Introduction to Seismology: Lecture Notes

04 May 2005

TODAY’S LECTURE
1. Fault geometry
2. First motions
3. Stereographic fault plane representation
4. Moment tensor
5. Radiation patterns

FAULT GEOMETRY
The fault geometry is described in terms of the orientation of the fault plane and the
direction of slip along the plane. The geometry of this model is shown in Figure 1.

x2 x3
North

Strike angle φ1
x1
δ n
d Dip
angle
Foot wall block Slip angle λ
n. d = 0

FAULT PLANE

Figure by MIT OCW.

Figure 1. Fault geometry used in earthquake studies. [[Adapted from Stein and Wysession, 2003]

∧ ∧
n is normal vector of the fault plane. d is slip vector which indicates the direction of
motion of hanging wall block. x1 axis is in the fault strike so φ is strike angle. The dip
angle δ gives the orientation of the fault plane with respect to the surface. The slip
angle λ gives the motion of the hanging wall block with respect to the foot wall block.
The motion is called left-lateral for λ = 0 , right-lateral for λ = 180 , normal faulting
for λ = 270 , and reverse or thrust faulting for λ = 90 . Most earthquakes consist of
some combination of these motions and have slip angles between these values. Note
that the basic fault types can be related to the orientations of the principal stress
directions. Actual fault geometries can be much more complicated. Such complicated
seismic events can be treated as a superposition of the simple events.

1
Introduction to Seismology: Lecture Notes
04 May 2005

FIRST MOTIONS
The focal mechanism uses the fact that the pattern of radiated seismic waves depends
on the fault geometry. The simplest method is the first motion, or polarity, of body
waves. Figure 2 illustrates the first motion concept for a strike-slip earthquake on a
vertical fault.

Figure by MIT OCW.

Figure 2. The relation between the first motion and fault geometry [Adapted from Stein and Wysession, 2003]

The first motion is compression when the fault moves “toward” the station and
dilatation for “away from”. A vertical seismogram records are upward for compression
and downward for dilatation. A problem is that the first motion on actual fault plane is
the same as that on the auxiliary plane which is perpendicular to fault plane, so the first
motions alone cannot resolve which plane is the actual fault plane. This is a fundamental
ambiguity in inverting seismic observations for fault models. We need additional
geologic or geodetic information such as the trend of a known fault or observations of
ground motion.

STEREOGRAPHIC FAULT PLANE REPRESENTATION

The fault geometry can be found from the distribution of data on a sphere around the
focus. A stereographic projection transforms a hemisphere to a plane. The graphic
construction is a stereonet (Figure 3).

2
 Introduction to Seismology: Lecture Notes
04 May 2005

Figure 3. A stereonet used to display a hemisphere on a flat surface

Using the stereonet, we can obtain the projection. Figure 4 illustrates the stereographic
projection.

Projection Projection Stereographic projection

T
sphere sphere of dipping plane

Horizontal plane Projection

plane

O O

Primitive
circle

Dipping plane
Spherical projection
Spherical projection of dipping plane
of dipping plane

Figure by MIT OCW.

Figure 4. The stereographic projection.

[http://hercules.geology.uiuc.edu/~hsui/classes/geo350/lectures/earthquakes/earthquakes.html]

We can also plot planes perpendicular to a given plane by rotating the stereonet and
finding the point on the equator 90° from the intersection of the plane with the equator.
Any plane through this point can be perpendicular to the plane. Thus, we can get the
focal mechanisms for earthquakes with various fault geometries. Figure 5-(a) shows the
three ideal focal mechanisms. Various focal mechanisms can be possible according to
various geometry.

3
 Introduction to Seismology: Lecture Notes
04 May 2005

STRIKE-SLIP FAULT

Left-lateral
on this plane
Right-lateral
on this plane
DIP-SLIP FAULTS

Thrust fault
Focal sphere side view

Image removed due to

copyright consideration.

Normal fault
Focal sphere side view

Vertical dip-slip
Focal sphere side view

Figure by MIT OCW.

(a) (b)
Figure 5. (a) The stereonet of different types of faults. (b) Focal mechanisms and some

seismograms for three different earthquakes. Compressional quadrants are shown shaded.
[Adapted from Stein and Wysession, 2003]

In reality, we plot the points where rays intersect the focal sphere, so that the nodal
planes can be found, considering the ray as compression (upward first motion) or
dilatation (downward first motion). Figure 5-(b) illustrates the focal mechanisms and
seismograms for three different earthquakes.

MOMENT TENSOR
To know the source properties from the observed seismic displacements, the solution
of equation of motion can be separated as below

4
Introduction to Seismology: Lecture Notes
04 May 2005

r r r r
u i ( x , t ) = Gij ( x , t ; x0 , t 0 ) f j ( x0 , t 0 ) (1)

u i is the displacement, f j is the force vector. The Green’s function G ij gives the

displacement at point
x that results from the unit force function applied at point x 0 .
Internal forces f must act in opposing directions, − f , at a distance d so as to
conserve momentum (force couple). For angular momentum conservation, there also
exists a complementary couple that balances the forces (double couple). There are nine
different force couples as shown in Figure 6.

3 2 3 2 3 2

1 1 1

M11 M12 M13

3 2 3 2 3 2

1 1 1

M21 M22 M23

3 2 3 2 3 2

1 1 1

M31 M32 M33

Figure by MIT OCW.

Figure 6. The nine different force couples for the components of the moment tensor. [Adapted from Shearer, 1999]

We define the moment tensor M as

⎡ M 11 M 12 M 13 ⎤
M = ⎢⎢ M 21 M 22 M 23 ⎥⎥ (2)
⎢⎣ M 31 M 32 M 33 ⎥⎦

M ij represents a pair of opposing forces pointing in the i direction, separated in the

j direction. Its magnitude is the product fd [unit: Nm] which is called seismic moment.

