1
Determination of flow
patterns
in rocks
: an introduction and overview
Description of homogeneous flow
2.5
PT
a
.
Stages of progressive
homogeneous deformation.
Reference frame is attached to the lower
boundary of the
experiment (deformation

zone
boundary
)
.
b
.
Two subsequent
stages
are us
ed
to determine the velocity fi
eld at a particular time
.
c
.
Marker points in the flow pattern can be connected by
lines.
d
.
Fo
r each line stretching rate (
̇
;
̇
=l
1
/l
0
) and angular velocity
(ω)
are defined
.
e
.
̇
and ω are plotted in curves again line
orientation
(
0

360°;
note: dextral rotation is positive!)
.
S
pecial
di
rections are:
ISA = instantaneous stretching axes
: two lines
(in 2D)
,
along which the stretching rate (
̇
) has its
maximum
and
minimum values; they are always orthogonal.
Irrotational
material lines
:
can have any orientation
.
̇
k
: t
he amplitude of the
̇

curve
is
a mea
s
ure of the
st
rain rate
.
Vorticity: the
elevati
on
of the s
ymmetry line
of the
ω

curve
.
f
.
Orientations of ISA and irrotational lines can be read from graphs
.
2
Types of flow
2.6PT
Isochoric flow: if the stretching rate
(
̇
is symmetrically arranged with
respect to the zero stretching

rate
axis, no
area change is involved
in the flow
(area in
/de
crease involves extra stretch in all
directions: the curve is shifted upwards
/downwards
)
Coaxial flow: if in a reference frame fixed to the ISA
,
the angular velocity curve is symmetrically
arranged with respect to the angular velocity axis, no ‘bulk rotation’ is involved in the flow, and
lines of zero angular velocity (irrotational lines) are orthogonal.
Flow is said to be coaxial
because a pair o
f lines that is irrotational is parallel to the ISA. This pure shear flow has
orthorhombic
shape symmetry.
General n
on

coaxial flow: if all material lines are given an identical extra angular velocity, the
angular velocity curve is shifted upwards (dextral
rotation; see definition of + and
–
in Fig.
2.
PT
5) or downwards (sinistral rotation). In both
(sinistral, dextral)
cases, flow is non

coaxial
since irrotational lines are no longer parallel to the ISA.
All non

coaxial
flows
have a monoclinic
symmetry.
Vor
ticity: the deviation of the angular velocity curve from
zero angular velocity
is a measure of the
irrotational character of the flow.
Simple shear: when the angular velocity curve touches the zero angular velocity axis and only one
irrotational line exist
s
.
Flow descriptions
(see parameter definitions in Fig. 2.6
PT)
α
k
: angle between one of the ISA and the side of the shear box (shear zone boundary)
̇
k
=
̇
1

̇
2
= ω
1

ω
2
; a measure of strain rate (amplitude of the stretching
/rotation
rate curve)
W
k
= (ω
1
+ω
2
)/
̇
k
; kinematic vorticity number
A
k
= (
̇
1
+
̇
2
)/
̇
k
;
a measure of areal change with time
E.g.: s
imple shear: W
k
= 1, A
k
= 0.
(
mit
ω
:
W
k
= 1+
0/1

0
= 1; ω
1
= x, ω
2
= 0; all material lines rotate
e.g. dextrally
; with
̇
k
: take scales, e.g.
ω
1
=
2
, ω
2
= 0
,
̇
k
= (1

(

1 )= 2
)
pure shear: W
k
= 0, A
k
= 0
3
Concept of vorticity and spin
Box 2.4PT
a
.
If
the velocity
of
a river
is fastest in the
middle
,
paddle whee
ls in the r
iver will ro
t
a
t
e in
opposite
direc
t
ion at
the
si
de
s
, but will not rotate in
t
he middle; th
ey reflect
the vor
t
icit
y
of
flow in
the river at
t
hree di
f
fe
rent sites
.
b
.
Vorticity is defined as the s
um of the angular
v
elocit
y
with re
spect
to ISA of a
n
y pair of
or
t
hogonal
ma
t
erial
line
s
(
s
uch a
s
p
and
q)
; additional ro
t
a
t
ion of ISA
(and all
t
he other line
s
and vector
s)
in
an
external ref
e
rence fra
me
is known as s
pin
.
4
Types of deformation (not flow)
2.7PT
Homogeneous
deformation
(instead of flow)
can be envisaged by the distribution patterns of stretch
and rotation
.
Note that deformation is normally composed of strain (which describes a
change in
shape) and a rotation; thus deformation ≠ strain.
a
.
Two
st
age
d of a
de
f
o
rmat
ion
sequence.
b
.
T
he
def
ormation p
attern
.
c
.
Sets of marker
points c
an be
connec
ted b
y
materia
l
l
ines and the rotation (
r
) and stretch (s) of
each
l
ine
is
monitor
ed.
d
.
The
s
e
can be p
l
otted
agai
ns
t
initia
l or
ientation of
the
li
n
e
s.
In the
curves,
p
r
incipa
l
s
trai
ns can
be
distingu
ished
.
e
.
Finit
e def
ormation as deduced from
the
s
e
curves
contai
ns el
e
ments
of
st
r
ain and
rotation
(
ρ
k
)
.
β
k
def
in
es the orienta
t
ion of a material
l
ine i
n
the
undeform
ed state that i
s
to
becom
e paralle
l
to the
l
ong
a
xi
s
of the
s
train
ellipse
i
n
the
defo
rm
ed
s
tate
.
5
Progressive and finite deformation
2.8PT
H
omogeneous
finite
deformation carries no information on the deformation path or on progressive
deformation. However, the stretch and rotation
history
of material lines does depend on the flow type
by which it accumulated. In the case of
inhomogeneous
deformation on s
ome scales, as is common in
deformed rocks, pure shear and simple shear progressive deformation can produce
distinct, different
structures.
The difference is best expressed in the symmetry of fabric structures.
The ef
fect of
def
or
mation
hi
s
to
r
y
a
.
Two
identical squa
r
es of material with two
m
a
r
ker
line
s (
red a
n
d gree
n
l
ine
s)
are defo
rm
ed
u
p
to
the same
fi
nite
str
ai
n
va
lue in simpl
e shea
r
and pure
s
hear progre
ssi
ve deformation
, res
pe
c
tive
ly
.
The
initia
l
orientation of the
squares
is
chosen
such
th
at the
s
hape and orie
nt
ation of defo
rmed
s
quare
s
is
identic
al
.
b
.
The
finite
stretch and
r
elative o
rientation of both marker
l
ine
s is
ide
ntica
l
in both
cases,
but
the
h
istory
of
st
retch and rotation of each l
ine
i
s
different.
c
.
C
ir
cul
a
r
diagra
m
s
show
the
di
s
tribu
ti
on
of all mate
ri
al
l
ines in
the
s
quare
s of a
. O
rn
amenta
tion
shows whe
re
li
ne
s
are
s
h
o
rte
n
ed
(s),
exte
nd
ed
(e) or
f
i
rst
shortened,
the
n
extended
(se)
for each
s
tep
of
progre
ss
ive
deformation. The orie
n
tation of
I
SA
is
i
ndi
cated
.
6
Concept of the
fabric attractor
If flow (patterns like 2.6PT) works on a material for some time, material lines (e.g. long axis of finite
strain) rotate toward an axis, which coincides with the extending irrotational material line; this axis
‘attracts’ material lines in
progressive deformation.
2.9PT
In
both pure shear and simple shear deformation, material lines rotate
towards
and concentrate near
an attractor direction. The line is
the fabric attractor
(FA).
Both foliation and lineation rotate
permanently
toward
the attractor.
7
Concept of fabric attractors: flow apophyses or flow eigenvectors
(e.g. Passchier 1987)
In a body deforming by homogeneous isochoric plane

strain flow, the rate of displacement (‘X
i
) of
particles at X
i
in a fixed Cartesian co

ordinate system
is
described by the velocity gradient tensor L'
(Malvern 1969) as:
[
]
[
]
[
]
where
W
is the vorticity of the flow and
S
is a scalar defining the stretching rate of the fl
ow. L' can be
expressed as the sum of a symmetric tensor D' and an anti

symmetric tensor W':
[
]
[
]
[
]
The eigenvectors d
i
of D' are the orthogonal instantaneous
stretching axes of the
flow and the
eigenvalues d
1
, d
2
and d
3
are instantaneous stretching rates with values 0,
S/2
and

S/2.
The vorticity
vector of the flow is parallel to d
1
(Fig. 1).
Fig.
1 (Foss
en; Passchier ’87).
Left: d
efinition of the vorticity vector. Right:
s
chematic representation
of a homogeneous plane

strain non

coaxial flow with vorticity number W
k
= 0.76 described by L'.
X’
i
—
co
ordinate system, d
i
—
eigenvectors of D', inst
antaneous stretching axes. l
i
—
eigenvectors
of
L',
=
flow apophyses. Ornamented surfaces
are traced by material lines parallel to X
’
i
. Arrow around X
’
i
indicates
sense of shear.
8
Eigenvectors l
i
of L' do not usually coincide with those of D', except for pure shear. l
1
is parallel to d
1
,
and l
2
and l
3
lie in the X’
2
–
X’
3
plane
,
symmetrically arranged with respect to d
2
and d
3
(Fig. 1)
(Bobyarchick 1986, Passchier 1986). The
instantaneous
non

coax
iality of the flow can be expressed
by a vorticity number:


