test the relation of outer rise 11/28/94 All have thrust fault plane solutions oriented approximately perpendicular to the trench axis. This
stresses to interplate coupling. -20˚ -20˚ could be a rare opportunity to study the compressional stress regime of the lower part of the bending
-19.5˚ -19.5˚ Capricorn plate. As more focal mechanisms are obtained for this region, we hope to constrain the nodal

nch
-21˚ -21˚ Seamount surface depth, where the shallow extensional stresses change to compressional. The focal

Lau Ridge

Tre
km mechanism for the 11/4/94b CMT event is tensional, so we can propose the nodal surface to be
Background
-22˚ -22˚

ga
-18.8˚ -18.8˚
between 30 and 40 km in depth.

Ton
0 100 200
-20˚ -20˚ km
-23˚ -23˚
176˚ 177˚ 178˚ 179˚ 180˚ -179˚ -178˚ -177˚ -176˚ -175˚ -174˚ -173˚ -172˚ -171˚ -170˚ -169˚ km 0 5 10

0 50
∑ The other cluster of 7 events to the north of the Capricorn seamount contained two of the three CMT
-172.6˚ -172.4˚ -172.2˚ -172˚
∑ Shallow earthquakes at subduction zones occur primarily in two regions: the seismogenic zone and The SPaSE and LABATTS deployments consisted of 12 broadband -20.5˚ -20.5˚
events in this region during this time period. These events are stronger, on average, than the ones in
the outer rise. Studying the spatial and temporal distribution of these earthquakes with their fault seismographs from November 1993 to December 1995 and 25 OBS's -175˚ -174.5˚ -174˚ -173.5˚ -173˚ -172.5˚ -172˚ the Capricorn cluster, and finding focal mechanisms for them was easier than for the Capricorn
plane solutions can provide information on the nature of interplate coupling and the stresses affecting from September to December 1994 [12]. events.
Earthquakes detected during the OBS deployment, with High-resolution multibeam bathmetry of the Capricorn
the plates. relative locations determined with multiple-event Seamount reveals numerous normal faults on the object
Depth (m)
hypocenter inversion algorithm [6]. A cluster of as it enters the Tonga Trench [11]. Focal mechanisms for a ∑ We currently use IASP91 for the multiple-event relocation. The strong lateral heterogeneity in the
∑ Intraplate earthquakes at subduction zones occur within both plates. Events in the overriding plate earthquakes north of the Capricorn Seamount were swarm of events on the Capricorn Seamount also region severely limits, the depth resolution. We are currently in the process of modifying the code to
500 1500 2500 3500 4500 5500 6500 7500 8500 9500
are very shallow (< 15 km) and reflect local stresses in the accreted forearc sediments [10]. Intraplate mostly shallow thrust-faulting events. correspond to a compressional stress regime. use multiple structure models based on the station location. This should allow for even better
Earthquakes OBS Stations Land Stations
bending earthquakes in the outer rise region of the seaward plate occur near the trench and can often constraints on depth and focal mechanism solutions.
be easily distinguished from the interplate activity. Focal mechanisms of these earthquakes can
provide constraints on the regional stress state of the downgoing plate and how it relates to interplate ∑ With enough well-constrained event locations and focal mechanisms, a more definite estimate of the
coupling and the subduction process in general. nodal surface depth within the outer rise can be found. Taking into account the stress regime
suggested by the seismicity and correlating the hypocenters of the events to bathymetry in the region
∑ Hypocenter locations obtained from teleseismic arrivals are not reliable due to the tradeoff between [11], lithospheric models can be tested to study the nature of the downgoing plate. Simple elastic
depth and origin time [2, 5], misidentifying the phases pP and pwP [5], and systematic bias if all flexure and elastic-plastic deformation are two possible types to test [8].
nearby stations are landward of the event, such as at oceanic-continental subduction zones. Joint-
hypocentral relocation techniques [6, 10] can increase relative depth precision, but they are limited by
the amount and quality of the teleseismic data. To better constrain oceanic earthquake locations, an
Focal Mechanisms ∑ Finally, it will be interesting to see what effect the subduction of the Capricorn Seamount has on the
system, since it should increase the local interplate coupling and hinder seismogenesis [9].
ocean bottom seismograph (OBS) array can be used, since the OBSís can be placed at a variety of
distances and azimuths near the source. Model Velocity (km/s)
1 1 P- and S-wave velocity
2 3 4 5 6 7 8 9 models used for synthetic
0 1
0
∑ The bending of the plate is associated with tension in the upper 20 km [3], and this is evident with Outer Rise and Large Events in the Study Area seismogram generation in
the focal mechanism
seismic reflection and bathymetric observations of abundant normal faults on the seaward side of the
100 determination process.
trench axis [11]. Most tension axes are perpendicular to the strike of the trench, independent of the Date Time Latitude Longitude Depth Notes Fault Plane Solution
IASP91 was used for
References
Plane 1 Plane 2 Moment (dyn-cm)
direction of relation motion between the two plates, indicating the failure is controlled by bending VS VP teleseismic stations only.

