GFZ Helmholtz Centre for Geosciences, Potsdam, Germany

Bernhard Steinberger

A new approach to estimate oceanic upper mantle viscosity

Earthquake Research Institute, Tokyo, Japan

Kiyoshi Baba

1

Methods to estimate mantle viscosity:

• Postglacial rebound (Haskell, 1935); mostly continental, most sensitive in upper mantle

2

2

• Fitting geoid (and dynamic topography, plate motions, anisotropy …)

(J. Braun, 2010)

3

3

• Slab sinking and matching subduction history with tomography

Čížková et al., 2012

4

4

Fitting hotspot tracks:

• Slow relative hotspot motions require high-viscosity lower mantle (Steinberger and O‘Connell, GJI, 1998)

​

• Sharpness of Hawaii-Emperor bend requires low-viscosity upper mantle (Richards and Griffiths, 1988)

​

5

5

• Hawaii-Emperor track explained by a combination of southward hotspot motion and Pacific plate motion change (Steinberger et al., 2004; Doubrovine et al., 2012; Torsvik et al., 2017)

​

Computed motion of plume in mantle

6

6

• Plume material emplaced below Emperor chain sheared off depending on viscosity structure: Analytical results

strong coupling

weak coupling

7

7

• Plume material emplaced below Emperor chain sheared off depending on viscosity structure: Analytical results

8

8

Numerical results computed with finite-element code ASPECT.

• Modelling domain is a cartesian box of length x width x depth = 4620 km x 3960 km x 660 km, i.e. with aspect ratio 7:6:1.
• Plume influx is prescribed at the bottom of the box.
• Time-dependent plate motion(s) prescribed at surface.
• Viscosity is both explicitly depth-dependent (for a given adiabatic reference temperature profile) and temperature-dependent for deviations from the adiabatic profile.
• Besides computing present-day temperature distribution in the mantle, we also compute dynamic topography and melt generation through time, and hence the resulting hotspot track, allowing further comparisons with observations.

9

9

red line

dark green

10

10

11

11

Stronger anomalies at shallower depth beneath Hawaiian chain (weak coupling model; dark green)

12

12

Example for a result of crustal thickness inferred from melt generation and displacement through plate motions

(see also Gassmöller et al., 2016, …)

13

13

The planned experiment with ocean bottom seismometers and electro-magnetometers:

• Deployment October 2025 with Japanese vessel Hakuhomaru
• Collection April/May 2026 with „Sonne“.

14

14

Vancouver

Honolulu

Tokyo

Taipei

Frankfurt

Mumbai

Kolkata

Hanoi

Istambul

Graz

Bengaluru

15

15