Seismological studies on mantle upwelling in NE Japan: Implications for the genesis of arc magmas

1
Seismological studies on mantle upwelling in NE
JapanImplications for the genesis of arc magmas
IBM Workshop in Honolulu

• Junichi Nakajima Akira Hasegawa
• Research Center for Prediction of Earthquakes
  Volcanic Eruptions
• Graduate School of Science, Tohoku University,
  JAPAN

November 8, 2007
2
Outline
Miller et al. (2006)

• Review previous results on mantle-wedge structure
  in 2000s and their interpretations.
• Introduce recent (preliminary) results of
  velocity structure in Tohoku and Hokkaido

Study area
3
Seismological observations in 1990s
Seismology Hasegawa et al. 1991 Zhao et al.,
1992-1994
Hasegawa et al. 1991
Zhao et al. 1992
4
Travel-time tomography in NE Japan
5
Seismic tomography study - Eqs. and stations
(Nakajima et al., 2001, JGR)

• 169,712 P-wave arrivals 103,993 S-wave arrivals
• Method Zhao et al. 1992, JGR
• Grid separation 15-25 km in both horizontal and
  vertical directions

Neqs. 4338
6
Inclined low-Vs zones in mantle wedge
Inclined low-V zone 50 km above the slab
Velocity reductions of 4-6 in Vp and 6-10 in
Vs
Pacific plate
Low-F event
Volcano
Nakajima et al. 2001
7
Velocity structure at a 40 km depth (below the
Moho)
dVp
Vp/Vs
dVs
Nakajima et al. JGR, 2001
8
Numerical simulation
Flow pattern (wedge)
(Eberle et al., PEPI, 2002)
Flow pattern
Upward flow (high-T) is generated in the mantle
wedge.
Karato 1993, GRL
Predicted low-V zone is consistent with the
observation.
Inclined low-V zone upwelling flow induced by
slab subduction
9
Question

• What causes an inclined low-velocity zone ?
• - thermal heterogeneity?
• - melts?
• - chemical heterogeneity?

10
Qp structure in NE Japan
(Tsumura et al., 2000)
11
Conversion from Qp to TemperatureNakajima and
Hasegawa, GRL, 2003
Simple relationship between Q, temperature,
pressure and frequency e.g., Karato, 2004
( f frequency Hz,P pressure GPa, T
temperatureK, H activation enthalpy kJ/mol)
Given T0, P0 and Q0 as reference values.
References T0 1025? (40 km depth) Kushiro,
1987 Q0-1 0.0035 Tsumura et al., 2000
a0.20, H(P)500 16P kJ/mol, H0500 kJ/mol
Karato, 2004
12
Thermal structure Nakajima and Hasegawa, GRL,
2003
Wet solidus of peridotite
13
Correction of thermal effect

• Observed low-velocity anomalies
• -gt 4-6 in Vp and 6-10 in Vs
• Expected velocity reductions from thermal
  anomalies
• -gt 1-2 in Vp and 2-3 in Vs
• Residuals of velocity anomalies
• -gt -dlnVp0.03-0.04
• -dlnVs0.04-0.07
• dlnVp/dlnVs 12

Karato (1993)
14
Takeis model (Takei, JGR, 2002)
dlnVs/dlnVp (Velocity reduction rate)
physical properties of fluids
Aspect ratio(a)
Volume fraction from dlnVs(f)
Takei (2002)
15
Melt distribution in low-V zone
Nakajima, Takei and Hasegawa (2005, EPSL)
Partial melting with fractions of 0.3-5 vol in
low-V zone.
16
Interpretation of depth variation in pore shapes
Accumulation below Moho?
dike/crack
Migration?
dike/crack
Generation?
Depth variation in aspect ratio of melt-filled
pores
17
Which direction does mantle upwelling flow?

• Seismic velocity/attenuation structures are the
  present-day snap shot and do not provide the
  direction of mantle flow.

Shear-wave splitting could provide an important
and independent information on mantle dynamics.
18
Results of shear-wave splitting
(Nakajima and Hasegawa, EPSL, 2004)
Assuming A-type olivine in back arc, flow
direction is inferred to be EW.
19
A model of return flow in NE Japan
Hasegawa Nakajima (2004)
20
Summary

• An inclined-low-velocity zone in the mantle wedge
  sub-parallel to the slab
• Temperatures in the mantle wedge of 1000-1300 C
• Depth variation in aspect ratio of melt-filled
  pores and melt fractions of 0.05-5 vol in the
  low-velocity zone
• Flow direction parallel to the slab dip

21
Recent tomographic results in NE Japan
_at_ Update previous results by Nakajima et al.
(2001) _at_ Obtain clearer images of inclined
low-velocity zone _at_ Understand whole fluid
circulation
22
Data set
23
Comparison with Nakajima et al. (2001)
This study
Nakajima et al. (2001)
dVp
dVs
Central part of Tohoku
24
Results
dVs
dVp
Sheet-like low-velocity zone Larger velocity
reductions in S wave than P wave (-dlnVp3-6,
-dlnVs5-10) Thickness of low-velocity zone of
10-30 km with an along-arc variation (seems to be
thinner in C and D)
25
Low-velocity zone beneath back-arc volcanoes
Diapirs from the upwelling?
26
Path of fluids from slab to mantle
Low-V zone at a dept of 150 km -gt Supply of
fluids from slab to mantle?
27
Summary of recent results
4
5
1
3
2

• 1 Low-velocity zone corresponding to
  oceanic crust down to a depth of 100 km (Tsuji et
  al., unpublished).
• 23 Low-velocity zone at a depth of 150 km.
• Supply of fluids to mantle there?
• 34 Sheet-like low-velocity zone
• -dlnVs gt -dlnVp.
• Thickness of 10-30 km with along-arc
  variation
• 5 Segregated diapirs from upwelling?.
• Source of magmas of back-arc volcanoes?

28
Low-velocity zone in mantle wedge
Inclined low-V zone (from back-arc to the VF)
Hasegawa and Nakajima (2004), AGU Geophys. Monog.
NE JapanZhao et al. (1992), Nakajima et al.
(2001) Alaska Aleutian Abers (1994), Zhao et
al. (1995) KamchatkaGorbatov et al.
(1999) TongaZhao et al. (1997)
After 2004
HokkaidoWang and Zhao (2006) KyushuWang and
Zhao (2006) New ZealandReyners et al.
(2006) AlaskaEberhart-Phillips et al.
(2006) TongaConder and Wiens (2006)
Is inclined low-V zone a common feature in
subduction zones?
29
S-wave velocity structure -HOKKAIDO