Escolar Documentos
Profissional Documentos
Cultura Documentos
VINE
Global
Tectonics
SECOND
EDITION
GLOBAL TECTONICS
Global Tectonics
PHILIP KEAREY
Department of Geology
University of Bristol
FREDERICK J. VINE
School of Environmental Sciences
University of East Anglia, Norwich
SECOND EDITION
b
Blackwell
Science
© 1990,1996 by Blackwell Science Ltd,
a Blackwell Publishing company
BLACKWELL PUBLISHING
All rights reserved. No part of this publication may be reproduced, stored in a retrieval
system, or transmitted, in any form or by any means,electronic, mechanical, photocopying,
recording or otherwise, except as permitted by the UK Copyright, Designs,and Patents Act
1988, without the prior permission of the publisher.
9 2005
Kearey,P.
Global tectonics / Philip Kearey, Frederick J. Vine. — 2nd ed.
p. cm.
Includes bibliographical references(p. -)and index.
ISBN 0-86542-924-3
1. Plate tectonics. I. Vine, F. J. II. Title.
QE511.4.K43 1996
551.1^36—dc20 95-49623
CIP
ISBN-13: 978-0-86542-924-6
A catalogue record for this title is available from the British Library.
The publisher's policy is to use permanent paper from mills that operate a sustainable
forestry policy, and which has been manufactured from pulp processed using acid-free and
elementary chlorine-free practices. Furthermore, the publisher ensures that the text paper
and cover board used have met acceptable environmental accreditation standards.
We thank Professors H. Robert Burger, Smith Col fessor W. Jason Morgan advice on plate kinematics.
lege; Stephan P. Phipps, University of Pennsylvania; We also acknowledge the intellectual stimulation
and Eli A. Silver, University of California at Santa provided by our colleagues and students in the Uni
Cruz, for their most helpful reviews of the original versities of Bristol and East Anglia, and at Princeton
manuscript. Professor Tony Hallam provided useful University where FJV spent a sabbatical term during
comments on palaeontological matters and Pro the preparation of this book.
1 / Historical perspective
i
i
Fig. 1.2 Taylor's mechanism for the formation of Tertiary noted, both convection and Earth expansion being
mountain belts by continental drift (after Taylor, 1910). considered unlikely. It is interesting to note that
historically many southern hemisphere geologists,
such as du Toit, Lester King and S.W. Carey, have
curred in response to the centripetal force experi been advocates of drift, perhaps because the geologi
enced by the high-standing continents because ofthe cal record from the southern continents and India
Earth's rotation. Simple calculations showed the favours their assembly into a single supercontinent
forces exerted by this mechanism to be much too (Gondwana) prior to 200 Ma ago. The non-drifters
small. Although in the later editions of his book this had difficulty in explaining the present separation of
approach was dropped,the objections ofthe majority faunal provinces, for example, which could be much
of the scientific community had become established. more readily explained if the continents were for
Du Toit, however, recognized the good geological merly together, and their attempts to explain these
arguments for the joining ofthe southern continents apparent faunal links or migrations also came in for
and A. Holmes,in the period 1927-29, developed a some ridicule. They had to invoke various improb
new theory of the mechanism of continental move able means such as island stepping-stones, isthmian
ment(Holmes, 1928). He proposed that continents links or rafting.
were moved by convection currents powered by the Very little was written about continental drift
heat of radioactive decay (Fig. 1.4). Although dif between the initial criticisms of Wegener's book and
fering considerably from the present concept of con about 1960. In the 1950s, employing methodology
vection, Holmes laid the foundation from which suggested by Lord Blackett, the palaeomagnetic
modern ideas developed. method was developed and S.K. Runcorn and his
Between the world wars two schools of thought co-workers demonstrated that relative movements
developed—the drifters and the non-drifters, the had occurred between North America and Europe.
latter far outnumbering the former. Each ridiculed The work was extended by K.M. Greet into South
the other's ideas. The non-drifters emphasized the America and by E. Irving into Australia. Palaeomag
lack of a plausible mechanism, as we have already netic results became more widely accepted when the
4 CHAPTER 1
Panqaea
Panthalassa Panthaiassa
pper Carboniferous
-ower Quaternary
» X X X X X X X X
\ X X X X X X X X/
V'X X X X J \ X X X X /
^ X X X' ^ X X X f
Fig. 1.4 The concept ofconvection as suggested by contributes to the main convective circulation. The
Holmes(1928), when it was believed that the oceanic eclogite melts at depth to form basaltic magma, which
crust was a thick continuation ofthe continental 'basaltic rises in ascending currents at A, heals the gaps in the
layer',(a) Currents ascending at A spread laterally, place a disrupted continent and forms new ocean floor. Local
continent under tension and split it, providing the swells, such as Iceland, would be formed from old sial left
obstruction ofthe old ocean floor can be overcome. This behind. Smaller current systems, initiated by the buoyancy
is accomplished by the formation of eclogite at B and C, of the basaltic magma,ascend beneath the continents and
where sub-continental currents meet sub-oceanic currents feed flood basalts or, beneath 'old'(Pacific) ocean floor,
and turn downwards. The high density of the eclogite feed the outpourings responsible for volcanic islands and
causes it to sink and make room for the continents to seamounts (redrawn from Holmes, 1928).
advance,(b) The foundering ofeclogite at B and C
hundred million years ofthe age ofthe Earth and the ing of the origin and history of the ocean basins to
solar system itself. Their studies also revealed the have emerged. It transpires that, in contrast to the
importance ofvertical movements of the continental continents, the oceanic areas are very young geo
crust in that the record was one ofrepeated uplift and logically (probably no greater than 200 Ma in age)
erosion, subsidence and sedimentation. But as J. and that horizontal, or lateral, movements have been
Tuzo Wilson, a Canadian geophysicist, has said, this all-important during their history offormation.
is like looking at the deck of a ship to see if it is In I96I, following intensive surveying of the sea
moving. floor during post-war years, R.S. Dietz proposed the
mechanism of'sea floor spreading' to explain contin
ental drift, a concept originally conceived by H.H.
1.2 Sea floor spreading and the Hess of Princeton University, whereby the contin
birth of plate tectonics ents moved in response to the growth ofocean basins
Ifthere is a possibility that the continental areas have between them. It was suggested that oceanic crust is
been rifted and drifted apart and together, then created from the Earth's mantle at the crest of the
presumably there should be some record of this mid-ocean ridge system, a volcanic submarine swell
within the ocean basins. However, it is only since or rise which occupies a median position in many
World War II and notably since I960 that sufficient of the world's oceans (Fig. 1.5). Oceanic crust is
data have been obtained from the 60% ofthe Earth's much thinner than continental crust, having a mean
surface covered by deep water for some understand thickness of about 7km, compared with the average
CHAPTER 1
Continent
Mid-ocean
Trench
ridge
Mantle Warmer Nj
Cooler
continental thickness of about 35 km; is chemically by J.T. Wilson in 1965, of a new class of faults
different; and is structurally far less complex. The termed transform faults, which connect linear belts
lateral motion ofthe oceanic crust was believed to be of tectonic activity (see Chapter 7). The Earth was
driven by convection currents in the upper mantle in then viewed as a mosaic of six major and several
the fashion of a conveyer belt. In order to keep the smaller plates in relative motion. The theory was
surface area of the Earth constant, it was further put on a stringent geometrical basis by the work of
proposed that the oceanic crust is thrust back down D.P. McKenzie,R.L. Parker and W.J. Morgan in the
into the mantle and resorbed at oceanic trenches. period 1967-68 (see Chapter 5), and confirmed by
These are vast bathymetric depressions, situated at earthquake seismology through the work of B.
certain ocean margins and associated with intense Isacks, J. Oliver and L.R. Sykes (see Chapter 8).
volcanic and earthquake activity. Within this frame The theory has been considerably amplified by
work the continents are quite passive elements—rafts intensive studies of the geological and geophysical
of less dense material which are drifted apart and processes affecting plate margins. Probably the aspect
together by ephemeral ocean floors. The continents about which there is currently the most contention is
themselves are a scum of generally much older the nature of the mechanism which causes plate
material that was derived or separated from the motions (see Chapter II).
Earth's interior either at a very early stage in its Although the basic theory of plate tectonics is well
history or, at least in part, steadily throughout geo established, by no means is understanding complete,
logical time. Instead of blocks ofcrust, we now think and its implications will fully occupy Earth scientists
in terms of 'plates' of comparatively rigid upper for many decades to come.
mantle and crust, perhaps 50-100 km thick and
which we term lithosphere (a term originally coined
by R.A. Daly many years ago and meaning 'rock 1.3 Geosynclinal theory
layer'). Lithospheric plates can have both continental Prior to the acceptance of plate tectonics, the static
and oceanic crust embedded in them. model of the Earth encompassed the formation of
The theory ofsea floor spreading was confirmed in tectonically active belts, which formed essentially by
the period 1963-66 following the suggestion of vertical movements, on the site of geosynclines. A
F.J. Vine and D.H. Matthews that the magnetic review of the development of the geosyncline
lineations of the sea floor might be explained in hypothesis and its explanation in terms of plate
terms of sea floor spreading and palaeomagnetism tectonics is provided by Mitchell & Reading(1986).
(see Chapter 4). On this model the conveyor belt Geosynclinal theory envisaged elongate, geo
of oceanic crust is viewed as a tape recorder which graphically fixed belts of deep subsidence and thick
registers the reversal history of the Earth's magnetic sediments as the precursors of mountain ranges in
field. which the strata were exposed by folding and uplift of
A further precursor to the development of the the geosynclinal sediments (Dickinson, 1971). A
theory of plate tectonics came with the recognition. plethora ofspecific nomenclature evolved to describe
HISTORICAL PERSPECTIVE
the lithological associations of the sedimentary fill the distribution in space and time of the conditions
and the relative locations ofthe geosynclines. of different metamorphic facies, the scheme of
The greatest failing ofgeosynclinal theory was that deformation in mountain belts and the association of
tectonic features were classified without there being different types ofeconomic deposit.
an understanding of their origin. Geosynclinal no The fact that plate tectonics is so successful in
menclature consequently represented an impedi unifying so many aspects of geology should not be
ment to the recognition of a common causal taken to indicate that it is completely understood.
mechanism. The relation of sedimentation to the Indeed,it is the critical testing ofthe implications of
mobilistic mechanism of plate tectonics (Mitchell & plate tectonic theory which has led to modifications
Reading, 1969) allowed the recognition of two and extrapolations,for example in the consideration
specific environments in which geosynclines formed, of the relevance of plate tectonic processes in con
namely passive continental margins and subduction tinental areas and the more distant geological past. It
zones. Although some workers retain geosynclinal is to be hoped that plate tectonic theory will be
terminology to describe sedimentary associations employed cautiously and critically.
(e.g. the terms eugeosyncline and miogeosyncline
for sediments with and without volcanic members,
respectively), this usage is not recommended, and Further reading
the term geosyncline must be recognized as no
Frankel, H.(1988) From continental drift to plate tecton
longer relevant to plate tectonic processes. ics. Nature 335,127-30.
Hallam, A. (1973a) A Revolution in the Earth Sciences.
Clarendon Press, Oxford.
1.4 Impact of plate tectonics
Hallam, A.(1975) Alfred Wegener and the hypothesis of
Plate tectonics is of very great significance as it continental drift. Sci. Am. 232(2), 88-97.
represents the first theory which provides a unified Mitchell, A.H.G. & Reading, H.G.(1986) Sedimentation
explanation of the Earth's major surface features. As and tectonics. In: Reading, H.G. (ed.) Sedimentary
such it has enabled an unprecedented linking of Environments and Facies., pp.471-519. Blackwell
Scientific Publications, Oxford.
many different aspects of geology, which had previ
Rupke,N.A.(1970)Continental drift before 1900. Nature
ously been considered independent and unrelated. A
227, 349-50.
deeper understanding of geology has ensued from Stewart, J.A. (1990) Drifting Continents and Colliding
the intepretation of many branches ofgeology within Raradi£fms: Perspectives on the Geoscience Revolution. Indi
the basic framework provided by plate tectonics. ana University Press, Bloomington.
Thus,for example, explanations can be provided for Vine,F.J.(1977)The continental drift debate. Nature 266,
the past distributions of flora and fauna, the spatial 19-22.
relationships of volcanic rock suites at plate margins.
2/ Internal structure of the Earth
epicentre. In this case the focal depth is determined reduces the strain in the system virtually to zero, but
by triangulation in the vertical plane, using the P-S ifthe shearing forces persist,strain would again build
time difference to calculate the distance to the focus. up to the point at which fault movement occurs. The
elastic rebound theory consequently implies that
earthquake activity represents a stepwise response to
2.1.5 Mechanism of earthquakes
persistent strain.
Most earthquakes are believed to occur according to
the elastic rebound theory^ which was developed after
2.1.6 Focal mechanism solutions of
the San Francisco earthquake of 1906. In this theory
earthquakes
an earthquake represents a sudden release of strain
energy which has built up over a period oftime. The seismic waves generated by earthquakes, when
In Fig. 2.3a a block of rock traversed by a pre recorded at seismograph stations around the world,
existing fracture (or fault) is being strained in such a can be used to determine the nature of the faulting
way as eventually to cause relative motion along the associated with the earthquake, to infer the orienta
plane ofthe fault. The line AB is a marker indicating tion ofthe fault plane and to gain information on the
the state of strain ofthe system, and the broken line state of stress of the lithosphere. The results ofsuch
the location ofthe fault. Relatively small amounts of an analysis are referred to as afocal mechanism solution
strain can be accommodated by the rock (Fig. 2.3b). or fault plane solution. The technique represents a
Eventually, however, the strain reaches the level very powerful method of analysing movements of
at which it exceeds the frictional and cementing the lithosphere, in particular those associated with
forces opposing movement along the fault plane plate tectonics. Information is available on a global
(Fig. 2.3c). At this point fault movement occurs scale as most earthquakes with a magnitude in excess
instantaneously(Fig. 2.3d). The 1906 San Francisco of5.5 can provide solutions,and it is not necessary to
earthquake resulted from a displacement of 6.8 m have recorders in the immediate vicinity ofthe earth
along the San Andreas Fault. In this model,faulting quake, so that data are provided from regions which
may be inaccessible for direct study.
(a) (b)
Accordi^ig to the elastic rebound theory, the strain
energy released by an earthquake is transmitted by
the seismic waves which radiate from the focus.
tr Consider the fault plane shown in Fig. 2.4 and the
plane orthogonal to it, the auxiliary plane. The first
Tt-
0 i-Compression®? Dilation
(0 (d)
Auxiliary
Focus
Dilation |Compression|
Fault plane
Fault S
Fig.2.3 Elastic rebound mechanism of earthquake Fig.2.4 Quadrantal distribution of compressional and
generation. dilational P wave first motions about an earthquake.
INTERNAL STRUCTURE OF THE EARTH 11
seismic waves to arrive at recorders around the sional and dilational first motions ifthe waves gener
earthquake are P waves, which cause compression/ ated by the earthquake followed straight line paths.
dilation of the rocks through which they travel. The The actual travel paths, however, are curved and the
shaded quadrants defined by the fault and auxiliary surface intersection of the dashed line, correspond
planes are compressed by movement along the fault, ing to the path that would have been followed by a
and so the first motion ofthe P wave arriving in these wave leaving the focus in the direction of the fault
quadrants corresponds to a compression. This is plane, represents the actual nodal plane.
recognized by an initial upward displacement at the It is clear, then, that simple mapping of compres
Earth's surface and recorded on the vertical seismo sional and dilational first motions on the Earth's
graph. Conversely, the unshaded quadrants are surface cannot readily provide the focal mechanism
stretched or dilated by the fault movement. The first solution. However, the complications can be over
motion of the P waves in these quadrants is thus come by considering the directions in which the
dilational, and recognized by an initial downward seismic waves left the focal region, as it is apparent
displacement at the seismographs. The region that compressions and dilations are restricted to
around the earthquake is therefore divided into four certain angular ranges.
quadrants on the basis of the P wave first motions, A focal mechanism solution is obtained firstly by
defined by the fault plane and the auxiliary plane. No determining the location ofthe focus by the method
P waves propagate along these planes as movement outlined in Section 2.1.4. Then, for each seismo
of the fault imparts only shearing motions in their graph recording the earthquake,a standard model for
directions; they are consequently known as nodal the velocity structure ofthe Earth is used to compute
planes. the travel path of the seismic wave from focus to
Simplistically, then, a focal mechanism solution seismograph, and hence to calculate the direction in
could be obtained by recording an earthquake at a which the wave left the focal region. These directions
number of seismographs distributed around its epi are then plotted, using an appropriate symbol for
centre, determining the nature ofthe first motions of compressional or dilational first motion,on an equal
the P waves, and then selecting the two orthogonal- area projection of the lower half of the focal sphere.,
planes which best divide compressional from dila that is, an imaginary sphere of small but arbitrary
tional first arrivals, that is, the nodal planes. In radius centred on the focus (Fig. 2.5). An equal area
practice, however, the technique is complicated by net, which facilitates such a plot, is illustrated in
the spheroidal shape of the Earth and the progressive Fig. 2.6. The scale around the circumference ofsuch
increase ofseismic velocity with depth, which causes a net refers to the azimuth,or horizontal component
the seismic waves to follow curved travel paths be ofdirection, while dips are plotted on the radial scale
tween the focus and recorders. Consider Fig. 2.5. from 0° at the perimeter to 90° at the centre. Planes
The dotted line represents the continuation of the through the focus are represented on such plots by
fault plane, and its intersection with the Earth's great circles with a curvature appropriate to their dip;
surface would represent the line separating compres- hence a diameter represents a vertical plane.
Let us assume that,for a particular earthquake,the
fault motion is strike-slip along a near vertical fault
Compression plane. This plane and the auxiliary plane plot as
Nodal plane orthogonal great circles on the projection ofthe focal
sphere, as shown on Fig. 2.7. The lineation defined
by the intersection of these planes is almost vertical,
so it is apparent that the direction of movement
along the fault is orthogonal to this intersection,that
Fault planp^:Focal sphei is, near horizontal. The two shaded and two un
shaded regions ofthe projection defined by the nodal
planes now correspond to the directions in which
Fig. 2.5 Distribution of compressional and dilational first compressional and dilational first motions, respec
arrivals from an earthquake on the surface of a spherical tively, left the focal region. A focal mechanism solu
Earth in which seismic velocity increases with depth. tion is thus obtained by plotting all the observational
12 CHAPTER 2
270°
ence either compression or dilation as a result ofthe fault are compressed or dilated, so that an identical
fault movement. The directions in which compres focal sphere projection is obtained. Similar results are
sional first motions and C2 and dilational first obtained when the faulting is normal (Fig. 2.9). In
motions and D2leave the focus are shown and C2 theory the fault plane can be distinguished by making
and D2 are plotted on the projection of the focal use of Anderson's simple theory offaulting (Section
sphere in Fig. 2.8b, on which the two nodal planes 2.10.2) which predicts that normal faults have dips
are also shown. Because Fig. 2.8a is a vertical section, of more than 45® and thrusts less than 45®. Thus,fj
the first motions indicated plot along an east-west is the fault plane in Fig. 2.8 and f2 the fault plane in
azimuth. Arrivals at seismographs at other azimuths Fig. 2.9.
would occupy other locations within the projection It is apparent that the diflferent types of faulting
space. Consider now Fig. 2.8c, in which plane api can be identified in a, focal mechanism solution by
becomes the fault plane f2 and f^ the auxiliary plane the distinctive pattern of compressional and dila
ap2. By considering the movement along the thrust tional regions on the resulting focal sphere. Indeed,
plane it is obvious that the same regions around the it is also possible to differentiate earthquakes which
14 CHAPTER 2
(a)
have originated by a combination offault types,such Fig. 2.10a. However,for this simple model, which is
as dip-slip accompanied by some strike-slip move known as a type I, or single-couple, source,S waves,
ment. The precision with which the directions ofthe whose corresponding ground motion is shearing,
nodal planes can be determined is dependent upon should be restricted to the region of the auxiliary
the number and distribution ofseismographs record plane(Fig. 2.10b),Recording ofthe S wave radiation
ing arrivals from the event. It is not possible, how pattern should then make it possible to determine
ever, to distinguish the fault and auxiliary planes. the actual fault plane. It was found, however, that
At one time it was believed that distinction be instead of this simple pattern, most earthquakes
tween the nodal planes could be made on the basis of produce S wave radiation along the direction ofboth
the pattern ofS wave arrivals. P waves radiate into all nodal planes (Fig. 2.10c). This observation initially
four quadrants of the source region as shown in cast into doubt the validity of the elastic rebound
INTERNAL STRUCTURE OF THE EARTH 15
c / : \
V3
200
X, Distance(km)
methods. As well as the phases Pg and P^ he noted depth of water above it. The internal layering of
the presence of an additional phase P* (Fig. 2.15) oceanic crust and its constancy over very wide areas
which he interpreted as the critically refracted arrival will be discussed later (Section 2.4.4).
from an interface where the velocity increased from In studying the deeper layering of the Earth, seis
about 5.6 to 6.3 km s"^. This interface was subse mic waves with much longer travel paths are em
quently named the Conrad discontinuity. Conrad's ployed. The velocity structure has been built up by
model was readily adopted by early petrologists who recording the travel times ofbody waves over the full
believed that two layers were necessarily present in range of possible epicentral angles. By assuming that
the continental crust. The upper layer, rich in silicon the Earth is radially symmetrical, it is possible to
and aluminium, was called the SIAL and was be invert the travel time data to provide a model of the
lieved to be the source ofgranitic magmas, while the velocity structure. A modern determination of the
lower, silicon- and magnesium-rich layer or SIMA velocity-depth curve (Hart et 1977) for both P
was believed to be the source of basaltic magmas.It is and S waves is shown in Fig. 2.16.
now known, however, that the upper crust has a Velocities increase abruptly at the Moho in both
composition more basic than granite, and that the continental and oceanic environments. A low velo
majority of basaltic magmas originate in the upper city zone (LVZ) is present between about 80 and
mantle. Consequently,the petrological necessity ofa 300 km depth. This appears to be universally present
two-layered crust no longer exists and, where applic for S waves, but may be absent in certain regions for
able, it is preferable to use the terms upper and lower P waves, especially beneath ancient shield areas.
crustal layer. Unlike the Moho,the Conrad disconti Between 400 and about 670 km velocity increases
nuity is not always present within the continental rapidly in a stepwise fashion within the mantle trans
crust, although the seismic velocity generally in ition zone, which separates the upper mantle from
creases with depth. There are some regions, however, the lower mantle. Each velocity increment probably
where it is believed that a low velocity layer may be corresponds to a mineral phase change to a denser
present at the base of the upper crust (Mueller & form at depth (Section 2.8.5). Both P and S velocities
Landisman, 1966). increase progressively in the lower mantle.
The oceanic crust has principally been studied by The Gutenberg discontinuity marks the core-
explosion seismology. The Moho is always present mantle boundary at a depth of2891 km,at which the
and the thickness of much of the oceanic crust is velocity of P waves decreases abruptly. S waves are
remarkably constant at about 7km irrespective ofthe not transmitted through the outer core, which is
INTERNAL STRUCTURE OF THE EARTH 19
Crust Velocity(km s
Upper rriaritle- pllll"J" 1.1 ithngnhprp
iiiiif'-vz
Transition zone I -Astheno-
sphere
j: MesoH;
Sspherei;
Depth (km)
consequently believed to be in a fluid state. The Seismic data, combined with knowledge of the
geomagnetic field (Section 3.6.4) is believed to mass and moment of inertia of the Earth, have
originate by the circulation of a good electrical con revealed that the mean atomic weight ofthe Earth is
ductor in this region. At a depth of 5150 km the P about 27, with a contribution of 22.4 from the
velocity increases abruptly and S waves are once again mantle and crust and 47.0 from the core. No single
transmitted. This inner core is thus believed to be type of meteorite possesses an atomic weight of 27,
solid as a result of the enormous confining pressure. the various types ofchondrite being somewhat lower
There appears to be no transition zone between inner and iron meteorites considerably higher. However,it
and outer core, as was originally believed. is possible to mix the proportions of different met
eorite compositions in such a way as to give both the
correct atomic weight and core/mantle ratio. Three
2.3 Composition of the Earth
such models are given in Table 2.1.
All bodies in the solar system are believed to have It is apparent that at least 90% ofthe Earth is made
been formed by the condensation and accretion of up ofiron, silicon, magnesium and oxygen, with the
the primitive interstellar material which made up the bulk of the remainder comprising calcium, alumin
solar nebula. The composition ofthe Sun is the same ium, nickel, sodium and possibly sulphur.
as the average composition of this material. Gravita
tional energy was released during accretion, and
together with the radioactive decay of short-lived 2.4 The crust
radioactive nuclides eventually led to heating of the
2.4.1 The continental crust
proto-Earth so that it differentiated into a radially
symmetrical body made up of a series ofshells whose Only the uppermost part of the crust is available for
density increased towards its centre. The differentia direct sampling at the surface or from boreholes. At
tion prevents any estimate being made ofthe overall greater depths within the crust virtually all informa
composition of the Earth by direct sampling. How tion about its composition is indirect. Much of our
ever, it is believed that meteorites are representatives information about the crust has been derived from a
of material within the solar nebula and that esti knowledge ofthe variation of seismic velocities with
mates of the Earth's composition can be made from depth and how these correspond to experimental
them. The presence of metallic and silicate phases in determinations of velocities measured over ranges of
meteorites is taken to indicate that the Earth consists temperature and pressure consistent with crustal
ofan iron/nickel core surrounded by a lower density conditions. Pressure increases with depth at a rate of
silicate mantle and crust. about 30MPa km"\ mainly due to the lithostatic
20 CHAPTER 2
the correct velocity range. Dewey & Burke (1973) Table 2.2 Oceanic crustal structure (after Bott, 1982a
have used this idea to explain the development of with permission from Edward Arnold).
anorthositic lower crust in terms of a continent-
continent collision (Section 9.3.2) wherein the crust P velocity Average thickness
is thickened, heat flow increases and granitic melts (km s"^) (km)
migrate upwards, leaving a refractory lower crust. Water 1.5 4.5
The Conrad discontinuity is envisaged as the top of Layer 1 1.6-2.5 0.4
this layer, and it is interesting to speculate that, Layer 2 3.4-6.2 1.4
according to this model, the Conrad discontinuity Layer 3 6.4-7.0 5.0
may only be developed in terrains which have suf
fered orogenesis of this type. Moho
4 6 8 I—--—izTd Sediment
I 1 1 1
(L2A) 3. Fractured basalt
Massive basalt
(L2) 5.0 (L2B) 5.2 with dykes
Dykes with
(L2C) 6.1 massive basalt
Depth
Metagabbros and gabbros
below
(L3A) 6.8 with pockets of plagiogranite
basement
and protrusions of serpentinite
(km) 4 (L3) 6.7
n 11 Mantle
8.1 Serpentinized ultramafics
Harzburgite and Iherzolite
Anomalous mantle
(Close to ridge axis)
Fig.2.17 P and S wave velocity structure ofthe oceanic 2.4.6 Oceanic layer 2
crust and its interpretation in terms oflayered models
proposed in 1965 and 1978. Numbers refer to velocities Layer 2 is variable in its thickness, in the range
in km s"^. Dashed curve refers to gradational increase in 1.0-2.5 km. Its seismic velocity is similarly variable
velocity with depth deduced from more sophisticated in the range 3.4-6.2 km s~^. This range is attribut
inversion techniques (after Spudich & Orcutt, 1980 and able to either consolidated sediments or extrusive
Harrison & Bonatti, 1981). igneous material. Direct sampling and dredging of
the sediment-free crests of ocean ridges, and the
of the volcanic nature of layer 2. Within layer 1 are necessity of a highly magnetic lithology at this level
a number of horizons which show up as prominent (Section 4.2), overwhelmingly prove an igneous
reflectors on seismic reflection records. Edgar(1974) origin. The basalts recovered are olivine tholeiites
has described the acoustic stratigraphy in the North containing calcic plagioclase, and are poor in potas
Atlantic, where up to four suprabasement reflectors sium,sodium and the incompatible elements(Sun et
are found (Fig. 2.18). Horizon A corresponds to an aL^ 1979). They exhibit very little areal variation in
Eocene chert, although deep sea drilling indicates composition,with the exception oflocations close to
that it maintains its reflective character even when oceanic islands (Section 5.4).
little or no chert is present. In such locations it may Three subdivisions of layer 2 have been recog
correspond to an early Tertiary hiatus beneath the nized. Sublayer 2A is only present on ocean ridges
chert. Horizon A* occurs beneath A,and represents near eruptive centres in areas affected by hydro-
the interface between Late Cretaceous/Early Ter thermal circulation ofsea water, and ranges in thick
tiary metal-rich clays and underlying euxinic black ness from zero to 1 km.Its porous, rubbly nature, as
clays. Horizon P represents the base of the black indicated by a P wave velocity of 3.6 km s"^, permits
clays, where they overlie a Late Jurassic/Lower Cre such circulation. The very low velocities(2.1 km s"^)
taceous limestone. Horizon B may represent a sedi of the top of very young layer 2 located on the
mentary horizon,although it has also been identified Mid-Atlantic Ridge (Purdy, 1987) probably indicate
as basalt similar to that at the top oflayer 2. a porosity of 30-50%, and the much higher veloci
Reflectors similar to A and B have been identified ties of older layer 2 imply that the porosity must be
in the Pacific and Caribbean, where they are termed reduced quite rapidly after its formation.Sublayer 2B
A', B' and A",B", respectively. forms the normal acoustic basement oflayer 1 when
INTERNAL STRUCTURE OF THE EARTH 23
(a)
North America Atlantic Ocean
Acoustic Basement
Fig.2.18 (a) Major seismic reflectors in the western wave velocities. This notion can now be discounted,
Atlantic Ocean,(b) Corresponding lithoiogies determined however, as the value ofPoisson's ratio for layer 3A,
by deep sea drilling (after Edgar, 1974, with permission which can be estimated directly from a knowledge of
from Springer-Verlag). both P and S wave velocities, is much lower than
would be expected for serpentinized peridotite. In
sublayer 2A is not developed. Its higher velocity of fact, Poisson's ratio for layer 3A is more in accord
4.8-5.5 km s~^ suggests a louver porosity. With time with a gabbroic composition, which also provides
layer 2A may be converted to layer 2B by the infilling seismic velocities in the observed range, and gabbros
of pores by secondary minerals such as calcite, quartz have been encountered in this level in one DSDP
and zeolites. borehole. It is possible, however, that all or at least
Sublayer 2C is about 1 km thick, where detected, part of layer 3B, where recognized, consists of ser
and its velocity range of5.8-6.2 km s~^ may indicate pentinized ultramafic material.
a high proportion ofintrusive, basic rocks. This layer The concept ofa predominantly gabbroic layer 3 is
grades downwards into layer 3. in accord with models suggested for the origin of
oceanic lithosphere (Section 6.10). These propose
that layer 3 forms by the crystallization of a magma
2.4.7 Oceanic layer 3
chamber, with an upper layer, possibly correspond
Layer 3 is the main component of the oceanic crust ing to sublayer 3A, of isotropic gabbro and a lower
and represents its plutonic foundation (Fox & layer, possibly corresponding to 3B, consisting of
Stroup, 1981). Some workers have subdivided it cumulate gabbro and ultrabasic rocks formed by
into sublayer 3A, with a velocity range of 6.5- crystal settling. This layering has been confirmed by
6.8 km s"^, and a higher velocity lower sublayer 3B direct observation and sampling by submersible on
(7.0-7.7 km s"^) (Christensen & Salisbury, 1972), the Vema Fracture Zone in the North Atlantic
although the majority of seismic data can be ex (Auzende et al., 1989).
plained in terms of a layer with a slight positive Recent in situ evidence from an ODP hole on the
velocity gradient (Spudich & Orcutt, 1980). Costa Rica rift in the eastern Pacific has shown, at
Hess (1962) suggested that layer 3 was formed least for this location, that the layer 2/3 seismic
from upper mantle material whose olivine had boundary lies within the basalt and is associated with
reacted with water to varying degrees to produce gradual changes in porosity and alteration (Detrick^^
serpentinized peridotite, and, indeed, 20-60% ser- uL 1994).
pentinization can explain the observed range of P The origin of the relatively small amounts of
24 CHAPTER 2
oceanic crust which do not conform to the standard spreading centres, the presence of similar ore miner
layering is discussed in Section 7.3. als and the observation that the sediments were
formed in deep water (Moores, 1982). Salisbury &
Christensen (1976) have compared the velocity
2.5 Ophiolites structure of the oceanic lithosphere with seismic
The study of oceanic lithosphere has been aided by velocities measured in samples from the Bay of
investigations of characteristic rock sequences on Islands ophiolite complex in Newfoundland, and
land known as ophiolites (literally'snake rock', refer concluded that the determined velocity stratigra
ring to the similarity of the colour and texture to phies are identical. Figure 2.19 shows the correlation
snakeskin; see Nicolas, 1989, for a full treatment of between the oceanic lithosphere and three well-
this topic). Ophiolites usually occur in collisional studied ophiolite bodies.
mountain belts, and their association of deep sea At one time it seemed that investigations of the
sediments, basalts, gabbros and ultrabasic rocks sug petrology and structure of the oceanic lithosphere
gests that they originated as oceanic lithosphere and could conveniently be accomplished by the study of
were subsequently thrust up into their continental ophiolite sequences on land. However, in recent
setting by a process known as obduction (Dewey, years this simple analogy has been challenged, and it
1976; Ben-Avraham et 1982; see also Section has been suggested that ophiolites do not represent
9.5). The complete ophiolite sequence (Gass, 1980) typical oceanic lithosphere, and were not emplaced
is shown in Table 2.3. The analogy ofophiolites with exclusively during continental collision (Mason,
oceanic lithosphere is supported by the gross simil 1985).
arity in chemistry (although there is considerable Dating ofevents indicates that obduction of many
difference in detail), metamorphic grades corre ophiolites occurred very soon after their creation.
sponding to temperature gradients existing under Continental collision, however, normally occurs a
long time after the formation of a mid-ocean ridge.
