Journal of the Geological Society Fault rocks and fault mechanisms R. H. SIBSON Journal of the Geological Society 1977; v. 133; p. 191-213 doi:10.1144/gsjgs.133.3.0191
Email alerting service
click here to receive free email alerts when new articles cite this article
Permission request
click here to seek permission to re-use all or part of this article
Subscribe
click here to subscribe to Journal of the Geological Society or the Lyell Collection
Notes
Downloaded by
on 29 May 2007
© 1977 Geological Society of London
Fault rocks and fault mechanisms R. H. S I B S O N SUMMARY Physical factors likely to affect the genesis of the various fault rocks--frictional properties, temperature, effective stress normal to the fault and differential stress--are examined in relation to the energy budget of fault zones, the main velocity modes of faulting and the type of faulting, whether thrust, wrench, or normal. In a conceptual model of a major fault zone cutting crystalline quartzo-feldspathic crust, a zone of elastico-frictional (EF) behaviour generating random-fabric fault rocks (gouge--breccia-cataclasite series--pseudotachylyte) overlies a region where quasi-plastic (QP) processes of rock deformation operate in ductile shear zones
with the production of mylonite series rocks possessing strong tectonite fabrics. In some cases, fault rocks developed by transient seismic faulting can be distinguished from those generated by slow aseismic shear. Random-fabric fault rocks may form as a result of seismic faulting within the ductile shear zones from time to time, but tend to be obliterated by continued shearing. Resistance to shear within the fault zone reaches a peak value (greatest for thrusts and least for normal faults) around the EF/OP transition level, which for normal geothermal gradients and an adequate supply of water, occurs at depths of lO-15 km.
SINCE L A P W O R T H ' $ (I885) description of the type mylonite from the Moine T h r u s t in N W Scotland, there have been m a n y petrographic descriptions and classifications of the texturally distinctive rocks found associated with fault zones (e.g. Waters & C a m p b e l l 1935, Hsu 1955, Christie 196o , 1963, R e e d 1964, Spry I969, Higgins 1971 ). These rocks provide a tangible source of information on the processes which operate in m a j o r fault zones, b u t little has been done to correlate the various types of fault rock with different deformation environments and modes of faulting. I n this paper, I discuss the physical factors which m a y affect the genesis of fault rocks, and m a k e some tentative correlations between their textures and deformation environments. For the most part, the effects of faulting in crystalline quartzofeldspathic crust are considered.
I. Textures and occurrence of fault rocks I n this paper fault rocks is used as a collective t e r m for the distinctive rock types found in zones of shear dislocation at both high and low crustal levels, whose textures are t h o u g h t to arise at least in p a r t f r o m the shearing process. It is not denied t h a t similar textures m a y develop in association with other geological structures (e.g. protoclastic textures arising f r o m igneous intrusion). T h e t e r m cataclastic rocks, introduced by Waters & C a m p b e l l (1935) as a collective n a m e for all rocks of the g o u g e - - b r e c c i a - - c a t a c l a s i t e - - m y l o n i t e kindred, is not used because of the misleading implication t h a t such rocks have developed solely by cataclasis. I n the strict sense, cataclasis involves the brittle f r a g m e n t a t i o n of m i n e r a l grains with rotation of grain fragments a c c o m p a n i e d by frictional grain b o u n d a r y sliding
dTl geol. Soc. Lond. vol. I33 , 1977, pp. 191-213, 8 figs., 3 plates, 3 tables. Printed in Great Britain.
I92
R. H. Sibson
and dilatancy, and it is now clear that these are not the dominant processes leading to the formation of mylonite series textures. In quartzo-feldspathic mylonites, quartz grains have been reduced in grain size by the dynamic recovery and recrystallization of highly strained grains which have undergone intense intracrystalline plastic deformation (Bell & Etheridge I973, White I973). It is also apparent, as originally noted for the type mylonite (Lapworth I885, Teall I918), that extensive recrystallization and neomineralization occurs in the groundmass of most cohesive 'cataclastic' rocks. (A) TEXTURAL CLASSIFICATION Because changes in textural type tend to be gradational (P1. 3 a) and some fault rocks retain mixed textures resulting from polyphase deformation under different T A B L E I : Textural
classification of fault
RANDOM FABRIC
FOLIATED
-
FAULT (visible
fragments
(vlslble
fragments
FAULT
L
rocks
BRECCIA
>30% o f r o c k m a s s )
GOUGE 430% o f r o c k
mess)
w PSEUDOTACHYLYTE
CR~SH FINE
BRECCIA
CRUSH BRECCIA
CRUSH NICROBRECCIA
(frasments
• 0.5
cm) I
(O.lcm • frasso (fragments
< 0.Scm)
< 0 . I cm)
O PROTOCATACLASITE
PROTOHYLOKITE
!
N
..... ¢.o '-tO
O !
CATACLASITE
I~fLONITE
Z
ULTRACATACLASITE
ULT~ONITE
BLASTONYLONITE
~O
Loch
Eport
SCOTLAND
C~
%
Loch 0bisary // I/
",',',
~ E A V A L ' .
• 347m .
.
.
"," .
.
.'',"
.
.
.
.
' . ' o ' . ' ' . ' ' , ' - ' , .
|..
J . ~
KEY Thrust
r "
Crush zone Phylbniti¢ shear belt
~ .
•
Crush melange.
;~
~,~
...-
,
.
.
..:.,,.
0
,
~
.I
I
.
increasing
.
".
°
.
°
.
.
ltl
-
Mylonitic { 30-60°
b '
. •
I
foliatm ( 60_90o ..4" J
palaeotemperature
,
>
IX~eudofachylyte- uitraca~¢lasite crush zone
F z o . I M a p and schematic cross-section of the Outer Hebrides Thrust in Eaval block, N Uist.
194
R. H. Sibson T A B L E 2 "Fault rocks and style of faulting in the Outer Hebrides Thrust zone StTle of Faulting
Fault Rocks
Brittle shearing of intact rock~ Pseudotachylyte Sliding on existing planes
J
Cataclastic crush zones
Cataclaslte-Ultracataclasite (& some pseudotachylyte)
Crush Melange
Crush breccias, microbreccias & protocataclasite
Quasi-plastlc shear zones Crush Melange with crude mylonitic foliation
Phyllonltic mylonites & ultramylonites Protomylonites
crustal conditions (PI. 2), (Christie I96O) attempts to classify the textures of fault rocks in a pigeon-hole manner that can never be entirely satisfactory. The scheme outlined in Table I serves as a simple reference system, adequate for the following discussion. It is based largely on that put forward by Spry (I969, p. 229), modified to avoid genetic connotations, with some additional elements from Higgins' (i 97 I) classification. The whole has been rearranged to emphasize a division which I consider to be of great mechanical significance; the separation of those fault rocks with an essentially random shape (and crystallographic) fabric from those that are foliated, usually with a strongly inosculafing L-S shape fabric (fluxion structure). Another main division is made on the presence or absence ofprirnary cohesion, the cohesive fault rocks then being further subdivided on the nature of their matrix. Cohesive rocks in which tectonic reduction in grain size has dominated processes of grain growth form the bulk of the commonly recognized fault rocks formed at other than near surface conditions (P1. I). The term phillonite is retained as a useful descriptive name for hydrated, mica-rich mylordtes and ultramylonites which have the silky appearance of phyllites. (B) D I S T R I B U T I O N
WITHIN
FAULT
ZONE'S
Many major ancient fault zones are now exposed at erosion levels which correspond to considerable depths when the faults were active. They often consist of a mesh of shear zones enclosing lozenge-shaped areas of comparatively undeformed rock, the whole being perhaps a few kilometres or possibly tens of kilometres in width. Variations in both the style and the rock products of faulting occur across such zones when they have a large component of finite dip-slip. As an example we consider a section across the Outer Hebrides Thrust, a major dislocation of probable late Caledonian age which borders the eastern coastline of the Outer Isles in NW Scotland (Francis & Sibson 1973, Sibson 1975). This structure is especially suitable for discussion, because it disrupts crystalline Lewisian basement gneisses of fairly uniform bulk composition, and because fault rocks are generally best
J1 geol. Soc. Lond.