,
which for isochoric plane

strain flow equals the kinematic
vorticity number of Truesdell (1954):
‘
√
(
)
.
The planar surfaces through
l
1
,
l
2
and
l
1
,
l
3
, defined as
eigenvector planes (Passchier 1986)
,
have
special properties
for all types of instantaneous steady flow and for
progressive steady flows following
integration of L'.
→
For
W
k
≤
1 all particle paths in the flow defined by L' follow
hyperboloid
curves (Fig
. 1)
,
which
approach the eigenvector
planes asymptotically; particles within the eigenvector
planes approach or
depart from the X
’
1
axis along
paths within the planes. For these reasons, eigenvectors
l
2
and 1
3
have
been named flow apophyses by Ramberg
(1975a, b). For plane

strain flow, particle paths within
the
eigenvector planes are the only straight orbits in the
flow (Fig. 1
(no flow in X
1

direction)
).
[
Explanations
: (i)
the term asymptotic means approaching a val
ue or curve arbitrarily closely; (ii)
a
hyperbola is defined as
the loci of all points within the drawing plane, for which the difference of the distances to the given points F1 and F2 is
constant]
Flow apophyses are theoretic
al
planes that compartmentalize the flow pattern. The number of possible
apophyses in a flow system ranges from 1

3. For planar deformations (2D) the maximum number is 2.
Particles cannot cross
flow
ap
ophys
e
s.
hyperbola definition
eigenvector plane
9
Poles of rotating passive lines move along trajectories (great circles) in
a
stereoplot during simple
shear
(
Fig. cf.
Foss
en)
. There is only one flow apophysis (AP), which parallels the x

z plane.
(
Explanation
: here only ω is visualized, the
lines does not stretch or shorten; just rotate a line and trace its end points)
Stereographic projection of pole trajectories of passive lines during progressive pure shear. AP =
flow
apophyses
. The particles move in four quadrants separates by two orth
ogonal flow apophyses.
10
Stereographic projection of the pole trajectories of passive lines during
sub

simple shear
(=
general
shear
). The particles move in four compartments separated by two flow apophyses at an angle
between 0
°
(3D simple shear) and 90
°
(3D pure shear
)
.
11
Relationship between W
k
and pure and simple shear
.
Fig. 1FB
Scale relations between the W
k

value and percent simple shear. Zones of pure shear dominated,
general shear, and simple
shear dominated deformations (from Forte and Bailey, 2007).
Relationship between kinematic vorticity number W
k
and
components of pure and simple shear for
instantaneous 2D flow; pure
and simple shear componen
ts make equal contributions to flow at
W
k
=
0.
71 (arrow). W
k
=
cos
α
, where
α
is the angle between flow
apophyses (Bobyarchick 1986), and
varies from 0
°
(simple shear) to 90
°
(pure shear). Relative proportion of pure shear to simple shear is
given
by
α/
90
°
; at
α
=
45
°
there should be an equal contribution of pure and
simple shear (45°/
90
°
=
0
.
5); for simple shear there is no pure shear
component (0
°/
90
°
=
0).
(
Explanation
: W
k
= cosα, with α as the angle between the AP is the most common definition of
vorticity
; actually W
k
= 0.7
07
= cos45
)
calculation
below
12
D
eformation in various W
k
fields
. S
imple shearing produces displacement with
little
shortening/thinning, general shear is a combination of both
thinning and displacement, and pure shear
zones experience
a great deal
of thinning with modest displacement relative to
overall strain
(Fig.
2FB)
.
Fig. 2FB
a. The initial condition before high

strain zone formation. Dikes, points A and B, and the strain ellipse
are provided as reference.
b. Pure shear deformation with two eigenvectors orthogonal to each other. White arrows indicate zone
parallel stretching.
c. General shear deformation with a W
k

value of 0.9. The acute angle between the two eigenvectors is
correspondingly 26°
(cos26 = ~0.
9)
.
d. Simple shear deformation with one eigenvector.
(
Explanation
:
thickening/
thinning
(e.g. of dikes)
and lateral flow of material is not shown)
13
Movement of axially

symmetric rigid objects
(e.g. Passchier 1987)
General theory
—
i
s
ochoric
plane

strain flow
.
E
quations of the movement of rigid objects
in
homogeneous flow (e.g. Jeffery 1922) are least complex for axially symmetric ellipsoidal objects. The
axial ratio of such objects can be expressed by a parameter B, where ‘a’ is the length al
ong the
symmetry axis and ‘b’ the radius in the circular section (Fig. 2
; in this Fig. a is larger than b, a prolate
object
):
.
B can represent a material line (B = 1), a prolate ellipsoid (0 < B < 1), a sphere (B = 0), an oblate
ellips
oid (

1 < B < 0
; a flat disc with the short axis along the
object symmetry axis
(
OSA
) or a material
plane (B =

1).
If an
internal reference frame X
i
is chosen with X
1
fixed to the
OSA
, the rotation of the object
with
respect to the external reference frame X
’
i
is given
by the Eulerian angles
θ, φ, ψ, described by a
rotation tensor R.
T
he orientation of an axially symmetric object
is
defined by two angles only
,
which
reduces
φ
to the azimuth and
ψ
to the plunge of
the OSA
(Fig. 2a).
Fig. 2P
a. Reference frame for rotation of axially symmetric rigid ellipsoids. X’
i
and X
i
: external and internal
co

ordinate systems. φ and ψ: azimuth and plunge of object symmetry
axis (OSA; this is the axis
that contains the vorticity

vector component ω
1
, the angular velocity components of OSA around
X
1
.
̇
̇
̇
rate of change of Eulerian angles (which define the object orientation),
v:
instantaneous displacement rate of OSA
on a sphere around cent
er
of co

ordinate systems.
b. Stereographic projection of OSA and v.
D' and W' can be expressed in terms of the
internal reference frame by:
D
= RD'R
T
and
W = RW'R
T
.
14
The instantaneous angular velocities of the object around the
X
i
axes are (Freeman 1985):
ω
1
=

W
32
= ((S∙W
k
)/2)
∙
cosφ
∙
cosψ
ω
2
=

W
13

B∙D
32
= ((B∙S

W
k
∙S)/2))
∙
cosφ
∙
sinψ
ω
3
=

W
21
+ B∙D
21
= (S/2)·sinφ·(W
k
+ B·cos2ψ).
ω
l
describes the axial rotational velocity of the
object around its symmetry axis, zero if the OSA
lies in
the X
’
2
–
X
’
3
plane
(=flow plane)
and a maximum if it lies parallel to X
’
1
.
T
he instantaneous
displacement rate of the OSA along a sphere around the
center
of the co

ordinate
systems
,
calculated from
the ω
equations (
above
) for a number of OSA posi
tions
, is
plotted
as a
vector
v
in stereograms (Fig
s
.
2,
3
P
). S appears as
a
constant and its magnitude does not
influence the shape
of the flow patterns: in
Fig. 3P,
S = 10. A half stereogram is
sufficient for presentation of the
movement pattern
because
of its bilateral symmetry.
Fig. 3P
Wulff projection of displacement rate vectors of OSA at 343 positions. 10° intervals of φ and 5°
intervals
of ψ
. W
k
= 0.4, B = 1.0
(material line)
. The vectors in Fig. 3 describe the
instantaneous
movement
of the OSA.
(
Explanation
:
there two stable positions, somewhere intermediate between X’
3
and X’
2
).
15
OSA

trajectories for S = 10 are shown for a complete range of B and W
k
values in Fig. 4P.
Fig. 4P
Movement patterns of the object symmetry axis (OSA) for the range of possible W
k
and B values.
Upper hemisphere Wulff projection.
Representative flow types shown along the top, representative
object shapes on the left. Symbols in stereograms: open circles: sources of OSA; open squares: sinks
of OSA; solid circles: transient stationary positions of OSA; bold lines at W
k
=


:
planes
of transient
stable positions, d
i
:
eigenvectors
D’
(instantaneous stretching axes).
l
i
:
eigenvectors of
L'.
(
Explanations
:
e.g,
spheres are irrotational in 3D pure shear and rotate infinitely in all other deformations; there
are
one
to
three
stable positions)
Stable positions according to B and W
k
:
a. W
k
<


3 stable positions: a transient position parallel to X’
1
and a source
–
sink pair in the X’
2

X’
3
plane. Source and sink are symmetrically arranged with respect to d
i
, and their actual po
sitions depend
on the sense of vorticity, B, and W
k
. At W
k
= 0 source and sink coincide with eigenvectors of D' and L'
at 45° to X’
2
and X’
3
. For any W
k
, material lines (B = 1) follow a trajectory
towards a sink parallel to
l
2
;
the normal to material
planes (B =