Sediments Layer 1
4-
rLayer h
Basic volcanics, commonly| Layer
___
6-
pillowed, merging into ^ Layer 2
Basic sheeted dyke complex J 8-
Layer 3
High level intrusives 10-
S ..
Trondhjemites 1 Layer 3
Gabbros / 12-
Upper
mantle
14-
Layered cumulates
Olivine gabbros| — Moho — I Pillow Sheeted
Pyroxenitcs > 1 Sediment
' lava dykes
Peridotites J I Ultrabasic
Upper mantle
B Gabbro I ,. \yi Peridotite
I cumulates '—'
S-seismological Moho P- petrological Moho
Harzburgite, commonly
serpentinized ±Iherzolite, Fig.2.19 Comparison ofoceanic crustal structure with
dunite, chromititc ophiolite complexes (after Mason, 1985,with permission
from Blackwell Scientific Publications).
INTERNAL STRUCTURE OF THE EARTH 25
so that the age of the sea floor obducted should be on ocean ridges, and the observed metamorphism
considerably greater than that of the collisional oro within ophiolite sequences.
geny. Ophiolites consequently represent lithosphere Hydrothermal circulation takes place by convec-
that was obducted while young and hot. Geochem- tive flow, probably through the whole of the oceanic
ical evidence (Pearce, 1980; Elthon, 1991) has sug- crust (Fyfe 8c Lonsdale, 1981), and is of great signi
geted that the original sites of ophiolites were ficance. It influences models of heat production, as
marginal seas (Section 8.9; see also Cawood 8c Suhr, it has been estimated that approximately 25% of the
1992), or Red Sea-type oceans. heat escaping from the Earth's surface is vented at
There have been many different mechanisms pro the mid-ocean ridges. The circulation must modify
posed for ophiolite obduction, none of which can the chemistry of the ocean crust, and consequently
satisfactorily explain all cases. It must thus be recog will affect the chemical relationship of lithosphere
nized that there may be several operative mechanisms and asthenosphere over geological time because of
and that, although certainly formed by some type of the recycling oflithosphere that occurs at subduction
accretionary process, ophiolite sequences may differ zones. It is also responsible for the formation of
significantly from lithosphere created in the major certain economically important ore deposits, partic
ocean basins. ularly massive sulphides.
These hydrothermal processes are most conven
iently studied in the metamorphic assemblages of
2.6 Metamorphism of oceanic
ophiolite complexes, and the model described below
crust and upper mantle
has been derived by Elthon (1981).
Many of the rocks sampled from the ocean basins Hydrothermal metamorphism of pillow lavas and
show evidence of metamorphism, including abun other extrusives gives rise to low temperature
dant greenschist facies assemblages and alkali meta (<230''C) and greenschist facies assemblages
somatism: In close proximity to such rocks, however, (Fig. 2.20). The distribution of alteration is highly
are found completely unaltered species. irregular, and is controlled by the localized fissuring
It is probable that this metamorphism is accom ofthe extrusive rocks. Higher temperature metamor
plished by the hydrothermal circulation of sea water phism is widespread within the sheeted dyke com
within the oceanic crust. There is much evidence for plex, producing assemblages typical ofthe actinolite
the existence ofsuch circulation, such as the presence facies, although pockets of unaltered rock do occur.
of metalliferous deposits which probably formed by The highest metamorphic temperatures are achieved
the leaching and concentration of minerals by sea at the base of the sheeted dyke complex and the
water, observations of active hydrothermal vents upper part ofthe gabbroic section. Rarely, retrograde
Sediments Facies
Pi low lavas
Greenschist
retrograded to
Zeolite
Greenschist
Sheeted dikes
Actinolite
retrograded to
Fig. 2.20 Schematic model for Greenschist
rocks of the greenschist facies occur at this level. 3 A£fe. Continental crust may be as old as 3960 Ma,
Alteration decreases to only about 10% within the the age ofthe oldest rocks yet discovered. On a very
top kilometre of the gabbroic section and thereafter broad scale the oldest crust consists of Precambrian
metamorphism is restricted to the locality of fissures cratons or shield areas which are surrounded by
and dykes, although metamorphism does not com younger orogenic belts, both active and inactive.
pletely terminate at depth. According to this model, Oceanic crust, however, is nowhere older than
sea water circulation occurs extensively in the upper 180 Ma, and progressively increases in age outwards
3 km of the crust, producing the metamorphic as from oceanic ridges (Section 4.3). Oceans are conse
semblages and cooling the crust. High temperature quently viewed as essentially transient features of
metamorphism only occurs near the spreading the Earth's surface. About 50% ofthe surface area of
centre. At depth the circulation becomes diminished the oceans has been created during the last 65 Ma,
as secondary minerals are deposited within the flow implying that 30% of the Earth's surface has been
channels. created during the most recent 1.5% of geological
As the ridge spreads continuously, oceanic litho- time.
sphere is moved laterally from the heat source and 4 Tectonic activity. Continental crust may be exten
undergoes retrograde metamorphism. This depends sively folded and faulted and contains evidence of
upon an adequate water supply, as water distribution being subjected to multiple tectonic events. Oceanic
is the major control of metamorphic grade. The crust, however, appears to be very stable and has
absence ofsufficient water allows the preservation of suffered little deformation except at plate margins.
relict high temperature assemblages. The hetero 5 Igneous activity. Very little igneous activity occurs
geneous nature of the distribution of metamorphic on the great majority of the continental crust. The
facies is consequently explained by a similarly hetero only major locations ofactivity are mountain belts of
geneous distribution ofcirculating fluids rather than Andean type (Section 9.2). The activity within the
extreme temperature variations. oceans is very much greater. Ocean ridges and island
arcs are the location of the Earth's most active areas
of volcanic and plutonic activity.
2.7 Differences between
continental and oceanic crust
2.8 The mantle
On the basis of information presented in this and
following chapters, the major differences between
2.8.1 Introduction
continental and oceanic crust can be summarized as
follows: The mantle constitutes the largest internal subdiv
1 Layering. The large-scale layering ofthe continen ision of the Earth by both mass and volume, and
tal crust is ill defined and highly variable, reflecting a extends from the Moho at a mean depth of about
complex geological history. In places there is a broad 21 km to the core-mantle boundary at a depth of
subdivision by the Conrad discontinuity, but this is 2885 km. It is generally believed to be chemically
not globally developed. By contrast, the layering of homogeneous, apart from the abundances of minor
the oceanic crust is well defined into three distinct and trace elements, and formed of silicate minerals.
layers which are found in all the ocean basins. How The mineralogy and structure of the silicates change
ever, the nature of these layers, in particular layers 2 with depth and give rise to a transition zone between
and 3, may change quite markedly with depth. In the 400 and 670 km depth, which separates the upper
case of layers 2 and 3 this reflects the complex and lower mantle.
development oftheir parental magma chamber. Mantle materials are only rarely brought to the
2 Thickness, The thickness ofcontinental crust aver surface, in ophiolite complexes (Section 2.5), in
ages about 35-40 km but is quite variable, thinning kimberlite pipes and as xenoliths in alkali basalts.
to only a few kilometres beneath rifts and thickening Consequently, most of our information about the
to up to 80 km beneath young mountain belts. The mantle is indirect and based on the variation of
oceanic crust maintains a relatively constant thick seismic velocities with depth combined with studies
ness of about 7km, although layer 1 increases in of mineral behaviour at high temperatures and pres
thickness towards the ocean margins. sure, and in shock-wave experiments. Geochemical
INTERNAL STRUCTURE OF THE EARTH 27
studies of meteorites and uitramafic rocks are also indicated by estimates ofPoisson's ratio from P and S
utilized in making predictions about the mantle. velocities, and the presence of peridotites in the basal
sections of ophiolite sequences and as nodules in
alkali basalts. The density of eclogites is also rather
2.8.2 Seismic structure of the mantle
too high to explain the Moho topography of iso-
The uppermost part of the mantle constitutes a statically compensated crustal structures.
high velocity lid abut 80 km thick in which seismic The bulk composition of the mantle can be esti
velocities remain constant at a figure in excess of mated in several ways: by using the compositions of
7.9 km s~^ or increase slightly with depth. This part various uitramafic rock types, from geochemical
of the mantle makes up the lower portion of the computations, from various meteorite mixtures and
lithosphere (Section 2.12), and its behaviour is inti by using data from experimental studies. It is neces
mately related to that ofits overlying crust. That the sary to distinguish between undepleted mantle and
crust and lithospheric mantle are not decoupled is depleted mantle which has undergone partial melt
indicated by the strong correlation of upper mantle ing so that many ofthe elements which do not easily
velocity with overlying crustal type (Condie, 1982). substitute within mantle minerals have been
Beneath the lithosphere lies a low velocity zone removed and combined into the crust. Ringwood
extending from a depth of approximately 100 to (1975) has suggested a number of possible combin
300 km (Fig. 2.16). This appears to be present ations of compositions, the simplest of which con
beneath most regions ofthe Earth with the exception sists of three parts dunite (a monomineralic rock
of the mantle beneath cratonic areas. From the base containing olivine), which represents fully depleted
ofthis zone seismic velocities increase slowly until a mantle, and one part basalt, corresponding to the
major discontinuity is reached at a depth of about partial melt. This hypothetical rock type has been
400 km, marking the upper region of the transition termed pyrolite, and further work has led to pyrolite
zone. There is a further velocity discontinuity at a compositions containing more complex combina
depth of 670 km, the base of the transition zone. tions ofingredients.
Within the lower mantle velocities increase slowly Pyrolite can exist in four different forms depend
with depth until the basal 200 km where gradients ing upon the ambient temperature, pressure and
decrease or low velocities are present. water vapour pressure. The normal form, existing to
a depth of 70 km is pyroxene pyrolite, which trans
forms into garnet pyrolite at depths of 80-350 km.
2.8.3 Mantle composition
The other forms can only exist in environments very
The fact that much ofthe oceanic crust is made up of rarely encountered within the Earth.
material of a basaltic composition derived from the Although estimates of mantle composition vary in
upper mantle suggests that the upper mantle is com detail, it is generally agreed that at least 90% of the
posed of either peridotite or eclogite (Harrison & mantle by mass can be represented in terms of the
Bonatti, 1981). The main difference between these oxides FeO, MgO and SiOj. A further 5-10% is
two rock types is that peridotite contains abundant made up ofCaO,Ai203 and Na20,and at least 98%
olivine and less than 15% garnet, whereas eclogite ofthe mantle comprises these six oxides.
contains little or no olivine and at least 30% garnet.
Both possess a seismic velocity which corresponds to
2.8.4 The mantle low velocity zone
the observed upper mantle value of about 8 km s"^.
Several lines ofevidence now suggest very strongly The low velocity zone (Fig. 2.16) is characterized by
that the upper mantle is peridotitic. Beneath the low seismic velocities, high seismic attenuation and
ocean basins the velocity is frequently anisotropic, also, probably, a high electrical conductivity. The
with velocities over 15% higher perpendicular to seismic effects are more pronounced for S waves than
ocean ridges. This can be explained by the preferred for P waves. The low seismic velocities could arise
orientation of olivine crystals, whose long[100]axes from a number of different mechanisms, including
are believed to lie in this direction. None of the an anomalously high temperature, a phase change, a
common minerals of eclogite exhibit the necessary compositional change,the presence ofopen cracks or
crystal elongation. A peridotitic composition is also fissures and partial melting. All but the latter appear
28 CHAPTER 2
to be unlikely, and it is generally accepted that the mass of the mantle at this level and so this phase
lower seismic velocities arise because ofthe presence transition is probably responsible for the 400 km
of molten material. That melting is likely to occur in seismic discontinuity. A further contribution is the
this region is supported by the fact that it is at this phase change ofaluminous orthopyroxenes to garnet,
level that mantle material most closely approaches its which occurs over the depth range of350-400 km.
melting point (Section 2.12, Fig. 2.36). The P-phase Mg2Si04 transforms into a spinel
Only a very small amount of melt is required to structure over the depth range of 500-600 km, and
lower the seismic velocity of the mantle to the ob at a depth of 670 km the spinel structure breaks
served values and to provide the observed attenu down to a mixture of MgSiOs with perovskite struc
ation properties. A liquid fraction of less than 1% ture and periclase (MgO). This involves an increase
would, if distributed along a network of fissures at in density(measured at zero pressure) ofabout 11%.
grain boundaries, produce these effects (O'Connell At about the same depth pyrope garnet transforms to
& Budiansky, 1977). The melt may also be respons an ilmenite structure and thence to a perovskite
ible for the high electrical conductivity of this zone. structure at a depth ofabout 800 km.The spinel and
For the partial melting to occur, it is probable that a garnet transformations are thus probably responsible
small quantity of water is required to lower the for the 670 km discontinuity. The mineral assem
silicate melting point, and this is probably supplied blages of the upper mantle and transition zone are
from the breakdown of hydrous mantle phases. The summarized in Table 2.4.
base of the low velocity zone is probably controlled A weaker discontinuity at 520 km depth is also of
by the availability of water. global extent (Shearer, 1990). It probably represents
The mantle low velocity zone is of major import the transformation from modified spinel to spinel
ance to plate tectonics as it represents a low viscosity structure.
layer along which relative movements of the litho- The recognition of the transition zone as repre
sphere and asthenosphere can be accommodated. senting a series of phase changes rather than com
positional changes has important ramifications for
the mechanism of plate tectonics. It means that
2.8.5 The mantle transition zone
convection currents in the mantle are capable of
There are two major velocity discontinuities in the crossing the transition zone (Section 11.5.3) if con
mantle at depths of400 km and 670 km.The former vection is to occur across the full vertical extent of
marks the top ofthe transition zone and the latter its the mantle.
base. The discontinuities are rarely sharp and occur The lower mantle is probably relatively homo
over a finite range in depth,so it is generally believed geneous in its mineralogy, having an essentially
that they represent phase changes rather than perovskite structure. The lowest 200-300 km of the
changes in chemistry. Some workers, however, pro mantle, however, are characterized by a decrease in
pose that the transition zone represents a change in seismic velocity, which is probably related to an
composition between an upper mantle dominated increased temperature gradient at the mantle-core
by Mg2Si04 in the form of olivine, and a lower boundary. The lowest 200-300 km ofthe mantle are
mantle more rich in silica composed of MgSiOa with known as the D" layer (Knittle & Jeanloz, 1991).
perovskite structure. It is also possible that large scale Laboratory experiments suggest that the liquid iron
lateral shearing associated with mantle convection of the core reacts with mantle silicates in the D"
contributes to the seismic signature of the disconti layer, with the production of metallic alloys and
nuities (Jeanloz, 1993). non-metallic silicates from perovskite. Earthquake
High pressure studies have shown that the olivine seismology has shown that these reactions create a
in the mantle, which is 89% forsterite ((Mgo.89 very heterogeneous layer. It is probable that mantle
^^^0.11)2^104), when subjected to a pressure corres plumes (Sections 5.4, 5.6) originate in the D"layer.
ponding to a depth of400 km and a temperature of
1600®C, transforms into a distorted form of spinel
structure known as the P-phase, in which the atoms 2.9 The core
are more closely packed(Ringwood & Major, 1970). The core, a spheroid with a mean radius of 3486 km,
It is believed that olivine constitutes about 57% by occurs at a depth of2891 km and occupies the centre
INTERNAL STRUCTURE OF THE EARTH 29
Table 2.4 Possible mineral assemblages ofthe upper mantle and transition zone (after Brown & Mussett, 1981).
c. 400
r. 670
\ 1 / \
Strontium llmenite (str) Perovskite (str)
plumbate (str) + + 3.99
simple oxides simple oxides
c. 1050
ofthe Earth. The core-mantle boundary(Gutenberg by any form of transition zone. The inner core is
discontinuity) generates strong seismic reflections believed to be solid for several reasons. Certain oscil
and thus probably represents a compositional inter lations of the Earth produced by very large earth
face. quakes can only be explained by a solid inner core. A
The outer core, at a depth of2891-5150 km,does seismic phase has been recognized which travels to
not transmit S waves and so must be fluid. This is and from the inner core as a P wave, but traverses the
confirmed by the generation ofthe geomagnetic field inner core as an S wave. The amplitude of a phase
in this region (Section 3.6.4) by dynamic processes reflected off the inner core also suggests that it must
and by the long period variations observed in the have a finite rigidity and thus be a solid.
geomagnetic field. The convective motions respons Shock wave experiments have shown that the ma
ible for the geomagnetic field involve velocities of jor constituents of both the inner and outer core
'^lO'^ma"^, five orders of magnitude greater than must comprise elements of an atomic number
convection in the mantle. A fluid state is also indi greater than 23, such as iron, nickel, vanadium or
cated by the response ofthe Earth to the gravitational cobalt. Of these elements, only iron is present in
attraction of the Sun and Moon. sufficient abundance in the solar system to form the
The boundary between the outer core and inner major part of the core. Again, by considering solar
core at 5150 km depth is sharp, and not represented system abundances, it appears that the core should
30 CHAPTER 2
contain about 4% nickel. This iron-nickel mixture Strain (e) is defined as the fractional change in
provides a composition for the outer core which is dimension of a material when it is subjected to a
8-15% too dense and it must therefore contain a stress, for example, the change in volume per unit
small quantity of some lighter element. The inner volume. Materials act in an elastic manner according
core, however, has a seismic velocity and density to Hooke's law, in which strain is proportional to
consistent with a composition of pure iron. stress, until a critical stress, known as the elastic limit,
There are several candidates for the light element is reached. Beyond the elastic limit, brittle materials
present in the outer core, which include silicon, deform by fracturing and ductile materials deform
sulphur, oxygen and potassium (Brett, 1976). Sili by flow. The elastic limit is a function oftemperature
con requires an over-complex model for the forma and pressure, so the behaviour of materials is
tion of the Earth and sulphur conflicts with the idea governed by these variables.
that the interior of the Earth is highly depleted in In considering the deformation of Earth materials
volatile elements. Oxygen appears to be the most it is important to recognize the importance of the
likely light element as FeO is probably sufficiently length of time for which stress is applied (Park,
soluble in iron. The presence of potassium is specu 1983). The behaviour of rocks in the short term is
lative, but is interesting in that it would provide a significantly different from that of the same material
heat source in the core which would be active over stressed over geologically realistic durations of
the whole ofthe Earth's history. It would also help to months or years. This difference in behaviour arises
explain an apparent potassium deficiency in the Earth from different strain rates. Rocks exhibit higher
compared to meteorites. strength at high strain rates than at low strain rates.
This is typified by the behaviour of pitch. When
struck with a hammer,that is, subjected to 'instanta
2.10 Deformation in the crust
neous'strain, pitch shatters. When left for a period of
and mantle
months, however, a block of pitch deforms slowly by
2.10.1 Introduction
flowing. The long-term behaviour ofstrained materi
als is known as creep.
.When a solid is stressed, it becomes strained. The
relationship between stress and strain consequently
2.10.2 Brittle deformation
governs the form of the deformation experienced by
the solid. A comprehensive account of deformation Brittle fracture is believed to be caused by progressive
processes is given by Ranalli (1987). failure along a network of micro- and mesocracks
Stress (a) is defined as the force exerted per unit by the amplification of stress within those cracks
area of the rock surface, and is thus a pressure orientated in the most suitable directions. When the
measured in pascals (Pa). The confining pressure magnitude of an applied tensional stress exceeds the
experienced by rocks with increasing depth in tensile strength ofthe material, cracks orthogonal to
the crust and mantle increases at a rate of about the tension fail first and extensional fracture occurs.
33MPa km~^ (0.33 kbarkm"^) depending on the Below a depth of a few hundred metres all prin
density ofthe overlying rocks. Any stress acting upon cipal pressures are usually compressional. Any cracks
a surface can be expressed in terms ofa normal stress progressively close with increasing overburden pres
perpendicular to the surface and two components of sure, and are probably completely closed when a
shear stress in the plane of the surface. The state of depth ofabout 5 km is reached. This implies that the
stress within a medium is conveniently specified by compressive strength of a material at low confining
the magnitudes and directions ofthe three principal pressures is much greater than the tensile strength, as
pressures which act on the three particular planes in failure is facilitated by pre-existing lines of weakness.
the medium along which the shear stress is zero. For example, the compressive strength of granite at
These are orthogonal and are termed and atmospheric pressure is 140 MPa, and its tensile
<^min5 that is, the maximum, intermediate and mini strength only about 4 MPa. When all cracks are
mum principal stresses, respectively. In geology, closed,fracture is controlled by the inherent strength
compressive stresses are normally treated as positive ofthe material,the normal stress on a plane offailure
and tensional stresses negative. and the coefficient of friction on that plane, which
INTERNAL STRUCTURE OF THE EARTH 31
<5 max
icr mIn
l^int
(b) (c)
Fig.2.21 Three classes offault determined by the nature ation with depth might cause the lithosphere to have
ofthe vertical principal stress:(a) normal fault;(b)thrust a 'sandwich' type theological layering (Ranalli &
fault;(c) strike-slip fault. Murphy, 1987), with perhaps several ductile layers
separating brittle layers. For example. Brace & Kohl-
resists relative motion across it (the Coulomb- stedt (1980) have investigated the limits of litho
Navier hypothesis). Under this closed crack regime, sphere strength based on measurements on quartz
the type of faulting which results, according to the and olivine, which are typical ofthe continental crust
simple theory of Anderson, depends upon which of and upper mantle, respectively. Their findings are
the principal pressures is vertical (Fig. 2.21). Nor illustrated in Fig. 2.23 (Molnar, 1988). Within the
mal,thrust and strike-slip faults occur depending on oceanic lithosphere the upper brittle crust gives way
whether a„ Omin or Oinj, respectively, is vertical. to a region of high strength at a depth of 20-60 km
Faulting at depth, which would otherwise appear to depending on the temperature gradient. The
require very high shear stresses because of the high strength thence gradually decreases and grades into
norrnal stresses, is probably facilitated by the reduc
tion of the frictional coefficient and the normal
stresses across the fault by the presence ofpore waters Oi + P
and hence a pore fluid pressure which counteracts the
normal stress.
The strength of brittle rocks increases with con
fining pressure but decreases with temperature. In
I .1. I
the uppermost 10 km the former effect is dominant
and the strength increases. Below this depth the
effect of temperature takes over, and rocks progres p:Rollmg;|
sively weaken downwards.
The mechanism of deformation of brittle solids
may be that of cataclasis (Fig. 2.22)(Ashby & Ver-
rall, 1977). This results from repeated shear fractur
ing, which acts to reduce the grain size of the rock,
and by the sliding or rolling ofgrains over each other.
Strength Strength
Wa^r
Crust' Britt e-
Moho
ductile Crust
transition
Brittle
ductile Moho
Brittle-
transition
ductile
^ 50- ► Mantle transition
Mantle
Fig. 2.23 Schematic strength
Oceanic Continental
lithosphere
profiles through oceanic and
lithosphere
continental lithosphere (after
Molnar, 1988, using information
100-* 100^ from Brace & Kohlstedt, 1980).
that of the asthenosphere. Continental crust, how strain rate, that is, lower the apparent viscosity, while
ever, is much thicker than oceanic, and at the temper increased pressure produces a more sluggish flow.
atures of 400-700 ®C experienced in its lower parts Plastic flow (Fig. 2.24) occurs when the yield
the minerals are much weaker than the olivine found strength of the material is exceeded. Movement takes
at this depth in the oceanic lithosphere, so, whereas place in a manner similar to cataclasis, in that defor
the oceanic lithosphere behaves as a single rigid plate mation occurs by gliding motions along intersecting
because of its strong centre, the continental litho planes within crystals and movement along crystal
sphere is characterized by a layer of weakness at lower
crustal depths. This theological layering may explain
how the continental crust may detach from the Oi + p
mantle part of the lithosphere on continental colli
sion and create a Himalayan-type mountain range by
the stacking of thrust slices of crust (Section 9.3).
The mechanisms of flow in crystalline solids have XWVXXYWT"
Dislocation glide
been largely deduced from studies of the behaviour on intersecting ^
of metals, which have the advantage that they deform planes
in a ductile manner at lower temperatures and pres
sures than silicates. Suflicient is known of silicate
deformation, however, to validate deducing their
flow behaviour at high temperatures and pressures
from that of metals. It is now possible, using modern
apparatus, to determine directly the temperature and
pressure dependence of creep in mantle materials
(Borch & Green, 1987).
When the temperature is low, up to about half the
melting temperature {T^ in kelvin), materials react
to low stresses by transient creep, which is a very slow
flow in the solid state. Only a small amount of
deformation results and large permanent strains are
not generated.
There are three types of ductile flow which are
thought to occur beneath the brittle-ductile transi
tion (Ashby & Verrall, 1977). All are dependent Fig. 2.24 Deformation by plastic flow (redrawn from
upon the ambient temperature and, less markedly, Ashby & Verrall, 1977, with permission from the Royal
pressure. Increased temperature acts to increase the Society).
INTERNAL STRUCTURE OF THE EARTH 33
CTi + p +P
Cell Boundary
formation sliding
Boundary
diffusion
Tj_ Diffusion_x
Diffusive
fluxes
\ / Lattice
diffusion
Dynamic recrystallization
Qi + P
Oi + p
Fig.2.26 Deformation by diffusion creep (redrawn from
Fig.2.25 Deformation by power-law creep (redrawn Ashby & Verrall, 1977, with permission from the Royal
from Ashby & Verrall, 1977, with permission from the Society).