SIBSON
x33 , i977
a
•
. ~
'~,q~
~
e
::
PLATE I
Mylonite series (a--protomylonite, b--mylonite, c--ultramylonite) versus Cataclasite series (d--protocataclasite, e--cataclasite, f--ultracataclasite). All specimens from the Outer Hebrides Thrust zone (crossed polars except c & f, × 7)194
J1 geol. Soc. Lond.
I33 , 1977
SIBSON
(a)
(b)
PLATE 2 Polyphase texture from Seaforth Head, Lewis (NB 305 158). (a) Vein material forming breccia matrix and disrupting banded mylonite-ultramylonite, was probably randomfabric pseudotachylyte developed by seismic faulting within a ductile shear zone (plane polars, × 3.0). (b) Porphyroclasts within vein are now aligned in a strong shape fabric as a result of continued ductile deformation (plane polars, x 15 . 75)-
J1 geol. Soc. Lond. i33 , i977
SIBSON
(a)
(b)
PLAIE 3" (a) Gradational change in texture. Amphibolite -+ protomlyonite -~mylonite --~ ultramylonite (Ness, Lewis NB 521664) (plane polars, × 14.o ). (b). Phyllonitic shear belt cutting crush m61ange, east coast ofN Uist (NF 922597). Note marginal curve-in of schistosity and asymmetric chevron folds developed within the belt.
Fault rocks and mechanisms
195
developed around thrusts as a result of their greater capacity for storing elastic strain energy (Sibson 1974). The section described (Fig. I) lies along the eastern coast of North Uist south of Loch Eport, about midway along the thrust zone. West of the thrust base the Lewisian complex consists largely of biotite-hornblende-quartzo-feldspathic banded gneisses with varying amounts of amphibolite. Large Laxfordian fold structures have imposed a general NW-SE structural grain which is cut by the thrust (Coward et al. I97O). Within the thrust zone the parent assemblage is much the same, but blocks of pyroxene granulite occur locally. From west to east across the thrust zone there are progressive changes in the style of faulting and the associated fault rocks (Fig. I and Table 2). West of the thrust zone proper, pseudotachylyte has been generated by rapid transient sliding on extremely brittle faults which developed in localized failure zones within the gneiss complex (Sibson I975). These increase in number towards the thrust base where disrupted pseudotachylyte veining is also found in crush zones, here ranging up to 3° m in thickness, intensely microfractured and largely composed of cataclasite and ultracataclasite. East of the thrust base the style of faulting becomes more ductile, and the crush zones give way to mylonitic shear belts up to 5 ° m in thickness, lying in a braided network more or less concordant with the thrust zone envelope. This network penetrates a crush mdlange (crush breccias, microbreccias, protocataclasite and locally protomylonite) of crushed acid gneiss with a varying metabasite content. All of the rocks within this mdlange are to some extent metamorphically downgraded, and within the shear belts almost total retrogression to lower greenschist assemblages (e.g. quartz + albite + epidote + muscovite + chlorite ± actinolite) has occurred and the rocks are phyllonites with strong L-S tectonite fabrics (P1. 3b). Though the schistosity developed within these ductile shear zones can be seen to intensify and curve in from the margins in thrust sense, within the zones it is invariably folded into a series of asymmetric crenulations and chevron folds with subhorizontal axial planes and a consistent down-dip vergence, indicating that the main phase of thrusting was followed by some down-dip sliding. While direct field evidence for the contemporaneity of brittle and ductile thrust deformation is lacking, strong arguments can be made against one set of features having preceded the other.
2. Rates of faulting and the energy budget of fault zones The rate and manner of energy dissipation in fault zones must be key factors affecting the genesis of the different varieties of fault rock. It is therefore apposite to consider the range of slip- and strain-rates associated with tectonically active faults. (A) B I M O D A L
DISTRIBUTION
OF S L I P - R A T E S
Across currently active faults, shear displacements generally take place either by intermittent seismic failure or by aseismic fault creep (Scholz et al. x969). In the
196
R. H. Sibson
Io,:j~ 9
-.-.. . . . ..-..
8-
(W'n{2) 7
ili
654321 -
•
.
.
ilJ
0-
-2
FIG.
I'.',':
-12 -11 -10
I
I
-a -8
I
I
I
I
-7 -s -s -4
I
I
I
-3 -2 -I IogloV
I
a
2
Power dissipated per unit area of a f a u l t s u r f a c e (Q.) v e r s u s slip r a t e ( v )
for various values of shear resistance (*t) (see eqn. 2t). Stippled areas represent the d o m i n a n t velocity modes o f faulting.
I
1
( m . s -11
former case, transient slip-rates of perhaps IO-IOO cm s -z may be attained over periods of at the most a few tens of seconds, at intervals of tens to thousands of years (Brune 197o ). In the latter, observed slip-rates usually lie in the range o. i - i o cm yr -1, which corresponds to the field of relative velocities between interacting lithospheric plates. In the upper crust aseismic fault creep may be quasicontinuous or episodic, and appears to be a metastable process which in some circumstances can degenerate into seismic failure (Scholz et al. 1972). Both fast and slow movements on faults generally occur as localized shear dislocations which do not extend over the entire surface of existing geological faults. One may therefore expect a dominantly bimodal frequency distribution of sliprates, though all intermediate gradations must occur (Fig. 2).
100 km
•
w 10 km I km 100 m
"::::!::i::. tom -
~.
":i}!:!'iii'
lrn -
~f'O
" ....
~ " " ~>"
i :.~i~!:.,
FIo.
3 W i d t h of shear zone (w) versus rate of shear straining (~) for aseismic slip-rates.
lOmm -lmm --14
I -13
I -12
I
-1|
I -10
I -9
I -8 Iogto
-1 ({1) - 6
-5
197
Fault rocks and mechanisms (B)
STRAIN-RATES
ASSOCIATED
FAULTING
WITH
Because rapid faulting involves loss of continuity across the fault, it is probably not meaningful to talk of the strain-rates induced inside a fault zone by seismic slip. While the quasi-continuous slip ofaseismic fault creep may take place on discrete planes at high crustal levels, movement at depth is probably accommodated within a shear zone of finite width (w). If the rate of shear straining is assumed to be constant throughout the zones, crude estimates of the likely range of values can be obtained from the velocity field for aseismic creep using y
---
[I)
--
W
where v is the slip velocity (Fig. 3). At depth fault creep may be taken up by steady shear in planar zones only a few tens or hundreds of metres in width, corresponding for example to the phyllonitic shear belts and cataclastic crush zones associated with the Outer Hebrides Thrust. This would give rise to localized shear strainrates as high as lO -1° to lO -11 s -1. These estimates differ considerably from those of Whitten (1956) who derived a strain-rate of 3 x IO-1~ s -~, often quoted as geologically 'representative', from triangulation of a broad region around the San Andreas Fault. His measurements probably reflect the rate of elastic shear strain accumulation around locked portions of the fault. L INPUTOF STKAINJ
r
J
, l ~LEASE OF E~RGY
)
,,, ~TORE OF ELASTIC STRAIN ENERGY)
SEISMIC MODE
ASEISMIC MODE
/.Jj h
,
'
l
-
OI"SIFATIO'I
L'z£:::'"[ )'
I INT~CRYSTALLINE DEFOR~TION
k.
