Water distribution in the Earth’s mantle Inferred from Electrical Conductivity�implications for the global water cycle

Shun-ichiro Karato

Yale University

Department of Geology & Geophysics

New Haven, CT

​

*

1

• Electrical conductivity is a useful sensor for the water content in the mantle.
• Water content is both radially and laterally heterogeneous.
• A large contrast in water content between the upper mantle and the transition zone suggests partial melting at ~410-km.

🡪 Most of the upper mantle is partially melted (melt fraction is small and does not affect properties except for seismic wave velocities in the deep upper mantle).

🡪 Partial melting at 410-km stabilizes the ocean mass.

​

​

*

2

Conclusions

How to infer the distribution of water from geophysical observations?

*

3

X

X

X

X

X

?

*

*

*: mostly for the upper mantle

Properties involving thermally activated processes are sensitive to water content.

Lab studies are more complete for electrical conductivity than for Q and LPO.

​

seismic wave velocity versus water content

*

4

Seismic velocities are insensitive to water content.

Influence of water on seismic discontinuities

oli

5

oli

wad

wad

*

Topography of discontinuities is insensitive to water content (at high T).

electrical conductivity from geophysical studies

Kelbert et al. (2009)

6

Tarits et al. (2004)

Ichiki et al. (2006)

Baba et al. (2010)

*

Dai and Karato (2009b)

wadsleyite

olivine, orthopyroxene, garnet, wadsleyite, ringwoodite

*

7

Sensitivity of electrical conductivity to T, Cw, fO₂, Mg#

8

*

🡪 Electrical conductivity is sensitive to Cw, but not to other parameters.

Testing the model for the upper mantle

*

9

pyrolite (olivine+opx+pyrope), SIMS water calibration

​

[Dai and Karato (2009)]

Electrical conductivity and water in the mantle

*

10

Mineral physics model

Geophysical model

*

11

X

Water content is layered (+ lateral heterogeneity)

🡪 Partial melting at ~ 410-km

What happens after 410-km melting?

*

12

a thick low velocity layer

(due to complete wetting)

Most of the upper mantle

is partially melted (with a

small melt fraction).

*

13

thick low velocity regions above the 410-km (Tauzin et al. 2010)

*

14

🡪 410-km partial melting stabilizes the ocean mass.

No mid-mantle melting

With mid-mantle melting

conclusions

• Water content (Cw) in the transition zone/upper mantle can be mapped from electrical conductivity observations.
• Mantle water content is layered.
• ~0.01 wt% for the upper mantle, ~0.1 wt% for the transition zone
• partial melting at 410-km
• a majority of the upper mantle is partially melted.
• a thick low velocity layer above 410-km
• Ocean mass is buffered by partial melting at 410-km
• Need for experimental studies on lower mantle minerals
• Need for geophysical observations for the lower mantle

​

15

*

16

*

17

*

*

18

Dixon et al. (2002)

19

Ito et al. (1983)

MORB source region (asthenosphere): well constrained (~0.01 wt%)

OIB source regions: water-rich (FOZO) (~0.1 wt%)

How are they distributed?

localized?

global (layered)?

*

Influence of element partitioning

20

Fe

H

wadsleyite

*

Meier et al. (2009)

puzzling results <-- due to insensitivity of seismological properties to water content?

<-- radial heterogeneity in water content?

<-- influence of kinetics on phase boundary topography?

​

Water-temperature distribution from VP,S and MTZ thickness

*

21

Water may affect seismological observations

• T-effect and water-effect on seismic wave velocities
• T-effect and water-effect on the phase boundary

​

22

h

V

*

*

23