1) approaches a sink normal to l
2
. The
OSA of objects with
other B

values have a source and
sink at an angle
±β/2
from X
’
3
defined by:
cosβ =
W
k
/B.
(
E
xplanation:
β gets smaller with higher W
k
; β gets smaller with higher B)
.
b
.
W
k
=


. A 'stab
le plane' exists through X
’
1
and
X’
3
(B > 0) or X
’
2
(B < 0). All OSA within this
plane are
in transient equilibrium. Outside the stable plane
,
OSA
rotate along
straight planar paths
towards X’
3
or X’
2
. If
W
k
= 1, the stable plane coincides with the shearing plane of simple shear flow.
c
.
W
k
>


(these are objects with small axial ratios
).
Only a transient position exists along
X
’1
. In all other
positions the OSA rotate continuously
.
For general flow and axi
ally non

symmetric objects
,
the theory predicts that sink positions exist.
coordinates
X3
X2
X1
rotation
16
Application.
W
hich objects had reached a stable position, and
which
we
re still
rotating as deformation
stopped?
Rigid
porphyroclasts in a ductilely deforming matrix often
recrystallize along their margins
and produce tails of recrystallized material
that
stretch
es
out into the matrix.
Immobile objects with a
symmetry axis at a sink in the
X
’
2
–
X
’
3
plane
(normal to the flow plane)
of the flow show
straight
σ

type
tails with '
stair

stepping'. E
ll
ipsoidal objects rotate by periodic
accelerations and decelerations
,
which
also influence
recrystallization rates. Tail development will be
significant during the period of slow
rotation when the
long axes of the object are near the X
’
1
–
X
’
3
(
shear)
plane, and
these tails will
become distorted to
δ

types during the
subsequent fast rotation when the long axes are near the
X
’
1
–
X
’
2
(normal to the shear)
plane.
δ

type tails can also develop around
spherical rotating objects if
recrystallizati
on is slow.
Thus, complex and

type clast

tail systems are considered
to be indicative of
permanently rotating objects,
while objects with straight
σ

type tails are probably at
stable positions.
R
ecrystallized tails will reach parallelism with
l
2
at high
finite strains
(the outer tails do not rotate
anymore)
.
X3
X2
X1
17
Tails of recrystallized material will tend to rotate
towards the extensional eigenvector
l
2
of L'
throughout
the deformation. The angle
η
between the M
x
(long)

axis of
an
irrotational rigid object,
which lies at a sink
in the
X
’
2
–
X’
3
plane, and
l
2
or the straight domain of the tail
away from the object
is a function of W
k
and B
*
only:
√
√
}
, with
and M
x
and M
n
are the object’s long
and short symmetry axes in the
X’
2
–
X’
3
plane
.
Fig
.
8
P

a
shows η
for the entire range of
W
k
and B
*
values
.
η
increases with decreasing
B
*
up to the
value B
*
= W
k
, the 'cut

off point'. At still
lower
B
*
values, the objects are rotatin
g permanently
.
W
k
can be derived from these graphs in two ways (Fig.
8
P

b
): (a) from the value
B
*
c
ri
t
= W
k
,
which
separates the
σ

type immobilized part of the clast
population in a rock from the complex and
δ

type
rotational
part, and (b) theoretically from
η
and B
*
for
individual objects if sub

parallelism
of
recrystallized
tails with
l
2
during the last stages of the flow can be
proven.
Fig. 8P
a. Curves for the orientation of stable sink
positions of rigid objects in the X’
2
–
X’
3
plane of flow for a range of W
k
values, η
—
angle between the long axis of an object
cross

section and l
2
, marked by
recrystallized tails at high finite strain.
B
*
—
shape
factor. If W
k
> B
*
no stable
sink positions exist.
b. Example of the expected geometries of rigid

object recrystallized tail systems in cross

section parallel to X’
2
–
X’
3
at W
k
= 0.5. At
low B
*
values, objects rotate permanently
and generate δ

type and co
mplex tails. At
high B
*
values, to the right of the 'cutoff
point', objects have their long axis at a
stable sink positi
on and generate σ

type
tails. η
decreases with increasing B
*
.
freely rotating clasts
B*
crit.
for W
k
=0.1
X3
X2
X1
flow apophyses
18
T
he
following requirements should be met
for an application
. (1)
The fabric and general setting of the
samples should
indicate that deformation was reasonably homogeneous
on the scale of the sample
.
Samples from the
limb of a major fold are unsuitable, but
samples from a
straight, regular shear zone
may be useful. (2) Grain size
in the matrix should be significantly smaller than the size
of the objects
in order to make reasonable the assumption
of homogeneous flow. (3) High finite strains
accumulated
by homo
geneous flow are required to
rotate sufficient objects towards sink positions. (4)
Object shape
should be regular and closely approach
orthorhombic shape symmetry. Deviations of object
shape
from an ellipsoid are not expected to influence the
position of s
inks (Bretherton 1962). (5) A sample
should contain a large number of spatially well dispersed objects
with variable B
*
values.
Fig. 10P shows an application: All porphyroclasts
in a mylonite
with
approximately orthorhombic
shape symmetry and two
symmetry
axes in the plane of the section were analyzed: B* values in the
inferred
X
’
2
–
X
’
3
plane and
η
,
the angle between the long axis and the trace of the tail
away from the
clast, were plotted. Because of the high inferred finite strain values, the
tails are as
sumed to parallel
l
2
,
at least during the last
stages of deformation. Nearly all complex and
δ

type clast

tail systems (open
circles) plot left of the B* = 0.6 line. A dense cluster of
σ

type systems (dots), which dip in the
opposite
direction to the stair
stepping, plot to the right of this
line. The solid curves represent
theoretical
η
values for W
k
= 0.6 and 0.7 from Fig. 8
P
.
Data plot of K

feldspar clast

tail systems in
quartzite
mylonite
,
St. Barth
e
1emy Massif,
France, from
thin sections normal to the
inferred vorticity vector of the flow.
Orientation of object long axes
with respect to
trace of recrystallized tails, in degrees, plotted
against
B*, as in Fig. 8
P
. Open circles, clast

tail systems with complex or
δ

type
geometr
y
indicative of permanent rotation; dots,
σ

type
clast

tail
systems.
19
M
ethod of rotating porphyroclasts
—
elaborations
General considerations
.
The behavior of porphyroclasts depends on the flow type (W
k
; i.e. orientation
of the eigenvectors) and the axial ratio of the
porphyroclasts
(B
*
). During deformation, rigid
porphyroclasts with an aspect ratio greater than a critical value rotate towards material attractors
nearly coincident with the flow eigenvect
ors. Porphyroclasts with axial ratios smaller than this critical
value rotate independent of the bulk flow attractors. In the case of general shear, porphyroclasts either
rotate backwards or forwards towards the flow attractors. Porphyroclasts that
forward
rotate reach a
stable end position within the obtuse angle
field
(β)
between the two eigenvectors
.
Porphyroclasts that have reached a stable position can be identified by sigma tails
.
Sigma tails form
during
slow rotation
, which occurs as
porphyroclasts
approach
their stable posi
tions. Backward rotated
clasts
are identified by long axes orientations antithetic to
the overall sense of shear and the presence
of synthetic sigma
tails on antithetic porphyroclasts. Synthetic si
gma tails on
antithetically oriented
porphyroclasts must be produced through
back rotation of the porphyroclast, because forward rotation
of
an antithetic porphyroclast would inhibit tail formation
(Fig. 4).
Fig. 4FB
Tail formation on rotating porphyro
clasts.
a. Sigma tails growing on a back

rotating porphyroclast.
b. Growth of sigma tails is inhibited by forward rotation of a porphyroclast with a long axis oriented
antithetic to the sense of shear. Presence of sigma tails on antithetically oriented
porphyroclasts is
evidence of back rotation.
Porphyroclasts will only reach stable end positions
if sufficient amounts of strain have accumulated.
obtuse angle field
20
Orientations of vorticity vectors and fabric asymmetries
.
Orthorhombic deformation s
ymmetries are
charact
erized by
parallelism between the finite strain elements (foliation and lineation) and the high

strain zone boundary and an abundance of symmetric structures (W
k
= 0). Monoclinic deformation
produces an angular discordance between the foliation and shear
zone boundaries as well as
asymmetric structures normal to the foliation and parallel to the elongation lineation (Fig. 3FB).
Triclinic deformation is characterized by asymmetric structures on sections both normal and parallel to
elongation lineations. In
zones of heterogeneous triclinic deformation, elongation lineations may vary
between strike

parallel and dip

parallel orientations.
The vorticity vector is referenced relative to the plane orthogonal to the vector. Maximum rotation
within the flow occurs
within this vorticity profile plane (VPP) and the plane contains the shear
direction. Relations between the VPP, lineation, and foliation depend upon the geometry of the shear
zone (Fig. 5FB). General shear zones should have maximum (fabric) asymmetry in
the VPP.
Fig. 3FB
a. Block diagram of monoclinic shear. The
transport direction and correspondingly
the vorticity profile plane (VPP) are
parallel to the lineation. Maximum
symmetry is expected in the lineation