Royal Society).
boundaries. The yield strength of materials deform Power-law creep is believed to be the most important
ing in this way is thus controlled by the magnitude of form of deformation occurring in the mantle, and
the stresses required to overcome the resistance of governs convective flow within it(Weertman,1978).
the crystal framework to this form of dislocation. Diffusion creep becomes dominant as temperatures
The strain produced tends to be limited by strain- or exceed O.SST^^ and results from the migration
work-hardening. of atoms in a stress gradient (Fig. 2.26). Atomic
Power-law creep (also known as hot creep, recovery migration through the crystal lattices is known as
creep or Weertman creep) (Fig. 2.25) takes place at Nabarro-Herring creep, while migration along crys
temperatures in excess of 0.55T^. In this form of tal boundaries is known as Coble creep. In this form
creep the strain rate is proportional to the nth. power of creep the strain rate (k) is proportional to the
ofthe stress, where » ^ 3. Power-law creep is similar applied stress (a) so that 0= 11^ the constant of
to plastic flow, with deformation taking place along proportionality being the dynamic viscosity (ri). ^
crystal glide planes, but diflfers in that diffusion of increases as the square of the grain radius. Diffusion
atoms is permitted by the higher temperatures so that creep is believed to occur within the asthenospheric
any barriers to movement are removed immediately part of the mantle (Section 2.12).
as they form; hence, work-hardening does not occur
and steady state creep is facilitated. The dislocations
2.10.4 Creep in the mantle
take up a polygonal form whose size increases with
increased stress. Sliding along crystal boundaries In applying the creep laws for metals to behaviour
and recrystallization may also occur. Because of the in the mantle, it should be recognized that experi
higher temperature the yield strength is lower than mental data cannot duplicate the very slow strain
for plastic flow,and strain results from lower stresses. rates experienced naturally. In addition, the effect of
34 CHAPTER 2
pressure on creep rates is poorly known, rock lines were used as a vertical reference in the surveying
strength is significantly reduced by the presence of and it was recognized that a correction would have
water, and mantle material is probably not mono- to be applied for the horizontal deflection caused by
mineralic. Consequently, the strain rates calculated the gravitational attraction of the Andes. When this
for the mantle may be much in error. correction, based on the mass of the Andes above
Low temperature plastic flow is likely to be opera sea level, was applied, however, it was found that
tive immediately beneath the brittle-ductile transi the actual vertical deflection was less than predicted
tion (Carter & Ave'lallemant, 1970; Goetze, 1978), (Fig. 2.27). This phenomenon was attributed to the
and because of the reasonably high yield strength existence of a negative mass anomaly beneath the
stress differences of about ICQ MPa may be present. Andes which compensates, that is to say, supports,
As the temperature increases to 0.55 power-law the positive mass of the mountains. In the 19th
creep is expected to become dominant. Because of century similar observations were made in the vicin
the differing compositions this would be at about ity of the Himalaya and it was recognized that the
830 K in the subroceanic mantle, but at only 530 K compensation of surface loading at depth is a wide
in the continental crust. The presence of structural spread phenomenon.
evidence in mantle-derived olivine nodules suggests The presence of subsurface compensation is con
that power-law creep exists to a depth of at least firmed by the variation in the Earth's gravitational
200 km. field over broad regions. Bouguer anomalies (Kearey
The strain rate at which power-law creep occurs & Brooks, 1991)are generally negative over elevated
should increase drastically on nearing the mantle low continental areas and positive over ocean basins
velocity zone, where temperatures most closely ap (Fig. 2.28). These observations confirm that the
proach the melting temperature. At greater depths, positive topography ofcontinents and negative topo
diffusion creep may become increasingly prominent graphy ofoceans is compensated by regions at depth
as the stress diflferences decrease and the temperature with density contrasts which are, respectively, nega
rises. The rate ofdiffusion is believed to be controlled tive and positive and whose mass anomaly approxi
by oxygen, which is the slowest ofthe atomic species mates that of the surface features.
present in olivine. However,computations based on The principle of isostasy is that beneath a certain
both Nabarro-Herring and Coble mechanisms seem depth, known as the depth of compensation, the
to suggest that power-law creep is dominant, except pressures generated by all overlying materials are
for the slowest strain rates. everywhere equal; that is, the weights of vertical
columns of unit cross-section, although internally
2.11 Isostasy
2.11.1 Introduction
.'-'-Root':-
Dep^ of
compensation
swa
Lrr N^piPc
Fig.2.28 Inverse correlation of Bouguer anomalies with
topography indicating its isostatic compensation.
relatively low density material and ocean basins by weak as this implies, as large gravity anomalies exist
relatively high density material (Fig. 2.29b). Equat over igneous intrusions with ages in excess of
ing the weights of columns of unit cross-section 100 Ma. The lithosphere must therefore be able to
beneath a mountain range and a region of zero support stress differences of up to 20-30 MPa for
elevation gives: considerable periods oftime without the necessity of
local compensation.
- 3000 - — Observed
Computed D = 2 x 10^^ Nm
E 2000
o D = 6x1022Nm
D =4xlO"Nm
^ 1000
Distance(km)
the Gulfof Bothnia, where the land is rising at a rate where A^oot is the gravity anomaly of the compen
of over 10 mm a"^. Twenty thousand years ago the sating root. Since the root is broad compared to its
land surface was covered by an ice sheet about 2.5 km thickness, its anomaly may be approximated by that
thick (Fig. 2.32a). The lithosphere accommodated of an infinite slab, that is 2kG{p^ - p^)^, where p^, is
this load by flexing (Fig. 2.32b), resulting in a sub the density of the substrate. Combining the above
sidence of600-700 m and a lateral displacement of two equations:
asthenospheric material. This stage currently pertains
lA = FAA - 2nGp^h - 27cG(pc - Pn,)r
in Greenland and Antarctica where, in Greenland,
the land surface is depressed by as much as 250 m From the Airy criterion for isostatic equilibrium:
below sea level by the weight ofice. Melting ofthe ice
f =/;p,/(p„-pc)
was complete about 10000 years ago (Fig. 2.32c),
and since this time the lithosphere has been return Substitution of this condition into the equation
ing to its original position and the land rising in reveals that the isostatic anomaly is equal to the free
order to regain isostatic equilibrium. A similar situa air anomaly over a broad flat feature, and this repre
tion pertains in northern Canada where the land sents a simple method for assessing the state of
surface around Hudson Bay is rising subsequent to isostatic equilibrium.
the removal of an icecap. The rate of isostatic re More rigorous tests of isostasy over topographic
bound provides an estimate for the viscosity of the features of irregular form are made by accurate com
upper mantle of 10^^Pas (Pascal seconds), and putation of isostatic anomalies. This procedure in
measurements based on world-wide modelling of volves calculating the shape of the compensation
post-glacial recovery and its associated oceanic required by a given hypothesis ofisostasy,computing
loading suggest that this figure generally applies its gravity anomaly and then subtracting this from
throughout the upper mantle as a whole (Peltier & the observed Bouguer anomaly to provide the iso
Andrews, 1976). Compared to the viscosity of water static anomaly. Figure 2.33 shows free air, Bouguer
(10"^ Pa s) or a lava flow (4x10^ Pa s), the viscosity and isostatic anomalies over a feature with varying
of the sub-lithospheric mantle is extremely high and degrees ofcompensation.
its fluid behaviour is only apparent in processes with Gravity anomalies can thus be used to determine if
a large time constant. Knowledge of the viscosity of a surface feature is isostatically compensated at
the mantle, however, provides an important control depth. They cannot, however, reveal the form of
on the nature of mantle convection, as will be compensation and indicate which type of mechanism
discussed in Chapter 11. is in operation. This is because the compensation
occurs at a relatively deep level and the differences in
the anomalies produced by a root/antiroot (accord
2.11.6 Tests of isostasy
ing to the Airy hypothesis) or by different density
The state ofisostatic compensation ofa region can be units (according to the Pratt hypothesis) would be
assessed by making use of gravity anomalies. The very small. Moreover, the gravity anomalies over
isostatic anomaly^ lA^ is defined as the Bouguer anom most regions contain short wavelength components
aly minus the gravity anomaly of the subsurface resulting from localized, uncompensated geological
compensation. Consider a broad, flat plateau of structures which obscure the differences in the re
elevation h compensated by a root ofthickness r. The gional field arising from the different forms of
terrain correction of such a feature is small in the compensation.
central part of the plateau so that here the Bouguer Information on the geometric form of isostatic
anomaly,BA,is related to the free air anomaly,FAA compensation can be gained by a combined analysis
by the relationship: of gravity and seismic,refraction data, as the latter
BA^FAA-BC
technique can provide a reasonably detailed picture
of the subsurface structure of the region under con
where BC is the Bouguer correction, equal to sideration. Such studies have demonstrated that the
2nGp^h^ where is the density of the compensated broad isostatic equilibrium ofcontinents and oceans
layer. For such an Airy compensation: is mainly accomplished by variations in crustal thick
ness according to the Airy hypothesis. Figure 2.34
INTERNAL STRUCTURE OF THE EARTH 39
Bouguer
Bouguer
'0% compensation
70% compensation
Q
/-\ 50-^
Rn - : ■•
>,... •• 3 30 I
..V .. "i *50
(b) 0 1000 km
Airy isostatic
Airy isostatic
o 1500-
Bouguer
Bouguer
30% compensation 0% compensation
I 40-
Fig. 2.33 Free air, Bouguer and Airy isostatic anomalies Plains, which reach a mean height of 1 km, are
over an idealized mountain range (a) in perfect isostatic underlain by crust 40-50 km thick and the Basin and
equilibrium, (b) with 70% isostatic compensation, Range province, at 2 km elevation, is underlain by
(c) with 30% isostatic compensation, (d) uncompensated. 25-30 km thick crust. Clearly, the Basin and Range
Densities in Mg m"^. province must be partially compensated by a Pratt-
type mechanism resulting from the presence of low
shows seismic velocity sections from the western density material in the upper mantle. Similarly,
USA in which surface topography is largely compen ocean ridges (Chapter 6) owe their elevation to a
sated by Moho topography, although in several loca region of low density material in the upper mantle
tions density variations in the upper mantle must be rather than to a thickened crust.
invoked to explain the isostatic compensation. A There are regions of the Earth's surface which do
cross-section across the western USA (Fig. 2.35) re not conform to the concepts of isostasy discussed
veals, however, that crustal thickness is not necessar here. The lithosphere has sufficient internal strength
ily related to topographic elevation as the Great to support relatively small loads without the
40 CHAPTER 2
o-
O •
8 8 •
E -1000-
o
• Isostatic anomaly
o Bouguer anomaly
3.6 4.64 4.8 3.8
5.68.v. 5.6 ® °
>.11 6.6
6.07
(D
C
CO
c c
CO CO
c
8.07
CD
c
D3
C
(/)
CO
CO
c
hUta CO
T3
■a
c
<0
'c
CO
"c (0
CD
CO
'c
c
o
CO CO (/) CO k.
CO CO CO o _CD O)
E c N > 0 o o o
c c c c c CO CD U c
o C
o o O o o
< Z lo "co CO "cD crt
Fig. 2.34 Bouguer and isostatic
5 2 2 < CO u U 'c CJ CO
5 si 2 lU
"c
(/3 C/j o "co s:
gravity anomalies and their relation
z CO
to seismic velocity sections from the
western USA. Velocities in km .s"'
(redrawn from Garland, 1979),
Colorado
I .c >- c I
slillill Basin and Range Colorado
Il1|l Great
siMii
u
^West f
Province Plateau I CO tr 2 I Plains
,1 Jr.'.Cs', East Fig. 2.35 Section from San
E 2o4 Francisco, California to Lamar,
Colorado based on seismic
elastic manner. It is made up ofthe crust and upper the melting temperature of the mantle materials(Le
most mantle. The lithosphere is underlain by the Pichon et al.^ 1973). Beneath ocean ridges, where
asthenosphere, which is a much weaker layer and temperature gradients are high, the asthenosphere
reacts to stress in a fluid manner. The lithosphere is must occur at shallow depth. Indeed, since it is
divided into plates, of which the crustal component actually created in the crestal region (Section 6.10),
can be oceanic and/or continental, and the relative the lithosphere there is particularly thin. The gradi
movements of plates take place upon the astheno ent decreases towards the deep ocean basins, and the
sphere. lithosphere thickens in this direction, the increase
However, having made these relatively simple correlating with the depth ofwater as the lithosphere
definitions, examination of the several properties subsides as a result of contraction on cooling
which might be expected to characterize these layers (Section 6.9). The mean lithosphere thickness on
reveals that they lead to different ideas of their this basis beneath oceans is probably 60-70 km.
thickness. The properties considered are thermal, Beneath continents a substantial portion of the ob
seismic, elastic and temporal. served heat flow is produced within the crust(Section
Temperature is believed to be the main pheno 2.13), so the temperature gradient in the sub-crustal
menon which controls the strength of subsurface lithosphere must be considerably lower than in
material. Hydrostatic pressure increases with depth oceanic areas. It is probable that the mantle solidus is
in an almost linear manner,and so the melting point not approached until a significantly greater depth,so
ofrocks also increases with depth. Melting will occur that the continental lithosphere has a thickness of
when the temperature curve intersects the melting 110-150 km, being at a maximum beneath cratonic
curve (solidus) for the material present at depth areas.
(Fig. 2.36). The asthenosphere is believed to repre The depth of the LVZ for seismic waves agrees
sent the location in the mantle where the melting quite well with the temperature model oflithosphere
point is most closely approached. This layer is cer and asthenosphere. The LVZ progressively thickens
tainly not completely molten,as it transmits S waves, away from the crests of mid-ocean ridges (Forsyth,
but it is possible that a small amount of melt is 1975)(Fig. 2.37), reaching an average thickness for
present. The depth at which the asthenosphere old oceanic lithosphere of 70-80 km. The LVZ is
occurs depends upon the geothermal gradient and rather deeper beneath continents, and occurs at a
depth of 110-150 km. Within the LVZ attenuation
of seismic energy, particularly shear wave energy, is
2000
very high. Both the low seismic velocities and high
1600- Age(Ma)
10 20 30 40 50
a
o
(U 1200-
L-
D 30-
to
60-
800-
E
A Continent 90-
.c
B Precambrian shield
C Oceanic ridge
Q.
0)
Zone of partial melting 4.1 km s ^
400- Q 120-
D Ocean basin
150-
Depth (km)
Fig.2.37 Shear wave model ofthe thickening of oceanic
Fig. 2.36 Variation of temperature with depth beneath lithosphere with age. Velocities in km s~^ (redrawn from
continental and oceanic regions (redrawn from Condie, Forsyth, 1975, with permission from the Royal
1982,Plate Tectonics and Crustal Evolution^ with Astronomical Society). The 150 km transition may be
permission from Pergamon Press Ltd). somewhat deeper.
42 CHAPTER 2
attenuation are consistent with the presence of a such as the Basin and Range Province, the elastic
relatively weak layer at this level. As would be ex thickness may be as small as 4 km, which is less than
pected for a temperature-controlled boundary, the the crustal thickness (Bechtel etal., 1990). Such very
lithosphere-asthenosphere interface is not sharply thin elastic thicknesses are undoubtedly due to very
defined,and occupies a zone several kilometres thick. high geothermal gradients.
When the Earth's surface is loaded, the lithosphere The concept of the lithosphere as a layer of uni
is believed to react by downward flexure. This is formly high strength is seen to be over-simplistic
frequently observed on continental areas in response when the theological layering is considered. The
to loading by ice sheets, large sedimentary basins and upper 20-40 km of the lithosphere are brittle and
mountain ranges, but is less common in the oceans respond to stress below the yield point by elastic
where only the loading by seamounts occurs with any deformation accompanied by transient creep. Be
frequency. The amount of flexure depends upon the neath the brittle zone is a layer which deforms by
magnitude of the load, the flexural rigidity of the plastic flow above a yield point of about 100 MPa.
lithosphere and the lithospheric thickness. The mag The lowest part, which is continuous with the as-
nitude of the load can be calculated and the flexural thenosphere, deforms by power-law creep and is de
rigidity of the lithosphere is now known with reason fined as the region where the temperature increases
able accuracy. Consequently, flexure studies can be with depth from 0.55 to 0.85 T^.
used to determine the elastic thickness of the litho The type of deformation experienced depends
sphere. The elastic thickness of the oceanic litho upon the duration ofthe applied loads. Over pei iods
sphere deduced in this way is 40 km or less, and of a few thousand years, most of the region exhibit
increases in thickness with distance from ocean ing power-law creep does not deform significantly
ridges (Watts et al., 1980)(Fig. 2.38). Beneath con and consequently is included within the elastic litho
tinental shield areas the elastic thickness can vary sphere. Long term loading, however, occurring over
from 30 to 110 km. Under tectonically active areas. periods of a few million years, permits power-law
deformation to occur so that this region then be
longs to the asthenosphere.
Age of oceanic lithosphere(Ma) It can thus be appreciated that the lithosphen; can
40 80 120 160 be defined in a number of different ways w hich
provide different estimates of its thickness. This
must be borne in mind throughout any consider
ation of plate tectonic processes.
20- m
The asthenosphere is believed to extend to a depth
"2^ 35( of about 700 km. The properties of the underlying
Ela
Elastic
region are only poorly known. Seismic waves w hich
thickness cross this region do not suffer great attenuation
(Section 8.3), and so it is generally accepted that this
is a layer of high strength, termed the mesosphere. The
compositional and theological layering of the Earth
are compared in Fig. 2.39.
6 km Continent Inn
""" U35h5phe;^--2*{n
Mohorovicic\jJ^
discontinuity \ ^
400 krtr^
670 knA
!Lower mantle;; ;;Mesosphere;;
Gutenberg —
2891 km^
discontinuity
5150 km^
The vast majority of the heat affecting the Earth's exponential, so that during the early history of the
surface comes from the Sun, which accounts for Earth the concentration of radioactive isotopes
some 99.98% of the Earth's surface energy budget. would have been significantly higher than at present
Most of this thermal energy, however, is reradiated and the thermal energy available to power its internal
into space, while the rest penetrates only a few processes would have been much greater (Section
hundred metres below the surface. Solar energy 12.2.2). Currently accepted models for the forma
consequently has a negligible effect on thermal pro tion of the Earth require an early phase of melting
cesses occurring in the interior of the Earth. The and differentiation of its originally homogeneous
geothermal energy loss from heat sources within the structure. This melting is believed to have been
Earth constitutes about 0.022% of its surface energy powered in part by thermal energy provided by the
budget. Other sources of energy include the energy decay ofshort-lived radioactive isotopes such as AP®,
generated by the gradual deceleration of the Earth's Ee®" and CP^.The differentiation ofthe Earth would
rotation and the energy released by earthquakes, but also have contributed energy to the Earth arising
these make up only about 0.002% of the energy from the loss in gravitational potential energy as the
budget. It is thus apparent that geothermal energy is dense iron-nickel core segregated to a lower energy
the" major source of the energy which drives the state at the centre of the Earth.
Earth's internal processes. The heat flow through a unit area of the Earth's
It is believed that the geothermal energy is derived surface, H,is given by:
in part from the energy given off during the radio
active decay oflong-lived isotopes, in particular
U23S, and Th^®^, and also from the heat re
H.fK oz
leased during the early stages of the formation of the where 6T/5z is the thermal gradient perpendicular
Earth. These isotopes would account for the present to the surface and K the thermal conductivity of the
geothermal loss if present in proportions similar to medium through which the heat is flowing. The units
those of chondritic meteorites. Radioactive decay is ofJT are mW m"^.
44 CHAPTER 2
On land, heat flow measurements are normally the heat flow decreases with the age of the litho-
made in boreholes. Mercury maximum thermomet sphere (Section 6.9), with high values over the
ers or thermistor probes are used to determine the oceanic ridge system and active marginal seas and
vertical temperature gradient. Thermal conductivity low values over the deep ocean basins and inactive
is measured on samples ofthe core using a technique marginal seas.
similar to the Lee's disc method. Although appearing The average heat flux in continental areas is
relatively simple, accurate heat flow measurements 57 mW m"^, and in oceanic areas lOI mW m~^, of
on land are difficult to accomplish. The drilling of a which about 30% is contributed by hydrothermal
borehole necessitates the use of fluid lubricants activity at the mid-oceanic ridge system (Davies,
which disturb the thermal regime of the borehole so 1980). The continental value is about 60% of the
that it has to be left for several months to allow the oceanic; about 75% of the geothermal energy is
disturbance to dissipate. Porous strata have to be lost through the oceans, and 25% through the
avoided as pore water acts as a heat sink and distorts continents.
the normal thermal gradients. Consequently, it is
rarely possible to utilize boreholes sunk for the
purposes ofhydrocarbon or hydrogeological explora Further reading
tion. In many areas readings may only be undertaken Anderson, D.L. (1989) Theory of the Earth. Blackwell
at depths below about 200 m so as to avoid the Scientific Publications, Oxford.
transient thermal effects of glaciations. Ashby, M.F. & Verrail, R.A.(1977) Micro-mechanisms of
Heat flow measurements are considerably easier to flow and fracture, and their relevance to the rheology of
accomplish at sea. The bottom temperatures in the the upper mantle. Phil. Trans. Roy. Soc. Land. 288A,
oceans remain essentially constant and so no compli 59-95.
cations arise because of transient thermal perturba Bott, M.H.P.(1982)The Interior ofthe Earth. Its Structure,
tions. A temperature probe is dropped into the upper Constitution and Evolution (2nd edn). Edward Arnold,
London.
soft sediment layer of the sea bed and, after a few
Brett, R.(1976) The current status of speculations on the
minutes' stabilization, the temperature gradient is composition ofthe core ofthe Earth. Rev. Geophys. Space
measured by a series of thermistor probes. A corer Phys. 14, 375-83.
associated with the probe collects a sediment sample Brown, G.C. & Mussett, A.E. (1981) The Inaccessible
for thermal conductivity measurements; alterna Earth. Unwin Hyman,London.
tively, the role of one of the thermistors can be Condie, K.C.(1982)Plate Tectonics and Crustal Evolution.
changed to provide a source of heat. The change in Pergamon Press, Oxford.
the temperature of this probe with time depends on Edgar, N.T. (1974) Acoustic stratigraphy in the deep
the rate at which heat is conducted away from it, and oceans. In: Burk, C.A. & Drake, C.L.,(eds) The Geolo^
this enables a direct, in situ measurement of the of Continental Marpfins, pp. 243-6. Springer-Verlag,
Berlin.
thermal conductivity ofthe sediment to be made.
Elthon, D. (1981) Metamorphism in oceanic spreading
A large proportion of geothermal energy escapes
centres. In-. Emiliani, C. (ed.) The Oceanic Lithosphere,
from the surface by conduction through the solid The Sea 7, pp. 285-303. Wiley, New York.
Earth. In the region of the oceanic ridge system, Fowler,C.M.R.(1990)The Solid Earth. Cambridge Univer
however, the circulation of sea water plays a major sity Press, Cambridge.
role in transporting heat to the surface and about Francheteau,J.(1983)The oceanic crust. Sci. Am. 249(3),
25% of the geothermal energy flux at the Earth's 68-84.
surface is lost in this way. Green, D.H. & Ringwood, A.E.(1967) An experimental
The pattern of heat flow provinces on the Earth's investigation of the gabbro to eclogite transition and its
surface broadly correlates with major physiographic petrological applications. Geochim. Cosmochim. Acta. 31,
767-833.
and geological subdivisions. On continents the
Jacobs,J.A.(1991)The Deep Interior ofthe Earth. Chapman
magnitude of heat flow generally decreases from the
& Hall, London.
time of the last major tectonic event (Sclater et»/., Kay, R.W. & Kay, S.M. (1981) The nature of the lower
1980). Heat flow values are thus low over the continental crust: inferences from geophysics, surface
Precambrian shields and much higher over regions geology and crustal xenoliths. Rev. Geophys. Space Phys.
affected by Tertiary orogenesis. Within the oceans 20, 735-60.
INTERNAL STRUCTURE OF THE EARTH 45
Mason,R.(1985) Ophiolito. Geokjgy Today 1,136-40. Parson, L.M., Murton, B.J. & Browning, P. (eds)(1992)
Moores, E.M.(1982) Origin and emplacement of ophio- Ophiolites and Their Modem Oceanic Analogues. Spec.
lites. Rev. Geophys. Space Rhys. 20,735-60. Pub. 60, Geological Society ofLondon.
Nicolas, A.(1989) Structure of Ophiolites and Dynamics of Ringwood, A.E. (1975) Composition and Petrolopy of the
Oceanic Lithosphere. Kluwer Academic Publishers, Earth^s Mantle. McGraw-Hill, New York.
Dordrecht.
Park, R.G.(1988) Geolopfical Structures and Moving Plates.
Blackie, Glasgow.
3/ Continental drift
xis of rotation
3.2.2 Geometrical reconstructions of
ole of rotation
continents
ngle of rotation
Although approximate reconstructions can be per
formed manually by moving models of continents
across an accurately constructed globe, the most
rigorous reconstructions are performed mathemat
ically by computer, as in this way it is possible to
minimize the degree of misfit between the juxta
posed continental margins.
The technique generally adopted in computer-
Great circle or based continental fitting is to assume a series of poles
equator of rotation of rotation for each pair of continents arranged in a
Small circles or grid of latitude and longitude positions. For each
latitudes of rotation
pole position the angle of rotation is determined
which brings the continental margins together with
Fig. 3.1 Euler's theorem. Diagram illustrating how the the smallest proportion of gaps and overlaps. The fit
motion of a continent on the Earth can be described by an is not made on the coastlines, as continental crust
angle ofrotation about a pole of rotation. extends beneath the surrounding shelfseas out to the
CONTINENTAL DRIFT 47
continental slope. Consequently, the true junction vides the pole of rotation for which the continental
between continental and oceanic lithosphere is taken edges fit most exactly.
to be at some isobath marking the midpoint of the
continental slope, for example the 1000 m contour.
3.2.3 The reconstruction of
Having determined the angle of rotation, the good
continents around the Atlantic
ness offit is quantified by some criterion based on the
degree of mismatch. This goodness offit is generally The first mathematical reassembly of continents
known as the objective function. Values ofthe objec based solely on geometrical criteria was performed by
tive function are entered on the grid ofpole positions Bullard et al. (1965), who fitted together the conti
and contoured. The location ofthe minimum objec nents on either side of the Atlantic (Fig. 3.2). This
tive function revealed by this procedure then pro was accomplished by sequentially fitting pairs of
continents after determining their best fitting poles part, on its pattern of magnetic lineations. Palaeo-
of rotation by the procedure described in Section magnetic evidence (Section 3.6) suggests that the
3.2.2. The only rotation involving parts ofthe same Palaeozoic crustal blocks which underlie Central
iandmass is that of the Iberian peninsular with re America were originally situated within the region
spect to the rest ofEurope. This is justified because of now occupied by the GulfofMexico,an area existing
the known presence ofoceanic lithosphere in the Bay within the reconstruction (White, 1980). The North
ofBiscay which is closed by this rotation. Geological Atlantic started opening about 180 Ma ago, and the
evidence (Section 3.3) and information provided by South Atlantic somewhat later, about 135 Ma ago.
magnetic lineations in the Atlantic (Section 4.5) The poles of rotation ofthe North and South Atlan
indicate that the reconstruction represents the con tic were sufficiently different that the opening created
tinental configuration during early Jurassic times the space between North and South America now
some 200 Ma ago. occupied by the Caribbean. This also allowed a clock
Examination of Fig. 3.2 reveals a number of wise rotation of the Central American blocks out of
overlaps of geological significance, some of which the Gulf of Mexico to their present locations. About
may be related to the process ofcontinental splitting 80 Ma ago the poles of rotation of the North and
(Section 10.3). Iceland is absent because it is of South Atlantic changed to an almost identical loca
Tertiary to Recent age and its construction during tion in the region of the present north pole so that
the opening ofthe Atlantic postdates the reconstruc from this time the whole Atlantic Ocean effectively
tion. The Bahama Platform appears to overlap the opened as a single unit.
African continental margin and mainland. It is prob
able, however, that the platform represents an accu
3.2.4 The reconstruction of Gondwana
mulation ofsediment capped by coral which formed
after the Americas separated (Dietz & Holden, Geometrical evidence alone has also been used in the
1970). Similarly,the Niger Delta ofAfrica appears to reconstruction of the southern continents which
form an overlap when in fact it also developed after make up Gondwana. The first such reconstruction
rifting. was performed by Smith & Hallam (1970) and is
A major criticism of the reconstruction is the illustrated in Fig. 3.3. The shapes of the continental
overlap of Central America on to South America and edges of the east coast of Africa, Madagascar, India,
absence of the Caribbean Sea. This must be viewed, Australia and Antarctica are not quite so well suited
however,in the light ofour knowledge ofthe history to fitting as the circum-Atlantic continents, and so
of the opening of the Atlantic based, for the most the reconstruction is more controversial. It does.
however, represent the best least-squares fit with a between these must now be based on geological
minimum of overlap and gaps. The main uncertain (Section 3.3), palaeomagnetic (Section 3.6) and sea
ties are the locations of Madagascar and India within floor spreading (Section 4.6) criteria.
the reconstructions, and Fig. 3.4, from Powell et al.
(1980), illustrates some of the alternative arrange
ments which have been suggested. Differentiation 3.3 Geological evidence for
continentai drift
The continental reconstructions discussed in Sec
tions 3.2.3 and 3.2.4 are based solely on the geomet
rical fit of continental shelf edges. If they represent
the true ancient configurations of continents it
should be possible to trace continuous geological
features from one continent to another across the
fits. The matching of features requires the rifting of
the supercontinent across the general trend of geo
logical features. This does not always occur as the
location ofthe rift is often controlled by the geology
of the supercontinent, and takes place along lines of
weakness which may run parallel to the geological
grain. However, there remain many geological fea
tures which can be correlated across juxtaposed con
tinental margins, some of which are listed below.
1 Fold belts. The continuity of Palaeozoic fold belts
across the assembled Laurasia and Gondwanaland
supercontinents is illustrated in Fig. 3.5. An espe
cially well-studied case is the continuation of the
Appalachian fold belt ofwestern North America with
the Caledonian fold belt ofnorthern Europe. Within
the sedimentary deposits associated with fold belts
there is often further evidence for continental drift.