HEAT GENERATED
) HEAT ABSORBED IN
I
HEAT GAINED FROM
'21%1=: HEAT FLUX IN FAULT ZONE
F~G. 4 T h e energy budget of fault zones.
198
R. H. Sibson (G) T H E
ENERGY
BUDGET
The energy budget of crustal fault zones is illustrated schematically in Fig. 4 for the two dominant modes of faulting. Input of elastic strain energy is derived from differential movements, usually as a direct result of lithospheric plate interaction. For aseismic fault creep to be a truly steady-state process, input of energy must be exactly balanced by its dissipation in the various sinks, so that the amount stored around the zone remains constant. If a fault zone 'sticks', the store of energy increases until some failure criterion is exceeded. Seismic slip may then occur by Reid's (191o) mechanism of elastic rebound, with sudden partial release of the stored energy. In earthquake faulting, the proportion of energy radiated as seismic waves is given by E8 = n E
(2)
where E is the total energy released and ~ is the seismic efficiency. Some idea of the radiation energy for an earthquake with surface-wave magnitude Ms can be obtained from the empirical Gutenberg-Richter relationship, logloE, = 1"5 M, + 11"8
(3)
Richter (I958), but estimates of the efficiency factor (and thus the total energy release) for crustal earthquakes are notoriously uncertain. Most lie in the range o. I-I o per cent (Pshennikov 1965, King 1969), though there have been suggestions that ~ approaches unity for very large events (Brune 197o ). Apart from the wave energy, possible sinks for energy dissipation are essentially the same for seismic as for aseismic faulting. Energy may be expended against or gained from gravity, depending whether the centre of gravity for the displaced rock and fluid on both sides of the fault is raised or lowered with respect to gravitational equipotentials. The amount involved will be minimal for strike-slip faulting. Proportionately the most energy may be expended against gravity during reverse faulting which on a large scale leads to crustal thickening, while energy is usually gained from gravity during normal faulting. Pshennikov (1965) suggests that for large earthquakes, the work done against gravity is rarely as great as the radiated energy. Thus, if the low estimates for seismic efficiency are correct, much of the energy released during earthquake faulting (perhaps 9 ° per cent or more) goes into rock deformation within the fault zone. Partitioning of dissipated energy between the various groups of rock deformation processes can be expected to vary with slip-rate. The importance of the essentially bimodal distribution is that seismic slip-rates may induce transient phenomena such as rapid localized rises in temperature and/or fluid pressure (Sibson 1973, I975) , which can be expected to leave textural imprints on the rock products of fast faulting, rendering them distinguishable from the products of slow aseismic movements. In this regard, the very fast dissipation of energy accompanying seismic faulting must lead to rapid gains in local entropy, and intense disordering of existing fabrics within fault zones. In contrast, the deformation required to
Fault rocks and mechanisms
I99
Differential Stress
yield point
f
. . . . . .
I "~me",g
intact crust
/
/
/
dastic region.,
I
~
] AA A ~ ~ " ~ o w V$1VV~' .....
,,
,
IStrainor -'-'-"- ]disp,l~.ent
~ weakeni,
~._
Strainor displacement
FIG. 5" Conditionfor localizationof major fault zones.
accomodate aseismic fault creep takes place at strain-rates compatible with flow by intracrystalline plasticity and mass diffusion in silicate rocks. A considerable proportion of the dissipated energy may then be expended in these processes.
3- Localization of fault zones Fault zones arise through the local concentration of deformation. This means that with increasing displacement the shear resistance of the zone either maintains a constant value less than that of the surrounding crust, or progressively decreases as the zone undergoes strain or displacement weakening (Fig. 5). Such ideas are consistent with plate tectonics theory, which suggests that deformation and seismic activity is concentrated at lithospheric plate boundaries, not because they are regions of comparatively high shear stress, but because they represent zones of persistent weakness extending through the lithosphere (Sykes & Sbar 1973). Thus, while it can be argued that cold, intact sialic crust can support differential stresses of at least 1.5 kb (Jeffreys I959) , seismological and other estimates suggest that within established fault zones, shear stress has at the most a value of a few hundred bars (Ambraseys I969, Brune et al. 1969). Processes of rock deformation within fault zones must therefore exhibit strain or displacement weakening, and the rate at which this occurs at different crustal levels must determine the ultimate width to which a fault zone evolves.
4. Crustal fault mechanisms and rock deformation The processes of rock deformation which accompany faulting must change with depth in the crust as a result of varying temperature (T), fluid pressure (PI) and confining pressure (Pc), and will also be affected by the velocity mode of the faulting. Deformation maps have been constructed for quartz by Rutter (I976) and White (i976). Such maps attempt to define the temperature/stress/strain-rate[ grain-size fields over which various steady-state flow mechanisms are dominant, but their application to polymineralic rock is as yet uncertain. However, a point of great importance is that along the P~ T gradients of intermediate facies series (Barrovian) metamorphism, it is not until lower greenschist-facies conditions are
200
R. H. Sibson
attained that penetrative crystallographic fabrics appear in quartzo-feldspathic rocks (Spry I969, p. 5). From the presence and nature of such quartz c-axis fabrics in rocks metamorphosed under medium- and high-grade conditions, one may infer the dominance of thermally activated dislocation processes (Wilson 1973, i975, White x976). Thus the isotherm defining the onset of greenschist-facies conditions, estimated by Turner (1968 , p. 366) to lie in the range 25o-3oo°C for P l o a d ----PH~o, probably marks the lower temperature boundary for extensive crystallographic fabric development in quartz. On this basis, a broad division can be made in the fault mechanisms which can accomodate shear in quartzo-feldspathic crust. At levels above the greenschist transition, where rock particles behave in an almost elastic manner to failure, and are not able to absorb large strains by crystal plasticity (or mass diffusion processes), deformation of rock masses is friction-dominated, so that the regime may be deemed elastico-frictional (EF). Lower down, where a major rock constituent (usually quartz) can readily deform by crystal plasticity, deformation behaviour may be said to be quasi-plastic (QP). (A) ELASTICO-FRICTIONAL (EF) MECHANISMS
(i) Brittle shearfailure of intact rock. It is now well established that the brittle strength of intact materials is largely determined by stress concentrations around included flaws (Griffith 1924, McClintock & Walsh I962), so that a rock mass can be treated as 'intact' if the dimensions of flaws are small compared with the mass under load. Because of the common presence of such features as bedding, foliation, minor faults and joints, the extent to which the upper crust may be considered intact is debatable, and for large-scale faulting this region may best be considered as a blocky aggregate, 'glued' together by the friction across planar disconfinuties (Birch I964). Evidence from laboratory triaxial tests suggests that, to a first approximation, the brittle shear failure of homogeneous intact rock following elastic straining is adequately described by the Mohr-Coulomb criterion, modified to take account of fluid pressure (Price I966, Mogi 1973). The criterion is • = C -b m ( a , -
Pt)
(4)