normal plane with a
zero W
k

value.
b. In triclinic shear, the VPP and transport
direction are not parallel to the lineation.
Therefore, both lineation

parallel and
lineation

normal planes should have non

zero W
k

values, but neither are the VPP.
In
monoclinic general shear zones, the maximum asymmetry
plane (and VPP) should be the lineation

parallel foliation

normal
plane.
T
he plane normal to both foliation
and lineation is expected to have
maximum symmetry, because
material will not rotate in this
plane and will record only the pure
shear
component of the general shear deformation (Fig. 3
FB
).
Triclinic shear is analogous to multiple instantaneous
monoclinic deformations superimposed on the
previous deformation,
but between each incremental
monoclinic deformation,
the shear direction is
changed, and the end result yields a single
triclinic deformation. In triclinic shear zones, the lineation
is
not expected to be parallel to the shear direction, but rather
oriented between the ISA and the fin
ite
strain axes. Orientation of lineations would also be
expected to change
with respect to the vorticity
vector throughout
the shear zone. For a triclinic deformation
,
the VPP is no longer parallel to fabric
elements in the rock (Fig. 3
FB
).
In the
field,
the identification of triclinic deformation relie
s
on
the
presence of a wide variation of lineation orientations within
a shear zone and a noticeable
porphyroclast asymmetry in both
lineation

normal and lineation

parallel planes.
21
Fig. 5FB
Relations
between the location of the VPP, lineation, and foliation in monoclinic shear zones. Dotted
lines are an aid to the visualization of the three

dimensional shapes of the figures. White arrows
indicated shear directions and directions of shortening and exten
sion.
a. Transtension, the VPP is parallel to both foliation and lineation; the zone widens with increasing
deformation.
b. Monoclinic general shear, the VPP is parallel to lineation and orthogonal to foliation; the zone can
widen or shorten.
c. Transpr
ession, the VPP is orthogonal to both lineation and foliation; the zone shortens with
increasing deformation.
22
The porphyroclast hyperbolic distribution
(PHD)
method
(Simpson and De Paor,
1993, 1997)
.
It
is
based on the premise that
the orientation of th
e long axes of backward rotated grains
within the acute
angle field between the flow eigenvectors delineates the orientation of the unstable eigenvector. The
stable
eigenvector is assumed
to be parallel with foliation.
Porphyroclasts in a given plane of a
sample
are identified as either forward or backward rotated
based on the orientation of long axes relative to
the overall sense
of shear. The angle
between the long axis of the grain and the normal to foliation
is
the phi (
φ
) angle, with positive
φ
values
indicating forward
rotated grains and negative
φ
values
indicating back

rotated
grains.
Axial ratios (R) of the porphyroclasts are also measured.
Both the
φ
and R

values are plotted on a
hyperbolic stereonet
(De Paor, 1988). A hyperbola is drawn to inc
lude all of the
back

rotated grains,
and the angle between the two limbs of
the hyperbola represents the acute angle between the two
eigenvectors,
such that the cosine of this angle (
ν
) yields
W
k
.
Simplification:
Plotting the data on a hyperbolic net
is not necessary because the porphyroclast with
the lowest φ angle always defines the kinematic vorticity number. W
k
is given by: W
k
= cos (90

φ),
where the φ is the smallest angle made with the normal to foliation by back

rotated grains. Grain
orientation
s are more easily visualized with a radial distribution plot than a hyperbolic
net (Fig. 6
FB
).
Porphyroclasts with small axial ratios (
<
1.4) were removed from consideration because sub

spherical
grains are not actually back

rotated, but rather are continuo
usly
forward rotated
.
Although an axial
ratio
of 1.4 is an arbitrary cutoff, clasts below this ratio are sub

spherical,
commonly difficult to
measure, plot close to the
origin on the hyperbolic net, and do not affect the determined
opening angle
of the hyp
erbola.
23
Fig. 6FB
a. Hyperbolic stereonet plot of axial ratios and long axis orientations of both forward and backward
rotated porphyroclasts. Solid circles are backward rotated and hollow diamonds are forward
rotated.
b. Radial distribution plot of backward rotated porphy
roclasts with maximum opening angle defined
by the solid black line.
24
Additional comments.
The rotational behavior of rigid elliptical porphyroclasts is controlled by the
bulk kinematic vorticity (W
k
), the axial ratio of the mineral grains (R), and
the orientation of their long
axes with respect to a fixed reference frame (φ). Three reference frames are used to evaluation
vorticity: (1) the finite strain axes; (2) the infinitesimal strain axes; (3) the shear zone boundary. The
shear zone boundary and
its normal are employed as the most reliable frame of reference. Axially
asymmetric porphyroclasts whose long axes are inclined ‘‘downstream’’ (a downstream dip

direction)
of the bulk transport direction at an orientation that falls within the acute angle
between the two
eigenvectors of the non

coaxial flow field will rotate opposite to the bulk shear sense within a
mylonitic shear zone (Fig. 2KN).
Fig. 2KN
Schematic diagram of mantled
porphyroclasts
that are inclined ‘‘upstream’
’ or ‘‘down

stream’’ relative to the bulk
sense of transport.
25
Backward

rotated porphyroclasts
(see e.g. x in Fig. 3KN)
are inclined ‘‘downstream’’ relative to the
bulk sense
of shear and exhibit
σ

type asymmetric tails of recrystallized
material attached to the broad
or long sides of the elongate grain
(Figs. 2
,
3
KN)
. Forward

rotated
porphyroclasts
a
re distinguished
by
:
(1)
Approximately
equate
or
spherical
δ

grains indicating continuous forward
rotation
and
(2)
σ

grains that are inclined ‘‘upstre
am’’
that exhibit recrystallized material attached to their
narrow
ends.
Klepeis et al. (1999) described two variations of backward

rotated
grains based on the following
criteria:
(1) ‘‘upstream’’ or
‘
‘downstream’’ inclined porphyroclasts exhibiting a sense of shear
contrary to the
bulk direction of transport with
σ

type tails of
recrystallized material attached to either the narrow or
broad sides
of the grain (
β
1
grains; Fig. 3
KN
); and
(2)
σ

type
porphyroclasts
inclined ‘‘downstream’’ exhibiting asymmetric tails attached to the
broad
sides of the grain and a rotational direction concurrent with
the bulk flow field (
β
2
grains; Fig. 3
KN).
Fig. 3KN
x
Schematic diagram illustrating fields of forward and backward rotation in a
dextral general shear re
gime, as well as, various micro
structures used in PHD analyses. Dashed lines
represent directions of maximum angular shear strain rate.
Those grains incline
d downstream with
sinistral tails on their narrow ends are probably near their stable orientations.
= delta clast
key clast
26
Application example
(Law et al. 2004).
Th
e orientation and
aspect ratio of porphyroclasts that have
either forward rotated or
back rotated
are
r
ecord
ed
o
n a hyperbolic net
,
the porphyroblasts are
cod
ed
w
ith respect to the type of recrystallization tail
(
σ
and
δ
). The
hyperbola that encloses all back

rotated
sigma

type porphyroclasts,
and separates them from all other types, is chosen. One
limb of this
hyperbola is asymptotic to the foliation, and the
mean kinematic vorticity number W
m
is given by the
cosine of
the acute angle between the two limbs of the hyperbola.
T
he
choice of the
most acute
hyperbola
available on the hyperbolic net
usually
reduc
es
th
e pure shear
component to a minimum
; t
he
estimated W
m
value is therefore regarded as
a maximum value.
Fig. 12L
Porphyroclast hyperbolic
distribution polar plot. Estimated orientation of flow
apophyses is given by
the hyperbola that
encloses all
back

rotated sigma

type
porphyroclasts, and separates them from all
other types. One limb of this hyperbola is
asymptotic to the foliation, and the mean
kinematic vorticity
number W
m
is given by
the cosine of the acute angle between the
two limbs of the hy
perbola. Average
orientation of shear bands in this sample
approximately bisects
acute angle between flow apophyses
(limbs
of hyperbola), and central segment of
leading edge of quartz c

axis fabric is
orthogonal to
average shear band
orientation
(
Comment
:
the plot is somewhat ridicules
but good
, as a 360° representation is given and data are mirrored)
.
For an explanation of the hyperbolic net see De Paor (1988)
27
Rigid grain net method
and updating of other rigid grain methods
Nomenclature
W
m
mean kinematic vorticity number
R porphyroclast aspect ratio (long axis/short axis)
B* shape factor of Bretherton (1962)
M
n
short axis of the porphyroclast
M
x
long axis of the porphyroclast
θ angle between long axis and the foliation
(or shear zone bounda
ry)
X’
2
–
X’
3
plane normal to the rotational axis X’
1
β angle between the stable

sink and source

sink in the X’
2
–
X’
3
plane
R
c
critical threshold between grains that rotate infinitely and those that reach a stable