The grain size and composition ofthe sediments can
be used to determine the nature and direction of
their source. The source of sediments in the Cale-
donides of northern Europe lies to the west in a
location now occupied by the Atlantic, indicating
that, in the past, this location must have been occu
pied by continental crust. The source ofthe 1 Ga old
Roraima Formation sediments of Guyana lies to the
east and appears to be situated in northwest Africa.
Diamonds at the base of this formation increase in
size to the east, and in West Africa diamonds in
similar sediments also increase in size eastwards to
their source in the Sudan (Tarling & Tarling, 1971).
2 A£fe provinces. The correlation of the patterns of
ages across the southern Atlantic is shown in
Fig. 3.6, which illustrates the matching of both Pre-
Fig.3.4 Possible alternative reconstructions of cambrian cratons and rocks of Palaeozoic age (Hal-
Gondwana (after Powell et 1980, with permission lam, 1975).
from Elsevier). Note the different positions of Madagascar 3 Ipfneous provinces. Distinctive igneous rocks can be
and India. traced between continents as shown in Fig. 3.7. This
50 CHAPTERS
Mountain belts
Younger
formed earl ier
than 260 m llion
years ago
^ ( Ural ^
/ Mountains
Greenland
^Europe]
Caledonian Mountains
Appalachian Mountains
tP a
■HpF South Pole Tethys Ocean
440 Ma ^>
South ^ /T
America 1) Africa ^ Andia ■
Australia,
Antarctica
s
[yl Mesozoic dolerite
[Hm Cratons Limit of Permo - Carboniferous glaclation
1: :. j Younger mobile belts • Precambrian anorthosite
Fig. 3.6 Correlation of cratons and younger mobile belts Fig. 3.7 Correlation of Permo-Carboniferous glacial
across the closed southern Atlantic Ocean (redrawn from deposits, Mesozoic dolerites and Precambrian anorthosites
Hallam, 1975). between the reconstructed continents of Gondwana.
CONTINENTAL DRIFT 51
O flj
j*: O
< B
(D "O c 2 0) 2
O)
<
^
— £
to 3 -C
o ^
CO <
O
X <
C
a<
|l
5 < LU <
Jurassic
Middle-
Upper
Triassic
Lower
Triassic
Upper
Triassic
Permian
Carboniferous
Devonian
ro
■1000 m
1 Carbonates and reef deposits. These deposits are of palaeotemperatures based on oxygen isotope
restricted to warm water and occur within 30® ofthe measurements. In certain marine organisms which
equator at the present day where temperatures fall in secrete calcium carbonate under equilibrium condi
the narrow range 25-30 ®C. tions with the sea water, the isotopic composition of
2 Evaporites. Evaporites are formed under hot arid oxygen in the calcite or aragonite depends upon
conditions in regions where evaporation exceeds sea temperature (Hoefs, 1980). Consequently, by mak
water influx and/or precipitation, and are usually ing assumptions about the isotopic composition of
found in basins bordering a sea with limited or the sea water, measurement of the ratio
intermittent connection to the ocean proper (Sec provides an estimate of water temperature. This
tion 12.3.4). At the present day they do not form method is not directly applicable to terrestrial depos
near the equator, but rather in the arid subtropical its, as the composition of the meteoric water from
high pressure zones between about 10® and 40® which calcium carbonate is precipitated is highly
where the required conditions prevail, and it is be variable. However, information is available from
lieved that fossil evaporites formed in a similar latitu studies of isotopic compositions of oxygen and car
dinal range (Windley, 1984). bon in snail shells and in cave deposits of calcium
3 Red-beds. These include arkoses, sandstones, carbonate (Faure, 1986).
shales and conglomerates which contain haematite. The results of applying these palaeoclimatic tech
They form under oxidizing conditions where there is niques strongly indicate that continents have
an adequate supply ofiron. A hot climate is required changed their latitudinal position throughout geo
for the dehydration oflimonite into haematite,and at logical time. For example, during the Permian and
presentthey are restricted to latitudes ofless than 30 ®. Carboniferous the Gondwana continents were
4 Coal and oil. These deposits are formed by the experiencing an extensive glaciation (Martin, 1981)
accumulation and degradation of organic remains, and must have been situated near the south pole
which require a warm, humid climate (Section (Fig. 3.9). At the same time in Europe and eastern
12.3.3). Both are predominantly formed in latitudes USA,coal and extensive reef deposits were forming,
ofless than 30®. which subsequently gave way to hot deserts with
5 Phosphorites. At the present day phosphorites form evaporite deposits. The northern continents were
within 45 ® ofthe equator along the western margins thus experiencing a tropical climate in equatorial
of continents where upwellings of cold, nutrient- latitudes.
rich, deep water occur, or in arid zones at low
latitudes along east-west seaways.
6 Bauxite and laterite. These aluminium and iron
3.5 Palaeontological evidence for
oxides only form in a strongly oxidizing environ continental drift
ment. It is believed that they only originate under the Continental drift has affected the distribution of
conditions of tropical or subtropical weathering. ancient animals and plants (Briggs, 1987) by creat
7 Desert deposits. Care must be employed in using ing barriers to their dispersal (Hallam, 1972). An
any of these deposits because desert conditions can obvious example of this would be the growth of an
prevail in both warm and cold environments. How ocean between two fragments of a supercontinent
ever, the dune bedding of desert sandstones can be which prevented migration between them by terres
used to infer the ancient direction of the prevailing trial life-forms. The past distribution of tetrapods
winds. Comparison ofthese with the direction ofthe implies that there must have been easy communica
modern wind systems found at their present latitude tion between all parts of Gondwana and Laurasia.
can indicate if the continent has undergone any Remains of the early Permian reptile Mesosaurus are
rotation. found in Brazil and southern Africa. Although
8 Glacial deposits. Glaciers and icecaps, excluding adapted to swimming, it is believed that Mesosaurus
those oflimited size formed in mountain ranges, are was incapable of travelling large distances and could
limited to regions within about 30® of the poles at not have crossed the 5000 km of ocean now present
the present day. between these two localities.
In addition to these sedimentological indicators Oceans can also represent dispersal barriers to
of climate, it is possible to make direct estimates certain animals which are adapted to live in relatively
CONTINENTAL DRIFT 53
Fig. 3.9 Use of palaeoclimatic data to Areas of tropical coal forests at 300 Ma which
control and confirm continental m some 50 Ma later became vast hot deserts
reconstructions (redrawn from Tarling &
Tarling, 1971, in Continental Drifts with Areas of glaciation between 250 and 300 Ma which
permission from Unwin Hyman Ltd). with arrows indicating known directions of ice movement
shallow marine environments. The widespread dis (Hurley, 1968; Plumstead, 1973), which are be
persal of marine invertebrates can only occur in their lieved to be cold climate forms. At the same time a
larval stages when they form part of the plankton varied tropical flora existed in Laurasia (Fig. 3.10).
(Hallam, 1973b). For most species the larval stage is After fragmentation, however, the flora of the indi
too short-lived to exist for the duration of the cross vidual continents diversified and followed separate
ing of a large ocean. Consequently, ancient faunai paths of evolution.
province boundaries frequently correlate with su A less obvious form of dispersal barrier is climate,
tures, which represent the join lines between ancient as the latitudinal motions of continents can create
continents brought into juxtaposition by the con climatic conditions unsuitable for certain organisms.
sumption of an intervening ocean. The distribution Indeed, relative continental movements can modify
of Cambrian trilobites strongly suggests that in the pattern of ocean currents, mean annual temper
Lower Palaeozoic times there existed several conti ature, the nature of seasonal fluctuations and many
nents separated by major ocean basins. The similarity other factors (Valentine & Moores, 1972). Also,
between ammonite species now found in India, plate tectonic processes can give rise to changes in
Madagascar and Africa indicates that only shallow topography which modify the habitats available for
seas could have existed between these regions in colonization.
Jurassic times. The diversity ofspecies is also controlled by conti
Palaeobotany similarly reveals the pattern ofconti nental drift. Diversity increases towards the equator
nental fragmentation. Before break-up, all the Gond- so that the diversity at the equator is about ten times
wana continents supported, in Carboniferous times, that at the poles. Consequently, drifting in a north-
the distinctive Glossopteris and Gangamopteris floras south direction would be expected to control the
54 CHAPTERS
i ! i mMi:
\ : i-i
/
^ ,-r
/ ..
■ flora
TropicalwithLauraslan
many
LV\^
Vvw
Polar Gondwanan
flora with few species
Tethys marine
foramlnlfera
species and areas of eurydesma fauna
of identical reef-forming
corals followed later by
tropical coal forests
orders on separated continents: the antbears (Eden record of the past direction of the Earth's magnetic
tata) ofSouth America, the pangolins (Pholidota) of field. These minerals are all pamm(i£fnetic^ that is,
northeast Africa and southeast Asia, the aardvarks they contain atoms which possess an odd number of
(Tubulidentata) of central and southern Africa and electrons. Magnetic fields are generated by the spin
the spiny,anteaters(Monotremata) of Australia. and orbital motions of the electrons. In shells with
Continental suturing leads to the homogenization paired electrons, their magnetic fields essentially
offaunas by cross-migration (Hallam, 1972)and the cancel each other. The unpaired electrons present in
extinction ofany less well-adapted groups which face paramagnetic substances cause the atoms to act as
stronger competition. Conversely, continental split small magnets or dipoles.
ting leads to the isolation of faunas which then When a paramagnetic substance is placed in a weak
follow their own distinct evolutionary development. external magnetic field, such as the Earth's field, the
For example, marsupial mammals probably reached atomic dipoles rotate so as to become parallel to the
Australia from South America in the Upper Creta external field direction. This induced ma£fnetization is
ceous along an Antarctic migration route (Hallam, lost when the substance is removed from the field as
1981) before the Late Cretaceous marine transgres the dipoles return to their original orientations.
sion flooded west Antarctica and closed the route for Certain paramagnetic substances which contain a
the later evolving placental mammals. Sea floor large number of unpaired electrons are ferro
spreading then ensured the isolation of Australia magnetic. The magnetic structure ofthese substances
when the sea level dropped, and the marsupials tends to devolve into a number of mapfnetic domains.,
evolved unchallenged until the Late Tertiary when within which the atoms are coupled by the interac
the collision of Asia and New Guinea allowed the tion ofthe magnetic fields ofthe unpaired electrons.
colonization of placental mammals from Asia. This interaction is only possible at temperatures
below the Curie temperature., as above this temper
ature the energy level is such as to prohibit inter
3.6 Palaeomagnetism atomic magnetic bonding and the substance then
behaves in an ordinary paramagnetic manner.
3.6.1 Introduction
Within each domain the internal alignment of
The science of palaeomagnetism is concerned with linked atomic dipoles causes the domain to possess a
studies of the fossil magnetism that is retained in net magnetic direction. When placed in a magnetic
certain rocks. If this magnetism originated at the field the domains whose magnetic directions are in
time the rock was formed, measurement of its direc the same sense as the external field grow in size at the
tion can be used to determine the latitude at which expense ofdomains aligned in other directions. After
the rock was created. If this latitude differs from the removal from the external field a preferred direction
present latitude at which the rock is found, very resulting from the growth and shrinkage of the
strong evidence has been furnished that it has moved domains is retained so that the substance exhibits
over the surface of the Earth. Moreover, if it can be an overall magnetic directionality. This retained
shown that the pattern of movement differs from magnetization is known as permanent or remanent
that ofrocks ofthe same age on a different continent, mapfnetism.
relative movement must have occurred between
them. In this way, palaeomagnetic measurements
3.6.3 Natural remanent magnetization
demonstrated that continental drift has taken place,
and provided the first quantitative estimates of rela Rocks can acquire a natural remanent magnetization
tive continental movements. For fuller accounts of (NRM) in several ways. If the NRM forms at the
the palaeomagnetic method,see Tarling (1983) and same time as the rock it is referred to as primary; if
Piper(1987). acquired during the subsequent history ofthe rock it
is termed secondary.
The primary remanence ofigneous rocks is known
3.6.2 Rock magnetism
as thermoremanent magnetization (TRM). It is ac
Palaeomagnetic techniques make use ofthe phenom quired as the rock cools from its molten state to
enon that certain minerals are capable of retaining a below the Curie temperature, which is realized after
56 CHAPTER 3
solidification. At this stage its ferromagnetic minerals domains relax and acquire the external field direc
pick up a magnetism in the same sense as the geo tion.
magnetic field at that time which is retained during Some CRM may be acquired soon after formation,
its subsequent history. for example during diagenesis, or during a metamor-
The primary remanence in clastic sedimentary phic event of known age, and hence preserve useful
rocks is known as detrital remanent magnetization palaeomagnetic information.
(DRM). As the sedimentary particles settle through CRM, TRM and DRM tend to be 'hard', and
the water column, any ferromagnetic minerals pre remain stable over long periods of time, whereas
sent align in the direction of the geomagnetic field certain secondary components of NRM, notably
(Fig. 3.12). On reaching bottom the particles flatten VRMs,tend to be 'soft' and lost relatively easily. It is
out, and if of elongate form preserve the azimuth of thus possible to destroy the 'soft' components and
the geomagnetic field but not its inclination. After isolate the 'hard' components by the technique of
burial, when the sediment is in a wet slurry state, the magnetic cleaning. This involves monitoring the ori
magnetic particles realign with the geomagnetic field entation and strength of the magnetization ofa rock
as a result of microseismic activity, and this orienta sample as it is subjected either to an alternating field
tion is retained as the rock consolidates. of increasing intensity or to increasing temperature.
Secondary NRM is acquired during the subsequent Having isolated the primary remanent magnetiza
history of the rock according to various possible tion, its strength and direction are measured with
mechanisms. Chemical remanent ma£fnetization either a spinner magnetometer or superconducting
(CRM)is acquired when ferromagnetic minerals are magnetometer. The latter instrument is extremely
formed as a result of a chemical reaction, such as sensitive and capable of measuring NRM orienta
oxidation. When ofa sufficient size for the formation tions of rocks with a very low concentration of
of one or more domains the grains become magnet ferromagnetic minerals.
ized in the direction of the geomagnetic field at the
time of reaction. Isothermal remanent magnetization
3.6.4 The past and present
(IRM)occurs in rocks which have been subjected to
geomagnetic field
strong magnetic fields, as in the case of a lightning
strike. Viscous remanent magnetization (VRM) may The magnetic field of the Earth approximates the
arise when a rock remains in a relatively weak mag field that would be expected from a large bar magnet
netic field over a long period oftime as the magnetic embedded within it inclined at an angle ofabout 12®
Direction of
Magnetic .geomagnetic
particle Y field
Wm
Water
Unconsolidated
sediment
Consolidated
sediment
Fig.3.12 Development of detrital
remanent magnetization.
CONTINENTAL DRIFT 57
to the spin axis. The actual cause ofthe geomagnetic position has moved over the Earth's surface.
field is certainly not by such a magnetostatic process, It is important to recognize that remanent magnet
as the magnet would have to possess an unrealisti- ization directions cannot provide an estimate of
cally large magnetization and would lie in a region palaeolongitude, as the assumed dipole field is axi-
where the temperatures would be greatly in excess of symmetric. There is a consequent uncertainty in the
the Curie temperature. ancient location of any sampling site, which could
The geomagnetic field is believed to originate from have been situated anywhere along a small circle,
a dynamic process, involving the convective circula defined by the palaeolatitude, centred on the pole
tion of electrical charge in the fluid outer core, position.
known as magnetohydrodynamics. However, it is Ifa palaeomagnetic study provides a magnetic pole
convenient to retain the dipole model as simple position different from the present pole, it implies
calculations can then be made to predict the geo either that the magnetic pole has moved throughout
magnetic field at any point on the Earth. geological time, that is, the magnetic pole has wan
The geomagnetic field undergoes progressive dered relative to the rotational pole, or if the poles
changes with time, resulting from variations in the have remained stationary that the sampling site has
convective circulation pattern in the core, known as moved, that is, continental drift has occurred. It
secular variation. One manifestation ofthis phenom appears that wandering of the magnetic pole away
enon is that the direction of the magnetic field at a from the geographic pole is unlikely because the
particular geographic location rotates irregularly Coriolis force is thought to be crucial in determining
about the direction implied by an axial dipole model the axial symmetry ofthe dipole field. Consequently,
with a periodicity of a few thousand years. In a palaeomagnetic studies can be used to provide a
palaeomagnetic study the effects of secular variation quantitative measure ofcontinental drift.
can be removed by collecting samples from a site An early discovery of palaeomagnetic work was
which span a stratigraphic interval of many thou that in any one study about half of the samples
sands of years. Averaging the data from these speci analysed provided a primary remanent magnetiza
mens should then remove secular variation so that tion direction in a sense 180® diflferent from the
for the purposes of palaeomagnetic analysis the geo remainder. Although the possibility ofself-reversal of
magnetic field in the past may be considered to rock magnetism remains, it is believed to be a rare
originate from a dipole aligned along the Earth's axis phenomenon, and so these data are taken to reflect
of rotation. changes in the polarity ofthe geomagnetic field. The
Palaeomagnetic measurements provide the inten field can remain normal for perhaps a million years
sity, azimuth and inclination of the primary reman- and then, over an interval of a few thousand years,
ent magnetization, which reflect the geomagnetic the north magnetic pole becomes the south magnetic
parameters at the time and place at which the rock pole and a period of reversed polarity obtains. Polar
was formed. By assuming the axial geocentric dipole ity reversals are random, but obviously affect all
model for the geomagnetic field discussed above,the regions of the Earth synchronously so that, coupled
inclination I can be used to determine the palaeo- with radiometric or palaeontologic dating, it is poss
latitude (|) at which the rock formed according to ible to construct a polarity timescale. This subject
the relationship tan
| (|) = tan I. With a knowledge of will be considered further in Chapter 4.
the palaeolatitude and the azimuth of the primary
remanent magnetization, that is, the ancient north
3.6.5 Apparent polar wandering curves
direction, the apparent location of the palaeopole
can be computed. Such computations, combined Palaeomagnetic data can be displayed in two ways.
with age determinations ofthe samples by radiomet- One way is to image what is believed to be the true
ric or biostratigraphic methods, make possible the situation,that is, plot the continent in a succession of
calculation of the apparent location of the north positions according to the ages of the sampling sites
magnetic pole at a particular time for the continent (Fig. 3.13a). This form of display requires the as
from which the samples were collected. Palaeo sumption of the palaeolongitudes of the sites. The
magnetic analyses of samples of a wide age range other way is to regard the continent as remaining at
can then be used to trace how the apparent pole a fixed position and plot the apparent positions of
58 CHAPTER 3
'Triassic-Jurassic
Present
jf Ifc ;
r7«:S« L Permian
^ / lS*dSl i
V
Pennsvlvanian
Cambrian
Devonian
Fig.3.13 Two methods of displaying paiaeomagnetic Subsequent work has modified the detail ofthe
data:(a) assuming fixed magnetic poles and applying movements shown. Note that the south pole has been
latitudinal shifts to the continent;(b) assuming a fixed plotted (redrawn from Creer, 1965, with permission from
continent and plotting a polar wandering path. the Royal Society).
Rifting and
creation of
ocean
Sutured continent
Single continent
Fig. 3.15 Palaeomagnetic signature of plate divergence method of continental reconstruction whereby the
and convergence (redrawn from Piper, 1987, with pole and angle of rotation necessary to bring the
permission from Open University Press). APW paths of two continents together are deter
mined. Indeed, modern palaeomagnetic measure
sents the full Wilsonian cycle (Section 10.4) of the ments,summarized in Van der Voo(1990),confirm
opening and closure of an ocean basin between two the precision ofthe reconstruction ofthe continents
continents. Before rifting the two segments A and B around the Atlantic of Bullard et al. (1965). As
ofthe initial continent have similar APW paths. They
are unlikely to be identical as it is improbable that the
initial rift and final suture would coincide. After
rifting the two segments describe diverging APW
paths until the hairpin at time 8 signals a change in
direction of motion to one of convergence. After
suturing at time 12 the two segments follow a
common polar track. Figure 3.16 shows part of this
seqiience illustrated by the continents of Pangaea.
Gondwana and Laurasia converged during Silurian
times. From then until the Mesozoic they followed a
common APW path, indicating that they formed the
single supercontinent Pangaea. During the Meso
Antarctica
zoic,Pangaea split into the present continents, which
each followed a separate APW path.
if-S. Arrierica.
1
3.6.6 Continental reconstructions
based on paiaeomagnetism
America
APW paths ofNorth America and Eurasia are similar Fig. 3.16 APW path of the south pole during the
in shape, showing that after a suitable rotation these Phanerozoic (redrawn from McElhinny, 1973,
continents would have followed an identical track Ptilaeoma£fnetism and Plate Tectonics^ with permission
over the Earth's surface. This, then, represents a from Cambridge University Press).
60 CHAPTER 3
Africa
Australia
Australia
Africa
Africa
Africa
Australia
South America * Australia
Africa—
Fig.3.17 Gondwana reconstructions based on polar Reconstructions shown as insets are based on the studies
wandering curves. Africa is in its present position and the of different workers (redrawn from Tarling, 1983;
other APW curves have been rotated to be coincident with Palaeoma£fmtism—Principles and Applications in Geologyy
particular geometrical fits of Gondwana:(a)the fit of Geophysics and Archaeology^ with permission from
Africa and South America;(b)-(e) eastern Gondwana fits. Chapman & Hall, Ltd).
stated before, palaeomagnetic data provide no in is confirmed by geometrical arguments. The match
formation on palaeolongitude, and so the east-west ing of APW paths for the other Gondwana con
separation of the continents is indeterminate. How tinents, however, is more subjective, and four
ever, by making use of the criterion that the conti possibilities are shown in Fig. 3.17b-e. The history
nental margins should be as close together as possible of continental movements since Devonian times,
without overlapping, it is possible to make adequate based on palaeomagnetic data, has been summarized
reconstructions. More accurate reconstructions can by Irving (1979).
be made for Mesozoic and Tertiary movements by
making use of oceanic magnetic lineations (Section
4.6). However, for Palaeozoic and Precambrian Further reading
times, such palaeomagnetic methods represent the
Bullard,E.G.,Everett, J.E.&Smith,A.G.(1965).The fit of
only quantitative methods of reconstruction avail
the continents around the Atlantic. Phil. Trans. Roy. Soc.
able.
Lond. 258A,41-51.
Figure 3.17, from Tarling (1983), shows recon Dewey, J.F. (1972) Plate tectonics. Sci. Am. 226(5), 56-
structions of Gondwana accomplished by APW path 68.
matching. There is wide agreement on the fit be Frakes, L.A. (1979) Climates Throughout Geologic Time.
tween South America and Africa (Fig. 3.17a), which Elsevier, New York.
CONTINENTAL DRIFT 61
Hallam, A.(1973b) Provinciality, diversity and extinction Powell, C.McA., Johnson, B.D. & Veevers, J.J. (1980) A
of Mesozoic marine invertebrates in relation to plate revised fit of east and west Gondwanaland, Tectonophys.
movements. In Tarling, D.H. & Runcorn, S.K. (eds) 63,13-29.
Implications of Continental Drift to the Earth Sciences^ 1, Tarling, D.H.(1983)Palaeomapfnetism. Chapman & Hall,
pp.287-94. Academic Press, London. London.
Piper, J,D.A.(1987)Palaeomapfnetism and the Continental Tarling, D.H.& Tarling, M.P.(1971) Continental Drift: a
Crust. Open University Press, Milton Keynes. Study ofthe Earth^s Movin£[ Surface. Bell, London.
4/ Sea floor spreading
^Cape:
iim/fwimm aii JVIendocIno
Mendocino
Pioneer
Francisco
run parallel to the crests of the mid-ocean ridge determined by both the geometric form ofthe source
system (Chapter 6), and are symmetrical about the and the orientation of its magnetization vector.
ridge axes (Fig. 4.2). Oceanic layer 2 maintains a relatively constant depth
The source of these linear magnetic anomalies and thickness and any anomalies arising because of
cannot be oceanic layer I, >vhich is made up of rugged topography on the top of the layer would
non-magnetic sediments. They cannot originate at a attenuate too rapidly to account for the amplitude of
depth corresponding to layer 3 as sources solely the anomalies observed on the surface 3-7 km above
within this layer would be incapable of generating the sea bed. Consequently, the lineations must arise
the steep anomaly gradients.The source ofthe anom because adjacent blocks oflayer 2 are magnetized in
alies must therefore be, at least in part, in oceanic different directions. Figure 4.3 shows an interpreta
layer 2. This conclusion is consistent with the basal tion of magnetic anomalies observed over the Juan
tic composition of layer 2 determined by dredging de Fuca Ridge in the northeastern Pacific. Layer 2 has
and drilling (Section 2.4.6), since basalt is known to been divided into a series of blocks running parallel
contain a relatively high proportion of magnetic to the ridge crest which have been assigned magneti
minerals. The magnetic lineations therefore confirm zation vectors which are either in the direction ofthe
that layer 2is everywhere composed ofthis rock type. ambient geomagnetic field or in the reversed direc
If magnetic lineations are generated by a layer of tion. The interpretation shows that the observed
homogeneous composition, how do the magnetic anomalies are simulated by a model in which the
contrasts originate that are responsible for the juxta intensities of the magnetization vary, and that rela
position of large positive and negative magnetic tively high values of some 10 A m"^ are required to
anomalies.^ The shape of a magnetic anomaly is produce the necessary contrasts.