where • is the shear stress and an the normal stress on the eventual fault, C is the long-term cohesive strength (typically a few hundred bars for crystalline rocks), a n d / ~ is the coefficient of internal friction (generally 0. 5 < / ~ < i.o). It predicts that shear fractures should develop in planes containing the intermediate principal stress (a~) and lying at an acute angle to the maximum principal stress (~i) given by 0i -- ½ tan -1 (I[/xi) (5) In terms of the principal compressive stresses (al > a~ > as), the criterion may be rewritten as ( a , - - a3) = =
~Cv/~ + (K, - - 1 ) ( a 3 - ~o + (~r~ - 1 ) ( a 8 - P s )
Pt)l
t
(6)
Fault rocks and mechanisms
201
where ao is the uniaxial compressive strength of the rock, K, -----
I + sin ¢~ i - - sin ¢~
(7)
tan -1 ~
(8)
and the angle of friction, ¢4
=
Equation 6 predicts a linear increase of differential stress at failure with increasing confining pressure. It is found experimentally that this is accompanied by a decrease in the stress drop at failure to some value of residual frictional strength, and an increase in ductility which in physical terms is represented by a widening of the fault zone that develops (Griggs & Handin I96O). Ultimately, the brittle-ductile transition is reached when frictional resistance becomes equal to the shear strength of the rock, and cohesive cataclastic flow occurs throughout the specimen without stress drop (Orowan 196o). At confining pressures which are less than those at the brittle-ductile transition, the release of strain energy accompanying stress drop ensures that shear failure of intact rock is a seismic process. (ii) Sliding on existing planes. It is generally accepted that frictional sliding on existing planes plays an important role in fault zones down to depths at which the frictional resistance to slip exceeds the ductile yield strength of the rock mass (Orowan 196o ). Experimental work suggests that to a sufficient approximation, the criterion for frictional failure down to depths of perhaps 2o km is adequately described by Amonton's Law, modified to take account of fluid pressure (Dieterich i974). For sliding to occur, the applied shear stress must equal the frictional resistance (,f), given by rf = t~(a, -- Ps)
(9)
where ~, the coefficient of static friction, typically has a value of about 0.75 (Byerlee 1968 ). Once static friction is overcome and slip begins, the kinetic shear resistance is
(io) where/z~ (usually < t~) is the kinetic coefficient of friction. Two situations may arise; either stable sliding occurs at constant load (one possible mechanism for aseismic fault creep), or transient slip takes place accompanied by a partial release of stress. In the latter case, the cycle may be repeated when the stress again builds up to failure, giving rise to a 'stick-slip' oscillation (Jaeger & Cook 1969, p. 64). Stick-slip is now usually cited as the instability mechanism for shallow earthquakes (Brace & Byerlee 1966), and is favoured by comparatively low temperature ( < 3oo°C, say), a high effective stress normal to the fault, the presence of 'brittle' minerals such as quartz and feldspar, and a loading system of low stiffness (Byerlee & Brace I968 , Stesky et al. I974). (iii) Shear across a cataclastic crush zone. Crush zones may develop by the localized
2o~
R. H. Sibson
ductile yielding o f a rock mass through cataclastic flow, or by the progressive accumulation of cataclastic detritus on a fault surface. In the latter case, the transition between stable sliding on a plane and steady-state shear across a crush zone is gradational, but can be said to occur when the crush or gouge zone has widened to an extent that there is no longer interference between asperities on opposing walls. Cataclastic shear-flow under constant or nearly constant load, though a ductile process, is still controlled by friction and is highly sensitive to effective normal stress. Treating an existing crush zone as a granular aggregate, the failure criterion for cataclastic flow may still be approximated by a generalized linear criterion of Mohr-Coulomb form (Eqn. 4), where cohesion, C, has a low value probably increasing slightly with depth, and the frictional coefficient, ~, has a value somewhat less than that for sliding between clean faces of similar material (Engelder et al. 1975)A very important feature of the ~ield condition for friction-dependent cataclastic flow, is that the yield stress is virtually independent of strain-rate, and hence of time (Donath & Fruth t97x). The inherent long-term stability of cataclastic flow is, however, open to question. The behaviour of a crush zone under shear may be governed by a non-linear constitutive law, in which case shear is likely to be metastable (Stuart I974). Indeed, Byerlee & Brace (i969) have reported mechanical instabilities leading to stick-slip behaviour in crushed aggregates under pressures well above their brittle-ductile transition. Cataclastic flow is also accompanied by dilatancy so that fluid pressure is likely to fluctuate considerably; il) has been shown experimentally for a gouge-laden fault in sandstone that slight decreases in effective normal stress can switch behaviour from stable cataclastic flow of the gouge to stick-slip sliding at the gouge-sandstone contact (Engelder et al. t975). By analogy, it is to be expected that, when natural aseismic shear across a crush zone degenerates into seismic failure, slip will be concentrated on nearly planar discontinuities as increasing strain-rates lead to more brittle behaviour. Riedel shears have also been observed to develop in crushed aggregates deforming by steady-state shear (Jaeger & Gay I974) , and bands of more intensely crushed material occur in experimentally produced gouge zones (Engelder et al. I975). (iv) Limiting condition f o r frictional failure. At a depth, z, in the crust the effective vertical stress is usually given by -- a, -- PI = pgz(I -- ~v)
(ii)
where p is the mean crustal density, g is the acceleration due to gravity, and the pore fluid factor
PI pgz
By assuming the vertical stress to coincide with one of the principal stresses, Anderson (x95I) successfully explained the orientation of thrust (~v = o3), wrench ( ~ ---- cr2) and normal ~ = ~a) faults using the Mohr-Coulomb criterion (though any linear frictional criterion, with or without a term for cohesive strength, gives similar results). In the E F rdgime, the limiting condition for frictional failure on existing faults may be expressed on the basis of Eqn. 9 as the
Fault rocks and mechanisms
203
minimum differential stress required to initiate slip at a given depth and fluid pressure (Sibson 1974). Thus,
(.~_
~,) >f
(R'-
1)
(q(R'-- i) + 1}
pgz(~ - ~)
(iS)
where q -- o for thrust faults, o < q = k = {(as -- as)/(al -- as)} < I for wrench faults, q -- I for normal faults, and R'
=
(VI
+
/Z 9" - - /z) - 2
(14)
These equations have profound implications for the relative development of fault rocks in the three types of 'Andersonian' fault zone (see below). • (B) Q U A S I - P L A S T I C
({~P) MECHANISMS
Once a major mineral constituent of a rock mass can deform extensively by intracrystalline plasticity, perhaps aided by mass diffusion processes, planar deformation zones may evolve through localized yielding followed by heterogeneous simple shear of the continuum. (i) Quasi-plastic shear zones. Intracrystalline plasticity is a thermally activated process, rather insensitive to confining pressure (Edmond & Paterson 1972 ). The condition for yield may therefore by approximated by the standard Von Mises criterion (Mogi 1972). This has the physical interpretation that yield occurs when the concentration of distortional strain energy, E~ =
(~1 6G - ~,)~
{1 - - k + k 2}
(15)
(where k is as defined for Eqn. 13 and G is the rigidity modulus) reaches some value characteristic of the material for a given temperature (7") and strain rate (~,); that is
Ea----f(~, T)
(16)
Once yield has occurred, steady-state flow of rocks by intracrystalline processes may be expected to obey a relationship of the form = A exp(--
E / R T ) (~1 -- ~3)"
(17)
