using gps and insar to study tectonics, deformation, and earthquakes gps displacements, velocities...

Using GPS and InSAR to study tectonics, deformation, and

earthquakes• GPS displacements, velocities (and transients)

• InSAR displacements

Position to within about 1 cmGreat spatial coverageSensitive to vertical motionBad for high slopes or treed regionsSatellite not always looking at what you want when you wantNot so good for very large earthquakes

InSAR Displacements

McCaffrey et al., 2004http://ees2.geo.rpi.edu/rob/www/gps/gps2004.htm

GPS displacements and velocities

CGPS has sub-mm precisionCan detect rate and direction changesGood for rates and transientsNot so precise at polesVertical less precise than horizontalOnly point measurements

My research: Use GPS displacements, velocities, and transients to figure out how fault zones work at depth.

Information from seismology, geology, geochemistry and lab experiments is

also needed to build and verify models.

Method: numerical modeling.

GPS velocities in the eastern Mediterranean region

Fitting the GPS velocities with moving, rigid blocks

which ones are missing?

It’s not as simple as a bunch of rigid blocks...

D

D’

D D’

North Anatolian Fault Zone (NAFZ) has M ~7.5

earthquakes about every 300 years

fault slip!0 to 20 km depth

Non-seismic creep: 20 km depth to asthenosphere

This allows rigid translationof one plate past the other

coseismiccoseismic////postseismicpostseismic “interseismic”“interseismic”

dates and earthquake size from paleoseismology:

recent geological slip rates

dates and earthquake size from paleoseismology:

recent geological slip rates

Snapshots: GPS slip rates

Snapshots: GPS slip rates

Within an individual earthquake cycle

Within an individual earthquake cycle

Over many earthquake cycles

Over many earthquake cycles

Deformation around a plate boundary fault at different time scales

Deformation around a plate boundary fault at different time scales

Here is how interseismic deformation around a fault looks with InSAR

from Yuri Fialko’s Science paper of last summer

Colors scale with surface velocity

High strain rate means elastic stresses are building up fast

Next earthquake will be soon (or big)

Of course, last year’s EOSC 352 students knew this already....

Shear strain rate and strain rate axis orientations

From EOSC 352 Homework #5

SCEC GPS velocity field version 3

Lots can be learned from modeling interseismic deformation(beyond today’s scope)

The 1999 Izmit, Turkey earthquake:InSAR and GPS displacements

wrapped (arghh) interferogram

InSAR: like having thousands of not-too-

precise GPS sites, measuring just one

direction

Green’s function for surface displacements due to slip on a

subsurface dislocation

0 km

32 km

Slip (meters)

Slip along the NAFZ in the M = 7.5Izmit, Turkey earthquake

When the Izmit earthquake happened it

built up stresses in some areas

We can actually calculate this stress change and • model the Earth’s response to it• calculate changes in earthquake probability

on local faults

Earth models must connect episodic earthquakes with steady relative plate

motion

‘rigid’ down to asthenosphere with

localized shear zones?

creeping below mid-crust?

• This is done with creeping goo• Distribution and properties of this goo control the Earth’s response (i.e., surface motions, stress evolution)

Test the hypotheses using finite-element models

Models must reproduce the pattern and decay of

“postseismic” deformation

postseismic velocities one year after the Izmit earthquake

Model misfit to data after one year

Three years after the 1999 earthquake

Earth is still responding to the earthquake stress perturbation

Model performance: decay of postseismic velocitieslinear viscoelastic lower crust

linearly viscous shear zonevelocity-strengthening shear zone

Total modeled afterslip after a yearTotal modeled afterslip after a year

Distance along fault (km)Distance along fault (km)

Dep

th

(km

)D

ep

th

(km

)

About twice this slip would be required to fit the postseismic GPS displacements after 1 year.

About twice this slip would be required to fit the postseismic GPS displacements after 1 year.

Anatolia-Eurasia plate boundary (central NAFZ)

Eurasia

Anatolia

Moho

?

tricky rheology required?

Postseismic strain (and stressing) rates in the Marmara Sea

Postseismic strain (and stressing) rates in the Marmara Sea

coseismiccoseismic

900 days900 days

Southwest BC: Our local active faultsQueen Charlotte Fault M 8

Cascadia Subduction Zone Fault M 9.?

Faults in the subducting slab M < 7?

Shallow crustal faults M < 8

Henton et al., 2001

GPS velocity field: Vancouver Island