64 CHAPTER 4
• Calculated
Am ' 0
Fig. 4.3 Interpretation of a magnetic anomaly profile magnetic field it generates an electric current which
across the Juan de Fuca ridge, nonheastern Pacific Ocean, has a magnetic field associated with it. Thus, once
in terms of normal and reversed magnetizations of two- 'excited', the dynamo becomes self-perpetuating as
dimensional rectangular blocks of oceanic layer 2. The long as there is a primary energy source to maintain
arrow marks the ridge crest(redrawn from Bott, 1967,
the convection currents. The process is complex, and
with permission from the Royal Astronomical Society).
analytic solutions are only available for the very
simplest configurations, which cannot approach the
field observed at the surface, and subsiirface temper true configuration in the core. The field is thought
atures would be well in excess of the Curie point, to be maintained by convection in the outer core
even given that its dependence upon pressure is which is thermally or gravitationally driven (Jacobs,
largely unknown. The temporal variation of the 1975).
internally generated field would also be inexplicable Since mathematical solution ofthe magnetohydro
with such a model. dynamic dynamo is not yet possible, although nu
The geomagnetic field is believed to originate by merical modelling is now possible (Glatzmeier &
magnetohydrodynamic processes within the fluid Roberts, 1995),the geomagnetic field has been mod
(outer) part of the Earth's core, magnetohydro- elled in two different ways. One approach is to derive
dynamics being that branch of physics concerned a complete mathematical solution for a simple me
with the interaction offluid motions and electromag chanical analogue of the geomagnetic field. A widely
netic fields. Indeed,this process is also believed to be utilized analogue is a disc dynamo consisting of two
responsible for the magnetic fields of other planets rotating conducting discs interconnected by coils of
and certain stars. The process requires the celestial wire. Under certain conditions the electric currents
body to be rotating and to be partly or completely flowing in the wire change direction and reverse the
composed of a mobile fluid which is a good electrical magnetic field direction. Although this to some ex
conductor. The turbulent or convecting fluid consti tent mimics the behaviour ofthe geomagnetic field, a
tutes a dynamo, because if it moves in a pre-existing major drawback to the analogy is that this type of
66 CHAPTER 4
Reversed Normal
Fig.4.4 Possible mechanism for reversal of the controlled by topography or temperatures at the base
geomagnetic field via an axial quadrupole transition field. of the mantle. These latter effects are controlled in
Shaded areas on the left side ofeach sphere represent turn by convection in the mantle.
positive radial fields, lines on the right the magnetic field It is estimated that a geomagnetic reversal takes
lines (after Fuller, 1987, with permission from Nature about 5000 years, and is accompained by a progres
326,132. Copyright © 1987 Macmillan Magazines Ltd).
sive decrease in field strength. However, the mea
surement of changes in magnetization direction
dynamo depends for its operation on its asymmetry, within a suite of thin, rapidly cooled lava flows(Coe
while the core is radially symmetrical. et 1995) has produced results which can be
Another approach is to consider the magnetic interpreted to suggest that, at times, the geomag
reversals as a random, stochastic process, and to netic field may have changed direction at a rate of6®
attribute reversals to large convection cells in the per day over several days. The capture of a geomag
core. These cells develop, as a result of the Coriolis netic reversal by a rock sequence in this way is very
force produced by the Earth's rotation, into spirals rare, but if this result is true, it will have very
similar to atmospheric cyclones (Parker, 1969). The important ramifications for our models of magnetic
cyclones remove the radial symmetry from the field generation.
system as they rotate in opposite directions in the The periodicity of geomagnetic reversals is highly
northern and southern hemispheres. Links between variable. The number of reversals per million years
the cyclones and magnetic reversals assume that the averaged over 10 Ma intervals has not remained con
former are randomly distributed, and various work stant throughout geological time (Fig. 4.5). There
ers have suggested the critical conditions under has been a gradual increase in reversal frequency
which reversals occur. Thus, for example, reversals during the Cenozoic (Livermore et 1984; Sto-
might be immediately preceded by the random at thers, 1986). This increase started in the Late Creta
tainment ofa highly symmetrical configuration or by ceous, after the field had been ofdominantly normal
a concentration of spirals near the geographic poles. polarity for 35 Ma. Land-based investigations have
The most recent theories of the mechanisms of
geomagnetic reversals (Fuller, 1987) postulate that
the process begins in a localized region of the core
(Hoffman, 1979). Bloxham &Gubbins(1985) have
produced models of the magnetic field at the core-
mantle boundary for the period 1715-1980 by con
0
tinuing downward the observed field at the Earth's
surface. Gubbins (1987) has used such models to
1
o
suggest that the present decay of the geomagnetic
c
field is caused by the localized development of re 0
3
C7
versed magnetic flux in the southern hemisphere. He 0
Age(Ma)
0 4 3 2 1 0 1 2 3 4 '
I 1—pJ—, I I ■ ■ ' . '. '
1 Gilbert .Gauss Matu- Brunhesj Matu- [Gauss Gilbert
I reverse|normal Innrmai
I normal !j vama
yarna ^ normal reverse
"reverse ^reverse
Layer 2
10-60 mm Layer 3
-60 mm a
Moho
Oceanic
Lithosphere
Upper
mantle
Asthenosphere
Fig.4.6 Sea floor spreading and the generation of thus seen to be controlled by the thickness ofthermal
magnetic lineations by the Vine-Matthews hypothesis boundary layers either side ofthe Gutenberg discon
(redrawn from Bott, 1982a, with permission from Edward tinuity. The model explains well the correlation
Arnold Ltd).
between the hotspot activity and reversal frequency
noted by Vogt(1975).
revealed a similar prolonged period of reversed Larson & Olson (1991) have noted the inverse
polarity in Upper Carboniferous and Permian time. correlation between magnetic reversal frequency and
Stothers (1986) has suggested that there is a period mantle plume activity (Section 5.4), as measured by
icity of reversals of about 30 Ma superimposed on the volume production rates of plume products. The
the general trends of Fig. 4.5. There is also strong mechanism proposed involves the cooling of the
evidence that the ratio ofthe mean lengths ofnormal outermost core as plume material rises from the
and reversed polarity intervals has changed through core-mantle boundary, which results in more rapid
out geological time (McFadden & Merrill, 1984). convective overturn in the outer core to replace the
Cox(1975)attributed changes in the frequency of lost heat. Above a certain level,this increased convec
magnetic reversals to changes in the conditions at the tion might result in a magnetic reversal.
core-mantle boundary. Modern theories have ex
tended this idea and propose that there exists a
coupling between the core and mantle in such a way 4.5 The Vine-Matthews
that the magnetic reversal frequency can be linked to hypothesis
convection in the mantle and hence to the mechan It is perhaps surprising to note that magnetic maps of
ism underlying plate tectonics. Courtillot & Besse the oceans showing magnetic lineations (Section
(1987)have proposed a model in which the thickness 4.2) were available for several years before the true
of the boundary layers on either side of the core- significance of the lineations was realized. The hy
mantle boundary affect convective processes in both pothesis of Vine & Matthews (1963) was of elegant
the outer core and mantle. In the region above the simplicity and combined the notion of sea floor
core-mantle boundary, a thermal boundary layer spreading (Section 4.3) with the phenomenon of
releases heat pulses to the mantle at a rate dependent geomagnetic field reversals (Section 4.4).
on its thickness. These thermal pulses ascend along The Vine-Matthews hypothesis explains the
paths influenced by mantle convection cells, eventu formation of magnetic lineations in the following
ally to reach the lithosphere as mantle hotspots. way. New oceanic crust is created by the solidification
Below the core-mantle boundary cold thermal of magma injected and extruded at the crest of an
pulses are emitted, again at a rate proportional to its ocean ridge (Fig^4'!6). On further cooling, the tem
thickness. These thermals affect the pattern of circu perature passes through the Curie point below which
lation within the outer core and moderate the rever ferromagnetic behaviour becomes possible (Section
sal rate. The coupling between core and mantle is 3.6.2). The solidified magma then acquires a
68 CHAPTER 4
tude is shown in Fig. 4.8. In general, the amplitude quences,and confirmed that at least 11 geomagnetic
of magnetic anomalies decreases as the latitude de field reversals had occurred over the last 3.5 Ma.
creases and as the strike of the ridge progresses from Subsequent work on other cores extended the rever
east-west to north-south. The symmetry of the sal history back to 20 Ma(Opdyke et«/., 1974).
anomalies is most apparent for ridges at high mag Pitman & Heirtzler(1966) and Vine (1966) used
netic latitudes (e.g. greater than 64®, which is equi the radiometrically dated reversal timescale to com
valent to geographic latitudes greater than 45®), pute the magnetic profiles that would be expected
north-south trending ridges at all latitudes and east- close to the crestal regions of mid-ocean ridges. By
west trending ridges at the magnetic equator. varying the spreading rate it was possible to obtain
very close simulations of all observed anomaly se
quences (Fig. 4.10), and consequently to determine
4.6 Magnetostratigraphy the spreading rates. A modern compilation is shown
Once the geomagnetic reversal timescale has been in Table 4.1. Extensions of this work show that the
calibrated, oceanic magnetic anomalies may be used same sequence of magnetic anomalies, representing
to date oceanic lithosphere. The method has been nearly 100 Ma of spreading and reversals of the
progressively refined so that it is now possible to Earth's magnetic field, can be observed over many
deduce ages back to mid-Jurassic times with an accu ridge flanks (e.g. Fig. 4.11). Later work has shown
racy of a few million years. that similar linear magnetic anomalies are developed
The Vine-Matthews hypothesis explains the over oceanic crust dating back to the Jurassic. Al
sequence of magnetic anomalies away from ocean though there is no oceanic crust older than this,
ridges in terms of normal and reversed magnetiza palaeomagnetic investigations on land have shown
tions of the oceanic crust acquired during polarity that geomagnetic reversals have occurred at least
reversals ofthe geomagnetic field. Verification ofthe back to 2.1 Ga.
hypothesis was provided by the consistency of the That spreading rates have varied with time is appar
implied reversal sequence with that observed inde ent from an examination of magnetic profiles from
pendently on land. Cox et a>L (1967) had measured different oceans. Examples are given in Fig.4.12 in
the remanent magnetization of lavas from a series of which the spreading rate in the south Atlantic is
land sites. The lavas were dated by a newly refined assumed to be constant and the distances to various
potassium-argon method, which allowed the con magnetic anomalies from ridge crests in other oceans
struction of a reversal timescale back to 4.5 Ma. The are plotted against the distance to the same anomaly
timescale could not be extended to earlier ages, as the in the South Atlantic. Inflection points in the curves
errors involved in K-Ar dating become too large. for the other oceans indicate when the spreading
Similarly, polarity events of less than 50 000 years rates changed there if the implicit assumption that
duration could not be resolved. The timescale, to the spreading rate has remained constant in the
5 Ma before present as later refined by Cande & Kent South Atlantic is correct.
(1992), is given in Fig.4.9. In magnetostratigraphic However, spreading rates may have changed with
terminology, polarity chrons are defined with dura time in all oceans; hence, independent dating is
tions of 10^-10^ years. Chrons may be dominantly required at points of different age in order to trans
of reversed or normal polarity, or contain mixed form the distance from the ridge axis into a temporal
events. scale and, therefore, the anomalies into a geo
Further verification of the geomagnetic reversal magnetic reversal timescale. This calibration was first
timescale was provided by palaeomagnetic investiga performed using dates obtained from the oldest
tions ofdeep sea cores(Opdyke etal.^ 1966). Unlike sediments overlying layer 2 penetrated in Deep Sea
lava flows, these provide a continuous record, and Drilling Project (DSD?) holes, but a more modern
permit accurate stratigraphic dating from their technique, made possible by superconducting mag
microfauna. This method is most conveniently ap netometers, is the detection of magnetic reversals in
plied to cores obtained in high magnetic latitudes thick limestone sequences on land which can be
where the geomagnetic inclination is high, because dated from their fossil content.
the cores are taken vertically and are not oriented The first long-term geomagnetic timescale, which
azimuthally. Excellent correlation was found be covered the last 80 Ma, was constructed by Heirtzler
tween these results and those from the lava se et al. (1968). They made the assumption that
CHAPTER 4
BRUNHES
Jaramillo
MATUYAMA
Oiduvai
Reunion
GAUSS
Kaena
2An2
Mammoth
GILBERT
Cochiti
Nunivak
Sidufjall
Thvera
Fig. 4.9 Geomagnetic polarit)'
timescale for the Plio-Pleistocene
(after Cande & Kent, 1992).
Numerical chrons are based on the
numbered sequence of marine
magnetic anomalies.
SEA FLOOR SPREADING 71
0-
500 nT
500 nT
Model 15mma~^
Model 15 mm a ^
km 50 0 50
—I 1 i_^S.L km 50 — CT
-3.3 3 L
"-5 km -3.3
500 nT
^3 km
Profile reversed
Juan De Fuca Ridge
Models lOmma Profile reversed 46° N
km 50 '0 50 50 '0 50 p |_
500 nT 500 nT
^4 km
Model 29 mm a Model 44 mm a
km 100 100
S.L
3.3
80 70 60 50 40 30 20 10 0 Age (Ma)
Table 4.1 Spreading rates at mid-ocean ridges (^spreading ity were numbered from 1 to 32 with increasing
rate' is defined as the accretion rate per ridge flank). time.
The Heirtzler timescale was checked using in
Observed Predicted
formation collected during the DSDP (Section 4.7),
rate rate
Anomaly number
-70 14- S Pacific Ocean
Distance from
N. Pacific
ridge X S. Indian
Ocean
10-2(km) Ocean
^ -40 in S. Atlantic
6 8 10 12 14 16 20 22 24 26 28 30
Fig. 4.12 iJCop) Relationship between the distance to a Fig.4.13 {Bottom) Magnetic reversal timescale from mid-
given anomaly in the South Atlantic and the distance to the Jurassic times to present(redrawn from Harland et«/.,
same anomaly in the South Indian, North Pacific and South 1982,A Geolo£fical Time Scale^ with permission from
Pacific Oceans. Numbers on the right refer to magnetic Cambridge University Press).
anomaly numbers (redrawn from Heirtzler a/., 1968).
OMa- r 55 Ma ■ -113Ma
25
Q Pleistocene -115
26
Aptian
0
Pliocene O)
0 Palaeocene -60 MO
1
0
27
-120
a. 28 Barrcmian
.29 Ml
30
M3 ■ 125
31
0
D)
Miocene Maastrichtian
■70 0 Hauterivian M5 I
O
0 32 I <j M7
> 130
M9
M10
20 33 '75 M10N
Valanginian lynif ■135
w Campanian
•80
MI2I
-25
M13|
-140
o
Berriasian M14j
M15
30 § Santonian •85
Oligocene ■M16 . 145
Coniacian M17
Tithonian
Turonian -90 M18
0 -35
c
0 M19 • 150
O)
13
O £ M20
0
Cenomanian
iS
40
•95 Kimmeridgian [^21 I
0 0
Q.
CO 155
-i i—— |\/|22
-100 M23
45
Oxfordian [^241 160
Eocene
M25 I
-105"
M25A|
50 0
(j
Albian
o M27I
165
(0 n n '
Callovian M29
■110 -7
■o 170
■o
■113
74 CHAPTER 4
Fig.4.14 Relative positions of Europe and Africa with the third leg ofthe DSDP (later to become ODP,the
respect to North America illustrating their separation Ocean Drilling Program). A series of holes was
during the Mesozoic and Cenozoic. Ages of drilled in the South Atlantic along a traverse at right
reconstruction shown in millions of years (redrawn from angles to the Mid-Atlantic Ridge (Fig. 4.15a) The
Pitman & Talwani, 1972, with permission from the age of the oceanic crust would ideally have been
Geological Society of America).
determined by radiometric dating of the layer 2
basalts which were penetrated in each hole. However
the basalts were too weathered for this to be possible,
of plate configurations can be made for different and so their ages were determined, albeit slightly
times, and the whole history of ocean basins deter underestimated, by palaeontologic dating of the
mined (Scotese et al., 1988). Figure 4.14 shows this basal sediments of layer 1. In Fig. 4.1Sb oldest sedi
method applied to the Mesozoic and Cenozoic his ment age is plotted against distance from the ridge
tory of the North Atlantic, while a much more axis, and it is readily apparent that there is a remark
complex example ofthe evolution ofthe Greenland- able linear relationship, with crustal age increasing
Iceland-Scotland region is described by Nunns with distance from the ridge. The predicted ages
(1983). agree well with the reversal timescale, with a half
spreading rate in this region of 20 mm a"^.
These results provide independent confirmation of
4.7 Deep sea drilling results the predictions of the Vine-Matthews hypothesis,
The Vine-Matthews hypothesis and geomagnetic and indicate the reasonable accuracy of the reversal
reversal timescale were verified independently during timescale to ages of at least 80 Ma.
SEA FLOOR SPREADING 75
Further reading
Courtillot, V. & Besse, J. (1987) Magnetic field reversals,
polar wander and core-mantle coupling. Science 237,
1140-7.
Cox,A.& Hart, RB.(1986)Plate Tectonics. How It Works.
0 400 800 1200 1600 Blackwell Scientific Publications, Oxford.
Distance from ridge crest(km) Cox,A.,Dalrymple,G.B.& Doell,RR (1967)Reversals of
the Earth's magnetic field. Sci. Am. 216,44-54.
Fig.4.15 (a) Location map of drilling sites on Leg 3 of
Fuller, M.(1987) A mechanism for reversals? Nature 326,
the DSDP in the South Atlantic,(b) Relationship between
132-3.
greatest sediment age and distance from the Mid-Atlantic
Gubbins, D.(1987) Mechanism for geomagnetic polarity
Ridge crest (after Maxwell el al., 1970, Science 168,
reversals. Nature 326,167-9.
1047-59, with permission from the AAAS. Copyright ©
Jacobs, J.A. (1975) The Earth's Core. Academic Press,
1970 by the AAAS).
London.
Kearey,P.& Brooks, M.(1991) Introduction to Geophys
ical Exploration (2nd edn). Blackwell Scientific Publica
tions, Oxford.
4.8 Introduction to Pitman, W.C. Ill & Heirtzler, J.S.(1966) Magnetic anom
transform faults alies over the Pacific-Antarctic ridge. Science 154,1164-
71.
The theory of sea floor spreading proposes that
Stothers, RB.(1986) Periodicity of the Earth's magnetic
oceanic lithosphere is created at mid-ocean ridges reversals. Nature 322,444-6.
and is balanced by the complementary destruction of Vine, F.J. (1966) Spreading of the ocean floor: new evi
oceanic lithosphere at subduction zones. While this dence. Science 154,1405-15.
theory neatly explains the geometry of lithospheric Vine, F.J. 8c Matthews, D.H.(1963) Magnetic anomalies
behaviour in two dimensions, a problem arises when over oceanic ridges. Nature 199,947-9.
5 / Theframework of plate tectonics
Lithosphere Lithosphere
Asthenosphere
Mespsphere
Fig.5.1 Block diagram summarizing the principal where focal mechanism solutions indicate tensional
features of plate tectonics. Arrows on lithosphere events associated with plate accretion and strike-slip
represent relative motions. Arrows in asthenosphere may events where the ridges are offset by transform faults.
represent complementary flow in the mantle (redrawn On land, shallow tensional events are also associated
from Isacks et 196S^ Journal of Geophysical Research with the Basin and Range Province of the western
73, with permission from the American Geophysical
USA,the East African rift system and the Lake Baikal
Union).
graben system. The first two examples appear to
originate where an ocean ridge penetrates beneath a
period of observation, the relatively rapid motions
continent, and its effects may extend beneath the
experienced by plates generate very large numbers of
continent beyond the rift system (Wilson, 1988).
earthquakes over a short interval oftime.
Earthquakes are classified according to their focal
depths: 0-70 km shallow focus, 70-300 km inter Fig.5.2 World-wide distribution of epicentres of large
mediate focus, greater than 300 km deep focus. magnitude earthquakes(mb>4)for the period 1961-67
An important belt of shallow focus earthquakes (after Barazangi & Dorman, 1969, with permission from
follows the crest ofthe ocean ridge system (Fig. 5.2), the Seismological Society of America).
0 20 40 60 100 120 140 160 180 -160 -140 -120 -100 -80 -60 -40 -20 0
78 CHAPTER 5
The majority of intermediate and deep events are determined by constructing great circles at right
associated with destructive plate margins. The north angles to the trends to transform faults affecting their
ern, eastern and western Pacific Ocean is ringed by a common margin and noting their point of intersec
belt of earthquakes which lie on planes occasionally tion. The most convenient type of plate margin to
offset by transform faults dipping at an angle of which to apply this technique is the accretive type
about 45" beneath the neighbouring plates. These (Fig. 5.3), as ocean ridges are frequently oflfset later
planes of earthquake foci, known as Benioff (or ally by transform faults (Section 7.2). Because of
Benioff-Wadati) zones, are typically associated with inaccuracies involved in mapping oceanic fracture
volcanic activity at the surface. The deepest events zones, the great circles rarely intersect at a single
recorded lie at a depth of about 680 km. Collisional point. Consequently, statistical methods are applied
mountain belts such as the Alpine-Himalayan chain which are able to predict a circle within which it is
are similarly characterized by intermediate and deep most probable that the relative rotation pole lies.
focus earthquakes although,since there is no longer a A second method is based on the variation of
Benioff zone present in such regions, the seismic spreading rate with angular distance from the pole of
activity occurs within a relatively broad belt. Careful rotation. Spreading rates are determined from mag
examination of epicentre locations has revealed, netic lineations (Section 4.6) by identifying anoma
however, that some of the shallow events lie on lies ofthe same age (usually number 3 or less so that
arcuate transcurrent fault zones(Section 9.3.6) asso the movement represents an instantaneous rotation)
ciated with the orogenic deformation. on either side of an ocean ridge and measuring the
The intraplate areas are relatively aseismic on this distance between them. The velocity of spreading is
timescale, although occasionally large magnitude at a maximum at the equator corresponding to the
earthquakes do occur. Although insignificant in their Euler pole and thence decreases according to the
release of seismic energy, intraplate earthquakes are
important as they can indicate the nature and direc
tion ofstress within plates (Section 11.7).
Euler pole
5.3 Relative plate motions
The motion of plates over the Earth's surface can be
described by making use ofEuler's theorem (Section
3.2.1), which says that the relative motion between
two plates is uniquely defined by an angular separa
tion about a pole of relative motion known as an
Euler pole. The pole and its antipole are the two
unique points on the surface of the Earth which do
not move relative to either of the two plates. An
important aspect of relative plate motion is that the
pole of any two plates tends to remain fixed relative
to them for long periods oftime. Plate velocities are
similarly constant for periods ofseveral million years
(Wilson, 1993).
There are three methods by which the pole of
relative motion for two plates can be determined.
The first, and most accurate, is based on the fact that
for true tangential motion to occur during the rela
tive movement of two plates, the transform faults
along their common boundary must follow the
traces of small circles centred upon the pole of Fig.5.3 Determination ofthe Euler pole for a spreading
relative motion(McKenzie & Parker, 1967;Morgan, ridge from its offsetting transform faults which describe
1968). The pole ofrotation oftwo plates can thus be small circles with respect to the pole.
THE FRAMEWORK OF PLATE TECTONICS 79
■-23 J 60°
\\ N y ^
EURASIAN rx- . >72 i\NORTH AMERICAN
y Nt
y •r^25 I
\ Yellowstone/ .•■-—i
^57
, rCARIBBEAN'S.ae
AFRICAN^
H \ ^ ✓''V
^78 PACIFIC "* l^^^cocbs) '"■••. —
^ •' T" 'I ■■■••••■•. - 0°
_^X5r'^^-rr66 south)
y2 "~
"/Reunion 80 -t ^;l n
MacDonaid /y NAZCA
94-— y AMERICAN \-
INDIAN f
r~'2 -^0
'•^es ^
X--J85 ^
•.... 106 ,y
J
X., ^^^40
. _ono
30
; ;■ J, gg I ^ Martin ;
/ Vaz
■•"i Kerguelen
.../.Tl
' ■•-.166 /
/
ANTARCTIC: ANTARCTIC^
Fig. 5.5 Directions and rates (in mm a") of relative (Section 10.4). Subduction at the Aleutian Trench
motion of the Earth's major plates shown at selected becomes more and more oblique westwards. The San
points on their boundaries. Circled numbers represent Andreas Fault represents a conservative boundary at
absolute motions (in mm a"') of plates relative to the which there is relative dextral movement between
hotspot reference frame (redrawn from Bott, 1982a, with
the North American and Pacific plates. The most
permission from Edward Arnold, based on data in Chase,
1978).
recent analysis of global plate motions predicts that
the rate of relative motion is 48 mm a~^ (DeMets et
al., 1990). This is significantly higher than the
35 mm a"^ suggested from geological investigations.
Alpine-Himalayan mountain chain. In these areas It strongly suggests that a significant fraction of the
relative motions also occur by internal deformation, movement between these two plates is accommod
extensional and compressional, respectively, of the ated elsewhere, notably within the Basin and Range
plates. In the Himalaya, data are consistent with a Province and a shear zone in the Mojave Desert.
convergence between India and Asia in a north- Plates containing parts of the Atlantic and Indian
south direction at a rate of approximately 50 mm a~^. oceans appear to be growing, in contrast to the
The convergence between Africa and Europe at the shrinking Pacific, Nazca and Cocos plates. A corol
Mediterranean-Alpine boundary is in the same sense lary of this is that the Mid-Atlantic Ridge and Oarls-
but somewhat slower at 10 mm a^^. berg Ridge of the western Indian Ocean must be
The greatest rates of subduction are found at the moving apart. This has important implications for
margins of the Pacific Ocean where the values plotted the nature of the driving mechanism of plate tec
fall in the range 65-106 mm a~^. The rate of accre tonics discussed in Chapter II. Not all ocean ndges
tion of oceanic lithosphere to each plate at the East spread in a direction perpendicular to the strike of
Pacific Rise, however, is typically lower than this at their magnetic lineations. It may be significant that
33-93 mm a"^, so that the ocean is shrinking in this the major obliquities of this type are found in the
area. In 300 Ma the Pacific would close completely, more slowly spreading seas, in particular the North
although this is unlikely to happen because events Atlantic, Gulf of Aden, Red Sea and southwestern
elsewhere might cause a reversal of this situation Indian Ocean.
THE FRAMEWORK OF PLATE TECTONICS 81
The analyses of Chase and Minster & Jordan Moreover, many of these island chains appear to be
imply that there is a slow compressionai movement mutually parallel. Where the volcanic centres are
between North and South America about a pole to closely spaced, aseismic ridges are constructed, such
the east of the West Indies at approximately 22°N. as the Ninety-East Ridge in the Indian Ocean and
Some of this movement is probably taken up in the the Greenland-Scotland Ridge in the North Atlantic
compressionai ridge/trench regime of the southern and the Rio Grande and Walvis ridges in the South
Caribbean Plate (Kellog & Bonini, 1982). To the Atlantic. These island chains and ridges are associ
east of the Caribbean, however, there is no obvious ated with broad crustal swells which currently occupy
seismic manifestation ofthe plate boundary between about 10% ofthe surface ofthe Earth, making them a
the Americas, and it has been suggested that defor major cause of epeirogeny (Crough, 1979).
mation takes place over a relatively wide area so that The island chains are invariably younger than the
it is aseismic. ocean crust on which they stand. The lower parts of
The analyses are consistent with extension across these volcanic edifices are believed to be formed
the East African.rift system at the very slow rate of predominantly of tholeiitic basalt, while the upper
about 6 mm a"^. The significance of this slow exten parts are alkali basalts (Karl et al., 1988) enriched in
sion is discussed further in Section 10.3. Na and K and, compared to mid-ocean ridge basalts,
have higher concentrations of Fe, Ti, Ba,Zr and rare
earth elements (REE) (Bonatti et al., 1977). Their
5.4 Hotspotsand composition is compatible with the mixing of juve
absolute plate motions nile mantle material and depleted asthenosphere
The major part of the Earth's volcanic activity takes (Schilling 1976)(Section 6.7). They are under
place at plate margins. However, a significant frac lain by a thickened crust but thinned lithosphere, and
tion occurs within the interiors of plates. In oceans represent a type of anomalous feature that will even
the intraplate volcanic activity gives rise to linear tually become welded to a continental margin as a
island and seamount chains such as the Hawaiian- suspect terrain (Section 9.5).
Emperor and Marquesas in the Pacific (Fig. 5.6). An example of an oceanic island chain is the
North America
%■
I ^
0 o
o d.
o <? :
T'jLti/o*
•-V °
O'
:
Jan Mayen
Iceland
Bowie ^
Ye bwstone
Azores
Cobb
Bermuda
Canary
Hawaii New
Sooerro England
Cape*
Galapagos
Verde I. Caroline
Marquesas Fernando
Samoa Ascension Comores
Pitcairn
San Felix
Society St Helena
Easter Reunion
Martin Vas Lord
Austral Juan •
Howe
Fernandez
Tristan
Marion
AustraliaA ^ yf
Discovery Kerguelen Tasman ^
Crozet
Louisviile
Shona Bouvet
Balleny
Fig.5.7 World-wide distribution of hotspots (after was proposed by Wilson(1963).It was suggested that
Duncan & Richards, 1991). the islands formed as the lithosphere passed over a
hotspot. The hotspots are now believed to originate
from mantle plumes rising from the lower mantle
Hawaiian-Emperor chain in the north-central Pacific which thin the overlying lithosphere(Olson & Nam,
Ocean (Fig. 5.6). This chain is some 4000 km long 1986). The volcanic rocks are then derived from
and shows a trend from active volcanoes at Hawaii in pressure-release melting and differentiation within
the southeast to extinct, subsided guyots in the the plume. Such plumes represent material of low
northwest. Dating of the various parts of the chain seismic velocity and can be detected by seismic tomo
(Jackson etcd.^ 1972)confirms this trend, and shows graphy (Section 2.1.8; Nataf & VanDecar, 1993).
that the volcanic activity has progressed from north Although the mantle plume mechanism has been
west to southeast at a rate of about 99 mm a"^ widely adopted,some workers (e.g., Turcotte & Ox-
(McDougall & Duncan, 1980), although Jackson burgh, 1978; Pilger, 1982) have questioned the ne
(1976) disputes the conclusion that voicanism has cessity for mantle hotspots and suggest that magmas
moved regularly along the chain. The chain conse simply flow to the surface from the asthenosphere
quently cuts across the older Cretaceous structural through fractures in the lithosphere resulting from
patterns of the sea floor. The change in direction of intraplate extensional stresses. This mechanism obvi
the chain correlates with a major change in spreading ates the problem of maintaining a mantle heat source
direction at about 43 Ma. The Hawaiian-Emperor for long periods. It does not, however, explain why
chain runs parallel to other chains on the Pacific fractures in the same plate should trend in the same
Plate, along which voicanism has progressed at a direction and develop at similar rates(Condie,1982).
similar rate. A useful summary of knowledge about hotspots is
A possible explanation ofthe origin ofisland chains provided by Duncan & Richards(1991).
THE FRAMEWORK OF PLATE TECTONICS 83
Morgan (1971, 1972a) proposed that mantle on the plate can then be computed and compared
plumes remain stationary with respect to each other with the actual tracks. This type of calculation indi
and the lower mantle,and are oflong duration. Ifso, cates that there have been no significant relative
the hotspots represent a fixed frame of reference by movements during the last 10 Ma, when instantan
which absolute motions of plates can be determined. eous motions are probably representative.
The most recent estimate of the present number of More rigorously, the consistency ofhotspot trajec
hotspots is about 40 (Stothers, 1993)(Fig. 5.7). tories on several plates can be tested by reconstructing
Morgan (1981) has used this hypothesis to de the relative positions of the major plates at various
scribe plate motions over the mantle for the past times in the past and locating one ofthe plates over its
200 Ma (Fig. 5.8). According to this analysis the hotspots. If the hotspot positions are indeed fixed,
hotspots appear to be fixed with respect to each other they should then all underlie the chains for which they
or to move at relative velocities of only a few milli are responsible. This test requires precise knowledge
metres per year. Significantly, many of the tracks, of relative plate positions in the past, and has been
such as that of the Madeira hotspot between Green applied with increasing accuracy as models of plate
land and Labrador, and of the St Helena hotspot motion have become more refined. Thus,Burke etal.
between the south coast of West Africa and the north (1973)and Molnar&Atwater(1973)concluded that
coast ofBrazil, appear to be the sites offuture rifting. there has been at least 15 mm a~^ relative movement
It is thus possible that hotspots weaken the litho- between hotspots over the last 120 Ma,and Molnar&
sphere so that at future dates rifting will take place Stock(1987) have determined interhotspot motions
along their trajectories. of up to 20 mm a"L
Predicted hotspot traces in the Indian Ocean The most recent analyses, summarized in Duncan
(Duncan & Richards, 1991) are shown in Fig. 5.9, & Richards (1991), suggest that the maximum
superimposed on volcanic structures on the sea floor possible relative motion between hotspots has been
and on land. The correlation between the two is only 2-5 mm a"^ over the last 120 Ma. This is con
excellent. For example, the Reunion hotspot began, siderably slower than typical plate velocities, so that
and was responsible for the Deccan Traps flood the concept of a hotspot reference frame is not
basalts of western India; India's northwards motion seriously degraded.
was then recorded by a north-south ridge in the A different approach to absolute plate motions has
western Indian Ocean and the Mascarene Plateau. been applied by Kaula (1975). He defined an abso
The hotspot is currently beneath a seamount 150 km lute reference frame by minimizing the lateral
west ofthe volcanically active island of Reunion. motions of plate boundaries, that is, by assuming
Continental plates also pass over hotspots al that plate margins move as little as possible with
though the effects are not as pronounced as on respect to the deep interior of the Earth. His model
oceanic plates. They give rise to domal uplifts with agrees quite well with that based on hotspots and
associated alkaline igneous activity, and may be re provides the following result: oceanic plates move at
sponsible for areas of high heat flow and geothermal about 50mma~^; continental plates move at about
activity. Crough et al, (1980) have proposed that 15 mm a" oceanic-oceanic plate boundaries move
kimberlite intrusions are formed by hotspots, and at about 20 mm a" and continent-oceanic plate
Fig. 5.10 shows the relationship of the Trindade boundaries move at about 15mma"^. Kaula con
hotspot trajectory to the occurrence of kimberlites cludes that the forces exerted by the lithosphere are
and alluvial diamond deposits in Brazil. significant in controlling the locations of accretive
The notion of fixed hotspots is important because and destructive plate margins.
plate movements have destroyed all other fixed co
ordinate systems, with the exception of the geo
magnetic field. The proposition can be tested in 5.5 True polar wander
several ways. For a single plate, the pole of rotation In Section 3.6 it was demonstrated that palaeo-
can be determined from a single hotspot and its rate magnetic techniques can be used to construct appar
of rotation found using the measured rate of migra ent polar wandering paths which track the motions
tion of volcanicity along the chain (McDougall & of plates with respect to the magnetic north pole and
Duncan, 1980). The trajectories of other hotspots hence, using an axial geocentric dipole model, the
84 CHAPTER 5
Iceland
""^abrador
0 Eifel
Bermuda
Madeira
Meteor q Ahagga
Canary , Tibesti
0
Verde
•■A
Jebel Marra
o Fernando
Cameroon
Guyana h Ascension
^ St Helena ) Comoros
0
Trindade
Tristan ®
0 Vema
« Discovery
Bouvet
1 <$
i
Fig. 5.8 Reconstruction of plates around the Atlantic reference (redrawn from Morgan, 1983, with permission
from 180 Ma to present in a fixed hotspot frame of from Elsevier).
