1 Kobe Univ., 2 JAMSTEC, 3 Flinders Univ.,

Imaging mantle structure of the central Mariana subduction-arc-back arc

system using marine magnetotelluricsN. Seama 1, 2, A. White 3, A. D. Chave 4, K. Baba 5, T. Goto 2, T. Matsuno 1, R. L. Evans 4, G. Boren 3, A. Yoneda 5, H. Iwamoto 1, R. Tsujino 1, Y. Baba 5,

H. Utada 5, G. Heinson 6, and K. Suyehiro 2 1 Kobe Univ., 2 JAMSTEC, 3 Flinders Univ.,

4 WHOI, 5 ERI, Univ. of Tokyo, 6 Univ. of Adelaide

Contents:* Observation & Data analysis* 2-D resistivity models & their interpretations

Observation Line

●: New data from KR05-17 deployment and KR06-12 recovery cruises (Kairei, JAMSTEC)

●: Previous study (Filloux, 1983; Goto et al., 2003; Baba et al., 2005; Seama et al., 2007)

Pacific Plate

TrenchFore Arc

Back Arc Spreading Axis

Remnant Arc Volcanic Arc

Observation sites near spreading axis12 sites in 55km -line (1-6km sites spacing)

spreading axis

Australian OBEM

Australian OBM

(Type 1)

Australian OBM (Type 2)

US OBEM

US OBE

ERI-OBEM (Type 2)

IFREE/JAMSTEC-OBEM

ERI-OBEM (Type 1)

Kobe OBEM

Data Analysis

1) Clean up the raw time series data (9 months)

2) Estimate the magnetotelluric impedance tensor responses (MT responses) from the time series data# BIRRP(Chave and Thomson, 2003, 2004)

3) Correct the MT responses for the effect of 3-D seafloor bathymetry# Nolasco et al. (1998), Matsuno et al., (2007)# FS3D (Baba and Seama, 2002)

4) Estimate 2-D resistivity (or conductivity) structure models to fit the corrected MT responses

Inversion methods for estimating 2-D resistivity models

1) Data Space Occam inversion (Siripunvaraporn and Egbert, 2000)

We modified this algorithm for the MT responses at ocean bottom.

2) Anisotropic inversion (Baba et al., 2006)

These inversion algorithms find optimally smooth sets of resistivity models that fit the corrected MT responses to a desired level of misfit.

X

YSites used for 2-D inversions

●: New data (26 sites)●: Previous study (8 sites)

2-D Resistivity Modelwith Hypocenters (Shiobara, personal comm.)

2-D Resistivity Model

Slab lithosphereImposedResistivity: 3000Ohm-mThickness:60kmbased on the results from EPR (Baba et al., 2006)

Fitting the corrected MT responsesData (dots) Model (red lines)

RMS misfit (tm app): 1.94RMS misfit (all): 1.77

2-D Resistivity Model

Resistivity Model

Forward Modeling Test (1)

Low resistivity beneath the fore-arc

Resistivity value of the low resistivity region beneath the fore-arc

20 Ohm-m

Resistivity value: 20 Ohm-m

5 4 3 2 1

Extent of the low resistivity region beneath the fore-arc

Low resistivity can be due to:1) high water contents2) existence of melt3) high temperature4) low resistivity rock

Forward Modeling Test (2)

Low resistivity beneath the volcanicarc

Resistivity value of the low resistivity region beneath the volcanic arc

20 Ohm-m

Low resistivity region beneath the volcanic arc

Low resistivity can be due to:1) high water contents2) existence of melt3) high temperature4) low resistivity rock

?

Conder , personal comm.

Takahashi et al., 2007

Forward Modeling Test (3)

Connection between the slab and the volcanicarc

Resistivity value of the region connected between the slab and the volcanic arc

Not enough resolution?

Resistivity Model

Anisotropic models beneath the back-arc basin

xx

yy

zz

z

yx

Conder , personal comm.

Standard Olivine 2 (SO2 model; Constable et al., 1992)

Dry Olivine

Characteristic features of the low resistivity region beneath the spreading axis xx

yy

zz

# Anisotropic featureExistence of melt

z

yx

Characteristic features of the low resistivity region beneath the spreading axis xx

yy

zz

spreading axis

# Asymmetric features

??

??

??

+ Location+ Shape

Existence of melt

z

yx

Asymmetric features of the low resistivity region beneath the spreading axis

spreading axis

yy

Conder et al., 2002 (Lau back-arc spreading)

MBA: Kitada et al., 2006

xx

yy

zz100km

Dry Dry

Wet Anisotropic

Wet Anisotropic

Anisotropic layered resistivity structure beneath the back-arc basin

z

yx

xx

yy

zz

60km

100km

Baba et al., 2006

Mariana vs EPR

Dry Dry

Wet Anisotropic

Wet Anisotropic

Resistivity profile with depth

Black: IsotropicBlue: Parallel to

spreading directionGreen: vertical directionRed: Perpendicular to

spreading direction

PT=1300C3000H/106Si=0.02wt%

Modified from Seama et al., 2007

Melt beginning depth

Grey: Olivine with different water contents

Summary (our results are initial, but probably show the first order of the nature)

# Existence of the low resistivity region beneath the fore-arc (probably due to water from the slab)

# Existence of the low resistivity region beneath the volcanic arc (probably due to low resistivity of the volcanic arc crust and of the upper most mantle)

# Existence of the asymmetric low resistivity region beneath the back-arc spreading axis (probably due to melt affected by the dynamics)

# Existence of the anisotropic layered resistivity structure beneath the back-arc basin (probably due to differences in water contents affected by the dynamics)

Forward Modeling Test (4)

Low resistivity beneath the back-arc spreading axis

Resistivity value of the low resistivity region beneath the spreading axis

10-30 Ohm-m

Isotropic models using different inversion algorithms

Data Space Occam inversion (Siripunvaraporn and Egbert, 2000)

Anisotropic inversion (Baba et al., 2006)