where E is an activation enthalpy, and A, R and n are constants (Heard 1976 ). From the concentration of ductile deformation within narrow shear zones, one may infer that initial yielding is accompanied by marked strain-softening. Mead (I 925) suggested that some cataclasis occurs at yield, the accompanying dilation drawing water into the shear zone. The manner in which this may promote strain-softening is discussed below. With quartz first beginning to deform extensively under lower greenschist-facies conditions, it is under these and higher grades of metamorphism that QP shear
R. H. Sibson
204
zones may develop in quartzo-feldspathic crust. Cataclastic processes may, however, persist to different depths for other mineral constituents. In mylonites derived from a quartzo-feldspathic host under greenschist-facies conditions, feldspar typically survives as ovoid porphyroclasts, highly resistant to deformation, while quartz may flow into ribbons by extreme plastic straining. It is this marked contrast in mechanical behaviour between different mineral components that gives rise to the characteristic inosculating textures of the mylonite series. Ramsay & Graham (I97O) have described the development of L--S fabrics in QP shear zones cutting isotropic country rocks. Schistosity lies anti-symmetric across the zones, curving in and intensifying away from the margins where it first appears at c. 135 ° to the shear direction, to lie sub-parallel to the walls in the central high strain regions. Thus schistosity follows the planar X T trajectories of finite strain, with a stretching lineation tracing out the X direction. (ii) Transient seismic shearing. The deformation maps of Rutter (I 976) and White (I976), and the work of Elliot (I 973) suggest that seismic rates of shear cannot be accommodated by mass diffusion processes, and probably not by the dislocation processes ofintracrystalline plasticity, even at the highest crustal temperatures. QP shear zones may therefore be taken to have developed under the more or less steady strain-rates of aseismic shear. However, transient shear fractures may sometimes propagate downwards from the EF regions of fault zones, and seismic instabilities may also occur within the QP r4gime if frictional constraints are overcome, perhaps by creep runaway leading to shear melting (Orowan 1960, Griggs & Baker 1969), or through an increase in fluid pressure. Textures developed by transient shear fractures in Q.P shear zones are likely to be largely random-fabric and of a cataclastic nature, though friction melting may possibly occur. Such rapidly imprinted textures are unlikely to persist, because the continuing processes of plastic crystal deformation, recrystallization and neomineralization within the shear zones will tend to obliterate them (P1. 2). (C) R O L E
OF MASS D I F F U S I O N
PROCESSES
Of the mass diffusion processes, Nabarro-Herring creep probably plays a significant role only in very high temperature deformation (Elliot 1973). Grain boundary diffusion with some sliding may assist dislocation processes and probably becomes the dominant mechanism once grain size drops towards IO t~m, since at constant stress the strain-rate for flow by grain-boundary diffusion is inversely proportional to grain size raised to the third power (Rutter i976 ). The lower temperature bound for grain-boundary diffusion (especially water assisted) is not well established, though pressure solution of quartz may sometimes occur even under diagenetic conditions (Kerrich I974). The question arises: to what extent can these processes occur in the EF regions of fault zones, once significant reduction of grain size has occurred by cataclasis ? This is important, because extensive pressure solution at high levels in fault zones would favour steady-state aseismic shearing. The deformation maps for quartz (Rutter 1976 ) and (White I976 ) suggest that while for grain diametres of IOO t~m
205
Fault rocks and mechanisms
T
FIG. 6. Mechanical effect of fluid pressure fluctuations within the EF r6gime.
C O'3b O'3a
O'lb
(Y
O'la
pressure solution may be inhibited by the comparatively high strain-rates associated with aseismic faulting, it may become important well above the E F - Q P transition as grain size approaches IO ~m. D)
WATER-INDUCED
WEAKENING
OF
FAULT
ZONES
Water in fault zones can potentially bring about weakening in a variety of ways. Firstly, as discussed above, the presence of water lowers the temperature at which grain-boundary diffusion can cause significant deformation by inducing pressure solution. Secondly, there is accumulating evidence that water concentrations in quartz and other silicates act as plasticizing agents, greatly reducing yield strength and increasing ductility by a process known as hydrolytic weakening (Griggs 1967). Thus the transition to intracrystalline plasticity in quartz above say 3oo°C, is dependent on the presence of sufficient water to promote hydrolytic weakening (Jones 1975). Thirdly, water together with other fluids may play a direct mechanical role in accordance with the law o f effective stress put forward for rocks by Hubbert & Rubey (1959). For the usual situation where the compressibility of the rock mass is much greater than that of individual grains, the effective stress (in summation notation) is
(I8)
= mj -- PS'3~j
where mj is the applied stress and 3~j is Kronecker's delta. Thus high fluid pressures may be expected to offset the effects of increasing confining pressure with depth, lowering intergranular frictional resistance and promoting EF behaviour. In particular, if fluid pressure inside a fault zone is greater than that outside, considerable relative weakening of the zone may result. From equations 6 and 7, by taking differentials, we obtain A(al - - %) =
- - 2 sin ~ i -sin~
APf
(I9)
This may be applied to any linear frictional failure criterion with a slope, ~ = t a n - i F in the Mohr diagram (Fig. 6). For a frictional coefficient, F = 0.75, the expression becomes
~(~
-
~.) =
-s',~Pt
(~o)
206
R . H . Sibson
Thus an increase in fluid pressure can cause a correspondingly greater decrease in the differential stress required for further shear failure. Further weakening may also occur by the formation of en echelon extension fractures within a fault zone, if at some stage fluid pressures and the prevailing stress field satisfy the conditions for hydraulic fracturing described by Secor (i965) and Price & Hancock (I972). Finally, there is evidence from both active and ancient fault zones suggesting the presence of high fluid pressures at the time of faulting (e.g. Berry I973, Beach & Fyfe i972 ).
5. Thermal effects of faulting (A) S T E A D Y
HEATING
FROM ASEISMIC SHEAR
Steady aseismic shearing across a crush or shear zone with shear resistance, rl, gives rise to a constant heat flux per unit volume of the zone, i-I
Q" -
-
r, . v
-
(2i)
tl)
where v is the slip-rate and w is the width of the zone. T h e resulting temperature rise both in the zone and its surroundings can be calculated using the standard conduction equations for constant heat production in a strip of width, w, lying in an infinite solid with the same thermal properties (Carslaw & Jaeger I959, section 2.I I, p. 80). Fig. 7 shows the temperature distribution around a i o m wide shear zone across which a displacement of I km has taken place at I-IO cm yr -1 against shear resistances of o . i - i . o kb. Points to be noted are: (I)
For these slip-rates, rise in temperature occurs over a region much broader t h a n the shear zone and is relatively insensitive to its width. (2) At the centre of the shear zone, for a given displacement and slip-rate, A T oc .rf .
(3) For a given displacement and shear resistance at the centre of the zone, AT
oc V'~..
(4) T h e steepness of the thermal gradient adjacent to the shear zone increases with slip-rate. Tf - 1 kbar
100 bar 5
5O AT -4
(°C)
-3
2O ~ 10-
-2
FIG. 7.
-1
T e m p e r a t u r e distribution a r o u n d a steady-state shear zone (thermal diffu-
1cm.w"1 ~ [, 100
I
I
I
200
300
400
I 500
I 600
I 700
I 800 d
I 900 1000 (m)
sivity, ~-= o. 007 cm%-l).