THE FRAMEWORK OF PLATE TECTONICS 85
1 1
&
86 CHAPTERS
Indian Plate
Arabian
Plate
African
Plate
Comores
r Chagos /j'j \
\ 1^1-62 Ninety-East
^M^scarene 58~^ /
2 °
Reunion /' \
Madagascar
Broken Ridge
Ridge
Australian Plate
Kergueien
Kergueien
Plateau
Antarctic Plate
Fig.5.9 Hotspot tracks in the Indian Ocean. Solid circles several workers, including Livermore et al. (1984),
show present locations of hotspots. Numbers refer to Andrews(1985) and Courtillot & Besse (1987).
radiometrically dated locations in millions of years. Dark The method employed to investigate TPW is as
stipple marks flood basalt provinces (redrawn from follows. Palaeomagnetic pole positions for the past
Duncan & Richards, 1991).
200 Ma are complied for a number of continents
which are separated by spreading oceans so that their
spin axis ofthe Earth. In Section 5.4 it was suggested relative motions can be reconstructed from magnetic
that hotspots are nearly stationary in the mantle, and lineation data (Section 4.6). The pole positions are
so their trajectories provide a record of the motions then corrected for the rotations relative to a single
of plates with respect to the mantle. A combination continent (usually Africa) experienced as a result of
of these two methods can be used to test if there has sea floor spreading since the time for which they
been any relative movement between the mantle and apply. In this way a composite or global apparent
the Earth's spin axis. This phenomenon is known as polar wander path is obtained. This is then compared
true polar wander (TPW), and has been studied by with the track of an axis of the hotspot reference
THE FRAMEWORK OF PLATE TECTONICS
1202v,122
/ ♦ *80
Trindade
(Fig. 5.12e,f), and also, possibly, since Permo- for the majority of TPW. It has also been suggested
Carboniferous times. They noted that times ofslow that TPW is excited by the mass redistributions
ing reversal rate are periods ofslow polar wander and associated with subduction zones (Spada et «/.,
vice versa, and have suggested that the correlation 1992), mountain building and erosion (Vermeersen
between TPW and reversal frequency changes may & Vlaar, 1993).
indicate a coupling between the mantle and the core. TPW clearly has important implications in the
They also noted that the major change in direction of study ofmantle inhomogeneities, mantle convection
the TPW path at about 30-40 Ma appears to corre and palaeoclimatology. The differences between
late with the collision of Africa, Arabia and India TPW paths derived by different authors (Fig. 5.11)
with Eurasia, and so there may also be coupling will probably diminish with the future availability of
between the mantle and the lithosphere. This in accurate pole positions and age determinations.
formation has led Courtillot & Besse to propose a
model in which mantle and core dynamics are linked
via thermal exchange across the core-mantle 5.6 Superplumes
boundary. The hotspots described in Section 5.4 are the surface
The relative motion between the rotation axis and manifestation of plumes of hot material ascending
the mantle as illustrated by the TPW path may be from the deep mantle. These are of moderate size and
interpreted as a shifting of the whole or part of the can be considered to form part ofthe normal mantle
Earth in response to some form of internal mass convecting system. It has been proposed, however,
redistribution which causes a change in the direction that at least once during the history ofthe Earth there
about which the moment ofinertia ofthe mantle is a has been an episode of much more intense volcanic
maximum (Andrews, 1985). For example,Anderson activity. The cause has been ascribed to a phenomena
(1982) relates TPW to the development of geoid termed superplumes^ large streams of overheated
highs by the insulation from heat loss of mantle material rising buoyantly from the D" layer at the
niaterial beneath supercontinents, which cause a shift base of the mantle (Section 2.8.5) which derived its
of the rotation axis. It is possible that only the heat from the core. These spread laterally at the base
lithosphere or the mantle or both lithosphere and of the lithosphere to affect an area ten times larger
mantle together shift during polar wander. It is than more normal plume activity.
highly unlikely that the lithosphere and mantle are Larson (1991a; 1991b; 1995) has proposed that
sufficiently decoupled to move independently, superplumes were responsible for the widespread
and so it appears probable that shifting oflithosphere volcanic and intrusive igneous activity that affected
and mantle as a single unit takes place during TPW. abnormally large amounts of ocean floor during the
Indeed,ifthere is coupling between core and mantle, mid-Cretaceous. One manifestation of this activity
the whole Earth may be affected. Andrews's inter was the creation of numerous chains of seamounts
pretation ofTPW is supported by astronomical data and ocean plateaux (Section 9.5) in the western
which shows that during the 20th century the loca Pacific (Fig. 5.6) at a rate some five times greater
tion ofthe Earth's rotational axis has moved at a rate during this period than at other times. Similarly
similar to that computed from palaeomagnetic data, there were extrusions ofthick, areally extensive flood
namely about l®Ma"^. This suggests that at least basalts on the continents, such as the Parana Basalts
part of the mass redistribution takes place in the of Brazil.
mantle, as the continents do not move this rapidly. Phenomena attributed to the mid-Cretaceous
Sabadini & Yuen (1989) have shown that both superplume episode are illustrated in Fig. 5.12. At
viscosity and chemical stratification in the mantle are 120-125 Ma the rate of formation of oceanic crust
important in determining the rate of polar wander. doubled over a period of 5 Ma,decreased within the
Another mechanism proposed for driving TPW is the next 40-50 Ma and returned to previous levels about
surface mass redistribution arising from major glaci- 80 Ma ago (Fig. 5.12c). The additional production
ations and deglaciations (Sabadini et al.^ 1982). ofcrust required increased subduction rates, and it is
However, mantle flow is required to explain TPW significant that the Andes ofSouth America and the
during periods with no evidence of significant con Sierra Nevada mountains ofthe western USA formed
tinental glaciation, and, indeed, may be responsible at this time by the subduction process (Section 9.2).
THE FRAMEWORK OF PLATE TECTONICS
Superplume
High-latitude surface '
temperature
Sea level
Oil resources
Oceanic crust
30 - production
Reversal rate
TRW velocity
Coupled to the increased crust production, and slabs. The episode would thus be viewed as a period
caused by the consequent general rise in the level of when mantle ascended passively as a result ofchang
the sea floor, was a world-wide increase in sea level to ing plate motions.
an elevation some 250 m higher than at the present Large mantle plumes are also believed to be re
day (Fig. 5.12a). The surface temperature of the sponsible for the formation of the volcanic margins
Earth also increased by about 10®C, as shown by of certain continents. They are described in Section
oxygen isotope measurements made on benthic fora- 10.3.6.
minifera from the North Pacific. This effect was
probably caused by the release of large amounts of
carbon dioxide during the volcanic eruptions, which 5.7 Direct measurement of
created an enhanced 'greenhouse' effect. During the relative plate motions
superplume episode the rates of carbon and carbon It is now possible to monitor the relative motion
ate sequestration in organisms increased due to the between plates by periodic resurveying of baselines
shallowing ofthe seas and the increased temperature, extending across plate boundaries (Stein, 1987).
which caused plankton to thrive. This is reflected in Before about 1980 the only methods available for
the presence of extensive black shale deposits at this this type ofinvestigation were the standard terrestrial
time (Force, 1984) and in the estimated oil reserves geodetic methods of baseline measurement using
of this period (Tissot, 1979; Fig. 5.12c), which may optical techniques or laser ranging instruments such
constitute about 50% of the world's supply. Also of as the geodolite (Thatcher, 1979). These methods
economic significance is the placement of a large are certainly sufficiently precise to measure relative
percentage of the world's diamond supply at this plate motions of a few tens of millimetres a year.
time, probably as a result of the diamonds' having However,in some regions the strain between certain
been translated to the surface by the rising plumes. plates is not all dissipated across the boundary, and
During the plume episode the rate of geomagnetic may extend into the adjacent plates for distances of
reversals (Section 4.4) was very low (Fig. 5.12e), up to about 100 km. In order to study these large
with the field remaining in normal polarity for some scale problems it is necessary to be able to measure
40 Ma. This indicates that activity in the core, where very large distances to very great accuracy. Terrestrial
the geomagnetic field originates (Section 3.6.4), was methods are extremely time consuming on land, and
low, perhaps related to the transfer of considerable impossible to use across major oceans. During the
quantities of heat to the mantle. last decade, however, the measurement of very long
The correlation ofTPW with the rate of magnetic baselines using extraterrestrial methods has become
reversals was considered in Section 5.5. Courtillot possible via the application ofspace technology.
& Besse (1987) correlated the period of very slow Three independent methods of extraterrestrial
TPW from 170 to 110 Ma with continental fragmen surveying are available. These are very long baseline
tation and the decrease in reversal rate during the interferometry, satellite laser ranging and satellite
Late Jurassic and Early Cretaceous (Fig. 5.12f). radio positioning.
Shortly after the start of the mid-Cretaceous mag The technique ofvery long baseline interferometry
netic quiet zone,TPW velocity increased suddenly to (VLBI) makes use of the radio signals from extra-
close to its present, relatively high rate. Courtillot & galactic radio sources or quasers (Niell et«/., 1979;
Besse (1987) related the slow TPW and decreasing Carter & Robertson, 1986; Clark et aL, 1987) as
reversal rate to reduced convective activity in the shown in Fig. 5.13. The signal from a particular
mantle. quasar is recorded simultaneously by two or more
Acceptance of a mid-Cretaceous superplume epi radio telescopes at the ends ofbaselines which may be
sode is not universal. For example, Anderson (1994) up to 10 000 km long. Because of their different
suggests that the phenomena of this period were locations on the Earth's surface, the signals received
caused by a general reorganization of plates on a at the telescopes are delayed by different times, the
global scale associated with the break-up of Pangaea magnitude of the delays between two stations being
and reorganization of the Pacific plate. The mantle proportional to the distance between them and the
upwelling in the latter may then have been a passive direction from which the signals are coming. Typi
reaction to plates^being pulled apart by their attached cally, during a 24-hour experiment, 10-15 quasars
THE FRAMEWORK OF PLATE TECTONICS 91
Extragalactic
random radio Difference in
signals time of arrival
antenna
Baseline
Portable
antenna
Calibrations Magnetic j
tape 1
I
Simultaneous
First order
vector baseline
Fig.5.13 Schematic outline of geodetic
solution
point
a VLBI system (redrawn from {General purpose Delays
positioning Cross-correlation
Niell etai^ 1979, with computer)
signal detection
permission from Elsevier). (minicomputer)
are each observed 5-15 times. This scheme provides three-dimensional method by which the relative
estimates of baseline length which are accurate to positions of instruments at the ends of baselines are
about 20 mm (Lyzenga et 1986). The usefulness determined from the signals received at instruments
of this system has been greatly enhanced by the from several satellites. The simultaneous observation
development of portable radiotelescopes which frees of multiple satellites makes extremely accurate mea
the technique from the necessity ofusing fixed obser surements possible with small portable receivers,
vatory installations. with relative positions determined to about 1 part
The technique of satellite laser ranging (SLR) per million ofthe baseline length. This appears to be
calculates the distance to an orbiting artificial satellite the most efficient method of establishing geodetic
(e.g. LAGEOS,the Laser Geodynamics Satellite) or a control on both local and regional surveys, and it is
reflector on the Moon by measuring the two-way probable that an accuracy of O.lp.p.m. will be
travel time of a pulse of laser light reflected from the achieved. The full constellation of 21 GPS satellites
satellite (Cohen & Smith, 1985). The travel time is has now been implemented.
subsequently converted to range using the speed of Stein (1987) has summarized the determinations
light. If two laser systems at different sites simultan of relative plate motions determined by these meth
eously track the same satellite, the relative location of ods. Generally, plate velocities averaged over a few
the sites can be computed by using a dynamic model years ofobservation agree remarkably well with those
of satellite motion, and repeated measurements pro averaged over millions of years (Gordon & Stein,
vide an accuracy ofabout 80 mm.Periodic repetition 1992). The methods were first applied to the mea
of the observations can then be used to observe surement of the rate of movement across the San
relative plate motions (Christodoulidis ,1985). Andreas Fault in California. Smith et al, (1985),
The accuracy of the measurements is determined by using SLR, have reported that a 900 km baseline
the knowledge of the satellite orbit, which in turn which crosses the Fault at an angle of 25 ® has been
depends on the forces acting upon the satellite. Of shortened at an average rate of 30 mm a"^. Lyzenga
these the Earth's gravitational field is the most im et aL (1986) have used VLBI to measure the length
portant, and improved knowledge of this field will of several baselines in the southwestern USA and
lead to improved accuracy in SLR. have found that over a period of4 years movement
The technique of satellite radiopositioning makes on the Fault has been 25±4mma~^. If the SLR
use ofradio interferometry from the Global Position results are not in error, it is probable that the differ
ing System (GPS) satellites (Dixon, 1991). It is a ence in rates provided by the two methods will
92 CHAPTER 5
provide constraints on the regional distribution of These techniques of direct measurement are
tectonic strain in this area. clearly extremely important in that they provide
Direct measurements of the rate of displacement estimates of relative plate movements that are inde
across the San Andreas Fault are lower than the pendent of plate tectonic models. It is probable that
48 mm a"^ predicted from global models of plate their accuracy will continue,to improve, and that
movements(DeMets 1987). However,during observations will become more widely distributed
the last half decade of observation, no major earth over the globe. The determination of intraplate de
quakes have occurred. Over longer time intervals, the formation and its relationship to intraplate stress
discrete jumps in fault movement associated with fields, earthquakes and magmatic activity should also
the elastic rebound mechanism of large earthquakes become possible. Important new findings are antici
(Section 2.1.5) would contribute to the total dis pated over the next few decades.
placement and provide a somewhat higher figure for
the average rate of movement. In addition, motion
between the Pacific and North American plates may 5.8 Instantaneous and
be occurring along other major faults which run finite rotations
parallel to the San Andreas Fault (Section 7.4). The motions of the plates described in Section 5.3
Tapley et al. (1985) have measured changes in are termed instantaneous as they refer to movements
length of four baselines by SLR between Australia taking place at a single instant in time or averaged
and the North American and Pacific plates, and over a very short period of geological time. Instant
have found that the rates differ by no more than aneous rotations cannot,therefore, provide informa
3mma"^ from average rates over the last 2 Ma. tion on the paths followed by the plates in arriving at
Similarly Christodoulidis et al (1985) and Carter & the point at which the instantaneous motion is mea
Robertson(1986)have measured the relative motion sured. Although it is a basic tenet of plate tectonics
between pairs of plates and have found a strong that poles of rotation remain fixed for long periods
correlation with the kinematic plate model of of time, consideration of the relationships between
Minster & Jordan (1978)(Section 5.3). plates forming an interlinked spherical shell reveals
Herrings?/>/.(1986)have reported VLBl measure that this cannot be the case for all plates (McKenzie
ments between various telescopes in the USA and & Morgan, 1969).
Europe. They have determined that the present rate Consider the three plates on a sphere A, B and C
of movement across the Atlantic Ocean is 19± shown in Fig. 5.14a. ^bc ^ac represent
10 mm a"^. The uncertainty in this estimate will un Euler poles for pairs of plates which describe their
doubtedly decrease in the future as the model ofthe instantaneous angular rotation. Let plate A be fixed.
atmosphere is improved and multiple telescope obser Clearly the poles and P^c can remain fixed
vations are made,but even so it agrees with the rate of with respect to the relevant pairs of plates. Thus,for
23 mm a~^ averaged over the past 1 Ma. example, any transform faults developing along
BC
BC BCi
BC^
I Fig.5.14 The three plate problem.
Pac5 Pbc and Pba refer to
instantaneous Euler poles between
plates A and C,B and C,B and A
BA ^ respectively, and a'-c, b'^c and b«a
to their relative rotation vectors. In
(b)P'bc is the present location of
Pbc- See text for explanation.
THE FRAMEWORK OF PLATE TECTONICS 93
Pi' N
Fig.5.15 (a) Rotation of
plates A and B about pole Pi
\ 33 it 30 /
I-*- A
'T"!
produces arcuate fracture A 11 /
zones with a radius of FZ
Ifz^M
curvature of 30°;(b) a jump
of the pole of rotation to P2
causes the fracture zones to
assume a radius of curvature
of90°. P'l represents the
positions of pole Pi after
rotation about pole P2(after
Cox & Hart, 1986, with
permission from Blackwell
Scientific Publications).
common plate margins would follow small circles small circles centred on a pole at 79°N,111°E. Ifthe
centred on the poles. Consider now the relative fracture zone patterns are analysed in more detail,
movements between plates B and C. It is apparent however,it can be seen that the fracture zones in fact
that if A,Pba and P^c remained fixed, the rotation consist of five different segments with significantly
vector ofC relative to B (b"c) acts through P^c and is different orientations which can be correlated be
given by the sum ofthe vectors and which act tween adjacent fracture zones. The apparent gross
about Pba and Pac-> respectively. Thus, Pgc lies small circle form of the fractures only represents the
within the plane ofPgA and Pac and is fixed relative third phase of movement. It is thus apparent that in
to A.Such a point, however, does not remain station the northeastern Pacific sea floor spreading has taken
ary with respect to B and C. Consequently, relative place about a pole of rotation that was continually
motion between B and C must take place about a changing position by small discrete jumps. This pro
pole that constantly changes position relative to B gression has been analysed and illustrated in greater
and C (Fig. 5.14b). Transform faults developed on detail by Engebretson ft a/. (1985).
the B-C boundary will not then follow simple small Changes in the direction of relative motions of
circle routes. plates do not cause large scale deformation of the
Even when a moving pole is not a geometrical plate boundaries but rather result in geometric ad
necessity, it is not uncommon for Euler poles to justments of transform faults and ocean ridge crests.
jump to a new location (Cox & Hart, 1986). In This may be a consequence ofthe lithosphere's being
Fig. 5.15 the pole of rotation of plates A and B was thin at accretive margins and consequently ofsmaller
initially at Pj,and gave rise to a transform fault with mechanical strength (Le Pichon et al., 1973). That
a small circle of radius 30°. The new pole location is the adjustments are only minor, however, is appreci
Pj,60° to the north of Pj, so that the transform fault ated from continental reconstructions such as shown
is now 90° from P2, that is, on the equator of this in Fig. 5.17, where the earthquake foci associated
pole. The occurrence of this pole jump is easily with present day activity are superimposed on the
recognizable from the abrupt change in curvature of predrift reconstruction. The coincidence ofshape of
the transform fault. the initial rift and modern plate margins indicates
Menard 8c Atwater (1968) have recognized five that there has been little post-drift modification of
different phases of spreading in the northeastern the latter.
Pacific. In Fig. 5.16 it is shown that the numerous The past relative positions of plates can be deter
large fracture zones of this region appear to lie on mined by the fitting of lineaments which are known
94 CHAPTERS
^
■
Mendocino —:==^
^ — ' tV \
—V. c\af'°!L-
••"■^ h
r.alap^gos .1 /■ -"
— '^^M^qpesas C rest of Fig. 5.16 Fracture zones in the
J East Pacific Rise northeastern Pacific showing
20° L I — ' 1 1- r / trends corresponding to five
160° 140° 120° ' 1?Qo possible spreading episodes, each
®0° with a new pole of rotation
(redrawn from Menard &
Atwater, 1968, with permission
from Nature 219, 463-7.
Copyright © 1968 Macmillan
Magazines Ltd).
to have been originally juxtaposed. One approach is Their trace, however, normally approximately
to fit former plate margins. Fossil accretive margins follows a small circle route, with any deviations
are usually readily identified from their symmetrical from this marked by characteristic tectonic activity
magnetic lineations (Section 4.2), and fossil trans (Section 7.4). Ancient destructive margins can be
form faults from the offsets they cause of the linea recognized from their linear belts of calc-alkaline
tions. Ancient transform faults on continents are magmatism, granitic batholiths, paired metamorphic
more difficult to identify, as their direction may be belts, and, possibly, ophiolite bodies.
largely controlled by the pre-existing crustal geology. The features most commonly used for determin-
ing earlier continental configurations are continental progression of finite rotations between the plates
margins and oceanic magnetic anomalies. The which led to their present position. Thus, although
former are obviously used to study the form of any relative movement between two plates can be
pre-drift supercontinents (Section 3.2.2). Because described in terms of an angular rotation about a
magnetic anomalies can be reliably dated (Section pole, the plates do not necessarily follow the route
4.6), and individual anomalies identified on either predicted by a single rotation, and may have followed
side oftheir parental spreading ridge,the locus ofany a much more complex path. Indeed, many long
particular anomaly represents an isochron. Fitting oceanic fracture zones appear to fit more closely a
together pairs of isochrons then allows reconstruc route consisting of a series of small circles about
tions to be made of plates at any time during the poles which suffered small shifts in position during
history of their drift. their history. Each change in pole position corre
Finally, reconstructions can be accomplished by sponds to a change in the direction of plate motion.
the fitting of polar wandering curves (Section 3.6.6),
because if two continents were once juxtaposed they
must have shared the same polar path. This tech 5.9 Triple junctions
nique is not particularly reliable, as polar wandering The stability of the boundaries between plates is
paths are rarely based on anything but a series of dependent upon their relative velocity vectors. If a
discrete measurements, each of which is subject to a boundary is unstable it will exist only instantan
wide range of uncertainty. Consequently, for the eously and will immediately devolve into a stable
most recent 200 Ma of Earth history this method is configuration.
only used as an independent test of the continental Figure 5.18a shows an unstable boundary between
drift suggested by other methods. Before 200 Ma, two plates where plate X is underthrusting plate Y at
this is the only technique available. be in a northeasterly direction and plate Y is under-
If the relative positions of plates are well con thrusting plate X at ab in a southwesterly direction.
strained during drift by oceanic magnetic anomalies The boundary is unstable because a trench can only
and fracture zones,it may be possible to compute the consume in one direction, so to accommodate these
I I I I hi- Trench
=Ridge
— Transform fault
(a) (b)
Fig.5.24 (a) Ridge (R)-trench (T)-transform fault(F) The FFF triple junction is never stable, as the velocity
triple junction between plates A,B and C.(b)Its lines coincide with the vector triangle, and, of
representation in velocity space. As the velocity lines ab, course, the sides of a triangle never meet in a single
be and ac do not intersect at a single point, the triple point. The other possible triple junctions are only
junction must be unstable.
stable for certain particular orientations ofthe juxta
posed plate margins.
Fig. 5.25a is unstable, as the velocity lines represent Only six types oftriple junction are present during
ing the trenches do not intersect at a single point the current phase of plate tectonics. These are RRR
(Fig. 5.25b). In time the system evolves into a stable (e.g., the junction ofEast Pacific Rise and Galapagos
configuration (Fig. 5.25c) in which the new triple Rift 2k)ne), TTT (central Japan), TTF (junction of
junction moves northwards along trench AB. The Peru-Chile Trench and West Chile Rise), FFR(pos
dashed lines show where plates B and C would have sibly at the junction of Owen Fracture Zone and
been if they had not been subducted. The point X Carlsberg ridge), FFT(junction ofSan Andreas Fault
(Fig. 5.25a,c) undergoes an abrupt change in relative and Mendocino Fracture Zone) and RTF(mouth of
motion ais the triple junction passes. This apparent Gulfof California).
change in underthrusting direction can be distin The evolution of the San Andreas Fault illustrates
guished from a global change as it occurs at different the importance of the role of triple junctions. In
times and locations along the plate boundary. In Oligocene times (Fig. 5.27a), the East Pacific Rise
order to be stable, the plate configuration shown in separated the Pacific and Farallon plates. The trans
Fig. 5.25a must be as in Fig. 5.25d. When plotted in form faults associated with this ridge have been
velocity space (Fig. 5.25e) the velocity lines then simplified, and only the Mendocino and Murray
intersect at a single point. fracture zones are shown. The Farallon Plate was
McKenzie & Morgan (1969) have determined the being underthrust beneath the North American
geometry and stability of the 16 possible combina Plate, and, since the rate of consumption exceeded
tions oftrench,ridge and transform fault(Fig. 5.26), the rate ofspreading at the East Pacific Rise,the ridge
taking into account the two possible polarities of system moved towards the trench. The first point of
trenches, but not transform faults. Ofthese, only the the ridge to meet the trench was the eastern extrem
RRR triple junction is stable for any orientation of ity of the Mendocino Fracture Zone. A quadruple
the ridges. This comes about because the associated junction existed momentarily at about 28 Ma, but
velocity lines are the perpendicular bisectors of the this devolved immediately into two triple junctions
triangle ofvelocity vectors, and these always intersect (Fig. 5.27b). The more northerly was of FFT type,
at a single point (the circumcentre of the triangle). the more southerly ofRTF type,and both were stable
THE FRAMEWORK OF PLATE TECTONICS 99
bc~._ C
(b)
ab
bC^^C
A""-
aC
(d) (e)
Fig.5.25 (a) Triple junction between three trenches consumed are shown as dashed lines,(d)Stable
separating plates A,B and C.(b)Its representation in configuration ofa trench-trench-trench triple junction,
velocity space, illustrating its instability,(c) The positions (e) Its representation in velocity space,((a) and (c)
plates B and C would have reached ifthey had not been redrawn from McKenzie & Morgan, 1969).
(insets on Fig. 5.27b). Because of the geometry of formed in response to the migration of these triple
the system the northern triple junction moved north junctions. The southerly migration of the southern
along the trench and the southern triple junction triple junction ceased as the eastern extremity of
moved south. Thus the dextral San Andreas Fault the Murray Fracture Zone reached the trench
Type Geometry Velocity triangle Stability Examples Type Geometry Velocity triangle Stability Examples
O
O
Stable if the
All orientations East Pacific o
angles between X
stable Rise and
ab and ac, be >
Galapagos Rift TTR(c) A IB respectively,
Zone -O
are equal,or if
Great Magnetic ac,be form a
Bight straight line X
C lab oi
Stable if ab, ac Stable if ac, be Intersection of
TTT(a) Central form a straight the Peru-Chile
form a straight Japan
line, or if be is line, or if C lies Trench and the
parallel to the on ab Chile Ridge
slip vector OA TTF(a)
Stable if the
complicated Stable if be, ab
TTT(b) general condition form a straight
for ab, bo and line, or if ac
ac to meet at goes through B
TTF(b)
a point is satisfied
Stable if ab, ac
Unstable form a straight
FFF line, or if ab,
TTF(c)
be do so
B
ab must go
RRT through circumcentre
of ABC Stable if C lies Owen Fracture
kC-^^ac on ab,or if Zone and the
FFR A B
ac, be form a Carisberg Ridge
straight line West Chile
Ridge and the
ac! lab Unstable, East Pacific Rise
A I B
RRF Aj-^- evolves to FFR
rr^bj Stable if ab, be San Andreas
1 1 FFT
form a straight Fault and Mendocino
line, or if ac, be Fracture Zone
be I
^"ac do so
Stable if ab
TTR(a) goes through C,
or if ac, be form
Stable if ab goes Mouth of the
>k .S: a straight line
A : B RTF(a) A B ab
lau through C, or Gulf of California
ab if ac. Deform a
straight line
A I
be
be Stable if complicated
general conditions
ar. \ are satisifed -.C
A B Stable if ac, ab
^"A ab B RTF{b)
cross on be
Fig.5.26 Geometry and stability of all possible triple junctions (redrawn from McKenzie & Morgan, 1969).
THE FRAMEWORK OF PLATE TECTONICS 101
illllii A
M Mendocino iiiiiSj!
iiliy
Pacific Plate North America Plate
(a)
(Fig. 5.27c). The triple junction changed to FFT type Gordon,RG.&Stein,S.(1992)Global tectonics and space
and began to move northwards. The Farallon Plate geodesy. Science 256,333-42.
continued to be subducted to the north and south of Isacks, B., Oliver, J. & Sykes, L.R (1968) Seismology
and the new global tectonics./. Res. 73, 5855-
the San Andreas Fault, until the geometry changed
99.
back to that shown in Fig. 5.27b when the East
Larson, RL.(1995) The mid-Cretaceous superplume epi
Pacific Rise to the south ofthe Murray Fracture Zone sode. Sci. Am.,Feb.,66-70.
reached the trench. The triple junction then reverted McKenzie,D.P.& Morgan,W.J.(1969)Evolution oftriple
to RTF type and changed to a southerly motion junctions. Nature 224,125-33.
along the trench. This represents the situation at the McKenzie, D.P. & Parker, RL.(1967) The North Pacific:
present day at the mouth of the Gulfof California. an example oftectonics on a sphere. Nature 216,1276-
80.