Depth (km)
stresses in the plate [1]. Approximately one compressional outer rise earthquake is confirmed per September 19, 1994 a 9:27:22.9 -18.38 -172.29 47 Capricorn group 50, 30, 110 207, 62, 79 5.20x10 23 200
The back arc and pacific
September 19, 1994 b 9:40:36.3 -18.42 -172.28 42 Capricorn group
year teleseismically [3]. These seem to occur systematically deeper than normal faulting outer rise September 19, 1994 c 9:47:33.1 -18.44 -172.29 48 Capricorn group
models were used for
earthquakes and have depths up to 50 km [2, 3]. The Tonga-Kermadec region is unusual because of September 19, 1994 d 10:24:16.3 -18.35 -172.27 45 Capricorn group 203, 70, 72 67, 27, 130 8.05x10 22 300 receiving stations in either
the large number of outer rise events. Half of these are near the northern Tonga trench, and both September 19, 1994 e 11:39:31.5 -18.34 -172.26 30 Capricorn group 220, 72, 88 46, 18, 96 6.02x10 22
the Fiji back arc or the 1. Chapple, W.M. and D.W. Forsyth, Earthquakes and Bending of Plates at Trenches, J. Geophys. Res.,
September 20, 1994 a 7:15:52.5 -18.37 -172.25 41 Capricorn group Pacific plate. The lau radial
tensional and compressional events have been found [3]. September 20, 1994 b 7:41:00.7 -18.40 -172.29 39 Capricorn group 400
and transverse models
84, 6729-6749, 1979.
September 20, 1994 c 10:27:07.3 -18.40 -172.25 51 Capricorn group
22 take into account some of 2. Chen, T. and D.W. Forsyth, A Detailed Study of Two Earthquakes Seaward of the Tonga Trench:
September 20, 1994 d 14:26:57.6 -18.35 -172.26 35 Capricorn group 35, 22, 60 247, 71, 101 2.27x10
Profile Along N30E Azimuth September 20, 1994 e 15:16:27.8 -18.27 -172.23 32 Capricorn group 88, 26, 113 243, 66, 79 3.20x10
22
500 the strong anisotropy in Implications for the Mechanical Behavior of the Oceanic Lithosphere, J. Geophys. Res., 83, 4995-
arc outer rise
September 20, 1994 f 18:28:22.8 -18.35 -172.28 27 Capricorn group
IASP91 back arc lau radial lau transverse pacific
the Lau Basin [12]. 5003, 1978.
0 September 20, 1994
September 21, 1994
g 23:04:31.0
9:24:04.5
-18.45
-18.43
-172.29
-172.30
40
35
Capricorn group
Capricorn group 217, 71, 55 102, 39, 149 1.53x10 22 3. Christensen, D.H. and L.J. Ruff, Seismic Coupling and Outer Rise Earthquakes, J. Geophys. Res., 93,
October 11, 1994 8:26:40.7 -18.32 -172.20 56 Capricorn group 13421-13444, 1988.
ne extension
zo -50 October 13, 1994 20:20:35.9 -18.46 -172.65 49 Capricorn group 4. Crawford, W.C., et al., Tonga Ridge and Lau Basin Crustal Structure From Seismic Refraction Data., J.
ic
en
October 14, 1994 16:16:11.7 -18.47 -172.47 64 Capricorn group
og October 24, 1994 4:44:32.8 -17.69 -172.08 26 north group 36, 16, 41 266, 80, 102 7.05x10
23 -174˚ -173˚ -172˚ -171˚ -174˚ -173˚ -172˚ -171˚ Geophys. Res., in press, 2001.
ism compression October 25, 1994 2:13:09.7 -20.67 -173.01 91
-16˚ -16˚ -16˚ -16˚
5. Forsyth, D.W., Determination of Focal Depths of Earthquakes Associated with the Bending of Oceanic
se -100
October 29, 1994 12:57:57.4 -18.25 -172.07 65 north group
Plates at Trenches, Phys. Earth Planet. Int., 28, 141-160, 1982.
Depth (km)

