January 26, 2010 1

Is magma generated due to retreating ice caps, likely to erupt? Constraints on lower-crustal stress

from InSAR

Andy Hooper1, Benedikt Ofeigsson2, Freysteinn Sigmundsson2, Halldór Geirsson3 , Pall

Einarsson2 and Erik Sturkell4

1Delft University of Technology2University of Iceland3Icelandic Meteorological Office4University of Gothenburg

Iceland Volcanic Systems

January 26, 2010 2

Iceland Volcanic Systems

January 26, 2010 3

Ice Loss and Magma Generation

• Estimated 14 million m3 extra magma generated per year by decompression melting, due to ice loss (Pagli and Sigmundsson, 2008)

• What proportion is likely to erupt?

• Can some be accommodated in crust due to stress change induced by melting ice?

1/26/2010 4

1/26/2010 6

Maximum deviatoric stress (tensile)

Instantaneous Stress Change

S

N•Integration of point loads over ice sheet

1/26/2010 7

Maximum deviatoric stress (tensile)

Instantaneous Stress Change

S

N•Integration of point loads over ice sheet

1/26/2010 8

Maximum deviatoric stress (tensile)

•Integration of point loads over ice sheet

•Ignores ductile response of lower crust (Maxwell time ~ 1-80 years?)

Instantaneous Stress Change

S

N

Ductile

Iceland Volcanic Systems

January 26, 2010 9

KverkfjöllVolcanic System

1/26/2010 10

2007-2008 Intrusion

(Relocated earthquakes, Jakobsdóttir et al, 2008)

1/26/2010 11

Maximum deviatoric stress (tensile)

Instantaneous Stress Change

Effect of relaxation through ductile flow below 10 km?

Ductile

1/26/2010 12

Decomposed InSAR + GPS

~Up East -6 cm 8.5-1.5 cm 6.5

1/26/2010 13

Boundary Element Model

• Planar dike with irregular outline

• Lithospheric

pressure

• Hydrostatic overpressure (=250 kg/m3),

• Uniform deviatoric

stress

• Prior probability density for strike and dip from seismicity

• Vertically varying elastic stiffness

• Bayesian Markov-chain Monte Carlo inversion

1/26/2010 14

Model Results

Maximum Likelihood solution (surface projection)

Plate spreading

1/26/2010 15

Model Results

Mean opening (surface projection)Distribution of dike position (profile)

1/26/2010 16

Comparison to DataData

Model

Residual

Up

East

-6 cm 8.5

-1.5 cm 6.5

1/26/2010 17

Tensile stress results

• Strike of maximum tensile normal stress: 129±9°(consistent with plate-spreading direction).

• Magnitude of deviatoric stress ~106 Pa, implying relaxation time >100 yr (loading ~104 Pa/yr).

• Implies that deviatoric stress due to ice-loss will build over time, leading to extra accommodation space for magma in lower crust.

Conclusions• Dike not emplaced perpendicular to maximum

tensile stress, as usually the case (=> significant shear on dike walls).

• Large deviatoric stress in lower crust consistent with plate-spreading.

• At least some of the extra magma generated will be captured in the lower crust, not erupted.

1/26/2010 18