5
Introduction to Seismology: Lecture Notes
04 May 2005

For angular momentum conservation, the condition M ij = M ji should be satisfied, so

the momentum tensor is symmetric. Therefore we have only six independent elements.
This moment tensor represents the internally generated forces that can act at a point in
an elastic medium.

∧
The displacement for a force couple with a distance d in the x k direction is given by
r r
r r r r r r ∧ r ∂Gij ( x , t ; x0 , t 0 ) r
ui ( x , t ) = Gij ( x , t; x0 , t 0 ) f j ( x0 , t 0 ) − Gij ( x , t; x0 − x k d , t 0 ) f j ( x0 , t 0 ) = f j ( x0 , t 0 )d (3)
∂xk
The last term can be replaced by moment tensor and thus
r r
r ∂Gij ( x , t; x 0 , t 0 ) r
ui ( x, t ) = M jk ( x 0 , t 0 ) (4)
∂x k

There is a linear relationship between the displacement and the components of the
moment tensor that involves the spatial derivatives of the Green’s functions. We can
see the internal force f is proportional to the spatial derivative of moment tensor
when compared equation (1) with (4).
∂
fi ~ M ij (5)
∂x j
Consider right-lateral movement on a vertical fault oriented in the x1 direction and the
corresponding moment tensor is given by

⎡ 0 M 12 0⎤ ⎡ 0 M0 0⎤
M = ⎢⎢ M 21 0 0⎥⎥ = ⎢⎢ M 0 0 0⎥⎥ (6)
⎢⎣ 0 0 0⎥⎦ ⎢⎣ 0 0 0⎥⎦

where M 0 = µD s called scalar seismic moment which is the best measure of

µ is shear modulus, D = D ( x) / L is average

earthquake size and energy release,
displacement, and s is area of the fault. M 0 can be time dependent, so

M 0 = µD (t ) s (t ) . The right-hand side time dependent terms become source time

function, x (t ) , thus the seismic moment function is given by

M (t ) = M 0 x(t ) = µD(t ) s (t ) (7)
We can diagonalize the moment matrix (6) to find principal axes. In this case, the
principal axes are at 45° to the original x1 and x2 axes.

6
Introduction to Seismology: Lecture Notes
04 May 2005

⎡M 0 0 0⎤
M ′ = ⎢⎢ 0 − M0 0⎥⎥ (8)
⎢⎣ 0 0 0⎥⎦

Principal axes become tension and pressure axis. The above matrix represents that x1′
coordinate is the tension axis, T, and x ′2 is the pressure axis, P. (Figure 7)

X2 T axis X2 X1
P axis X2

X1 X1

Figure by MIT OCW.

Figure 7. The double-coupled forces and their rotation along the principal axes. [[Adapted from Shearer, 1999]

RADIATION PATTERNS
P-wave potential in spherical coordinate is given by
− f (t − r / α ) − f (τ )
φ (r , t ) = = (9)
r r
where α is the P wave velocity, r is the distance from the point source, and τ is time
residual. Therefore, the displacement field is given by the gradient of the displacement
potential u = ∂φ / ∂r
⎛ 1 ⎞ ⎛ 1 ⎞ ∂f (τ )
u (r , t ) = ⎜ 2 ⎟ f (τ ) + ⎜ ⎟ (10)
⎝r ⎠ ⎝ rα ⎠ ∂τ
The first term in the right hand side is near field displacement because of the decay as
1 / r 2 and the last term is far field displacement with the decay as 1 / r . When we
consider the relation between internal force and moment tensor given by equation (5),
we can find that the near field term has no time dependence but the far field term has
time dependence. The relations are given by
∂M ∂f ∂M
f ~ , ~ ~ M& (t ) (11)
∂x ∂τ ∂τ
Therefore, the near field term represents the permanent static displacement due to the
source and the far field term represents the dynamic response or transient seismic
waves that are radiated by the source that cause no permanent displacement. Figure 7

7
Introduction to Seismology: Lecture Notes
04 May 2005

represents the near and far field behaviors.

Near-field
M(t) Displacement

.
M(t) = u(t)

Far-field
Displacement
.u(t)
Far-field
Velocity

Figure by MIT OCW.

Figure 7. The relationships between near-field and far-field displacement and velocity. [Adapted from Shearer, 1999]

In spherical coordinates, far field displacement is given by

1
ur = M& (t − r / α ) sin 2θ cos φ (12)
4πρα 3 r
The first amplitude term decays as 1 / r . The second term reflects the pulse radiated
M& (t ) , which propagates away with the P-wave speed α and arrives at
from the fault,
a distance r at time t − r / α . M & (t ) is called the seismic moment rate function or
source time function. Its integration form in terms of time is given by equation (7). The
final term describes the P-wave radiation pattern depending on the two angle (θ , φ ) . At
θ = φ = 90 o , the displacement is zero on the two nodal planes. The maximum
amplitudes are between the two nodal planes. Figure 8 represents the far-field
radiation pattern for P-waves and S-waves for a double-couple source.

8
Introduction to Seismology: Lecture Notes
04 May 2005

Fault Normal

Slip Vector

Tension
Axis

Pressure Axis
Figure by MIT OCW.

Figure 8. The far-field radiation pattern for P-waves (top) and S-waves (bottom) for a double-

couple source. [Adapted from Shearer, 1999]

Reference
Stein S. and M. Wysession, An introduction to seismology, earthquakes, and earth
structure, Blackwell Publishing, 2003.
Shearer P.M., Introduction to seismology, Cambridge University Press, 1999.