Minster, J.B. & Jordan, T.H. (1978) Present-day plate
Further reading motions.J.^eophys. Res. 83,5331-54.
Bott, M.H.P.(1982) The Interior ofthe Earth, its Structure, Molnar,P.& Stock,J.(1987)Relative motions of hotspots
Constitution and Evolution (2nd edn). Edward Arnold, in the Pacific, Atlantic and Indian oceans since late
London. Cretaceous time. Nature 327, 587-91.
Burke, K. & Dewey, J.F. (1973) Plume generated triple Morgan, W.J. (1968) Rises, trenches, great faults and
junctions: key indicators in applying plate tectonics to crustal blocks.J.pfeophys. Res. 73,1959-82.
old rocks./. Geol. 81,406-33. Morgan, W.J. (1971) Convection plumes in the lower
Chase, C.G. (1978) Plate kinematics: the Americas, East mantle. Nature 230,42-3.
Africa, and the rest of the world. Earth planet. Sci. Lett. Morgan, W.J.(1972) Deep mantle convection plumes and
37, 355-68. plate motions. Bull. Am. Ass. Petrol. Geol. 56, 203-13.
Courtillot, V. & Besse, J. (1987) Magnetic field reversals, Morgan, W.J. (1981) Hot spot tracks and the opening
polar wander and core-mantle coupling. Science 237, of the Atlantic and Indian oceans. In Emiliani, C.(ed.)
1140-7. The Oceanic Lithosphere. The Sea 7, pp.443-87. Wiley,
Cox,A.& Hart,RB.(1986)Plate Tectonics. How It Works. London.
Biackwell Scientific Publications, Oxford. Olson,P. & Nam,l.S.(1986)Formation ofseafloor swells
Duncan, RA.& Richards, M.A.(1991) Hotspots, mantle by mantle plumes./.^eophys. Res. 91,7181-91.
plumes, flood basalts, and true polar wander. Bsv. Geo- Stein, RS.(1987) Contemporary plate motion and crustal
phys. 29, 31-50. deformation. Rev. Geophys. Space Phys. 25,855-63.
6/ Ocean ridges
103
104 CHAPTERS
FVVF
(EPR3°S)
vents
F ViV F
Intermediate
(ERR 21° N)
FVVF
Slow
VE'>'4x
Axis (MAR 37° N)
—I— —1—
20 10 0 10 20 30
km
Fig.6.1 High resolution bathymetric profiles ofocean region within 8 km ofthe ridge axis, and is asymmet
ridges at fast, intermediate and slow spreading rates. EPR, rical with the greater intensity on the eastern flank.
East Pacific Rise; MAR,Mid-Atlantic Ridge. Neovolcanic The half extension rate due to the faulting is
zone bracketed by Vs,zone offissuring by Fs, extent of 4.1 mm a"^, compared to 1.6 mm a"^ observed on
active faulting by Ps(redrawn with permission from the Mid-Atlantic Ridge in the FAMOUS area.
MacDonald, 1982,Annual Review ofEarth and Planetary
Sciences 10. Copyright © 1982 by Annual Reviews Inc.).
6.3 Broad structure of the
decreases in width and tectonic extensional features
lithosphere below ridges
become increasingly important until volcanic activity Gravity measurements have shown that free air
wanes completely and tectonic extension is the anomalies are broadly zero over ridges (Figs 6.4,
dominant process affecting the median valley. The 6.5), indicating that they are in a state of isostatic
applicability of the flow chart depends upon the equilibrium (Section 2.11.6), although small scale
small spreading cells beneath the axial region being topographic features are uncompensated and cause
stable in time, and it will be modified as more positive and negative free air anomalies. The small,
information becomes available. long wavelength, positive and negative free air anom
Bicknell et al. (1988) have reported on a detailed alies over the crests and flanks, respectively, ofridges
survey ofthe East Pacific Rise at 19®30'S. They have are a consequence of the compensation, with the
found that faulting is more prevalent than on slow positives being caused by the greater elevation ofthe
spreading ridges, and conclude that faulting ac ridge and the negatives from the compensating mass
counts for the vast majority of the relief. They ob deficiency. The gravitational effects ofthe compensa
served both inward and outward facing fault scarps tion dominate the gravity field away from the ridge
which give rise to a horst and graben topography. crest, and indicate that the compensation is deep.
This differs from slower spreading ridges, where the Seismic refraction experiments by Talwani et aL
topography is formed by back-tilted, inward-facing (1965)over the East Pacific Rise have shown that the
normal faults. Active faulting is confined to the crust is slightly thinner than encountered in the main
West Western
wall marginal Eastern Fig.6.2 Diagrammatic cross-
..... high Western Central Eastern marginal section ofthe inner rift valley of
500 m trough high trough high the Mid-Atlantic Ridge at
36°50'N in the FAMOUS area
(redrawn from Ballard & van
Andel, 1977, with permission
1 0 Tkm from the Geological Society of
Vertical and horizontal scale America).
OCEAN RIDGES 105
20 km
'
Greenstones 8(
* Hydrothermal vents '"Vl metagabbros
Heat flow
-500
3000-
Bougueranomaly
. 2000-
1000-'
500n
Free air anomaly
Fig.6.5 Gravity anomalies and
0 crustal structure defined by
-500-• seismic refraction across the
W ^Seismic line
0- J=t- '-'hh Mid-Atlantic Ridge at about
31°N. Bouguer anomaly
5- reduction density 2,60 Mg m"^,
Layer 2
^(6.5) 7.3 7 38.2 P wave velocities in km s"^
10- Layer 3 (redrawn from Talwani et al.
Moho l96Sy Journal ofGeophysical
Anomalous mantle
Mantle Research 70, with permission
15
from the American Geophysical
500 500 1000 km Union).
ocean basins, and that the upper mantle velocity mantle by a Pratt-type mechanism (Section 2.11.3).
beneath the crestal region is anomalously low Talwani et al.(1965) proposed that the anomalously
(Fig.6.4). Oceanic layer 1 rocks (Section 2.4.5) are low upper mantle velocities detected beneath ridges
only present within topographic depressions, but correspond to the tops ofregions oflow density. The
layers 2 and 3 appear to be continuous across the densities were determined by making use of the
ridge except for a narrow magma chamber at the Nafe-Drake relationship between P wave velocity
crest. A similar structure has been determined for and density (Nafe & Drake, 1963), and a series of
the Mid-Adantic Ridge (Fig. 6.5). The suggestion of models produced which satisfied both the seismic
this work that layer 3 is not continuous across the and gravity data. One of these is shown in Fig.6.6,
ridge has now been disproved (Whitmarsh, 1975; and indicates the presence beneath the ridge of a
Fowler, 1976). body with a density contrast of -0.25 Mgm~^ ex
As the crust does not thicken beneath ridges, tending to a depth ofsome 30 km. This large density
isostatic compensation must occur ^vithin the upper contrast is difficult to explain geologically, and it is
OCEAN RIDGES 107
Observed
Calculated
D 1
L-
0 3000-
3
>
O)
D CO ■
O E
GO o 2000-
c
CO
2.60 2.90
0-
Crust
Anomalous mantle
20-
a Mantle
0
Q
40-
Fig.6.6 Possible model ofthe structure beneath the Mid- ployed, and the anomalous body is considerably
Atlantic Ridge from gravity modelling with seismic larger, extending to a depth of 200 km.
refraction control. Densities in Mg m"^(redrawn from Given the ambiguity inherent in gravity model
Talwani et \96S^ Journal ofGeophysical Research 70, ling, the two interpretations shown probably
with permission from the American Geophysical Union). represent end members ofa suite ofpossible interpre
tations. Indeed, seismic tomography (Section 2.1.8)
probable that the Nafe-Drake relationship does not suggests the low velocity region beneath ocean ridges
apply to rocks which may contain significant frac extends to a depth of 100 km (Anderson et al,^
tions of partially molten material. An alternative 1992). They demonstrate without ambiguity, how
interpretation, constructed by Keen & Tramontini ever, that ridges are underlain by large, low density
(1970), is shown in Fig.6.7. A much lower, more bodies in the upper mantle whose upper surfaces
realistic density contrast of -0.04 Mgm~^ is em slope away from the ridge crests.
about 0.006 Mgm"^. Extended over a depth range 6.10). Evidence for the presence of such a magma
of 100 km this density contrast would support a chamber has been provided from detailed seismic
relative ridge elevation of0.25 km.It is very difficult surveys.
to assess the contribution ofthis source as the extent On the East Pacific Rise,a seismic refraction survey
of partial melting in the anomalous mantle is un using ocean bottom seismometers (Orcutt et al.^
known. However, theoretical relationships between 1976) detected a pronounced low velocity zone
seismic velocity and density for solid materials do not about 2km beneath the ridge axis, which is underlain
seem to apply to the anomalous mantle and so it is by upper mantle with the anomalously low velocity
possible that this discrepancy results from the pres of 7.7 km s"^. Away from the crest, the low velocity
ence ofa molten fraction. zone is absent beneath crust 2.9 Ma old but the
3 A temperature dependent phase change. The high anomalous upper mantle persists. By the time that
temperatures beneath ocean ridge crests might cause crust 5 Ma old is reached, normal oceanic layering
a transition to a mineralogy oflower density. Normal is encountered and normal mantle velocities are
upper mantle is made up of pyroxene pyrolite (Sec present.
tion 2.8.3), but at the high temperatures and low Herron et al. (1978; 1980), employing multi
pressures experienced within the top 40 km below a channel seismic techniques, detected reflections
ridge crest the lower density form, plagioclase pyro from the top of a magma chamber beneath the East
lite, would be more stable. This has a density up to Pacific Rise. This was 1.5-2.0 km beneath the ridge
0.07 Mgm"^ lower than pyroxene pyrolite, and, ex crest, and its width, varying from 2-8 km, appeared
tending over a depth range of20 km,would support to correlate with the varying width ofthe ridge crest
a ridge elevation of some 0.7 km. A problem with (Fig. 6.8). The sides ofthe magma chamber were not
this hypothesis is that the upper mantle in this area detected, being either too steep to provide reflections
has been depleted by removal ofthe molten fraction at near vertical incidence or too gradational to pro
so that plagioclase may not be able to develop in vide a reflecting horizon. This work has been ex
great quantities. tended by Detrick et al. (1987), who similarly
The major factor contributing to the uplift of
ridges is thus thermal expansion and contraction,
which can also account for the topography of the
ocean basins (Section 6.9). Partial melting and phase
changes may also contribute to a lesser, unknown
extent. Certainly, the model shown in Fig. 6.6 ap
Moho
pears unreasonable because too high a density con
trast is required, and the model shown in Fig. 6.7 is
preferred. However, the model of Keen & Tramon-
tini(1970)can also be criticized in that the densities vC-.^
detected the top of a low velocity magma chamber The most recent studies, summarized by Smith &
1.2-2.4 km deep and less than 4-6 km wide beneath Cann (1993), suggest that no steady state magma
the axis of the East Pacific Rise. Moreover, by run chamber exists beneath ocean ridges; rather, discrete
ning a reflection profile along the ridge axis they were pockets of magma rise to feed small volcanoes in the
able to demonstrate that the magma chamber is inner rift. Contrary to the melt-dominated magma
present as a continuous feature along at least 60% of chambers described in Section 6.3, Sinton & Detrick
the crestal region surveyed. Further work on the (1992) have proposed a model in which magma
records(Mutter 1988a) has demonstrated that chambers comprise narrow, hot, crystal-melt mush
the magma chamber is often displaced laterally from zones. Magma chambers are viewed as composite
the rise axis. Tomographic imaging (Section 2.1.8) structures comprising an outer transition zone made
ofthe crestal region ofthe East Pacific Rise.(Burnett up of a hot, crystal framework with only small
et «/., 1989) has shown that the actual magma amounts of interstitial melts and an inner zone of
chamber(> 50% melt)is quite small, being no more crystal mush with sufficient melt for it to behave as a
than a few hundred metres thick and 4km wide very viscous fluid. Pure melt only develops in fast
supplied by a narrow feeder. Surrounding the cham spreading ridges where sufficient magma is present to
ber is a zone with a velocity up to 1 km s~^ lower than form a thin melt lens on top of the mush zone
normal. This is interpreted as a reservoir of very hot (Fig.6.9a). This lens may be tens of kilometres long
rock containing little melt, whose low density iso- along the ridge crest, but only tens or hundreds of
statically supports the ridge topography. Kent et al metres in thickness. Slow spreading ridges have in
(1990)have further decreased the estimate ofthe size sufficient magma supplies for a melt lens to develop
of the magma chamber, suggesting a width of 1 km (Fig.6.9b) and eruptions only occur when there are
and a thickness of 100 m. Harding et al. (1989), periodic influxes of magma from the mantle. Such a
however, suggest that magma is present within a model is consistent with seismic reflection data from
zone at least 6 km wide. ocean ridges and petrological observations which
The crest of the Mid-Atlantic Ridge has been less require magma to have been modified by crystalliza
intensively surveyed and the evidence for the pres tion or mixing within the crust, which could not
ence ofan axial magma chamber is less clear. Francis occur in a large, well mixed chamber. It also explains
& Porter (1973), from the recording of local earth why less fractionation occurs in the volcanic rocks of
quakes with ocean bottom detectors, located a low slow spreading ridges. A problem with this model,
velocity region about 3 km beneath the median valley however, is that it is not apparent how the layered
which they interpreted as a magma chamber. Whit- gabbros oflayer 3 might develop.
marsh (1975), using seismic refraction methods, Many early investigations of ocean ridges were
similarly inferred the presence of a low velocity zone essentially two-dimensional in that they were based
2-3 km wide extending through the upper 2.5 km of on quite widely spaced profiles oriented perpendi
crust. Calvert(1995) has isolated seismic reflections cular to their strike. More recently 'swath'-mapping
from a survey at 23®17'N which have been inter systems have been employed which provide complete
preted as arising from the top ofa 4 km-wide magma aerial coverage of oceanic features. These systems
chamber at a depth of 1.2 km. However, Fowler have been used to reveal variations in the structure of
(1976)disputed the presence ofany sizeable magma ocean ridges along strike. A full review of these
chamber as S waves propagate across the axis,indicat developments is provided in MacDonald et al.
ing the absence ofa fluid phase. (1988).
These investigations of the East Pacific Rise and Studies ofthe East Pacific Rise have shown that it is
Mid-Atlantic Ridge appear to indicate the presence segmented along its strike by the occurrences of
of an axial magma chamber a few kilometres wide at non-transform ridge axis discontinuities such as
shallow depth. This width, however, is an order of propagating rifts (Section 6.11) and overlapping
magnitude less than implied from some ophiolite spreading centres(OSC) which occur at local depth
models (Section 2.5) in which magma chambers are maxima,and by smooth variations in the depth ofthe
inferred to be up to 30 km wide. The seismic data, ridge axis. These features may migrate up or down
moreover, cannot distinguish between the presence the ridge axis with time.
ofliquid magma and a partially molten crystal mush. OSCs (MacDonald & Fox, 1983) are non-rigid
110 CHAPTERS
^Volcanics c
Mush
Transition
(b) 2
Rift valley
QJ e Transition
Q 6
Distance (km)
Fig.6.9 Interpretive models of magma chambers beneath weak. The OSC geometry is obviously unstable, and
a fast(a) and slow (b) spreading ridge (after Sinton & its development has been deduced from the behav
Derrick, 1992). iour of slits in a solid wax film floating on molten
wax, which appears to represent a reasonable ana
discontinuities where the spreading centre of a ridge logue (Fig. 6.10a). Tension applied orthogonal to
is offset by a distance of 0.5-10 km, with the two the slits (spreading centres) causes their lateral prop
ridge portions overlapping each other by about three agation (Fig. 6.10b) until they overlap (Fig. 6.10c),
times the offset. It has been proposed that OSCs and the enclosed zone is subjected to shear and
originate on fast spreading ridges where lateral off rotational deformation. The OSCs continue to ad
sets are less than 15 km,and true transform faults fail vance until one tip links with the other OSC
to develop because the lithosphere is too thin and (Fig. 6.10d). A single spreading centre then develops
OCEAN RIDGES 111
2500-
il III* _ ^1 ...
I
(a)
^ 3000-
TJ
/ Propagation of spreading
Melt segregation events A - E need not
centres along strike
3500- be synchronous
,/ -50-100 km
shallow axial magma chamber. Magma chambers short-lived, as their effects can only be traced for a
usually occur where the ridge is broad and where an few kilometres in the spreading direction. Second
axial rift is present. It has been suggested that this order segmentations, however, create off-axis scars
correlation, and the lateral segmentation of ocean on the spreading crust consisting of cuspate ridges
ridges, arises from the variation in space and time of and elongate basins which cause differential relief of
the availability of magma beneath the ridge axis. several hundred metres. The scars do not follow
Increasing magma supplies give rise to shallow small circle routes about the speading pole, but form
magma chambers and a consequent broadening axial V-shaped wakes at 60-80® to the ridge. This indi
bulge on which an axial rift is developed. Decreasing cates that the OSCs responsible for the segmentation
magma supply causes the presence of only a small migrate along the ridge at velocities of up to several
magma chamber, or its absence, the axial region to hundred millimetres per year. Figure 6.12 summa
be narrow, and the absence of an axial rift because rizes the three general cases for the evolution of
the magma chamber is too small to give rise to a ridge-axis discontinuities in terms of the movement
significant collapse structure. of magma pulses.
Segmentation of ocean ridges thus appears to The residual structures of second order disconti
be controlled by the distribution of partial melts nuities leave a surprisingly wide wake 30-50 km
beneath them (Toomey et aL^ 1990; Gente et al,^ wide and disrupt magnetic lineations over a zone
1995), which feed magma chambers at discrete loca 50-80 km wide. It has been estimated that about
tions along them and create local depth anomalies. 20% ofthe sea floor has been affected by such wakes.
The ridge model of Sinton & Derrick (1992), de All ofthe ridge-axis discontinuities responsible for
scribed above, precludes extensive mixing within the segmentation have their origin in the upwelling and
small axial magma chamber along the ridge, and segregation of molten mantle material beneath the
could explain the observed geochemical segmenta ridge and the subsequent propagation of pulses of
tion. With time the magma may migrate away from magma along it. Differences between the types of
its sources,creating a gradual increase in depth ofthe discontinuity may arise because oftheir relative size,
axis as the pressure within it gradually wanes. This duration, structural evolution, frequency and
phenomenon may explain the non-coincidence of whether they are related to major changes in direc
magma chamber and rise culmination noted by tion ofthe spreading ridge.
Mutter et al. (1988a). The brittle shell overlying the The fine scale mapping of ocean ridges which is
magma stretches and cracks and magma intrudes so now possible with modern equipment has clearly
that eruptions follow the path of magma migration. allowed new ideas to evolve about the detailed behav
After eruption the removal of supporting magma iour of the ridge crests. The ideas concerning lateral
gives rise to the formation ofan axial summit graben. segmentation were derived on the East Pacific Rise,
Evidence for the pulse-like, episodic spreading of but Sempere et al. (1990) and Gente et aL (1995)
ridges,in which sea floor spreading occurs alternately have demonstrated that a similar segmentation is
by stretching and fracturing of existing crust and present on the Mid-Atlantic Ridge. There, the seg
then dyke injection and copious volcanicity has been mentation is recognized by variation in the ridge-axis
provided by studies ofophiolites(Harper,1978)and morphology. It will be interesting to see if similiar
seismic reflection profiles across old Atiantic litho- segmentation is present on currentiy unexplored
sphere (McCarthy et aL^ 1988). Ridge axis discon regions ofthe ocean ridge system.
tinuities consequently occur where the magmatic
pulses finally come to rest. The variable bathymetry
6.6 Heatflow and hydrothermall
and associated geophysical and geochemical differ
circulation
ences imply that adjacent segments of ridge have
distinct, different sources in the mantle. First to third On a large scale it has been found that there is a
order segmentation is caused by the variable depth relationship between the heat flow from oceanic
associated with magma migration, and fourth order lithosphere and the inverse square root of its age,
effects are caused by the geochemical differences in which leads to the relationship between water depth
magma supply. and age discussed in Section 6.9. On a fine scale,
Third and fourth order segmentations appear to be however, there are discrepancies between observed
OCEAN RIDGES 113
Small or no
offset or
depth anomaly.
no discordant zone
on flanks
and expected heat flows for young lithosphere. In outcrops. The penetration of this convection is not
particular, there is a large scatter in heat flow magni known, but it is possible that it is crust-wide.
tude near the crests of ocean ridges. Thermal lows Hydrothermal circulation ofsea water in the upper
tend to occur in flat-floored valleys and highs within crust beneath ocean ridges transports about 25% of
areas of rugged topography (Lister, 1980). Blanket the global heat loss, and is clearly a major factor in
ing by sediment does not appear to be the cause of the Earth's thermal budget.
the low heat flow as the troughs are the least sedi-
mented areas of the ridge and also the youngest and
6.7 Petrology of ocean ridges
therefore hottest. To explain these phenomena it
has been proposed that the pattern of heat flow is Under normal conditions the peridotite ofthe upper
controlled by the circulation ofsea water through the mantle does not melt. However,the high heat flow at
rocks ofthe oceanic crust. ocean ridges implies that the geothermal gradient
Although the penetration of water through the crosses the peridotite solidus at a depth of about
hard rock ofthe sea floor at first seems unlikely,it has 50 km (Wyllie, 1981, 1988), giving rise to the
been shown that thermal contraction can induce a parental magma of the oceanic crust (Fig. 2.32).
suflficient permeability for efficient convective flow to A similarly high geothermal gradient is believed to
exist. The cracks are predicted to advance rapidly and exist beneath oceanic islands as oceanic lithosphere
cool a large volume ofrock in a relatively short time, traverses a mantle plume or hotspot(Section 5.4),so
so that intense localized sources ofheat are produced generating basaltic rocks by a similar mechanism.
at the surface. Active geothermal systems which are Ocean ridge basalts have the composition of oliv-
driven by water coming into contact with near- ine tholeiite (Kay et «/., 1970), and exhibit only
molten material are expected to be short-lived, but minor compositional variation caused by variable
the relatively gentle circulation ofcool water, driven alumina and iron contents. They may contain phe-
by heat conducted from below, should persist for nocrysts of olivine or plagioclase or, rarely, clinopy-
some time. The chemistry of hydrothermal springs roxene (Nisbet & Fowler, 1978). Three alternative
on the East Pacific Rise and Mid-Atlantic Ridge is sources have been suggested for the parental magma
remarkably similar, in spite of the great diflference of olivine tholeiite: a small degree of partial melting
in spreading rates of these two accretive margins of mantle peridotite at 75-90 km followed by precip
(Campbell 1988). itation ofolivine as the melt ascends; 20-30% partial
Further evidence that hydrothermal circulation melting at depths in the range 30-60 km; shallow
occurs comes from the presence of metalliferous melting with separation ofthe melt at 15-25 km.
deposits in the crestal area. The metals are those The simplest interpretation of the chemistry of
known to be hydrothermally mobile, and must have oceanic basalts, suggested from experimental petrol
been leached from the oceanic crust by the ingress of ogy, is that separation of the partial melt occurs at a
sea water which permitted their extraction in a hot, depth of 15-25 km. However, a wide range of alter
acidic, sulphide-rich solution (Rona, 1984). On native interpretations exist. The analysis of trace
coming into contact with cold sea water on or just elements reveals that much of the compositional
below the sea floor the solutions precipitate base variation in the basalts is explicable in terms of high
metal sulphide deposits. The presence of such de level fractionation. To explain the most extreme
posits is corroborated by studies of ophiolites (Sec variations, however, it is necessary to invoke the
tion 12.3.2). mixing of batches of magma. The frequent presence
Detailed heat flow surveys on the Galapagos Rift of xenocrysts of deep-level origin indicates that the
have shown that the pattern oflarge scale zoning and rocks only spend a very short time in a high level
wide range of individual values is consistent with magma chamber.
hydrothermal circulation. Small-scale variations are On a smaller scale, a detailed sampling ofthe East
believed to arise from variations in the near-surface Pacific Rise by Langmuir et al,(1986) has revealed a
permeability, while larger scale variations are due to series of basalts which are diverse in their major and
major convection patterns which exist in a permeable trace element chemistry. This compositional vari
layer several kilometres thick which is influenced by ation has been interpreted in terms of a series of
topography, local venting and recharge at basement magmatic injection centres along the crest of the
OCEAN RIDGES 115
ridge which correlate with bathymetric highs spaced The drag would then prevent the ascent of the
about 50-150 km apart. Magma moves outwards lithosphere in the crestal region while flanking litho
from the injection points along the ridge so that the sphere rises by movement along normal fault planes
temperature of eruption decreases regularly from (Fig.6.13). This model thus relates the formation of
maxima at the bathymetric highs, which correspond a median rift to the presence of high viscosity
to the centres of small spreading cells. Batiza et cd. magma, and its absence if the upwelling materials is
(1988) have sampled along the axis of the southern oflow viscosity.
Mid-Atlantic Ridge, and have shown that there are A different hypothesis involves 'steady state neck
regular patterns ofchemical variation along it caused ing' of the lithosphere. In this model the median rift
by differences in the depth and extent of partial forms by necking or thinning in a ductile layer
melting and degree of fractionation. They conclude beneath the rift in the manner of a plastic material
that these patterns imply the presence of a deep necking out under tension. The layer does not break
central magma supply, with no melt migration along as new material is constantly added from below. At
the axis and no large, well mixed magma chamber. slow spreading ridges the strength ofthe lithosphere
is considered to be sufficiently high that necking
creates a median rift. Fast spreading ridges are char
6.8 Differences between fast and acterized by young crust which is believed to be too
slow spreading ridges hot and weak for necking to be significant.
The morphology of ridges, and in particular the Davis (1984), by contrast, has attributed the for
development of a median rift, appears to correlate mation of a median rift to the superimposition of
with the rate of speading and magma supply (Mor anelastic deformation due to dyke intrusion/
gan 8c Chen, 1993). Spreading rates and morphol fracturing onto the isostatic topography predicted by
ogy also appear to correlate with the ridge's being plate cooling theory.
situated within an expanding or a contracting ocean. Flower (1981) has shown that differences in the
In expanding oceans,such as the Atlantic and Indian lithology and chemistry of basalts generated at mid-
Oceans,spreading rates are low,ridges tend to follow ocean ridges show a simple correlation with spread
the centre of the ocean basins and a median rift is ing rate. The differences are not related to processes
developed. In contracting oceans,such as the Pacific, in the upper mande, as the primary melts appear
spreading rates are high, ridges may be displaced to be identical. They are believed to reflect the
from the centre and no median rift is developed. fractionation environment after partial melting.
There are several hypotheses concerning the presence Slow spreading systems are characterized by a com
or absence of a median rift(MacDonald, 1982). plex magma chamber in which there is a widespread
The 'hydraulic head loss' model suggests that accumulation of calcic plagioclase, the presence of
viscous forces within a conduit tapping the magma phenocryst-liquid reaction morphologies and
source are sufficient to decrease the hydraulic pres pyroxene-dominated fractionation extracts. These
sure,causing a bathymetric depression over the ridge phenomena are consistent with fractionation at
axis. This hydraulic head loss then causes uplift of many different pressures in a chamber which appears
the rift valley walls relative to the floor. The loss of to be transient. This conclusion is in accord with the
hydraulic pressure is proportional to the upwelling pattern of rare earth elements in basalts sampled
velocity and inversely proportional to the cube ofthe from the Mid-Atlantic Ridge (Langmuir et al..
conduit diameter. At slow spreading ridges, magma
ascends through narrow conduits formed of cool,
old lithosphere. At fast spreading ridges there is Magma
chamber
much less head loss as the conduits are flanked by
hotter lithosphere, and a flat topography develops
instead of a rift.
A mechanism similarly dependent upon magma
viscosity (Cann, 1974) suggests that the ascending
asthenosphere beneath the ridge can exert a viscous Fig.6.13 Possible mechanism for the development of a
drag on the newly formed lithosphere on either side. median rift in terms of a magma body of high viscosity.
116 CHAPTERS
6.9 Depth-age relationship of Fig.6.14 Theoretical relationship of water depth and age
oceanic lithosphere or distance for a spreading rate of 10 mm a"^.
As newly formed oceanic lithosphere moves away
from a mid-ocean ridge, it becomes removed from mechanically rigid, and the lower unit is a viscous
underlying heat sources and cools. This cooling thermal boundary layer. As the lithosphere travels
has two effects. First, the lithosphere contracts away from a spreading centre, both units thicken and
and increases in density. Secondly, because the provide the linear d-yft relationship described
lithosphere-asthenosphere boundary is controlled above. However, the lower unit eventually thickens
by temperature (Section 2.12), the cooling causes to the point at which it becomes stable and starts to
the lithosphere to increase in thickness away from the convect. This brings extra heat to the base of the
mid-ocean ridge. This latter phenomenon has been upper layer and prevents it thickening at the same
confirmed by lithosphere thickness estimates derived rate. The age-depth relationship for oceanic litho
from surface wave dispersion studies in the Pacific sphere older than 80 Ma is then given by:
Ocean, which indicate that the thickness increases
from only a few kilometres at the ridge crest to 35 km d = 6400- 3200exp(-f/62.8)
at 15 Ma age and 70 km at 50 Ma. These age-depth relationships are shown in
The cooling and contraction of the lithosphere Fig.6.14, and appear to explain the observed
cause a progressive increase in the depth to the top of bathymetry in most areas (e.g., Renkin & Sclater
the lithosphere away from the ridge (Sclater & 1988).