sink position
R
cmin
minimum R
c
as defined by
Law et al. (2004)
R
cmax
maximum R
c
as defined by Law et al. (2004)
Models for the rotation of
rigid
elliptical objects in a fluid demonstrate
that during simple shear (mean
kinematic vorticity
number W
m
=
1) rigid objects will rotate infinitely, regardle
ss
of their aspect ratio
(R). With increasing contributions of pure
shear (0 <
Wm < 1), porphyroclasts will either rotate with
the
simple shear component (forward) or against it (backward)
until they reach a stable

sink
orientation that is unique to R
and
W
m
(Fig. 1
J).
Fig. 1J
Rotation of two simplified elliptical porphyroclasts within a regime of general shear.
Porphyroclast on
the left has an aspect ratio of 2 (B* = 0.6) and is in the stable

sink orientation of θ = 27° and
represents one of many possible original orientations that rotated forward to the stable

sink position.
The porphyroclasts on the right is b
ack rotated, due to the pure shear component, and has a long axis
at a negative angle (
θ
) to the foliation.
28
T
he
Passchier (1987)
(“Passchier plot”)
and Wallis (1995)
(“Wallis plot”)
methods
and the
“
porphyroclast hyperbolic distribution
”
(PHD) plot
(Simpson and De Paor,
1993, 1997)
are used for
practical applications (Fig. 3J)
.
Fig. 3J
Examples
of tailless porphyroclast data.
The Passchier plot uses the shape factor
(where M
n
=
short axis
and M
x
=
long axis of the porphyroclast) vs. angle between
porphyroclast long axis and foliation (
θ
) to define the critical threshold used to estimate W
m
. The
Wallis plot uses the aspect ratio (R
=
long axis/short axis) and angle from macroscopic foliation (
θ
) to
locate the critical threshold (R
c
). W
m
is calculated
using R
c
where
. Upper
and lower R
c
values are used to estimate a range in likely W
m
estimates. The PHD plot uses the
hyperbolic net to plot aspect ratio (R) and
θ
.
T
he cosine of
the opening angle (
β
) of the best

fit
enveloping
hyperbola yields
the
W
m
.
Review of techniques
.
During general shear, rigid grain analysis assumes that the
orientation of
porphyroclasts within a flowing matrix record
a critical threshold (R
c
) between
porphyroclasts that
rotate indefinitely
(low aspect ratio), and therefore do not develop
a preferred orientation, and those
that reach a stable

sink orientation
(higher aspect ratio). This unique combination
of W
m
, R or B* and
θ
define the value of R
c
betw
een these two
groups of rigid grains. W
m
to B* and
θ
are related by (see
also earlier):
{
}
(where M
n
=
short axis
and M
x
=
long axis of the porphyroclast)
Passchier

plot.
The θ

equation
generate
s
a hyperbolic
curve in
θ
vs. B* space that represents the ideal
distribution of
grains for a particular W
m
. The vertices of this hyperbola mark
the unique R
c
value
where W
m
=
B*. Assuming high
strain, a natural distribution of porph
yroclasts should define
a limb of
this hyperbola for a range of B* values that is greater
than B* at R
c
. With low strains, a misleading
distribution
of porphyroclasts has the potential to overestimate
the simple shear component because
high aspect ratio po
rphyroclasts
have yet to rotate into their stable

sink orientation. Porphyroclasts
with a B* < B* at R
c
will rotate infinitely and
should define a broad distribution with
θ
=
±
90
°
. In
contrast, porphyroclasts
with a B* > B* at R
c
are predicted to reach stable
sink
orientations with a
limited range in
θ
values (Fig. 3
J
A). Whether a porphyroclast will rotate forward or backward to
a
stable

sink position depends on the initial
θ
at a specific B* and W
m
. R
c
should be defined by the
ei
ther B* or R
and
an
abrupt change in range of
θ
values (Fig. 3
J
A
,
B).
Although the distribution of porphyroclasts on the Passchier plot can be informative
for the high
quality data sets (Fig. 3
J
A), without a reference frame for comparing
complex natural d
ata
with the
theoretical values established by
the θ

equation
, defining R
c
will remain ambiguous.
B* = W
m
at R
c
29
Wallis

plot.
The Wallis plot still uses
θ
on the Y

axis,
but replaces B* with the more intuitive
porphyroclast aspect
ratio (R
=
long axis/short axis) on the
X

axis (Fig. 3
J
B; Wallis,
1992, 1995). W
m
is calculated from the R
c
values separating
porphyroclasts that reach a stable

sink orientation (
θ
<
θ
at
R
c
)
from those that rotate continuously (
θ
>
θ
at R
c
). W
m
is
calculated
from
(Wallis et al., 1993):
.
The distribution of porphyroclasts often
defines a gradual transition between
continuously rotating
(random orientation) porphyroclasts and stable

to semi

stable porphyroclasts that define R
c
.
T
he
original Wallis plot
is improved
by
drawing
an enveloping surface to better

define the grain
distribution, and use a range in possible W
m
values (R
cmin
and
R
cmax
; Fig. 3
J
B; Law et al., 2004; Jessup
et al., 2006).
PHD

plot.
The porphyroclast hyperbolic distribution (PHD) method
estimates W
m
by using R and the
angle between the pole to
foliation and long axis of tailed porphyroclasts (Fig. 3
J
C);
plotted
using the
hyperbolic net (HN). Each hyperbola of the HN represents
the theoretically predicted orientation of
porphyroclasts for
a particula
r R and W
m
as plotted in
polar
coordinates
(see sign convention)
.
R
c
is defined as the vertices of the hyperbola. One limb of the hyperbola represents the stable
sink
orientation for porphyroclasts while the other is the metastable
position
(Fig. 3JC).
At
θ
> the
metastable orientation, porphyroclasts
will rotate forward until they define another semi

hyperbolic
cluster on the concave side of the same hyperbola. Assuming
significant shear, back

rotated clasts
with variable aspect
ratios, plo
tted on the HN, should define a semi

hyperbolic
cluster representing the
stable

sink orientation. The linear
cluster is then rotated to find the best

fit hyperbola whose limbs
represent the two eigenvectors of flow, one of which is asymptotic
to the foliat
ion (i.e., the source

sink
and stable

sink positions). The vertex of this hyperbola
separates the low aspect ratio porphyroclasts
with random
orientat
ion (i.e., infinitely rotating)
from higher aspect ratio
porphyroclasts with a narrow
range of orientation
s (
Fig. 3
J
C).
Supplementary information.
Table 1J:
Critical threshold values
R
W
m
θ at Rc
B*
β
cos(β)
1.1
0.1
42
0.1
84
0.1
1.21
0.2
39
0.2
78
0.2
1.3
0.3
36
0.3
73
0.3
1.5
0.4
33
0.4
66
0.4
1.7
0.5
30
0.5
60
0.5
2
0.6
27
0.6
53
0.6
2.4
0.7
23
0.7
46
0.7
3
0.8
18
0.8
37
0.8
4.4
0.9
13
0.9
26
0.9
R = aspect ratio (long axis/short axis).
W
m
= mean kinematic vorticity number.
θ = angle from foliation
(Fig. 1J).
R
c
= critical threshold.
B* = shape factor.
β = opening angle of hyperbola.
30
Fig. 2J
Plot showing the relationship between mean kinematic
vorticity number (W
m
), shape factor (B*), and aspect
ratio (R) at critical value
s.
31
The Rigid Grain Net (RGN)
.
Eq.
{
}
is
used to
calculate
semi

hyperbolas for a range of W
m

values that express
the relationship between
θ
and B*
(
Fig. 4
J
, location A). The
second set of curves represent the possible R
c
(vertices curves)
values for
when W
m
=
B* (
Fig. 4
J
, location B). Each
semi

hyperbola was calculated for a particular W
m
and a
series
of B* values.
T
he shape factor (B*
)
enables W
m
values to be obtained directl
y
from the RGN.
P
ositive and negative semi

hyperbolas are
plotted at 0.025 increments
(of B*)
for a range in W
m
(0.1

1.0) by
solving for
θ
using
the θ

equation.
For
a particular shape factor, when B*
=
W
m
and
θ
>
θ
at
R
c
, the
semi

hyperbolas transition
into vertical lines to define the
maximum B* value below which
grains rotate freely
(
Fig. 4
, location C).
To highlight the continuity
in R
c
values for the range in W
m
values represented by the
RGN, a second curve (vertices curve) links the R
c
values on
eac
h hyperbola
(
Fig. 4
, location B).
Fig. 4J
The Rigid Grain Net (RGN) using semi

hyperbolas. Location A is an example of a semi

hyperbola;
location B highlights the vertices curve; location C is an
example of a R
c
value when W
m
=
B*;
location D points to one of a series of aspect ratio (R) values included on the RGN to demonstrate its
relationship with the
less intuitive shape factor (B*); location E is a W
m
value for a semi

