Okinawa.pdf

Available online at www.sciencedirect.com

Tectonophysics 466 (2009) 281 – 288

www.elsevier.com/locate/tecto

Structure of the southernmost Okinawa Trough from

reflection and wide-angle seismic data

Frauke Klingelhoefer a,

⁎ , Chao-Shing Lee b , Jing-Yi Lin a , Jean-Claude Sibuet a

a Ifremer, Department of Marine Geosciences, BP 70, 29280 Plouzané cedex, France

b Institute of Applied Geophysics, National Taiwan Ocean University, 2 Pei-Ning Road, Keelung 202, Taiwan

Received 28 February 2006; received in revised form 19 September 2006

Available online 9 January 2008

Abstract

During a passive seismic experiment in the Okinawa Trough the shots of two reflection profiles were recorded by ocean bottom seismometers

(OBS). Both profiles include 3 ocean-bottom instruments, are about 65 km in length and located in the axial portion of the southwestern Okinawa

Trough. Processing of the reflection seismic data images recent deformation of the sedimentary units. Forward modelling of the wide-angle data

on both profiles reveals a 1 – 2 km thick sedimentary infill overlying an acoustic basement characterised by seismic velocities between 3.2 and

3.5 km/s. Crustal thickness could only be modelled on one profile and was determined to be around 10 km, thickening towards the Ryukyu Arc in

the south. Gravity modelling was used to additionally constrain both profiles especially the deep structure of Profile 1.

Keywords: Okinawa Trough; Wide-angle seismics; Crustal structure; Back-arc basin

1. Introduction

of seamounts which ends close to Taiwan in Kueishantao Island

( Sibuet et al., 1998 ).

Five other belts of basins corresponding to five different

rifting phases were identified in the East China Sea ( Sibuet and

Hsu, 1997 ). They are mostly oriented NE

The Okinawa Trough, bounded by the Ryukyu Arc to the south

and east and by the East China Sea continental shelf to the north

and west, is a curved backarc basin, located behind the Ryukyu

trench-arc system ( Fig. 1 ). Its maximumdepth (2300 m) is located

to the south and decreases to only 200 m towards the north, close

to Japan. It formed by extension either of continental lithosphere

( Uyeda, 1977 ) or of continental lithosphere previously intruded

by arc volcanism ( Sibuet and Hsu, 1997 ). Sibuet and Hsu (1997)

proposed, that the series of continental shelf basins located

parallel to the mainland China shoreline consist of several belts of

backarc basins separated by relic arcs. The Okinawa Trough in

this case represents an active backarc basin formed by extension

of a continental crust which contains an unknown proportion of

arc material. From Kyushu to Okinawa Island subaerial active

volcanoes coincide with the Ryukyu arc ( Sibuet et al., 1998 ).

Southwest of Okinawa Island, the volcanic front moves

progressively into the Okinawa Trough and consists of a series

SWand their tectonic

history shows that rifting occurred between late Cretaceous/

early Paleocene and middle Miocene ( Sibuet and Hsu, 1997 ). In

the Okinawa Trough, rifting probably started in middle Miocene

( Letouzey and Kimura, 1986 ).

During one of the first wide-angle seismic studies carried out in

the Philippine Sea area, three regional seismic transects were

acquired ( Murauchi et al., 1968 ). On the basis of these refraction

seismic measurements, Murauchi et al. (1968) deduced, that the

crust underlying the Okinawa Trough is generally similar to that of

continental crust. They found aMoho depth of 12 km in the eastern

Okinawa Trough rapidly thinning towards the northwest ( Fig. 1 ).

From reflection seismic data the sedimentary and upper crustal

velocities were determined by Ludwig et al. (1973) in the northern

Okinawa Trough around 31°N and Leyden et al. (1973) correlated

offshore refraction velocities with onland drilling and refraction

data to propose a crustal structure in the Okinawa Trough. Based

on models from wide-angle seismic data, Hagen et al. (1988)

–

⁎ Corresponding author. Tel.: +33 298 22 42 21; fax: +33 298 22 45 49.

E-mail address: fklingel@ifremer.fr (F. Klingelhoefer).

doi: 10.1016/j.tecto.2007.11.031

282

F. Klingelhoefer et al. / Tectonophysics 466 (2009) 281 – 288

Fig. 1. Contoured seafloor bathymetry of the study region (Sibuet, unpublished data). OBS positions are marked by black circles and the two black lines mark the two

profiles from this study. Black diamonds mark receiving ship position from Lee et al. (1980) , black dotted line represents the OBS profile from Hagen et al. (1988) and

black broken line represents Profile 1 from the TAICRUST experiment ( Wang et al., 2001, 2004 ). Inset: location of the study area. Inverted triangles mark refraction

measurement positions from Murauchi et al. (1968).