207
Fault rocks and mechanisms (B) T R A N S I E N T
HEATING
F R O M SEISMIC F A U L T I N G
From the preceding discussion, it follows that the very much greater rates of frictional dissipation associated with seismic faulting (Fig. 2) may lead to large transient increases in temperature on the fault plane. Indeed, from theoretical analyses, several workers (Jeffreys 1942, Anderson 195 I, McKenzie & Brune 1972) have suggested that friction melting with production of pseudotachylyte should occur on seismically active faults at depths of more than a kilometre or so. If it does occur, melting could lead to almost complete release of stress by lubrication of the fault plane, and is therefore of considerable interest seismologically. However, these analyses are founded on the assumption that frictional sliding on faults occurs under dry conditions. Sibson (I973) showed that the general scarcity of pseudotachylyte can be explained by an interaction between temperature rise and the intergranular fluid present in most established fault zones. If constant volume conditions are maintained, temperature rise is limited to about I oo°C at depths down to I O km or so. Other factors affecting the production of pseudotachylyte, particularly that associated with the Outer Hebrides Thrust, are discussed by Sibson (I975). C) R O C K D E F O R M A T I O N
BY,THERMAL
FRAGMENTATION
Apart from the extreme process of friction melting under dry conditions, disaggregation of rock by grain-boundary and transgranular cracking may" arise from the sudden imposition of steep thermal gradients adjacent to the fault. In part this results from the unequal expansion of individual grains across which there is a strong thermal gradient, but the effect will be enhanced by the presence of different minerals, particularly those with considerable anisotropy of thermal expansion. In a study of the thermal expansion of igneous rocks, Richter & Simmons (1974) pointed out that for a given rise in temperature, the number of cracks produced increases markedly once the heating rate exceeds 3 × IO-2°C s-l- Crack production is therefore likely to be very great during seismic slip, when heating rates may approach io3°C s -1 or more. Rapid heating may also promote overpressures within fluid inclusions, causing explosive decrepitation and grain shattering (Sibson 1975). All such processes of thermal fragmentation will tend to disorder existing rock fabrics. D) D I S T O R T I O N
OF E X I S T I N G
ISOTHERMS
A point of interest is the extent to which vertical displacement across a major fault zone, such as that of the Outer Hebrides Thrust, can distort existing isotherms. This structure has an outcrop width of at least 12 km in Lewis, and with an eastward sheet dip of 25 °, its true thickness (w) must exceed 5 km. The thermal constant for a slab of this thickness (giving a rough estimate of the time needed to conduct a significant amount of heat across it), is i2)2
t¢ , ~ - - = K
I'2 × I o e y r
(22)
~o8
R. H. Sibson
when the thermal diffusivity, ~ = 0.007 cm~s -~. For likely time-averaged slip-rates of I-IO cm yr -~, the reverse dip-slip across the Outer Hebrides Thrust (thought to be in the range i o + 5 km) could have been accomplished in a time, t~, of zo~-IO n years. If ts >> t,, horizontal thermal gradients across the fault zone would remain low throughout the thrusting episode, whereas if tl < t, little heat would be conducted across the zone and isotherms would be vertically displaced almost as passive markers. For the Outer Hebrides Thrust, where t~ ~ re, one may infer that during the thrusting episode there was a significant increase in temperature and a progressive change in metamorphic environment passing horizontally from the downthrown to the upthrown side of the fault, as evinced by the distribution of fault rocks.
6. Conceptual model for a major fault zone The preceding arguments lead to a two-layer model for a major fault zone. At high crustal levels, E F behaviour produces mainly random-fabric fault rocks, whereas at depth rocks of the mylonite series are generated by QP processes in ductile shear zones (Fig. 8). In quartzo-feldspathic crust, the EF[QP transition probably corresponds roughly with the 3oo°C isotherm, marking the lower temperature boundary to greenschist-facies conditions. Thus, for 'normal' geothermal gradients of ~o-3o°C]km, the transition will lie in the interval lO-15 km. Reverse dip-slip across the sheaf of dislocations making up such a fault zone would lead to distortion of isotherms and the horizontal zonation of fault styles and rocks seen in sections across the Outer Hebrides Thrust. (A) R O C K
PRODUCTS
r
OF T H E E F R E G I M E
Within this rdgime, faulting at both seismic and aseismic rates will tend to produce essentially random-fabric fault rocks by frictional processes. All sliding on planes
"~-~::;~.;'~","~:_,
er --
(pseudotachylyte if dry)
Cohesive.
"~i~.. :~..%....
r a n d 0 m - fabric
" ':~;~' i : ~ : :"~'~,~.-:~,. ,
crush breccias,rocks '~;i~:ii!.~..~;~i~'.;~!~:~" . of the cataclasite series,.: ',"i:i~i~.~...~.~;~.%1.'~!i~'~"...: pseudotachylyte if dry "' :.i~!!i~;~!~!i~i:.i:!~;;.'~::;i: ~;:~;.~',
10-15 kmI ~.
I
Cohesi vfoiiat e. ed rocks of the mylonite =a'~,~,~.\.~,~ ~ series& blastomylonites '~~ ~ ~
FIO. 8. C o n c e p t u a l m o d e l of a m a j o r fault zone.
/
Fault rocks and mechanisms
209
in quartz-rich rocks leads to frictional wear by brittle shearing of asperities, plucking of grains from sidewalls and ploughing of asperities into opposing walls. Additionally, sliding at seismic rates may induce thermal fragmentation, and pseudotachylyte may be generated by friction-melting under dry conditions at depths of more than a ldlometre or so. Aseismic fault slip may take place by stable sliding on planes and/or cataclastic shear flow, the latter probably becoming dominant with increase in effective confining pressure. This last factor is probably the most important in governing the rate of generation of cataclastic detritus for a given rock type. Formation of incohesive gouge and fault breccia is probably restricted to relatively near-surface conditions, and the most common products of the EF rdgime will be rocks of the cataclasite series. Within crush zones, cataclastic detritus is progressively comminuted by fracture and rotational attrition. As temperature increases with depth, marie minerals with ready glide planes may be dragged out and smeared along sliding surfaces (Stesky et al. I974) , while mortar texture may arise from the marginal recrystallization of quartz grains that have undergone slight plastic straining. Steady cataclastic flow within a crush zone may lead to the formation of crude shape-fabrics by the alignment of elongate porphyroclasts. If this elongation is crystallographically controlled, weak lattice preferred orientation may result. Such shape fabrics may be enhanced by pressure-solution processes once grain size is reduced. (B) R O C K P R O D U C T S
OF T H E QP RI~OIME
Within the QP r6gime, the products of aseismic shearing will be rocks of the mylonite series (and blastomylonites), and the sequence protomylonite -+ mylonite -+ ultramylonite may be observed passing from the margins to the interiors of ductile shear zones. Inside such zones, random-fabric fault rocks (cataclasite series and even pseudotachylyte) may be generated by seismic faulting from time to time, but rarely survive. (Cl) F A C T O R S A F F E C T I N Q
THE EF/QP
TRANSITION
In the Outer Hebrides Thrust zone, QP shear zones containing phyllonites penetrate essentially random-fabric crush m61ange. Bearing in mind the rather low lateral thermal gradients induced by aseismic shearing (Fig. 7), it is apparent that factors other than temperature affect the EF[QP transition which must cusp sharply upwards in the vicinity of these shear zones. Probably the most important of these is the availability of water to promote hydrolytic weakening in quartz. The very considerable hydration that has gone on in these shear zones (the water content of the phyllonites at circa 2 per cent by weight, is about twice that of the surrounding rocks) bears watness to their role as conduits for syntectonic aqueous flow. In regions deficient in water, the EF[QP transition may occur at greater temperatures and depths.
R.H. Sibson
2IO Thrust
Wrench
Normal
(k = [) 4
,
1.6
"
1
Ed
16
:
2.56
:
1
Depth to any stress value
1
"
2.5
:
4
(al
-
%)
TAB L E 3 : Ratios of differential stress (al-a3) , distortional strain energy (Ed) and depth to any particular stress level around "Andersonian" thrust, wrench and normal faults (;~ ----o. 75).
