PLATE MOTIONS: BASIC CONCEPTS

North Eurasian
American 20 mm/yr
plate
plate

Pacific
plate
35 mm/yr

North
American Iceland Spreading
plate Center, Thingvellir

San Andreas Transform Fault

Carrizo Plain
 BASIC CONCEPTS: RIGID PLATES

Earth's outer shell made up of ~15 major rigid plates ~ 100 km thick

Plates move relative to each other at speeds of a few cm/ yr (about

the speed at which fingernails grow)

Plates are rigid in the sense that little (ideally no) deformation occurs
within them,

Most (ideally all) deformation occurs at their boundaries, giving rise to

earthquakes, mountain building, volcanism, and other spectacular
phenomena.

Style of boundary and intraplate deformation depends on direction &

rate of motion, together with thermo-mechanical structure
 In most places PLATE BOUNDARIES: GENERALLY BUT
we know general NOT FULLY KNOWN
plate boundary
geometry from
geology,
topography, and
earthquakes

Ideal plate
boundaries very
narrow
Many real plate
boundaries -
especially
continental - are
deformation
zones up to 1000
Gordon & Stein, 1992
km wide, with
motion spread
beyond nominal In some places: Indian Ocean. Mediterranean, NW
boundary Asia, etc. we’re still trying to figure out plate geometry
 NW ASIA
Not clear where
North America ?
boundary is
May be Okhotsk ?
plate distinct
from North ?
America ? ?
May be Amuria
plate east of
Baikal rift
distinct from ?
Eurasia
May be North
China plate

Wei and Seno, 1988

 EULER VECTOR
Relative motion between two rigid plates on the spherical earth can be
described as a rotation about an Euler pole

At a point r along the boundary

between two plates, with latitude
λ and longitude µ, the linear
velocity of plate j with respect to
plate i , v ji , is given by the Linear velocity
vector cross product

v ji = ωj i x r
r
r is the position vector to the
point on the boundary

ωj i is the angular velocity vector

or Euler vector described by its

magnitude (rotation rate) |ωj i | Stein & Wysession, 2003

and pole (surface position) (θ, φ)

Direction of relative motion is a small circle
about the Euler pole ω21
2 wrt 1
First plate ( j) moves counterclockwise ( right
handed sense) about pole with respect to
second plate (i).

Boundary segments with relative motion

parallel to the boundary are transforms, small
circles about the pole

Segments with relative motion away from the

boundary are spreading centers

Segments with relative motion toward

ω12
boundary are subduction zones
1 wrt 2

Magnitude (rate) of relative motion increases

with distance from pole because
|v ji | = |ωj i | | r | sin γ , where γ is the angle
between pole and site

All points on a boundary have the same

angular velocity, but the magnitude of linear
velocity varies from zero at the pole to a
maximum 90º away. Stein & Wysession, 2003
 GPS DATA RIGID
NORTH AMERICAN
PLATE ROTATING
ABOUT EULER
POLE

Direction follow
small circles

Rates increase
as sine of
angular distance
from pole

Velocities differ
from these in
nonrigid
boundary zone

Stein & Sella, 2002

 CARTESIAN COMPONENTS OF ANGULAR VELOCITY ω
AND LINEAR VELOCITY v
At a point r on the
plate boundary, with
latitude λ and
longitude µ, linear
relative velocity v , is
given by the vector
cross product

v =ω x r

r is the position
vector to the point on
the boundary

ω is the angular
velocity or Euler
vector described by
its magnitude
(rotation rate) |ω |
and pole (surface
position) (θ, φ)
 LINEAR VELOCITY TYPICALLY DONE AS EITHER
NS, EW COMPONENTS OR RATE & AZIMUTH
Scalar (dot)
product with
unit vectors
in NS & EW
directions

Gives
NS & EW
components
of linear
velocity

And hence
rate and
azimuth
CONVERGENCE - BOUNDARY TYPE
ALEUTIAN TRENCH
54 mm/yr CHANGES WITH
ORIENTATION
PACIFIC -
NORTH AMERICA

PACIFIC wrt
STRIKE SLIP - NORTH
SAN ANDREAS AMERICA
pole

EXTENSION -
GULF OF CALIFORNIA

Stein & Wysession, 2003

BOUNDARY TYPE CHANGES WITH ORIENTATION

EURASIA - NUBIA (west Africa)