14 km deepening towards the south.

Lee et al. (1980) using two ship seismic refraction lines,

found that the Okinawa Trough is underlain by a roughly 9 km

thick crust, which is overlain by an acoustic basement layer and a

–

During the 2003 Okinawa Trough cruise performed on the

Ocean Researcher 1 vessel, two reflection seismic profiles were

shot along aligned ocean bottom seismometers (OBS) which

were deployed to record earthquake data. The main aim of the

marine experiment was the acquisition of passive seismological

data in order to constrain the deep crustal and mantle structure in

the region of an inferred tear of the Ryukyu slab ( Lin et al.,

2004b,a ). Results from inversion of the seismological data are

described by Lin et al. (this volume). The reflection and wide-

angle seismic data presented in this paper provide information

on the sedimentary and crustal velocity structure at a much

higher resolution than the models established from seismolo-

gical data. Both models are therefore complimentary and

presented in separate papers.

2 km thick layer of sediments ( Fig. 1 inset). During the

TAICRUST survey combined reflection and wide-angle seismic

profiles were acquired in southwestern Ryukyu subduction zone

( Fig. 1 inset). Profile 1 of this experiment, extends from the

southern Okinawa Trough, close to our study area, over the

Ryukyu Arc, the Nanao Forarc Basin, the Yaeyama Accretionary

Ridge, the Ryukyu Trench into the Huatung Basin ( Wang et al.,

2001, 2004 ). 8 OBS were deployed, on this 300 km long profile.

Forward modelling and tomography of these data yielded a

Moho depth of 22 km underneath the Ryukyu Arc, rising

towards the Okinawa Trough ( Wang et al., 2001, 2004 ).

Higher crustal thicknesses of up to 20

2. Data acquisition, quality and processing

23 km are found in

the northernmost Okinawa Trough ( Nakahigashi et al., 2004 ). A

regional velocity model has been constructed using available

datasets and applied to reduce the relocation error of local

earthquakes ( Font et al., 2003 ). Large-scale velocity models

from earthquake tomography image the subducting slab in the

upper mantle ( Lallemand et al., 2001; Nakamura et al., 2003;

Hsu, 2001 ).

–

The two reflection seismic profiles presented in this paper

were shot along the ocean-bottom instruments using a 600 m-

long 12 channel streamer and a 1275 in. 3 airgun array ( Fig. 1 ).

Each of the two profiles modelled in this study used three ocean

bottom seismometers of the Ifremer OBS pool. All instruments

are equipped with a 3-component externally deployed geophone

and one hydrophone (Auffret et al., 2004).

Because the main aim of this cruise was the recording of the

natural seismicity in the Okinawa Trough during 12 days, the

propose the nature of the crust east of Taiwan and south of the

Ryukyu Trench to be oceanic and continental north of the trench,

close to our study area ( Fig. 1 ). Moho depth in the continental part

of the model is around 12

–

F. Klingelhoefer et al. / Tectonophysics 466 (2009) 281 – 288

283

Fig. 2. (A) Part of reflection seismic Profile 2 crossing the Okinawa Trough. The data are bandpass filtered, stacked and migrated. An automatic gain control has been

applied. (B) Interpretation of the reflection seismic section.

interval between OBS stations was chosen to be about 20 km,

and not specifically designed for wide-angle seismic experi-

ments. The active seismic shots served two purposes: first to

allow an exact relocation of the OBS position and second the

acquisition of reflection seismic data using a streamer. Analyses

of active source data on the two profiles and the reflection

seismic data allowed us to determine the geometry of the se-

dimentary and upper crustal layers in the trough. Although the

number of instruments per profile was small and the reflection

seismic data are of only fair quality, these data present sub-

stantial additional constraints on the crustal structure in this area

where deep seismic data are otherwise scarce. They also provide

complimentary information to the lower resolution velocity

models from the passive seismic experiment Lin et al. (this

volume).

Processing of the multi-channel seismic data included spher-

ical divergence correction, bandpass filtering (filter corner

frequencies: 3

Pre-processing of the OBS data included calculation of the

clock-drift corrections (between 2 and 5 ms per day) to adjust

the clock in each instrument to the GPS base time. Instrument

locations were corrected for drift from the deployment position

during their descent to the seafloor using the direct water wave

arrival. Picking of the onset of first and secondary arrivals was

performed without filtering where possible. Different filters

were applied to the instruments where necessary, depending on

the quality of the data and offset to the source. Data quality of

the 3 instruments on Profile 2 is good with deep arrivals from

the crust-mantle boundary and the upper mantle on all three

instruments ( Fig. 3 ). On Profile 1 the data quality is much lower

and no crustal or mantle arrivals could be identified on all three

instruments.