As a consequence of the standard Von Mises criterion for yield (Eqn. I6), the
.EF/Q.P transition may be expected to occur at different levels for different types of faulting. Estimates of the ratios of differential stress and concentrations of distortional strain energy required to initiate frictional sliding on thrust, wrench and normal faults may be obtained from equations 13 and 15 (Table 3) (Sibson I974). The much greater concentrations of strain energy around thrust faults suggest that for a given thermal gradient, the EF/QP transition occurs at significantly higher levels for these structures than for normal faults, with the transition for wrench faults lying somewhere in between.
(D)
VARIATION
OF
SHEAR
RESISTANCE
WITH
DEPTH
For the simple two-layer model described above, one may infer from equation 13 that shear resistance within the EF rdgime of the fault zone is essentially frictional and, for a constant pore fluid factor, increases linearly down to the EF/QP transition. The rate of increase of frictional resistance with depth will be greatest for thrusts and least for normal faults. Two factors, considered in relation to the general flow law (Eqn. 17) expected to operate in QP shear zones, suggest that resistance to steady aseisrnic shear decreases below the transition. Firstly, at constant strain-rate, increase of temperature with depth must lead to a decrease in shear resistance. Secondly, there is evidence that ductile shear zones widen with depth (Bak et al. 1975). As the timeaveraged rate of displacement across a major fault zone is likely to remain constant throughout the lithosphere, the rate of shear straining must therefore decrease with depth. Decreasing strain-rate augments the effect of increasing temperature, ensuring that shear resistance decreases with depth after reaching a peak value in the vicinity of the EF/QP transition (Fig. 8). The concept is supported by experimental evidence (Rutter I97~ ). Clearly, the presence and amplitude of a peak shear-resistance must play key roles in determining the gross behaviour of a fault system, as the concentration of distortional strain energy is a function of the square of the differential stress (Eqn. i5). That the peak value is likely to be greatest for thrusts and least for normal faults (Table 3), is borne out by the global pattern of energy release from shallow earthquakes (Sibson I974). Significantly, most of the seismic energy released along the San Andreas Fault comes from the depth interval 5-IO km (Press & Brace I966 ).
Fault rocks and mechanisms
21 I
7. Conclusions (I) Variations occur in the style and rock products of faulting across the Outer Hebrides Thrust and other ancient fault zones. These can be explained in terms of a two-layer mechanical model in which a zone of elastico-frictional (EF) behaviour generating mainly random-fabric fault rocks, initially overlies a quasi-plastic (QP) r~gime where mylonite series rocks are developed in ductile shear zones. (2) In some cases, fault rocks produced by transient seismic faulting can be distinguished from those developed by steady aseismic shearing. (3) The EF/QP transition can be recognized and mapped in the field more easily than the brittle-ductile transition. (4) Shear resistance within major fault zones reaches a peak value, which is greatest for thrusts and least for normal faults, around the EF[QP transition. For normal geothermal gradients and an adequate supply of water, this lies at depths of Io-I 5 km in quartzo-feldspathic crust. ACKNOWLEDGEMENTS.Much of the work leading to this paper was carried out during tenure of a Royal Commission for the Exhibition of I8I 5 Overseas Scholarship, which is gratefully acknowledged. I thank Dr N. J. Price, Dr E. H. Rutter and Professor J. V. Watson for discussion and critical reading of the manuscript.
References AMBRASEYS,N. N. 1969 . Maximum intensity of ground movements caused by faulting. Proc. 4th Wld Conf. Earthq. Engng. x, I54-7I. Santiago. ANDERSON, E. M. I95 I. The dynamics of faulting. 2nd ed., Oliver & Boyd, Edinburgh. BAK,J., SORENSEN,K., GROCOTT,J., KORSTGARD,J. A., NASH, D. & WATTERSON,J. I975. Tectonic implications of Precambrian shear belts in western Greenland. Wature, Phys. Sci. 254~ 566-9. BEACH, A. & FYFE, W. S. 1972. Fluid transport and shear zones at Scourie, Sutherland: evidence for overthrusting ? Contr. Mineral. Petrol. 36, 175-80. BELL, T. H. • ETHERIDGE, M. A. 1973. Microstructure ofmylonites and their descriptive terminology. Lithos 6, 337-48. BERRY, F. A. F. 1973. High fluid potentials in the California Coast Ranges and their tectonic significance. Bull. Am. Assoc. Petrol. Geol. 5"/, I219-49. BIRCH, F. 1964. Megageological considerations in rock mechanics. In Judd, W. R. (ed.), State of stress in the earth's crust. Elsevier, New York, 55-8o. BRACE,W. F. & BYERLEE,J. D. 1966. Stick-slip as a mechanism for earthquakes. Science, ,IV.Y. i53 , 990-2. BRUNE, J. N. i97o. Tectonic stress and the spectra of seismic shear waves from earthquakes. 07. geophys. Res. 75, 4997-5009 • , HENYEY, T. L. & RoY, R. F. I969. Heat flow, stress and rate of slip along the San Andreas Fault, California. 07. geophys. Res. 74, 3821-7. BYERLEE, J. D. I968. Brittle-ductile transition in rocks. 07. geophys. Res. 73, 4741-5 o& BRACE,W. F. 1968. Stick-slip, stable sliding and earthquakes--effect of rock type, pressure, strain rate and stiffness. 07. geophys. Res. 73, 6o31-7& BRACE,W. F. I969. High pressure mechanical instability in rocks. Science, W.T. 164, 713-15. CARSLAW, H. S. & JAEOER,J. C. I959. Conduction of heat in solids. 2nd ed. Clarendon Press, Oxford. CHRISTIE,J. M. 196o. Mylonitic rocks of the Moine Thrust Zone in the Assynt district, north-west Scotland. Trans. geol. Soc. Edinb. 18, 79-93.
I~
--
R.H.