NORTH EURASIA
AMERICA

EXTENSION STRIKE-SLIP
TERCEIRA GLORIA
RIFT TRANSFORM OBLIQUE CONVERGENCE
NORTH AFRICA
NUBIA

SMALL CIRCLE ABOUT POLE NUVEL-1

Argus et al., 1989
+ EURASIA wrt NUBIA POLE
 FINDING EULER VECTORS
Until recently, done by combining different types of
data from different boundaries

Spreading rates from sea-floor magnetic anomalies

Directions of motion from orientations of transform faults and slip

vectors of earthquakes on transforms and subduction zones

Problems with resulting geologic plate motion models:

No way to measure rates at subduction boundaries

Data average over different time scales:
-magnetic anomalies typically 3 Myr
-transform azimuths millions of years
-slip vectors seconds
Data only at plate boundaries
Inversion assumes rigid plates
 SPREADING RATES ANOMALY: 2’ 2 CENTRAL

FROM MAGNETIC
ANOMALIES

Match observed
profiles to synthetics
for different spreading
rates
Time resolution limited
by magnetic reversal
history GULF OF CALIFORNIA
PACIFIC - NORTH AMERICA
NUVEL-1 uses anomaly
2’ (3 ma) and so
averages over that time
Can’t go finer than
central anomaly
corresponding to last
reversal (780 ka)

DeMets et al., 1987

 TRANSFORM FAULT AZIMUTH FROM BATHYMETRY AND
STRIKE-SLIP EARTHQUAKE FOCAL MECHANISMS

Measure
azimuth from
bathymetry PACIFIC
High-
resolution
(Seabeam,
Gloria) is best
Averages over
millions of
years
Earthquake
mechanisms
also used, less
ANTARCTIC
precise

Stein & Wysession, 2003

 SUBDUCTION
NORTH AMERICA
AZIMUTH FROM
TRENCH THRUST
FAULT
MECHANISM SLIP
CONVERGENCE - VECTORS
ALEUTIAN TRENCH
54 mm/yr

PACIFIC

Common problem: for oblique (not trench

Fault plane
normal) convergence
Forearc sliver moves distinctly from both
plates
Slip vectors record motion of sliver relative Stein & Wysession, 2003
to oceanic plate, not major plate motion
 FINDING EULER POLE FROM RELATIVE MOTION DATA

Geometric
conditions:

Slip vectors and

transform faults
lie on small circles
about the pole, so
pole lies on a great
circle at right angles
to them

Rate of plate motion

increases with sine
of distance from
pole

Cox & Hart, 1986

INVERSE PROBLEM - FIND EULER VECTORS FROM DATA
POSE INVERSE
PROBLEM -
FIND EULER
VECTORS
FROM DATA

Set up model
vector m (Euler
vectors)
and data vector d
(observed rates
and azimuths)
Form partial
derivative matrix G
LEAST SQUARES
SOLUTION TO
INVERSE
PROBLEM

Find change in
model vector ∆m
from starting model
(Euler vectors)
using partial
derivative matrix G
to minimize
misfit ∆d to data
vector (observed
rates and azimuths)
 IMPROVED PLATE
Pre-NUVEL models assumed GEOMETRY: DISTRICT INDIA
single Indo-Australian plate
& AUSTRALIA
Deformation in Central Indian
Ocean shown by large
earthquakes & folding

New model: distinct Indian and

Australian plates separated by a
diffuse boundary zone perhaps
formed by Himalayan uplift

New model better fits focal

mechanisms & magnetics

Improved fit statistically significant,

so two plates resolved

Subsequent studies refined model

and show that India and Australia
have been distinct for at least 3
Myr and likely longer.

Wiens et al., 1985

SUCCESSIVE MODELS FIT USE MORE DATA & FIT BETTER

DeMets et al., 1990

More data Smaller misfit

NUVEL-1A GLOBAL RELATIVE PLATE MOTION MODEL
Plate motions averaged over past 3 Ma

Demets, Gordon, Argus & Stein, 1994

 EULER VECTOR OPERATIONS
To reverse sense of motion, use negative (same rate and with antipole:
negative latitude, longitude +180°)
ωjk = -ω kj

We assume that plates are rigid, so all motion occurs at their

boundaries. We can then add Euler vectors

ω jk = ω ji + ω ik

because the motion of plate j with respect to plate k equals the sum
of the motion of plate j with respect to plate I and the motion of plate i
with respect to plate k

Thus from a set of vectors with respect to one plate, e.g. i

ωjk = ω ji - ω ki

we form any Euler vector needed.