24
October 31, 1994 23:04:05.9 -17.93 -172.74 37 CMT (not outer rise) 203, 27, 87 26, 63, 92 1.78x10
November 1, 1994 a 6:02:13.9 -17.53 -172.04 56 6. Jordan, T.H. and K.A. Sverdrup, Teleseismic Location Techniques and Their Application to
-150 November 1, 1994
November 4, 1994
b
a
14:29:05.4
1:35:32.2
-17.75
-17.76
-172.08
-172.09
30
18
north group
north group 175, 20, 215 52, 79, 287 2.05x10
25 11/4/94b 11/4/94b Earthquake Clusters in the South-Center Pacific, Bull. Seis. Soc. Am., 71, 1105-1130, 1981.
-17˚ 10/31/94 -17˚ -17˚ 10/31/94 -17˚
io n November 4, 1994 b 1:43:38.2 -17.81 -172.07 27 north group 19, 14, 42 248, 81, 100 7.80x10 23
Since the variation in crustal thickness across the trench axis (5.5 km in the 7. Lay, T. et al., Temporal Variation of Large Intraplate Earthquakes in Coupled Subduction Zones, Phys.
te ns -200 November 4, 1994 c 5:24:19.7 -17.80 -172.09 36 CMT, north group 350, 88, -150 259, 60, 358 4.10x10 23
Earth Planet. In., 54, 258-312, 1989.
ex November 4, 1994 d 10:12:41.1 -17.72 -172.03 29 CMT, north group 15, 18, -95 200, 72, 272 1.81x10 24 Pacific, 9 km in the arc) is important for the scale of this study, we first
November 8, 1994 12:45:36.8 -17.48 -171.97 17 constructed velocity models for the Pacific and backarc sides of the trench 8. Mueller, S., W. Spence, and G.L. Choy, Inelastic Models of Lithospheric Stress ñ Implications for
11/4/94a 11/4/94a
-250
November 9, 1994 21:03:07.3 -20.45 -173.22 18
22
[12]. The anisotropy in the Lau Basin dictated the creation of two more Outer-Rise Seismicity and Dynamics, Geophys. J. Int., 125, 54-72, 1996.
November 22, 1994 22:45:07.2 -18.44 -172.27 48 Capricorn group 56, 34, 113 209, 59, 75 7.73x10 models for the SH and SV components of the Lau Basin. These models -18˚ -18˚ -18˚ -18˚
November 28, 1994 2:47:28.2 -19.31 -172.51 36 140, 42, 160 245, 77, 50 1.24x10
23 9. Scholz, C.H. and C. Small, The Effect of Seamount Subduction on Seismic Coupling, Geology, 25,
December 8, 1994 15:09:32.9 -19.71 -173.00 58
were flattened and used to construct reflectivity synthetic seismograms for 487-490, 1997.
stations in the corresponding regions. The hypocenter locations were set
-300
-175 -174 -173 -172
December 12, 1994 10:47:50.4 -19.67 -173.02 68
by the mutiple-event hypocenter relocation previously performed. Long- 10. Ward, S.N., Body Wave Inversion: Momemt Tensors and Depths of Oceanic Intraplate Bending
Rotated Longitude peroid regional waveform inversion of these synthetics allowed for focal Earthquakes, J. Geophys. Res., 88, 9315-9330, 1983.
This table shows the outer rise earthquakes in the region of study. These origin times and locations were mechanism determination. By searching over a range of fault plane -19˚
-173˚ -172˚
-19˚ -19˚ -19˚ 11. Wright, D.J. et al., Bathymetry of the Tonga Trench and Forearc: A Map Series, Marine Geophs. Res.,
Simplified cartoon of an ocean-ocean subduction zone and the implied principal The relocated earthquake depths in this region of -174˚ -171˚ -174˚ -173˚ -172˚ -171˚
stress directions in the slab. The downgoing flexural outer rise is subject to tension determined using a multiple-event hypocenter relocation algorithm with phase arrival times [6]. The focal solutions the best-fitting solution was found. Amplitudes were weighted 21, 489-511, 2000.
study are plotted perpendicular to the strike of the more heavily than shape in order to better fit the surface wave arrivals. A CMT Solutions Our Solutions
and compression at shallow and deep depths. The seismogenic zone is the locus of trench. The shape of the slab and swarm of events on mechanisms were found from regional waveform inversion. The two main clusters of outer rise events are 12. Zhao, D. et al., Depth Extent of the Lau-Back-Arc Spreading Center and its Relation to Subduction
denoted by "Capricorn group" and "north group." The three events with CMT solutions during the OBS water layer was not accounted for. The four examples above represent the Processes, Science, 278, 254-257, 1997.
interplate coupling. Subduction of asperities like seamounts might be important the Capricorn Seamount are evident. Note that the
deployment are also included. data and synthetic for one vertical and three transverse records. Comparison of Harvard CMT solutions to our solutions for three events during the OBS
for increasing seismic coupling and controlling the local seismic cycle. relocation procedure currently uses the IASP91 model
only. When multiple models for the region are deployment. Agreement is exceptional for the 10/31/94 and 11/4/94b events. This
incorporated into the procedure, we will have more helped validify and calibrate our focal mechanism determination procedure.
confidence in these absolute depths.