hyperbola.
To relate hyperbolas on the HN and the RGN,
full hyperbolas
are plotted
on the RGN
;
highlight
ed
are
the critical curves
that
define W
m
=
0.6;
plotted
are
also
hypothetical distribution
s
of
porphyroclasts
(
Fig. 5
J
). The two hyperbolas that are included
on the simplified HN are highlighted in black on
the
RGN
(W
m
=
0.6) for positive and negative
θ
(
Fig. 5
J
A
,
B).
Fields of the HN and RGN plots that
represent the maximum
aspect ratio for porphyroclasts that are predicted to rotate
forward
infinitely
and thereby have a complete range in
θ
between
±90°
, ar
e represented by a circle of constant R for all
orientations that is tangential to the vertex of each hyperbola
and plotted on the center of the HN (i.e.,
when B* <W
m
;
Fig
.
5
J
A
,
C). On the RGN, this area includes all of
the potential range in B* between 0
and B* at R
c
(i.e., to the
left of the apex of the hyperbolas;
Fig. 5
J
B
,
D). An additional
section of the
RGN is highlighted on the HN that defines
porphyroclasts that will rotate to the vertices curve for R
c
values
≥
th
e ‘‘true’’ R
c
for the sample. On the HN, this curve
is defined by linking the R
c
value from
each potential hyperbola
greater than the ‘‘true’’ R
c
for this sample to generate
a small section of the
vertices curve as shown on the RGN
(
Fig. 5
J
A
,
C). This impo
rtant clarification shows that these
porphyroclasts must be considered as stable

sink orientations
when c
hoosing the best

fit hyperbola.
The RGN is available as
an Excel

worksheet. This enables to
monitor how the distribution of
porphyroclasts
is
developing during data acquisition.
32
Fig.5J
a.
Half of the HN simplified to graphically demonstrate the relationship between one hyperbola (W
m
=
0.60) and the vertices curve for
that hyperbola. The vertices curve is drawn using the
vertices of
several hyperbolas (dashed) for a range in W
m
values greater than the W
m
for the sample (0.60).
Gray circles with letters (a

i) on the hyperbola for W
m
=
0.60 are shown to compare how these
define t
he hyperbolas on the HN and RGN.
The circle th
at
defines the highest aspect ratio (R
=
2)
below which porphyroclasts are predicted to rotate infinitely is also included.
b.
The RGN with complete hyperbolas.
Highlighted in black are the positive and negative hyperbolas
that correspond to a W
m
=
0.60, as well as the section of the vertices curve for B* > B* at R
c
. A
series of gray circles represent equivalent points on the RGN and HN.
c.
The same plot as
a.
with an overlay of different types of hypothetical porphyroclasts
in their
predicted dist
ribution; gray squares are infinitely rotating, black crosses are limited rotation, gray
circles are stable

to metastable

sink positions.
d.
The
same plot as
b.
with hypothetical porphyroclasts distributed in various sections of the RGN.
vertices

curve
33
Application
of
the
RGN
.
“
The choice of rigid porphyroclasts
i
s highly selective, ignoring
all but the
most appropriate grains for estimating W
m
“
(Fig. 7J).
Fig. 7J
All R
c

based methods may tend to underestimate the vorticity number if clasts of large aspect ratio
are
not present, and therefore within individual samples the upper bound of this W
m
range is probably
closest to the true value.
34
SC’

type shear

band
method
Fig. 4KN
Rose diagrams generated from orientation data for synthetic and antithetic SC’

type shear bands and
other related structures such as synthetically and antithetically imbricated mineral grains are
combined with PHD diagrams. This combination is meant to ill
ustrate geometric relationships and
frequency distributions of shear band orientations relative to the inferred flow field determined from
PHD analyses. Porphyroclast axial ratio data relates to the scale along the base of each diagram.
Annular dashed half

circles measure shear band frequency data associated with rose diagrams.
Dotted lines indicate the maximum inclination of synthetic SC’

type shear band within two standard
deviations of the mean and are used to estimate kinematic vorticity
. Black X’s repr
esent plotted
backward

rotated porphyroclasts
while gray

filled circles depict plotted forward

rotated
porphyroclasts.
35
Synthetic shear bands are oriented either parallel to, or at an angle less than the acute bisector (AB;
Fig. 4KN). Antithetic shear ba
nds populations show a range of inclination with a mean inclination
lying near the obtuse bisector (OB). These data suggest that extensional SC’

type shear bands initially
form parallel to AB and OB.
Shear bands that are inclined at an angle less than AB a
nd OB are
the
result of either: (1) rotation towards the shear zone boundary
during progressive non

coaxial flow; (2)
formed under heterogeneous
non

steady

state conditions and/or varying bulk vorticity;
or (3) formed
during separate episodes of
deformation. Assuming
a steady

state general flow regime, synthetic and
antithetic
extensional shear bands are expected to rotate towards the stable
eigenvector and away from
AB and OB throughout progressive non

coaxial
deformation. Assuming steady

state
v
orticity during
progressive deformation, the most steeply inclined
shear bands may provide the best direct estimate of
AB and OB in
that shear bands at this orientation may not have been significantly
rotated.
Estimated
values for bulk vorticity
are
determ
ined
by utilizing the most steeply inclined shear band
orientation within
±
two standard deviations of the mean in order to eliminate outliers
(Fig.
4KN)
.
Samples may exhibit moderately to well

defined SC’

type shear bands that form an anastomosing
networ
k of quartz dominated
micro

shear zones and enclose large microlithons of feldspar
(Fig. 7
KN).
Fig. 7KN
Composite image of sample CC

6

30

1,
example of weakly recrystallized coarse to
medium

grained quartz

feldspar aggregates.
An
anastomosing network of quartz

rich,
micro

scale shear zones have accommodated
the majority of the strain within the rock mass.
Because these coarse to medium

grained quartz

feldspar
aggregates appear to have undergone
comparatively
smaller amounts of strain, the arithmetic mean of shear band
orientations has been used
to estimate bulk vorticity.
It is assuming that this average orientation roughly bisects the angle ν (Fig.
8KN).
Fig. 8KN
Rose diagrams illustrating
SC’

type shear band orientations. Dashed line represents the estimated
orientation of AB.
36
Quartz
c

axis fabric
s and oblique grain shape
vorticity method
s (review in Xypolias 2009)
These methods are based on two
assumptions:
(1) Under progressive simple,
pure
,
and general shear, the central
girdle segment of quartz c

axis

fabrics develops nearly
orthogonal to the flow/shear plane (A
1
)
.
Therefore, the angle,
β
, between the
perpendicular to the central girdle segment of quartz c

axis
fabric
and the foliatio
n (S
A
) is equal to the
angle between the
flow plane (A
1
) and the flattening plane of finite strain
(
Fig. 1
a,c;
Wallis, 1992
).
(2) The long axes of quartz neoblasts within an oblique

grain
shape
foliation
align nearly
parallel to
the extensional
instantaneous stretching axis (ISA
1
)
(
Fig. 1
a,b;
Wallis, 1995
). This is interpreted to be
the result of a complex process of
continuous nucleation, passive deformation
,
and rotation of the
recrystallized grains composing the S
B
.
Therefore, the maximum obser
ved angle,
δ
, between the S
B
and
the main (S
A
) foliation is equal to the angle between the ISA
1
and the largest principal axis, X, of
the finite strain (
Fig. 1
a,b;
Wallis, 1995
).
Fig. 1X
According to the oblique

grain

shape/quartz c

axis

fabric
method (referred to hereinafter as
δ
/
β

method), if both angles
δ
and
β
are known then an estimate of vorticity number can be obtained
using
the equation (
Wallis, 1995
):
W
m
= sin 2 (
δ
+
β
)
This e
q. only
holds for two

dimensional flow. The
δ
/
β

method
record
s
the last increments of plastic
deformation (
Wallis, 1995
).
Fig. 1. (a) Simplified sketch showing the
relative orientation o
f instantaneous flow
elements and their angular relationships in
real space for a dextral general shear flow.
A
1
and A
2
–
flow apophyses; ISA
1
and ISA
3
–
instantaneous stretching axes; X and Z
–
principal strain axes. The vorticity vector lies
perpendicular to the page. (b) The angle, δ,
between the oblique

grain

shape fabric (S
B
)
and the main foliation (S
A
) is inferred to be
equal to the angle between t
he ISA
1
and the
principal finite strain axis X. (c) The angle, β,
between the perpendicular to the central girdle
segment of quartz c

axis fabric and the main
foliation (S
A
) is inferred to be equal to the
angle between the flow apophysis A
1
(flow
plane) an
d the principal finite strain axis X.
37
R
XZ
/β

method
:
The strain

ratio/quartz c

axis

fabric method
estimates
W
m
(finite deformation)
as:
Fig. 13L
V
orticity analysis after Wallis (1995). Flow plane is inferred to be orthogonal to central segment of
the leading edge of quartz c

axis fabric (or a

axis maximum), measured in section oriented
perpendicular to foliation and parallel to lineation. Acute angle between foliation and inferred flow
plane (i.e. normal to central segment of fabric) defines β; complementary angle ψ between foliation
and central segment of fabric is a measure of externa
l fabric asymmetry. β is used in combination with
the ratio of principal
stretches, R
xz
=(1