Operations easily done using Cartesian components
 DIFFERENT TYPES OF EULER VECTORS

GLOBAL EULER VECTOR - derived using all data from all plate
boundaries, assumes all plates are rigid.

BEST FITTING VECTOR - for a plate pair using only data from that
pair of plates' boundary

CLOSURE FITTING VECTOR - using only data from the other plates’
boundaries

Ideally, if the plates were rigid and data perfect:

- all three vectors would be the same.

-for three plates meeting at a triple junction, the best fitting vectors for
each of the three plate pairs would sum to zero.

These provide tests for plate rigidity

 SPACE GEODESY & GEOLOGIC
PLATE MOTION MODELS
GENERALLY AGREE
Plate motions over a few years
observed by space geodesy very
similar to predictions of NUVEL-1
or similar geologic models
describing average motions over
past 3 Ma

Hence plate motions are

generally steady, presumably
because viscous asthenosphere
damps episodic motions at plate
boundaries

However, in places NUVEL-1

and space geodesy disagree.

Why?
Robbins et al., 1993
 GLOBAL PLATE CIRCUIT CLOSURE
Because we only have certain types of data for some boundaries, others
are inferred by vector summation assuming rigid plates. In particular,
convergence rates at subduction zones are estimated by global closure,
combining data from all plate boundaries.

Predicted rate at
which the Cocos
plate subducts
beneath North
America depends on
measured rates of
Pacific-North
America spreading in
the Gulf of California
and Cocos-Pacific
spreading on the
East Pacific Rise.

DeMets et al., 1990

 SOME BOUNDARIES - NO DIRECT DATA
In some cases, such as relative motion between North and South America,
no direct data were used because the boundary location and geometry are
unclear, so the relative motion is inferred entirely from closure

Motion is poorly
known

EU

NA

NB

SA

Motions of plate pairs based on both rate

Wysession et al., 1995 and azimuth data are best known
 Direction & rate of Juan de Fuca plate
AT TRIPLE JUNCTIONS, subduction beneath North America found
WHERE THREE PLATES by combining:
MEET, ADD LINEAR Direction & rate of Juan de Fuca-Pacific
VELOCITY VECTORS spreading at Juan de Fuca ridge

Direction & rate of North America-Pacific

motion from Gulf of California and San
Andreas (transform) fault

Plate motion showed subduction, despite

no trench or thrust fault earthquakes
 CASCADE VOLCANOES
INDICATE JUAN DE FUCA
SUBDUCTION BENEATH
NORTH AMERICA

Mt Saint Helens
1980 eruption

USGS
 EURASIA - NUBIA (West Africa) motion

Primarily derived from small differences in spreading

rate & direction between North America - Eurasia and
North America - Nubia motion at Mid-Atlantic Ridge
Eurasia - Nubia
spreading at
NORTH EURASIA Azores Triple
AMERICA junction
NUBIA

Argus et al., 1989

Relative motions between plates are most important
ABSOLUTE
In some applications important to consider absolute PLATE MOTIONS
plate motions, those with respect to the deep mantle

In general both plates and plate boundaries move

with respect to the deep mantle

For example, assume Africa were not moving with

respect to the deep mantle. If so, as lithosphere is
added by spreading at the Mid-Atlantic ridge, both
the ridge and South America move westward
relative to the mantle.

Conversely, as the African plate lost area by

subduction beneath Eurasia in the Mediterranean,
the trench "rolls backwards”, moving south relative
to the mantle.

