08/09/2007VLab-workshop1 Geophysical Anomalies in the Central Pacific Upper Mantle Implications for Water Transport by a Plume Shun-ichiro Karato Yale.

08/09/2007 VLab-workshop 1

Geophysical Anomalies in the Central Pacific Upper Mantle

Implications for Water Transport by a Plume

Shun-ichiro KaratoYale University

New Haven, CT, USA

08/09/2007 VLab-workshop 2

Outline

• Anomalies in the central Pacific upper mantle– Seismic anisotropy– Electrical conductivity

• Some mineral physics background– Seismic anisotropy and water– Electrical conductivity and water

• Conclusions– Plume-asthenosphere interaction

08/09/2007 VLab-workshop 3

Ekström and Dziewonski (1998)

Upper mantle in the central Pacific has unusually strong VSH>VSV anisotropy.

08/09/2007 VLab-workshop 4

Upper mantle in the central Pacific has unusually low conductivity.

Karato (2007)

Ele

ctric

al c

ondu

ctiv

ity,

S/m

08/09/2007 VLab-workshop 5

Cause for anomaliesHow do various laterally varying parameters affect seismic anisotropy and electrical conductivity?

Mineral physics of anisotropy and conductivity Temperature X Major element chemistry X Partial melting X O Water (hydrogen) O

08/09/2007 VLab-workshop 6

Temperature or major element chemistry (e.g., Fe/Mg) does not change the elastic anisotropy so much.

--> A change in LPO should be a cause.

Seismic anisotropy = elastic anisotropy of minerals + LPO (lattice-preferred orientation)

08/09/2007 VLab-workshop 7

LPO changes with physical/chemical conditions.

LPO is determined by the dominant slip system(s).

Conventional interpretation of anisotropy assumes one type of slip system ([100](010)).

If the dominant slip system changes, LPO will change (fabric transition), then the nature of seismic anisotropy will change.

How could the dominant slip system change with physical/chemical conditions?

olivine

08/09/2007 VLab-workshop 8

LPO is controlled by the relative strength of slip systems.Deformation with b = [001] slip systems is more enhanced by water than deformation with b = [100] slip systems.

Could fabric transition occur at higher water content? [Karato (1995)]-->simple shear deformation experiments (change in the slip direction)

at higher water fugacity (0.3 GPa-->15 GPa)

08/09/2007 VLab-workshop 9

Olivine LPO

Conventional interpretation of seismic anisotropy is based on the limited observations and assumed A-type fabric.

But lab studies in my group have shown that other LPOs dominate under different conditions.

Karato et al. (2007)

08/09/2007 VLab-workshop 10

Influence of water content and stress on deformation fabrics of olivine (at T~1500-1600 K (asthenospheric temperature))

Katayama et al. (2004)

Seismic anisotropy in the asthenosphere, plume roots is likely caused by A-, or E- or C-type olivine fabrics depending on the water content. E- or C-type

fabric.In the typical asthenosphere, dominant fabric is likely E-type.

08/09/2007 VLab-workshop 11

E-type

A-type

a

a

b

b

c

c

A-type olivine fabric causes strong VSH>VSV anisotropy.E-type olivine fabric causes weak VSH>VSV anisotropy.

Karato (2007)

08/09/2007 VLab-workshop 12

Influence of water (hydrogen) on electrical conductivity in olivine

Wang et al. (2006)

1. A large amount of hydrogen can be dissolved in olivine.

2. Hydrogen diffusion is fast.

Hydrogen may enhance conductivity?(Karato, 1990)

Ele

ctric

al c

ondu

ctiv

ity,

S/m

Water content

08/09/2007 VLab-workshop 13

Conductivity in “normal” asthenosphere can be explained by a typical water content (~0.01 wt%).Conductivity in the central Pacific corresponds to “dry” olivine.

08/09/2007 VLab-workshop 14

Anomalies can be attributed to the dry asthenosphere in the central Pacific. Why is the asthenosphere of the central Pacific dry? What is the role of the Hawaii plume on modifying the properties of the central Pacific asthenosphere?

08/09/2007 VLab-workshop 15

Roles of a plume to modify the composition of the asthenosphere

• Plume: enriched (undepleted) = more water– direct mixing does not explain “dry”

asthenosphere.– plume=wet + hot-> deep melting->

depleted materials

08/09/2007 VLab-workshop 16

deep melting in a plume

(from Hirschmann (2006))

08/09/2007 VLab-workshop 17

In a plume column, melting occurs in the deep asthenosphere, providing “depleted” (dry)

materials to the asthenosphere.

08/09/2007 VLab-workshop 18

A plume will feed “depleted (dry)” materials to theasthenosphere due to deep melting--> cause for geophysical anomalies in the central Pacific?

Karato et al. (2007)

08/09/2007 VLab-workshop 19

Conclusions• Both seismic anisotropy and electrical

conductivity in the central Pacific are anomalous.

• These anomalies can be attributed to a low water (hydrogen) content in the asthenosphere in this region.

• A plume supplies “depleted (dry)” materials to the asthenosphere due to deep melting.

08/09/2007 VLab-workshop 20

Couvy et al. (2004)A

E

B

C

A three-dimensional fabric diagram of olivine

E

C

Karato et al. (2007)

08/09/2007 VLab-workshop 21

Katayama et al. (2004)

At low stress and high T, A-, E- or C-type olivine fabrics will be important.