Sibson
1963 . The Moine Thrust in the Assynt region, north-west Scotland. Univ. Calif. Publs geol. Sci. 40, 345-419. COWARD, M. P., FRANCIS, P. W., GgA~AM, R. H. & WATSOn,J. V. i97o. Large-scale Laxfordian structures of the Outer Hebrides in relation to those of the Scottish mainland. Tectonophysics xo, 425-35. DIF.TE~ICH, J. H. I974. Earthquake mechanisms and modelling. In Donath, F. A. (ed.), Annual review of earth and planetary science, 2, Annual Reviews Inc., Palo Alto, Calif., 275-3oI. DONATn, F. A. & FRUa'H, L. S. 197 x. Dependence of strain-rate effects on deformation mechanisms and rock type. 07. Geol. 79, 347-7 I. EDMOND, J. M. & PAa'~RSON, M. S. x972. Volume changes during the deformation of rocks at high pressure. Int. 07. Rock Mech. & Mining Sci. 9, I6X-82. ELLIOT, D. I973. Diffusion flow laws in metamorphic rocks. Bull. geol. Soc. Am. $ ~ 9645-64. ENGELD~.R, J. T., LOGAN,J. M. & HANDIN,J. i975. The sliding characteristics of sandstone on quartz fault gouge. Pageophys. x x3, 69-86. FRANCIS, P. W, & SmsoN, R. H. I973. The Outer Hebrides Thrust. In Park, R. G. & Tarney, J. (ed.), The early Precambrian of Scotland and related rocks of Greenland, Univ. of Keele, 95-Io4. GRIFFIa'H, A. A. x924. Theory of rupture. Proc. 1st Int. Congr. appl. Mech. Delft, 55-63. GRmGs, D. T. I967. Hydrolytic weakening of quartz and other silicates. Geophys. 07. R. astr. Soc. x4, I9-3I. -& BAKFR, D. W. I969. The origin of deep-focus earthquakes. In: Properties of matter under unusual conditions. Interscience, New York, 23-42. & HIANDIN,J. 196O. Observations on fracture and an hypothesis for earthquakes. Mem. geol. Soc. Am. 79, 347--64. HEARD, H. C. x976. Comparison of flow properties of rocks at crustal conditions. Phil. Trans. R. Soc. Lond. A2839 I73-86. t-IIcGINS, M. W. I97x. Cataclastic rocks. Prof. Pap. U.S. geol. Surv. 687" Hsu, K. J. x955. Granulites and mylonites of the region about Cucamonga and San Antonio Canyons, San Gabriel Mountains, California. Univ. Calif. Publs. geol. Sci. 3o, 223-352. H~JBFRT, M. & RUB~Y, W. W. I959. Role of fluid pressure in mechanics of overthrust faulting. Bull. geol. Soc. Am. 7o9 x15-2o 5. JA~GEX%J. C. & COOK, N. G. W. I969. Fundamentals of Rock Mechanics. Methuen, London. -& GAY, N, C. i974. Behaviour of lightly confined granular material. Int. 07. Rock Mech. & Mining Sci. xx~ 295-3oI. JFFFRE'ZS, H. I942. On the mechanics of faulting. Geol. Mag. 79, 29I-5 • -I959. The Earth. 4th ed., Cambridge Univ. Press. JONES, M. E. I975. Water weakening of quartz and its application to natural rock deformation. J l geol. Soc. Lond. x3x ~ 429-32. KE~RICn, R. F. x974. Aspects of pressure solution as a deformation mechanism. Thesis Ph.D., Univ. London (unpubl.). KING, C. Y. I969. Seismic efficiency, J . geophys. Res. 74, 17o2-3 . LAPWORa'H, C. I885. The highland controversy in British geology; its causes, course and consequence. Nature, Lond. 32, 558-9 . MCCLINTOCK, F. A. & WALSH,J. B. i962. Friction on Griffith cracks under pressure. Proc. 4th U.S. Natl. Congr. appl. Mech. I o 15-~ ~. McKE~zI~, D. & B g u ~ , J. N. ~972. Melting on fault planes during large earthquakes. Geophys. J.R. astr. Soc. ~9~ 65-78. MEAD, W. J. ~925. On the geological role of dilatancy. J. Geol. 33~ 685-98. Mock, K. I97~. Fracture and flow of rocks. Tectonophysics x3, 54~-68. I973. Rock fracture. In Donath, F. A. (ed.), Annual review of earth and planetary science, x~ Annual Reviews Inc., Palo Alto. Calif., 63-84. O~OWAN, E. x96o. Mechanism of seismic faulting. Mem. geol. Soc. Am. 79~ 323-45. P~ss, F. & BRAC~, W. F. I966. Earthquake prediction. Science N.Y. x5~, I575-84. PRm~, N . J . I966. Fault and joint development in brittle and semi-brittle rock. Pergamon, London. & HANCOCK, P. L. I972. Development of fracture cleavage and kindred structures. Int. geol. Congr. 3, 584-92"
Fault rocks and mechanisms
213
PSHENmKOV, K. B. I965. The energy balance in the focal region of a strong earthquake. Izv. Acad. Sci. U.S.S.R., Phys. Solid Earth xo, 17-24. R~'~SAY,J. G. & GRAHAM,R. H. I97o. Strain variation in shear belts. Can.07. Earth Sci. 7, 786-813 • R~ED, J . J . 1964. Mylonites, cataclasites and associated rocks along the Alpine Fault, South Island, New Zealand. N.g. 07l Geol. Geophys. 7, 645-84 • REID, H. F. I9IO. The mechanics of the earthquake. In, The California earthquake of April 18, 19o6, Report of the State Investigation Commission, Part 2, I6-28, Carnegie Institution of Washington, Washington D.C. RICHTER, C. F. I958. Elementary Seismology, Freeman, San Francisco. RICHTER, D. & SIMMONS, G. I974. Thermal expansion behaviour of igneous rocks. Int. 3. Rock Mech. & Mining Sci. Ix, 4o3-11. RUTTER, E. H. I972. The influence of interstitial water on the rheological behaviour of calcite rocks. Tectonophysics 14, 13-33. I976. The kinetics of rock deformation by pressure solution. Phil. Trans. R. Soc. Lond. A283, 2o3-I9. SCHOLZ, C. H., WYss, M. & SMITH, S. W. I969. Seismic and aseismic slip on the San Andreas Fault. 07. geophys. Res. 74~ 2o49-69. , MOLNAR, P. & JOHNSON, T. I970. Detailed studies of frictional sliding of granite and implications for the earthquake mechanism. 07. geophys. Res. 74, 2o49-69. SECOR, D. T. I965. Role of fluid pressure in jointing. Am. 07. Sci. ~'63~ 633-46. SmSON, R. H. I973. Interactions between temperature and pore fluid pressure during earthquake faulting--a mechanism for partial or total stress relief. Nature, Phys. Sci. 243, 66-8. I974. Frictional constraints on thrust, wrench and normal faults. Nature, Phys. Sci. ~'49, 540-3. I975. Generation of pseudotachylyte by ancient seismic faulting. Geophys. 37. R. astr. Soc. 43, 775-94. SPRY, A. I969. Metamorphic Textures. Pergamon, London. STESKY,R. M., BRACE,W. F., RILEY, D. K. & ROBIN, P. Y. F. '974- Friction in faulted rock at high temperature and pressure. Tectonophysics 23~ 177-003. STUART, W. D. I974. Diffusionless dilatancy model for earthquake precursors. Geophys. Res. Letters x, 0 6 i - 4. SYKES, L. R. & SBAR, M. L. I973. Intraplate earthquakes, lithospheric stresses and the driving mechanism of plate tectonics. Nature, Phys. Sci. 245, 298-3o2. TEALL, J. H. H. i918. Dynamic metamorphism; a review, mainly personal. Proc. Geol. Assoc. 29, I-15. TURNER, F . J . 1968. Metamorphic petrology. McGraw-Hill, New York. WATERS, A. C. & CAMPBELL, C. D. 1935. Mylonites from the San Andreas Fault zone. Am. 07. Sci. 29, 473-5o3. WHITE, S. I973, The dislocation structures responsible for the optical effects in some naturally deformed quartzes, ft. Mater. Sci. 8, 49o-9 . 1976. The effects of strain on the microstructures, fabrics and deformation mechanisms in quartzite. Phil. Tram. R. Soc. Lond. A283, 69-86. WHITTEN, C. A. I956. Crustal movements in California and Nevada. Trans. Am. geophys. Un. 37, 393-8. WILSON, C. J. L. i973. The prograde microfabric in a deformed quartzite sequence, Mt. Isa, Australia. Tectonophysics 19, 39-81. I975. Preferred orientation in quartz ribbon mylonites. Bull. geol. Soc. Am. 86, 968-74. Received lO March 1976, revised typescript received 1 October I976. Read at 'The mechanics and effects of faulting' Ordinary General Meeting, at Burlington House on Io March 1976. RICHARD H. SIBSON,Department of Geology, Royal School of Mines,, Imperial College, London SW7 2BP.