ε
1
)/(1

ε
3
), to estimate mean kinematic vorticity number W
m
.
This method assumes plane strain deformation.
Plane strain conditions are indicated by the cross

gir
dle pattern of quartz c

axis fabrics, and the orthogonal relationship between the cross

girdle fabric
and sample Y direction (within foliation and perpendicular to lineation) also argues for monoclinic
rather than triclinic flow.
Tikoff and Fossen (1995)
i
nvestigated the effect of the third stretching
direction (Y

axis) on
vorticity

number estimates and demonstrated that the
2D
vorticity analysis can
overestimate the actual,
3D

W
m
by only a small amount (<0.05).
The overestimation will be greatest
for
nearly equal components of coaxial and non

coaxial deformation and will decline to zero as W
m
goes to zero or one.
V
orticity number estimates using R
XZ
/
β

method are very sensitive
to small
changes in the evaluated angle
β
.
I
n most quartz c

axis
fabric
diag
rams the angle
β
can be evaluated
with an error
±
2
°;
the method
becomes unreliable in high

strain samples (R
XZ
>10
–
15) with small
β
(<5
°).
R
XZ
/δ method
:
T
he maximum observed angle
δ
between S
B
and S
A
grain shape foliation is primarily
related to both
the
degree of non

coaxiality and the shape of the strain ellipse just
prior to the end of
deformation.
W
n
(instantaneous deformation) =
sin
(
2
δ) [(
R
XZ
+
1
)/(R
XZ
–
1
)]
which shows that W
n
is related to R
XZ
and
δ
.
38
This e
q. can be solved for
δ
and the result can be plotted for different W
n
values
.
δ = 1/2
sin

1
[
Wn
(R
XZ

1)/(
R
XZ
+
1
).
T
his plot (RXZ versus
δ
) is illustrated in the synthetic diagram of
Fig. 3
, which also includes an R
XZ
versus
β
plot
.
Fig. 3X
Fig. 3. Synthetic diagram
showing plots of the finite strain

ratio R
XZ
versus
β
and R
XZ
versus
δ
.
This diagram indicates
that during progressive steady

state deformation the angle
δ
attains a high
value at low strain (R
XZ
<
4) and then is subjected
to small changes. This observat
ion appears to be in
accordance with
experimental work
that indicates
that the oblique

grain

shape fabric forms early
during non

coaxial
shear attaining a stable orientation at low imposed shear
strain (
γ
<1.5). The p
lot of
R
XZ
versus
δ
also indicates that
the
W
n
curves
get broader spaced for increasing strain values (
Fig. 3
).
This implies
that, in contrast to the R
XZ
/
β

method,
W
n
estimates are
relatively insensitive to small
changes in the values of R
XZ
and
δ
.
The
R
XZ
/
δ method
describes instantaneous
deformation and, therefore, does
not require the assumption
of steady

state deformation.
Comparison to the particle methods
.
D
ifference could
be due to: (1) analytical problems
associated
with individual methods (e.g.
the particle methods tend
to
give a
minimum estimate of W
m
if rigid
grains of sufficiently
high aspect ratio are not present); (2) the two methods record
different parts of
the deformation history (i.e. they have different
lengths of ‘strain memory’); (3) the two methods
reflect a
contrast i
n synchronous flow behavior between rigid particle
rotation of the porphyroclasts
and plastic deformation of the
surrounding matrix quartz grains.
39
A vorticity nomogram
incorporates
all the parameters (
β
,
δ
, R
XZ
and
W
n
/W
m
) involved in the three
vorticity
analysis methods. The advantages of such nomographic
approach are
:
(1)
it instantly provides the vorticity number for all methods;
(2) it
allows to check, for all methods, the sensitivity of estimated
vorticity numbers values in changes
of input
parameters (
β
,
δ
, R
XZ
);
(3) it provides a rapid means of evaluating the consistency of values
estimated by the various methods;
(4) it permits the presentation
of uncertainties in vorticity number estimates as error bars;
(5) it
enables the rapid evaluati
on of mean strain level for a suite of
samples, where only
β

and
δ

angle data are available.
Fig. 4X
Fig. 4. Suggested nomogram for estimating W
m
(or W
n
) and R
XZ
in naturally deformed quartzites and
an application to five representative samples. Circ
les indicate estimates obtained using the best
assigned input data. Error bars were constructed using uncertainty in the estimations of the β

angle as
well as error in R
XZ
values.
40
Stretch across a plane

strain shear zone
(Law et al. 2004, Wallis et al. 1993)
Calculation of shortening value (S) measured perpendicular to flow plane, taking into account both
strain magnitude and vorticity of
flow (adapted from Wallis et al. 1993). Assuming plane strain
deformation, stretc
h measured parallel to the flow plane in the transport
direction is given by S

1
.
41
Non

perfectly matrix bounded rigid inclusions
(
M
ar
ques et al. 2007)
Overview of inclusion models.
1. Fundamental solutions: Aspect ratio, inclination, and material
property contrast
. If the interest is focused on slow deformation processes, two analytical solutions
address the problem of an inclusion immersed in a matrix of different property:
(1) Je
ffery’s model
for rigid ellipsoids in a deforming viscous matrix (Jeffery, 1922), and
(2) Eshelby’s model for a deformable elastic ellipsoid in a far

field loaded matrix with different
properties (Eshelby, 1957, 1959).
Common to these solutions is the id
entification of the material property contrast between
inclusion and matrix, µ
i
/µ
m
, the aspect ratio, A
r
, and the inclination φ as the governing
parameters that influence the inclusion behavior (Fig. 1M).
Fig. 1M
Sketch of an inclusion in a shear zone.
The factors that determine inclusion behavior are: inclination
φ (positive counter clockwise), aspect ratio A
r
, ratio W
r
between shear zone width (w) and inclusion
short axis (b), distance to other inclusions (d ), presence of a rim with thickness h that
controls the
degree of welding between inclusion and matrix, shape of the inclusion, viscosities of inclusion (
µ
i
),
rim (
µ
r
) and matrix (
µ
m
), and finally
the relative strength of pure to simple shear S
r
. Note that the
angle and quadrant
convention used (s
ee text) is related to the shear direction, i.e., in top to
the left
shear zones the above sketch has to be flipped horizontally.
The modeling
of inclusion behavior
in
ductile
shear zones used to assume that the matrix acts as a
viscous
fluid and the clas
ts as isolated perfectly bonded rigid inclusions,
therefore allowing for
application of Jeffrey’s solution.
F
actors that can change
the rotation behavior of rigid inclusions
are:
(i) the addition of a pure shear component
to sim
ple shear (variable
vorticity).
(ii) inclusion interaction
in multi

inclusion systems.
Inclusion interaction in multi

inclusion systems usually leads to tilling effects, which strongly affect
inclusion rotation and bring the inclusion to a stable equilibrium orientation.
(ii
i) sli
p at inclusion/matrix interface.
If the slipping contact is due to the existence of weak mantle phase then the inclination of the clast will
be controlled by the viscosity contrast between mantle and matrix and the rate at which the mantle
material i
s produced.
(iv) flow confinement.
Depending on the W
r
value, elliptical inclusions can rotate backwards from φ =
0° (opposite to Jeffery’s model) and stabilize at shallow positive angles (0≤φ<90°
)
(Marques and
Coelho, 2001).
These factors may furthermore
be affected by the actual shape of the inclusion.
42
First applications.
Fig. 3M et al. and Fig. 4M et al.
Asymmetrical inclusions comprise mainly
disrupted and strongly attenuated quartz

rich
material,
which mostly derives from older
quartz veins and segregations within the
phyllonitic gneiss
.
Field data is fitted with a power

law and
compared to the ice experiments of Marques
and Bose (2004), the equ
ivalent void derived
by Schmid and Podladchikov (2004), and the
transtension (S
r
=

0.3) case of combined pure
and simple shear (Ghosh and Ramberg, 1976;
Marques and Coelho, 2003).
Site 1: Given the φ

A
r
data and the fact that the
clasts are found in relative isolation, the explanation
for the SPO is that the clast

matrix interface was slipping. This can be seen from the ice in PDMS
(a
viscous polymer)
experiments that correspond very well to the field data and also from the equival
ent
void conjecture. Had the interface been welded,
then
the Ghosh and Ramberg’s (1976) model could be
applicable, which also predicts stable SPOs at positive angles for cases where the pure shear
component is extensional at 90° to the shear plane. However
, the stabilization trend for any such
negative S
r
case is opposite to the field data as illustrated with S
r
=