3. Velocity and gravity modeling

120 Hz), normal move-out correction and

stack ( Fig. 2 ). An automatic gain control was applied after the

stack. The shallow sedimentary layers are well imaged in the

seismic reflection section and the accoustic basement reflector

is distinguishable in regions of low sediment thickness.

5, 60

–

The data were modelled using the inversion and ray tracing

algorithm RAYINVR ( Zelt and Smith, 1992 ). Modelling was

performed using a layer-stripping approach, proceeding from the

top of the structure towards the bottom. The upper layers in the

model, not constrained by arrivals from within the same layer,

–

284

F. Klingelhoefer et al. / Tectonophysics 466 (2009) 281 – 288

Fig. 3. (A) Bandpass (3 – 5, 48 – 72) Hz) filtered vertical geophone data from OBS 15. The data are gain-adjusted according to offset and reduced to a velocity of 6 km/s.

(B) Data from OBS 14 processed as in (A). (C) Data from OBS 13 processed as in (A).

were adjusted to improve the fit of arrivals from lower layers. We

used a two-dimensional iterative damped least-squares inversion

of travel times ( Zelt and Smith, 1992 ). Arrival times of the main

sedimentary layers and basement were picked from the

reflection seismic data where possible. The arrival times were

converted to depth using the OBS data and velocities consistent

with those from velocity analysis of the reflection seismic data.

The depth and velocities of the crustal layers and the upper

mantle were modelled from the OBS data only. Based on the data

quality the picking uncertainties were taken to be 50 ms for the

direct water arrivals and 100 ms for all other arrivals.

The Profile 2 velocity model is comprised of 6 layers: the

water layer, one sedimentary layer, one underlying acoustic

basement layer, two crustal layers and the upper mantle layer

( Fig. 4 ). Each layer is defined by depth and velocity nodes. The

water velocity is found to be 1500 m/s from modelling, which

agrees with existing water velocity data from oceanographic

measurement compilations ( Dietrich et al., 1975 ). The seafloor

F. Klingelhoefer et al. / Tectonophysics 466 (2009) 281 – 288

285

Fig. 4. (A) Gravity anomaly from satellite altimetry ( Sandwell and Smith, 1995 ) (broken black line) and predicted gravity anomaly from conversion of the seismic

velocities to densities. (B) Final velocity model of Profile 2 including the model boundaries (solid lines) and iso-velocity contours every 0.25 km/s. OBS locations are

marked by black circles. Mean layer densities used for gravity modelling are annotated in italic. (C) Ray coverage of Profile 1 with every 5th ray plotted. (D) Fit

between the travel time picks (dark grey bars) and the predicted arrival times (black lines) from ray-tracing.

bathymetry was taken from a bathymetric map of the Okinawa

Trough, produced from a new grid (grid spacing 150 m) and

including all available swath bathymetric data from French and

Japanese oceanographic ships ( Sibuet et al., 2004b ). The sea-

floor model layer includes depth nodes at a spacing of 1.5 km

( Fig. 4 ). The sedimentary layer is modelled using the reflection

seismic data for layer geometry converted to depth using

velocities from the OBS data and sampled at the same node

spacing as the seafloor layer. Sediment velocities range from

1.9 km/s to 2.1 km/s. The acoustic basement layer displays

velocities between 3.2 and 3.5 km/s. The top of the crust has

also been modelled with a node spacing of 2.5 km, as it is not

well resolved everywhere on the profile in the reflection seismic

data. The crust is modelled by two layers of 4.5 to 5.6 km/s and

5.6 to 7.0 km/s. The lower-crustal layer and the Moho are

imaged at a lower resolution.

Thus we use a depth node spacing of 10 km. No arrivals from

the upper mantle have been modelled so a constant gradient was

assumed from 8.0 km/s to 8.2 km/s throughout the model for

gravity modelling ( Fig. 4 ).

As no deep arrivals could be used to constrain the deeper

velocity structure on Profile 1, a similar deep crustal structure

and Moho depth as found in Profile 2, have been assumed to

model Profile 1 ( Figs. 5 and 4 ). The deep part of this model is

thus constrained by gravity modelling only. Sedimentary velo-

cities range from 1.9 to 2.1 km/s and the acoustic basement from

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: