M. Ritzwoller, N. Shapiro, S. Zhong, & J. Wahr University of Colorado at Boulder

Thermal structure of the upper mantle beneath Antarctica:

Implications for heat flux and visco-elastic rebound

Thermal structure of the upper mantle beneath Antarctica:

Implications for heat flux and visco-elastic rebound

M. Ritzwoller, N. Shapiro, S. Zhong, & J. WahrUniversity of Colorado at Boulder

Why mantle temperatures beneath Antarctica?

• Information about tectonic history.• Surface heat flux -- boundary condition for ice sheet and ice

stream modeling.• Temperature tied to rheology -- affects solid earth’s response to

icesheet loading/unloading, with possible feedbacks to icesheet stability, sea level and climate change.

Visco-Elastic Response to Plausible Glacial Loading/Unloading 10 ky BP

Visco-Elastic Response to Plausible Glacial Loading/Unloading 10 ky BP

“West Antarctic” Rheology “East Antarctic” Rheology

• Maximum difference in total depression ~300 m.• Magnitude depends mostly on lithospheric thickness.• Rate of change depends mostly on absolute viscosity.

Extent of Basal Water 6 My into Coupled Climate - Ice Sheet Simulation

Extent of Basal Water 6 My into Coupled Climate - Ice Sheet Simulation

“W. Antarctic” Heat Flux (75.4 mW/m2)

“E. Antarctic” Heat Flux (37.7 mW/m2)

David Pollard, Robert DeConto, Andrew Nyblade,Sensitivity of Cz Ant. Ice Sheet Variations to GeothermalHeat Flux, Submitted to Global and Planetary Change, 2004.

Inferring Upper Mantle TemperaturesInferring Upper Mantle Temperatures

Problem 1: Seismic models alone do not reveal temperaturesfaithfully.

Approach 1: Add heat flux information to calibrate upper mantletemperatures -- works well for other continents.

Problem 2: No heat flow data for Antarctica.

Approach 2: Extrapolate heat flux data from other continents -- works well elsewhere. Use seismic model to guide extrapolation.

Problem 3: Poor horizontal resolution.

Approach 3: “Antarctic Array” & new observing methods.

from Jaupart and Mareschal (1999)

Typical cratonic temperature profile from thermal modelers

Problem 1: Mantle temperatures from seismic models alone don’t agree well with physical models of mantle

temperature structure

Problem 1: Mantle temperatures from seismic models alone don’t agree well with physical models of mantle

temperature structure

Approach 1. Constraining seismic inversions to fit surface heat flux

Approach 1. Constraining seismic inversions to fit surface heat flux

Data: surface wave dispersion maps

100 sec Rayleigh wave group speed

Local dispersion curvesLocal dispersion curves

All dispersion maps: Rayleigh and Love wave group and phase velocities at all periods

Inversion of dispersion curvesInversion of dispersion curves

Monte-Carlo sampling of model space to find an ensemble of acceptable models

All dispersion maps: Rayleigh and Love wave group and phase velocities at all periods

Heat flux: Constraint in Uppermost MantleHeat flux: Constraint in Uppermost Mantleseismically

acceptable models

Inversion with the seismic parameterizationInversion with the seismic parameterizationseismically

acceptable models

Inversion with the seismic parameterizationInversion with the seismic parameterizationseismically

acceptable models

Simple thermal parameterization of the continental uppermost mantleSimple thermal parameterization

of the continental uppermost mantle

Lithospheric thickness and mantle heat flow in CanadaLithospheric thickness and mantle heat flow in Canada

Power-law relation between lithospheric thickness and mantle heat flow is

consistent with the model of Jaupart et al. (1998) who postulated that the steady

heat flux at the base of the lithosphere is supplied by small-scale convection.

Problem 2: Little heat flow data for AntarcticaProblem 2: Little heat flow data for Antarctica

Heat flow data base: Pollack et al., 1993

Approach 2: Extrapolate heat flow measurements to Antaractica

Approach 2: Extrapolate heat flow measurements to Antaractica

Extrapolation is guided by a global seismic model.Produces a distribution of values on a 2 deg x 2 deg grid world-wide.Works well elsewhere in the world.

Mean and st dev much higherin West Antarctica.

Two points in Antarctica

Mean of the Extrapolated DistributionMean of the Extrapolated Distribution

A A’

Results: Vs and temperature across Antarctica

Results: Vs and temperature across Antarctica

temp

100 km depth

W. Ant. Rift

E. Ant.Craton

So.Pole

• W. vs E. Antarctica: @100 km > 1000 deg difference. @ 300 km. > 400 deg difference.• Along Transantarctic Mtns:

@ 100 km, > 1deg/km laterally.• E. Antarctic cratonic core > 300 km thick, but

much thinner nearer the coast.

Other cratons lithospheric thicknessResults: Lithospheric thickness compared

with other continentsResults: Lithospheric thickness compared

with other continents

Results: Lithospheric thickness vs mantle heat flow compared with other

continents

Results: Lithospheric thickness vs mantle heat flow compared with other

continents

Other Continents Antarctica

Results: Lithospheric thickness vs mantle heat flow compared with other

continents

Results: Lithospheric thickness vs mantle heat flow compared with other

continents

Other Continents Antarctica

Anomalous Mantle Structure Beneath Antarctica?

Anomalous Mantle Structure Beneath Antarctica?

Locations with relatively thin lithospherebut low heat flux.

Cause?Erosion of the continental rootscaused by Mesozoic rifting?

orSimply poor lateral resolution?

Problem 3: Low resolution.Approach 3: Improve instrumentation and methods.

Problem 3: Low resolution.Approach 3: Improve instrumentation and methods.

• Improving instrumentation: “Antarctic Array” --a vision for seismology on an ice-boundcontinent. In the planning stages now,for initial deployment during IPY.

www.antarcticarray.org

• Develop seismic methods to extract information about earth structure without using earthquakes

as the source -- needed because significantseismicity is remote to Antarctica.

Use surface waves emanating from microseismsand atmospheric fluctuations to estimate Greenfunctions between receivers.:

20 sec periodRayleigh wave

Record section: Cross-correlate1 month of ambient noise, Z

Bandpass centered on: 20 sec

Green Functions by Cross-Correlating Ambient Noise in Antarctica?

Green Functions by Cross-Correlating Ambient Noise in Antarctica?

Summary and ConclusionsSummary and Conclusions

Understanding mantle temperatures beneath Antarcticais particularly important, due to potential ties to icesheet/stream stability and sea level & climate changethrough heat flux and mantle rheology.

Combining information about surface wave dispersion with heat flow information extrapolated from other continents

is providing new information about the temperature structure of the mantle beneath Antarctica.

New methods of surface wave analysis based on non-earthquakesources promise improved resolution.

Great advances will require a new generation of seismic instrumentation such as that being proposed as partof the new Antarctic Array initiative.

Results: Mantle heat flow compared with other continents

Results: Mantle heat flow compared with other continents

conversion between seismic velocity and temperatureconversion between seismic velocity and temperature

Method of Goes et al. (2000)

β =

μ

ρ

Elastic parameters for one mineral:

μ ( P , T , X ) = μ0

+ ( T − T0

)

∂ μ

∂ T

+ ( P − P0

)

∂ μ

∂ P

+ X

∂ μ

∂ X

K ( P , T , X ) = K0

+ ( T − T0

)

∂ K

∂ T

+ ( P − P0

)

∂ K

∂ P

+ X

∂ K

∂ X

ρ ( P , T , X ) = ρ0

( X ) 1 − α ( T − T0

) +

( P − P0

)

K

⎡

⎣

⎢

⎤

⎦

⎥

ρ0

( X ) = ρ0 X = 0

∂ ρ

∂ X

α ( T ) = α0

+ α1

T + α2

T

− 1

+ α3

T

− 2

ρ - density P - pressureμ - shea r modulus X – iron contentK – bulk modulus α - coefficient of thermal expansionT - temperature

Followi ng parameters ar e defined fro mlaboratory experiment:

ρ0 X = 0

, μ0

, K0

,

∂ ρ

∂ X

,

∂ μ

∂ X

,

∂ K

∂ X

,

∂ μ

∂ T

,

∂ K

∂ T

,

∂ μ

∂ P

,

∂ K

∂ P

, α0

, α1

, α2

, α3

Voig -t Reuss-Hill averaging for a combinati on o f minera :ls

ρ = λi

ρi∑

μ =

1

2

λi

μi∑ +

λi

μi

∑

⎛

⎝

⎜

⎜

⎞

⎠

⎟

⎟

− 1 ⎡

⎣

⎢

⎢

⎤

⎦

⎥

⎥

K =

1

2

λi

Ki∑ +

λi

Ki

∑

⎛

⎝

⎜

⎜

⎞

⎠

⎟

⎟

− 1 ⎡

⎣

⎢

⎢

⎤

⎦

⎥

⎥

λI i s the volumetr icproportion of minera l i non-linear relation

computed with the method of Goes et al. (2000) using laboratory-measured thermo-elastic properties of main mantle minerals and cratonic mantle composition

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

2. constraints from thermal data (heat flow)

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

2. constraints from thermal data (heat flow)

3. randomly generated thermal models

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

2. constraints from thermal data (heat flow)

3. randomly generated thermal models

4. converting thermal models into seismic models

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

2. constraints from thermal data (heat flow)

3. randomly generated thermal models

4. converting thermal models into seismic models

5. finding the ensemble of acceptable seismic models

Monte-Carlo inversion of the seismic data based on the thermal description of model Monte-Carlo inversion of the seismic data based on the thermal description of model

1. a-priori range of physically plausible thermal models

2. constraints from thermal data (heat flow)

3. randomly generated thermal models

4. converting thermal models into seismic models

5. finding the ensemble of acceptable seismic models

6. converting into ensemble of acceptable thermal models

3D seismic model3D seismic model

Where the cratons are?Where the cratons are?Geological data

(Goodwin, 1996)

No informationabout mantle structure

Geophysical dataHeat flow

(Pollack et al, 1993)

Un-evenly distributedOver Earth’s surface

Where the cratons are?Where the cratons are?Geophysical data

Inversion of heat flow(Artemieva and Mooney, 1998)

Un-evenly distributedOver Earth’s surface

Geological data

(Goodwin, 1996)

No informationabout mantle structure

Seismic surface-wavesSeismic surface-waves

global set of broadband fundamental-mode Rayleigh and Love wave

dispersion measurements (more than 200,000 paths worldwide)

Group velocities 18-200 s.Measured at Boulder. Phase velocities

40-150 s. Provided by Harvard and Utrecht groups

1. Data2. Two-step inversion procedure

1. Surface-wave tomography: construction of 2D dispersion maps

2. Inversion of dispersion curves for the shear-velocity model

•Provide homogeneous coverage in the uppermost mantle

•Provide sensitivity to the thermal structure of the uppermost mantle

Where the cratons are?Where the cratons are?Geological data

(Goodwin, 1996)

No informationabout mantle structure

Where the cratons are?Where the cratons are?Geological data

(Goodwin, 1996)

No informationabout mantle structure