0.3. We can therefore rule out
perfect bonding and, given the good fit between the natural data and the simple shear only lubrication
models, we
can also rule out the necessity of an additional pure shear component to explain the data.
43
Fig. 5M et al., Fig. 7M et a
l.
Features: (1) σ

porphyroclast (1), with straight tails rooting with asymmetric development at the
crests, hence showing very high stair stepping and
top to left sense of shear. A dark biotite layer
developed around this clast at opposing faces (contraction
quadrants, marked by small black arrows),
while a white thin rim of quartz (marked by small white arrows) formed at the two other faces
(expansion quadrants).
(ii) Porphyroclasts with shapes that vary from elliptical (3 to 7), to lozenge (2) or skewed re
ctangle
(8).
(iii) Porphyroclasts at negative inclinations (1 and 9) or, more commonly, positive inclinations.
(iv)
Lack of well

developed recrystallization tails in most porphyroclasts.
(v) Pressure shadow tails as marked by white bigger arrow on the
left of porphyroclast 10.
44
‘‘Site
2
Data Fit’’, ‘‘ice experiments’’, and ‘‘equivalent void’’ models. Transpression is plotted for S
r
= 1 according to Ghosh and Ramberg’s model (1976).
‘‘FEM Slip S
r
= 1’’ represents finite element
calculations for slipping inclusions in pure and simple shear
.
Site 2
. T
wo distinct distributions
are evident
: (1) a set of positive
inclinations and (2) a set of negative
inclinations. In both
cases, larger aspect ratio clasts plot closer to the shear plane
than smaller ones.
T
he clasts with negative inclinations can be
fitted to Ghosh and Ramberg’s model with a shear zone
flattening
component that equals the simple shear c
omponent,
S
r
=
1. Note though that the
transpression line ceases to exist
below a minimum aspect ratio, i.e. clasts with smaller aspect
ratios
are not stable
(rotate freely)
under the given pure to simple shear ratio
and perfect interface bonding.
The
fini
te element experiments,
with
the addition of
a flattening pure shear component
,
shift the stable
inclinations
of lubricated clasts closer to the shear plane. The actual angles
are too shallow, yet the
trend in the data is reproduced.
V
arying the actual
amount of pure to simple shear
can vary the actual
position of the stable inclination curve.
Evidence for slipping boundaries around part of the clast
population is indicated
by:
t
hin rims of mica
and/or
quartz and/or fine

grained feldspar commonly surrou
nd feldspar
porphyroclasts, all typically
weaker phases than the porphyroclasts
.
These weak rims could have worked as effective lubricants
that kept the rigid feldspar in slipping contact with the mylonitic
matrix. Having excluded the
possibility of confin
ed flow,
the straight tails with very high stair stepping are also indicative
that the
inclusion was in slipping contact with the matrix. It is
therefore a reasonable conclusion that the clasts
exhibit slipping and non

slipping interfaces with
a far field
flow condition where simple shear and a
flattening
pure shear component are of approximately equal strength.
The total shear strain must be
more than 20.
Literature
Bobyarchick, A. R. 1986. The eigenvalues of steady flow in Mohr
space.
Tectonophysics 12
2,
35

51.
Bretherton, F.P., 1962. The motion of rigid particles in shear flow at low
Reynolds number. Journal of Fluid Mechanics 14,
284

301.
De Paor, D.G., 1988. R
f
/
φ
f
strain analysis using an orientation net. Journal of Structural Geology 10, 323

333.
Eshelby, J.D., 1957. The determination of the elastic field of an ellipsoidal inclusion, and related problems. Proceedings of
the Royal Society of London Series A

Mathematical and Physical Sciences 241 (1226), 376

396.
Eshelby, J.D., 1959. The elastic fie
ld outside an ellipsoidal inclusion. Proceedings of the Royal Society of London Series A.
Mathematical and Physical
Sciences 252 (1271), 561

569.
Forte, A.M., and C.M. Bailey, 2007. Testing the utility of the porphyroclasts hyperbolic distribution method o
f kinematic
vorticity analysis. Journal Struct. Geol., 29, 983

1001, 2007.
Fossen, H. Textbook, http://folk.uib.no/nglhe/StructModulesTextbook/Deformation.swf
Freeman, B. 1985. The motion of rigid ellipsoidal particles in slow
flows.
Tectonophysics
113,163

183.
Ghosh, S. K. & Ramberg, H. 1976. Reorientation of inclusions by
combination of pure and simple shear.
Tectonophysics
34,
1

70.
Jeffery, G. B. 1922. The motion of ellipsoidal particles immersed in a
viscous fluid.
Proc. R. Soc. Lond.
A102, 161

179.
45
Jessuo, M.J., Law, R.D., and Frassi, Ch., 2007. The rigid grain net (RGN): an alternative method for estimating mean
kinematic vorticity number (W
m
). J. Struct. Geol., 29, 411

421.
Klepeis, K.A., Daczko, N.R., Clarke, G.L., 1999. Kinematic vorticity a
nd tectonic significance
of superposed mylonites in a
major lower crustal shear zone, northern
Fiordland, New Zealand. Journal of Structural Geology 21, 1385
–
1405.
Kurz, G.A. C.J. Northrup, 2008. Structural analysis of mylonitic rocks in the Cougar Creek
Complex, Oregon
–
Idaho using
the porphyroclast hyperbolic distribution method, and potential use of SC’

type extensional shear bands as quantitative
vorticity indicators. Journal of Structural Geology 30, 1005
–
1012.
Law, R.D., Searle, M.P., Simpson, R.L., 2
004. Strain, deformation temperatures
and vorticity of flow at the top of the Greater
Himalayan Slab,
Everest Massif, Tibet. Journal of the Geological Society, London 161,
305

320.
Malvern, L. E. 1969.
Introduction to the Mechanics of a Continuous
Medium.
Prentice

Hall, Englewood Cliffs, New Jersey.
Marques, F.O., Coelho, S., 2001. Rotation of rigid elliptical cylinders in viscous simple shear flow: analogue experiments.
Journal of Structural Geology 23 (4), 609

617.
Marques, F.O., Coelho, S., 2003. 2

D sha
pe preferred orientations of rigid particles in transtensional viscous flow. Journal of
Structural Geology 25 (6), 841

854.
Marques, F.O., Bose, S., 2004. Influence of a permanent low

friction boundary on rotation and flow in rigid
inclusion/viscous matrix
systems from an analogue perspective. Tectonophysics 382 (3

4), 229

245.
Marques F.O., D.W. Schmid, T.B. Andersen, 2007.
Applications of inclusion behavior models to a major shear zone system:
The Nordfjord

Sogn Detachment Zone in western Norway. Journal
of Structural Geology, 29, 1622

1631.
Passchier, C.W. and Trouw, R.A.J.
, 2005.
Microtectonics, Springer
.
Passchier, C. W. 1987. Stable positions of rigid objects in non

coaxial
flow

a study in vorticity analysis.
J. Struct. Geol. 9,
679

690.
Passchier, C.
W. 1986. Flow in natural shear zones
and
the consequences
of spinning flow regimes.
Earth Planet. Sci. Lett.
77, 70

80.
Ramberg, H. 1975a. Superposition of homogeneous strain and progressive
deformation in rocks.
Bull. geol. lnst. Univ.
Uppsala 6,
35

67.
Ramberg, H. 1975b. Particle paths, displacement and progressive
strain applicable to rocks.
Tectonophysics
28, 1173

1187.
Schmid, D.W., Podladchikov, Y.Y., 2004. Are isolated stable rigid clasts in shear zones equivalent to voids? Tectonophysics
384 (1

4),
233

242.
Schmid, D.W., Podladchikov, Y.Y., 2005. Mantled porphyroclast gauges. Journal of Structural Geology 27 (3), 571

585.
Simpson, C., De Paor, D.G., 1993. Strain and kinematic analysis in general shear zones. Journal of Structural Geology 15, 1

20.
S
impson, C., De Paor, D.G., 1997. Practical analysis of general shear zones using the porphyroclast hyperbolic distribution
method: an example from the Scandinavian Caledonides. In: Sengputta, S. (Ed.), Evolution of Geological Structures in
Micro

to Macro

Scales, pp. 169

184.
Tikoff, B., Fossen, H., 1995. The limitations of three

dimensional kinematic
vorticity analysis. Journal of Structural Geology
,
17, 1771

1784.
Truesdell, C. 1954.
The Kinematics of Vorticity.
Indiana University
Press, Bloomington.
Wallis, S.R., Platt, J.P. & Knott, S.D. 1993. Recognition of syn

convergence extension in accretionary wedges with examples
from the Calabrian Arc and the Eastern Alps. American Journal of Science, 293, 463
–
495.
Wallis, S.R., 1995. Vorticity analysis and r
ecognition of ductile extension in
the Sanbagawa belt, SW Japan. Journal of
Structural Geology 17,
1077

1093.
Xypolias, P. 2009. Some new aspects of kinematic vorticity analysis in naturally deformed quartzites.
Journal of Structural
Geology
31, 3

10.
last modified:
June 16, 2011
Enter the password to open this PDF file:
File name:

File size:

Title:

Author:

Subject:

Keywords:

Creation Date:

Modification Date:

Creator:

PDF Producer:

PDF Version:

Page Count:

Preparing document for printing…
0%
Comments 0
Log in to post a comment