Increasingly, it seems that such motions may have

significant tectonic consequences

No direct way to measure absolute motions, need to

infer indirectly
NNR - NO NET
ROTATION
ABSOLUTE
MOTIONS

NNR reference frame obtained assuming no net rotation of the lithosphere

as a whole, so sum of the absolute motion of all plates weighted by their
area is zero
NNR reference frame similar to hotspot frame
Despite unresolved questions about the nature and existence of hot spots
and plumes, NNR reference frames often used to infer absolute motions
 ABSOLUTE MOTION CALCULATION
To compute absolute motions, recognize that motions in an absolute
reference frame correspond to adding a rotation to all plates

Use Euler vector formulation and treat absolute reference frame as

mathematically equivalent to another plate

For example, given NNR Euler vector relative to North America ωNNR-
NA its negative ω NA-NNR is the absolute Euler vector ΩNA for North
America in NUVEL-NNR reference frame

Hence absolute and relative motions related by

ωij = ΩI - Ωj
and linear velocity in absolute reference frame at point r on
plate i is
Vi = Ωi x r
RIDGES TYPICALLY MIGRATE WITH RESPECT TO THE
MANTLE
May have effects on topography, spreading process,
magma chemistry

Stein et al, 1977

 ABSOLUTE MOTION
Hot spot hypothesis assumes certain
linear volcanic trends result from plate FROM HOT SPOTS
motion over a hot spot, fixed source of
volcanism, causing melting in the
overriding plate.

If the overriding plate is oceanic, get

progression from active volcanism that
builds islands, to older islands, to
underwater seamounts as sea floor
moves away from hot spot, cools, and
subsides. Get topographic swell around
hotspot and volcanic age progression
away from it.

Direction and age of volcanic chain give

motion of plate with respect to hot spot.

Hot spot tracks beneath different plates

and assuming hot spots fixed with
respect to deep mantle (or move relative
to each other more slowly than plates),
yields hot spot reference frame.
Often assumed hot HOTSPOT / PLUME HYPOTHESIS
spots result from
plumes of hot material
rising from great
depth, perhaps core-
mantle boundary

Plumes would be
secondary convection
mode, ~ 5% of heat
transfer, bringing up
deep mantle material.

Would be important in
Earth’s thermal &
chemical evolution.

Would have tectonic

significance - heads
of new plume might
cause continental
breakup and flood
basalts
 HOTSPOT / PLUME CONTROVERSY

Concepts of hot spots and plumes are attractive and widely used, but the
relation between the persistent volcanism and possible deep mantle plumes is
under active investigation because of many deviations from what would be
expected:

Some hot spots move significantly

Some chains show no clear age progression

Oceanic heat flow data show little or no thermal anomalies at the swells

Seismological studies find low-velocity anomalies, but assessing their depth

extent and relation to possible plumes is difficult and controversial

Convection models of plumes rising from core-mantle boundary may not

correctly include pressure effects
 HOTSPOT TYPES:
MIDPLATE CONTINENTAL (Yellowstone,…)
MIDPLATE OCEANIC (Hawaii, Bermuda,…)
ON OR NEAR RIDGE (Iceland, Azores, Easter…)
YELLOWSTONE NATIONAL PARK, WYOMING USA

Type example of continental hot spot?

 AGE PROGRESSIVE VOLCANISM
Trend consistent with absolute motion of North America

Stein & Wysession, 2003

COMPLEXITY: additional
volcanic progression to
west - Newberry
volcanics

Proposed alternative:
forced mantle flow and
decompression melting
resulting from local plate
motions. Near subduction
zone upper mantle forced to
flow northwest because of
corner flow driven by
subducting plate. Yellowstone
and Newberry magmatism
follow these trends as fertile
mantle flows past residuum
and ascends (red-to-white
arrows).

Humphreys et al., 2000

 MIDPLATE OCEANIC
HOTSPOT: HAWAIIAN
EMPEROR CHAIN

Bend in the Hawaiian-Emperor

chain interpreted as indicating
Pacific plate changed direction
about 43 Ma, leaving bend as
plate moved over fixed hotspot
now under Hawaii

Mauna Loa
PROBLEM 1:
THE 43 Ma
“NONEVENT”

No evidence for
change in
relative plate
motions at 43
Ma, since
fracture zone
orientations
unaffected
PROBLEM 2:
HAWAIIAN
HOTSPOT
HAS NOT
BEEN FIXED

Fixed hotspot
would cause all
seamounts to
have same
paleolatitude
Hawaiian hotspot
actually drifted
southward
between 47 and
81 Ma

Tarduno and Cottrell

(1997)
 SUMMARY
Absolute motions can be defined relative to either No Net
Rotation or Hotspot reference frames
GPS data given in ITRF, designed to be like NUVEL-NNR
Absolute motions may have roles in tectonics but details
unclear
Hotspot/plume model has major problems and may need to
be discarded, but it’s not clear what the alternatives are