Francheteau, 1970), accompanied by a decrease in
heat flow.It follows that the width ofa ridge depends
6.10 Origin of the oceanic crust
upon the spreading rate, and so provides an explana
tion for the relative widths of the rapidly spreading A widely accepted model of the petrological pro
East Pacific Rise and more slowly spreading Mid- cesses occurring at ocean ridges has been derived by
Atlantic Ridge. Parsons & Sclater(1977) have deter Cann (1970; 1974). In this model (Fig.6.15), hot
mined the nature of the age-depth relationships of asthenospheric material ascends buoyantly (Nicolas
oceanic hthosphere, and have shown that the depth, et»/., 1994) suflSciently rapidly up a narrow zone to
d (metres) is related to age t(Ma)by: pass through the basalt melting curve and provides
an interstitial melt of basaltic composition. The mol
rf = 2500 + 350>/f ten fraction increases in volume as the asthenosphere
It has been found, however, that this relationship rises, and eventually departs the parental material to
only holds for oceanic lithosphere younger than ascend independently and produce a magma cham
80 Ma. For older lithosphere the relationship indi ber within the lower part of the oceanic crust at the
cates a more gradual increase of depth with age. In level oflayer 3. Part ofthis magma is erupted on the
order to explain this. Parsons & McKenzie (1978) ocean floor through fissures to produce a sequence of
suggested a model in which the cooling layer com pillow lavas, and immediately beneath is a zone of
prises two units rather than the single unit implied by dykes. These collectively make up layer 2. Kidd
Parsons & Sclater (1977). In this model the upper (1977) has modelled these processes of extrusion
unit, through which heat moves by conduction is and intrusion and compared them with observations
OCEAN RIDGES 117
Spreading axis
Pillow lavas
Layer2
Sheeted dyke
complex
3=4 Magma chamber Gabbro
Seismological • • **Cumulates4/>'
Moho t'Tr
antle Peridotite
Petrological
Moho
Hviagma rises in crystal mush
I 1 1
5 0 5 km
Horizontal and vertical scale
Fig.6.15 Diagrammatic cross-section of the structure of clase also crystallizes and layered olivine gabbros
the crust and upper mantle beneath the crestal region of a form.Much ofthe residual liquid, still volumetrically
mid-ocean ridge. The shape of the magma chamber is quite large, then solidifies over a very small temper
based on the seismic reflection model of Detrick et al. ature range to form an upper,'isotropic' gabbro. A
(1987)(after Bott, 1982a). small volatile-rich residuum of this diflferentiation
process, consisting essentially of plagioclase and
of ophiolite complexes. Layer 2C was found to quartz, is the last fraction to crystallize, sometimes
consist entirely of sheeted dykes, which were in intruding upwards to form veins and small pockets of
truded through zones less than 50 m wide, and 'plagiogranite' within the overlying sheeted dyke
which demonstrate a highly extensional regime. The complex. The abundance of volatiles, notably water,
dykes show so^e 10% more chilled margins on one in the uppermost part ofthe magma chamber may be
side than the -^wing that approximately 10% due, at least in part, to interaction with sea water
of the dyke Xu dykes, such that the percolating downwards and/or stoping of the over
margins o' ^ up on opposite lying, hydrothermally altered dykes into the magma
sides of *« sea floor chamber.
spreadir ause The two gabbro units, isotropic and layered, are
dyke the often correlated with seismic layers 3A and SB,
iSt. It respectively. The ultramafic cumulates, rich in oliv
^ dykes ine and pyroxene, would then account for the sub-
low less Moho seismic velocities. Thus, the Moho occurs
dykes are within the crystallized magma chamber at the base of
tst as they the mafic section. Oflf-axis, however, in a lower tem
a result of perature environment, the uppermost ultramafics
,o metamor- may become partially hydrated (i.e., serpentinized)
ibrate at high and as a result acquire lower seismic velocites more
iter. characteristic of layer SB. The seismic Moho would
from the crystal- then occur at a somewhat greater depth, within the
r(Fig.6.15). The ultramafic section. As a result of this uncertainty in
magma chamber, defining the seismic Moho, petrologists have tended
through the magma to define the base of the crust as the base of the
nite with occasional magma chamber, that is, the dunite/chromitite ho
accumuiationr^Ot-.,^^Jj^ With further cooling rizon. Hence, this level is termed the 'petrological
pyroxene crystallizes ahci-^ lulate peridotitic layers Moho'(see Fig.6.15).
(i.e., of olivine and pyroxene) are produced, giving The magma chamber is continuously replenished
way upwards to pyroxenites as the crystallization of by magma rising from the asthenosphere, which
pyroxene begins to dominate. Ultimately, plagio- limits the degree of fractionation possible and
118 CHAPTER 6
explains the preponderance of basaltic rocks within in another way by thermal modelling of the cooling
the volcanics dredged from the oceans. The highly magma chamber (Sleep, 1975; Kusznir & Bott,
fractionated varieties occasionally encountered in the 1976). Even though these models took no account
oceans and ophiolite bodies probably result from the of cooling of the top 2km by hydrothermal circula
isolation of magma bodies by lateral injection from tion of sea water, the calculations confirm that
the magma chamber. magma chambers can only exist at spreading rates in
Cann (1974) has also examined how oceanic excess of5 mm a"^. In contrast to the observations of
crustal structure would vary with spreading rate. No Detrick et aL (1987),they predict that the roofofthe
change in crustal thickness is predicted as the amount chamber initially dips steeply beneath the axial zone
ofmagma available depends only on the temperature and thence the slope of the roof decreases laterally
of the upwelling asthenosphere which controls the away from the axis. Crystal settling at the base ofthe
depth at which partial melting takes place. As this is chamber narrows it and flattens the floor so that the
thought to be the same for all ridges the same maximum horizontal dimension occurs at the junc
amount of magma will be produced and crust of tion between the isotropic and layered gabbros.
similar thickness should result. However,changes in There is very good agreement between the petro-
the internal structure of the crust are predicted as logical and thermal models of oceanic crust forma
speading rate increases. The interval between succes tion and the lithological succession within ophiolite
sive dyke intrusions decreases so that dykes cool less units (Section 2.5). All require the presence of a
between injections. This causes a decrease in the magma chamber as the necessary prerequisite of
thickness ofthe sheeted dyke unit,reaching the limit, oceanic crust formation. The only major problem,
at very high spreading rates, when lava continuously however, is that the shape of the magma chamber
extrudes without dyke intrusion. The resulting crust predicted by modelling does not seem to conform
would then consist of pillow lavas directly overlying with the form determined by the most recent seismic
homogeneous dolerite grading down into gabbro. reflection surveys (Detrick et 1987). Instead of
Also, as each dyke cools less before subsequent intru the teardrop shape required by modelling, a mush
sion, there is a greater tendency for dykes to intrude room shape appears to explain most closely the
the line of existing dykes, thus decreasing the width reflection results. Resolution of this apparent anom
ofthe zone ofintrusion. Because ofthe persistence of aly must await further investigation.
the magma chamber and the high thermal gradients
immediately off-axis, as spreading rate increases
6.11 Propagating rifts
stresses are concentrated in a thinner layer and frac
turing will occur more frequently so that the widths The direction ofspreading at an ocean ridge does not
of individual dykes become less with increased always remain constant over long periods of time,
spreading rate. but may undergo several small changes. Menard &
The lower crust is probably less affected by increas Atwater (1968) proposed that the northeastern
ing spreading rate, although the width ofthe magma Pacific had changed direction five times on the basis
chamber would be expected to increase. Oceanic of changes in the orientation of major transform
crustal structure evolving from processes in the faults (Section 5.8) and magnetic anomaly patterns.
magma chamber would not be expected to differ Small changes in spreading direction have also
greatly. been proposed as an explanation of the anomalous
There is a critical spreading (half separation) rate topography associated with oceanic fracture zones
of about 5 mm a~^ below which a spreading centre (Section 7.3).
cannot exist. This is because lateral heat conduction Menard & Atwater (1968) made the assumption
away from the axial zone is then sufficient to freeze that the reorientation of a ridge would take place by
dykes as they are intruded. Similarly, permanent smooth, continuous rotations of individual ridge
magma chambers could not exist because the rate of segments until they became orthogonal to the new
replenishment of heat by replenishment of the spreading direction (Fig. 6.16a). The ridge would
magma chamber is insufficient to compensate for then lie at an angle to the original magnetic anomaly
losses by conduction. pattern. Long portions of ridges affected in this way
The problem of axial structure can be approached might be expected to devolve into shorter lengths.
OCEAN RIDGES 119
(a) Pseudofault
Propagating
rift^
Failed \ Pseudofault
Doomed rift
New spreading
direction
(b)
two models.
4°N +
Hey et aL (1988) have reported the results of a
detailed investigation of the region where the direc
tion of spreading of the Pacific-Farallon boundary
changed direction at about 54 Ma,just north of the
major bend of the Surveyor Fracture Zone, using
side-scan sonar, magnetometry and seismic reflec
tion. They found that the change in direction ofsea
floor fabric revealed by sonar is abrupt,in accord with
the propagating rift model. Similar conclusions were
also reached by Caress (1988). Yi&yetal.(1980)
have described the results of a survey of an area west
of the Galapagos Islands at 96 ®W. They concluded
that here a new ridge is progressively breaking
through the Cocos plate, and the magnetic data in
particular (Fig.6.18) provide convincing evidence
that the ridge propagation mechanism is operative.
The results were confirmed in a later work(Hey et al. (b)
1986). Brozena & White(1990)have reported ridge
propagation from the South Atlantic, so this phe r+ + +
96° W 95° 94°
nomenon appears to be independent of spreading
rate. Fig.6.18 (a) Predicted magnetic lineation pattern
Engeln et aL (1988) have pointed out that the resulting from ridge propagation;(b) observed magnetic
growth of the region of overlap between old and anomalies near 96°W west ofthe Galapagos Islands
propagating rifts must take a finite amount of time (redrawn from Hey et al. 19S0^ Journal ofGeophysical
because propagation is not infinitely fast and the new Research 85, with permission from the American
ridge does not reach its full spreading rate instan Geophysical Union).
taneously. At small scales of a few kilometres, this
can lead to the formation of overlapping spreading Easter and Juan Fernandez microplates on the East
centres, such as have been described by MacDonald Pacific Rise (Searle etal.., 1989).
et al, (1987). At much larger scales, and during The cause of the initiation of ridge propagation is
propagation over millions of years, the mechanism unknown. It is probable that it is quite a common
can give rise to the formation of microplates several mechanism by which spreading centres adjust to
hundred kilometres in diameter between the two changing directions of the forces driving plate
rifts. Examples ofsuch features are believed to be the motions (Hey^f a^/., 1980). An important corollary
OCEAN RIDGES 121
of this is that the ridge-push force at spreading MacDonald, K.C., Sempere, J.-C., Fox, P.J. & Tyce, R.
centres (Section 11.6)is not a primary driving mech (1987) Tectonic evolution of ridge-axis discontinuities
anism as it appears to be quite easily overridden by the meeting, linking or self-decapitation of neighbor
during ridge propagation. ing ridge segments. Geology 15,993-7.
MacDonald, K.C. et al. (1988) A new view of the mid-
ocean ridge from the behaviour ofridge-axis discontinu
ities. Nature 335, 217-25.
Further reading Open University Course Team (1989) The Ocean Basins:
Atwater, T.(1981) Propagating rifts in seafloor spreading their Structure and Evolution. Pergamon Press, Oxford.
patterns. Nature 290,185-6. Parsons, B. & McKenzie, D.P. (1978) Mantle convection
Cann, J.R. (1974) A model for oceanic crustal structure and the thermal structure ofthe pldXts.].geophys. Res. 83,
developed. Geophys.J. Roy. astr. Sac. 39,169-87. 4485-96.
Fowler, C.M.R. (1976) Crustal structure of the mid- Parsons, B. & Sclater, J.G. (1977) An analysis of the
Atlantic ridge crest at 37®N. Geophys.J. Roy. astr. Sac. 47, variation of ocean floor bathymetry and heat flow with
459-91. 2LQC.J.ge<^hys. Res. 82,803-27.
Hey, R-N., Menard, H.W., Atwater, T.M. & Caress, D.W. Sclater, J.G. & Francheteau, J.(1970) The implications of
(1988) Changes in direction of seafloor spreading re terrestrial heat flow observations on current tectonic and
visited. Ros. 93,2803-11. geochemical models ofthe crust and upper mantle ofthe
Keen, C. & Tramontini, C. (1970) A seismic refraction Earth. Geophys. J. Roy. astr. Soc. 20, 509-42.
survey on the mid-Adantic ridge. Geophys. J. Roy. astr. Smith,D.K. & Cann,J.R.(1993)Building the crust at the
Sac. 20,473-91. Mid-Atlantic Ridge. Nature 365,707-15.
Kusznir, N.J. & Bott, M.H.P.(1976) A thermal study of Talwani, M., Le Pichon, X. & Ewing, M.(1965) Crustal
the formation of oceanic crust. Geophys. J. Roy. astr. Soc. structure of the mid-ocean ridges 2. Computed model
47,83-95. from gravity and seismic refraction data./. Res.
Lister, C.R.B.(1980) Heat flow and hydrothermal circula 70,341-52.
tion. Rsv. Earth planet. Set. 8,95-117. Wyllie,P.J.(1988) Magma genesis, plate tectonics and the
MacDonald, KC.(1982) Mid-ocean ridges: fine scale tec chemical differentiation of the Earth. Rev. geophys. 26,
tonic, volcanic and hydrothermal processes within a plate 370-404.
boundary zone.A»». Rop. Earth planet. Sci. 10,155-90.
7/ Transform and transcurrent faults
Dextral Sinistral
122
TRANSFORM AND TRANSCURRENT FAULTS 123
3'
if
(a)
I , I 11
Jc ..rM"
f uJT
(b)
Fig.7.2 (a) Six possible types of dextral transform fault: parallel fracture zones which appear to displace the
(i) ridge to ridge;(ii) ridge to concave arc: (iii) ridge to crest of the Mid-Atlantic Ridge in equatorial lati
convex arc;(iv) concave arc to concave arc;(v) concave tudes between Africa and South America were be
arc to convex arc;(vi) convex arc to convex arc. lieved to represent sinistral transcurrent faults which
(b) Appearance ofthe dextral transform faults after a displaced an originally straight crest. Wilson (1965),
period oftime (redrawn from Wilson, 1965, with
however, interpreted them as dextral transform
permission from Nature 207,334-47. Copyright(© 1965
faults. The evolution of these faults is intimately
Macmillan Magazines Ltd).
related to the rifting event which preceded the open
ing of the Atlantic Ocean. Figure 7.3a shows an
consistent with dextral transform motion. Events idealized continent containing two lines of weak
along the fracture zone beyond the ridge extremities ness. When the continent is subjected to tension
are rare, confirming the transform mechanism of approximately parallel to these lines, an incipient
faulting. fracture develops which is largely perpendicular to
the direction of tension, but, in the vicinity of the
lines ofweakness,develops preferentially along these
7.2 Origin of transform faults old lines (Fig. 7.3b). After rifting (Fig. 7.3c) the
Before the recognition of transform faulting, the mid-ocean ridge crest maintains the shape of the
124 CHAPTER?
"^Lines of old
weakness
(a)
(b)
New rift and seismiczone
initial rift, and the portions parallel to the lines of of Siberia by rotation about a fulcrum in the New
weakness develop into transform faults which are Siberian Islands (Fig. 7.4 inset).
only active between the offset ridge crests. These The different offsets observed across the Murray
transform fault locations are consequently inherited Fracture Zone from magnetic lineations are thought
from the initial rift and act in such a way that the to be due to a change in location ofthe ridge crest to
mid-ocean ridge can maintain a median position the south ofthe fracture zone about 40 Ma ago. The
within the expanding ocean. change in offset ofanomalies ofthe same age implies
An example of transform faulting is shown in a 'ridge jump' of approximately 500 km to the east
Fig. 7.4, which shows how the northern part of the (Harrison & Sclater, 1972). Similar but better docu
Mid-Atlantic Ridge is terminated. In Early Tertiary mented ridge jumps, which also greatly reduce the
times the Mid-Atlantic Ridge bifurcated to the south offset of crust of the same age on either side of a
of Greenland. The western branch, which is now fracture zone, occur in the extreme south of the
inactive, passed through Baffin Bay and terminated Atlantic Ocean (Barker, 1979). Here,to the south of
against the Wegener Fault,an extinct,sinistral ridge- the Falkland-Agulhas Fracture 2k)ne, the ridge crest
ridge transform fault. The active eastern branch has jumped westward on three occasions since the
passes through Iceland, and terminates southwest of opening of the South Atlantic, that is at 98,63 and
Spitsbergen at the De Geer Fault. This dextral ridge- 59 Ma. In so doing it has reduced the original offset
ridge transform fault connects to the northernmost of 1400 km to approximately 200 km. Other ridge
part of the Mid-Atlantic Ridge in the Arctic. This is jumps, producing major changes in ridge crest
itself transformed into the Verkhoyansk Mountains geometry within the past 10 Ma, have occurred to
TRANSFORM AND TRANSCURRENT FAULTS 125
%
•'A'S
Canadian Basin
90°W
Siberian
t: Basin
-
Wegener /w:-
De Geer
Fault
Fig.7.4 Northern termination of the Mid-Atlantic Ridge fracture zones, but on land their recognition in detail
(redrawn from Wilson, 1965, with permission from is considerably more complex. This difference is
Nature 207, 334-47. Copyright © 1965 Macmillan partly a function of the proximity of the fault to
Magazines Ltd). observation, as oceanic faults are commonly studied
from the sea surface several kilometres above the
ocean floor. Direct observations ofa fracture zone on
the north of Iceland (Vogt et 1970) and along the Mid-Atlantic Ridge (Choukroune et 1978)
the crest of the East Pacific Rise in the east central have shown, however, that it consists of a complex
Pacific (Herron, 1972). swarm of faults occupying a zone 300-1000 m in
The proposal that transform fault locations are width, and Searle (1983) suggests that these multi-
inherited from ancient lines of weakness is strongly fault zones are wider and more common on fast
supported by considering the continuation of trans spreading ridges such as the East Pacific Rise. More
form faults offsetting the Red Sea Ridge into the importantly, however, their differences reflect the
adjacent continents (Garson & Krs, 1976). The contrasting natures of the oceanic and continental
faults appear to correlate well with ancient fractures crusts in their response to predominantly strike-slip
present in the flanking Precambrian Arabian and shear failure.
Nubian shields. The geometry ofthe Red Sea has also
been controlled by northerly trending Precambrian
7.3 Oceanic fracture zones
structures(Dixon At/., 1987).
Transform faults occur in both oceanic and con Transform faults in the oceans are marked by fracture
tinental lithosphere. In the oceans their trace is well zones, which are long, linear, bathymetric depres
defined, in the absence of sedimentary cover, by sions which normally follow arcs of small circles on
126 CHAPTER?
the Earth's surface perpendicular to the offset ridge which comprised some 23% of the area surveyed,
(Bonatti & Crane, 1984). they encountered serpentinite with very few of the
Fracture zones mark both the active transform basaltic rocks which normally characterize oceanic
segment and its fossilized trace. It has been suggested layer 2(Section 2.4.6). The hypothesis proposed to
(Collette, 1979) that the fractures result from ther explain this extensive serpentinite follows from the
mal contraction in the direction ofthe ridge axis. The segmentation of ocean ridges and the occurrence of
internal stresses caused by contraction are much magma generation only at certain locations on
larger than the breaking strength ofthe rocks, and it slow spreading ridges (Section 6.5). The oceanic
is possible that fracture zones develop along the crust formed at magmatic centres would be of the
resulting lines of weakness. normal,layered type (Section 2.4.4). Sections distant
Dredging of fracture zones has recovered both from the centres would, however, receive an insuffi
normal oceanic crustal rocks and rocks which show cient magma supply to produce basaltic rocks and
much greater metamorphism and shearing. Very upper mantle-derived peridotitic material would be
commonly large blocks ofserpentinite lie at the bases brought to the surface, where it is altered by sea water
ofthe fracture zones. to serpentinite. As magmatic centres grow, migrate
Bonatti & Honnorez(1976)and Fox etal.(1976) along the ridge axis and decline, the normal oceanic
have examined specimens recovered from the thick crust would similarly migrate and would enclose
crustal sections exposed in the large equatorial Atlan those regions of serpentinitic crust which originate
tic fracture zones, which were found to consist of where magma was absent. This work is important
ultramafic, gabbroic and basaltic rock types and their as it implies that serpentinized peridotite is more
metamorphosed and tectonized equivalents. Serpen common in slow spreading oceans than previously
tinite intrusion appears to be quite common within recognized. There are wide ranging implications.
fracture zones, accompanied by alkali basalt volcan- Peridotite is much more reactive with sea water than
ism,hydrothermal activity and metallogenesis. Inves basalt and on weathering would release magnesium,
tigations ofthe Vema Fraction Zone(Auzende nickel,chromium and noble metals. Sepentinite also
1989) have indicated a sequence similar to normal contains far more water than altered basalt, which
oceanic layering. The islands of St Peter-Paul in the would account for much ofthe water supplied to the
equatorial Atlantic, which lie on a ridge associated mantle in subduction zones (Section 8.8).
with the St Paul Fracture 2k)ne, are composed of Ocean fracture zones must bring oceanic crust of
mantle peridotite. diflferent ages into juxtaposition. The depth ofthe sea
In the North Atlantic, fracture zone crust is very floor is dependent upon the square root of the age
heterogeneous in thickness and internal structure (Section 6.9), and so it would be expected that a
(Derrick et 1993). It is often thin(< 2km) with scarp would develop across the fracture zone from
low seismic velocities, and layer 3 is absent. The the younger, higher crust to the lower, older crust
crustal thinning may extend several tens of kilo (Menard & Atwater, 1969; DeLong et til.., 1977)
metres from the fracture zone. Geologically, this (Figs 7.5, 7.6b). The rate of subsidence of oceanic
structure may represent a thin, intensely fractured lithosphere is inversely dependent upon the square
and hydrothermally altered basaltic layer underlain root of its age (DeLong et at.., 1977), so the higher,
by serpentinized ultramafic rocks. The apparent younger crust subsides more rapidly than the lower,
thickness variations may then reflect different extents older side. The combination of contraction in the
ofserpentinization. The thin mafic crust is possibly a vertical plane and horizontally perpendicular to the
result of reduced magma supply at ridge oflTsets. direction of the ridge axis would result in a small
Work by Cannat et til. (1995) has suggested that component of dip-slip motion along the fracture
serpentinized rocks may be much more common in zone away from the active transform fault. DeLong et
the oceanic crust than previously expected, even in at.(1977) have suggested that this small amount of
areas distant from fracture zones. They dredged in dip-slip motion could give rise to fracture zone
the region of the North Atlantic Ridge at 22-24®N seismicity and deformation of rocks within the floor
over areas of positive gravity anomalies, which are and walls.
probably indicative of relatively thin crust, and over Transverse ridges are often found in association
areas with a normal gravity field. Over the former. with major fracture zones and can provide vertical
TRANSFORM AND TRANSCURRENT FAULTS 127
subsidence (Bonatti 8c Crane, 1984). There are certain oceanic transform faults in which
Lowrie et al. (1986) have noted that, in some the direction of the fault plane does not correspond
fracture zones, the scarp height may be preserved exactly to the direction ofspreading on either side so
even after 100 Ma. They accept that some parts that there is a component of extension across the
of fracture zones are weak, characterized by active fault. When this occurs the fault may adjust its
volcanism, and maintain the theoretical depths pre trajectory so as to become approximately parallel to
dicted for cooling lithosphere. Other parts, however the spreading direction by devolving into a series of
appear to be welded together and lock in their initial fault segments joined by small lengths of spreading
differential bathymetry. The differential cooling centre (Fig. 7.7a). A fault system in which new crust
stresses would then cause flexure of the lithosphere originates is termed a leaky transform fault(Thomp
on both sides of the fracture zone. Future work son & Melson, 1972; Taylor et al., 1994). A leaky
will reveal if there is any systematic pattern in the transform fault may develop where the location of
128 CHAPTER?
1 1
United States
Mexico
Pacific Ocean
(a)
faulting is determined by some existing crustal weak In strike-slip or wrench faults the primary motion is
ness. An example of this type is the Gulf of Califor horizontal along a vertical fault plane. Fundamental
nia, where the transform fault system developed is faults penetrate the whole lithosphere. Only trans
controlled by the presence of a small tract ofoceanic form faults mark active plate boundaries.
crust between the continental crust ofBaja California Examples of transform faults which displace con
and Mexico (Fig. 7.7b). An alternative mechanism tinental lithosphere are the dextral San Andreas
for leaky transform fault development occurs when Fault of California, the dextral Alpine Fault of New
there is a small shift in the position of the pole of Zealand and the sinistral Dead Sea fault system which
rotation about which the fault describes a small connects the Red Sea to the Bitlis Mountains of
circle. The fault would then adjust to the new small Turkey. Other fundamental faults which do not
circle direction by becoming leaky. An example of represent present plate boundaries (Reading, 1980)
this type is the Cayman Trough, a leaky transform include the Great Glen Fault ofScotland, the North
fault which defines the western part of the northern Pyrenean Fault Zone and the Insubric Line of the
boundary of the Caribbean Plate with the North Alps. An additional category is represented by those
American Plate. fundamental faults associated with indentation tec
tonics (Section 9.3.6).
Large strike-slip faults frequently produce distinc
tive topographic features. These may be a primary
TRANSFORM AND TRANSCURRENT FAULTS 129
fault scarp, a linear trough resulting from differential on the stress state of the region of the San Andreas
erosion of contrasting juxtaposed lithologies or ero fault system which has led to a new model for
sion of a fault breccia (Allen, 1981). The strike-slip continental strike-slip faulting. Various types of in
movement causes horizontal displacements of sur situ stress measurements and other geological and
face features such as streams, canyons and ridges. geophysical indicators demonstrate that, east of the
In contrast to oceanic fracture zones, which are Great Central Valley of California (Fig. 7.8), there is
characterized by a relatively simple linear trough, much strike-slip and normal faulting, but to the west
plate boundaries on continents are considerably the deformation is wholly compressional,dominated
more complex. This undoubtedly reflects the differ by folding, strike-slip and reverse faulting parallel to
ence in strength and thickness between oceanic and the San Andreas, so that the direction of maximum
continental lithosphere, and also the inhomogen- horizontal compression is orthogonal to the fault
eous nature of the continental crust, which may zone. Classical faulting theory cannot explain com
contain many ancient lines of weakness along which pression at such a high angle to a strike-slip fault as
a plate would rupture preferentially. It is thus neces the Pacific and North American plate motions are
sary to distinguish between the transform fault, the only slightly convergent. Moreover, Zoback et al.
principal rupture; the transform fault zone, a set of point to a paradox concerning major strike-slip faults
approximately parallel faults which splay from and in that heat flow observations show no frictionally
rejoin the main fault within a strip about 10 km in generated heat,so that the fault must slip in response
width; and the transform fault system, a set of high to very low shear stresses which are 10-20% ofthose
angle faults approximately parallel to the main fault, predicted by frictional faulting theory. They suggest
forming a belt some 500 km wide, and which may a common explanation for these phenomena in
include a subset striking at a significant angle to the terms of a model in which the San Andreas is an
main fault. extremely weak fault which can locally reorient the
The complex nature ofcontinental transform fault regional stresses (Mount & Suppe, 1987). Far from
systems is illustrated by the San Andreas fault system the fault, shear stresses are high and contained by the
in California (Fig. 7.8). The system developed in frictional strength of the crust, but shear stresses on
Oligocene times and its location within a continent planes parallel to the'weak'faults ofthe San Andreas
may have been controlled by an existing fracture system must be quite low. Consequently, the prin
network (Beck, 1986), as otherwise the more logical cipal stresses become reoriented so as to minimize
place for the fault to develop would be the locus shear stress on planes parallel to the San Andreas.
of the transition from continental to oceanic litho This requires a rotation such that the direction of
sphere at the continental margin. There has been maximum principal stress becomes nearly either per
about 1500 km ofdextral movement along the trans pendicular or parallel to the fault. The maximum
form fault zone since Oligocene times. The San horizontal compressive stress becomes nearly or
Andreas Fault is the only continuous fault in this thogonal to the fault ifthe regional compression is at
zone,and extends for some 1200 km,but only about an angle to the fault in excess of 45®, as occurs at
300 km of horizontal movement appears to have present. However if this angle is less than 45®, the
occurred along it. The remaining 1200 km move maximum horizontal compression becomes almost
ment must have occurred along other northwest parallel to the fault. The 20® clockwise change in
faults within the transform zone and in folding ofthe relative plate motions that occurred at 4-5 Ma would
rocks along it. There also exist major sinistral faults, have been preceded by a 90® change in stress dir
such as the Garlock and Big Pine, which strike north ection, and would explain the fault-parallel basin
easterly. Much of the present day seismicity occurs development and fault-normal extension prevalent at
along such oblique faults (Nicholson et«/., 1986a), that time.
which in places define small crustal blocks which are Zoback et al, suggest that the weakness of the San
rotating in a clockwise sense as a result ofthe regional Andreas fault system extends to a depth approxi
stress field. Much of the deformation caused by the mately equivalent to the whole lithosphere. A pos
transform fault system changes from brittle to duc sible mechanism for decreasing the strength of the
tile at 15 km depth, below which seismicity is rare. fault is the presence of abnormally high pressure
Zoback et al, (1987) have presented information fluids in the fault zone which lowers the effective
130 CHAPTER?
120°
Cape i 0 50100150
Menaocino
Ft Delgada
40'
Ft Arena
Fort Ross
Hayward F.
./ Mojave Desert
N -
White Wolf F
Nacimlento F.
35° ■•••
Gabriel
Banning F
^ Angeles'Tv Salton
Sea
Inglewood-:^ ^ ^
■ F. r.: ^
Diego
^
50 100 150
' isAexico 115°
Fig. 7.8 Generalized fault map of coastal and southern This revolutionary new model of very weak conti
California (redrawn from Crowell, 1962, with permission nental strike-slip faults clearly explains the conflict
from the Geological Society of America). ing geological data in central California, and would
be verified by direct in situ stress measurements at
depth in the fault zone. It may also be applicable to
normal stress across it. However, the pressure could other major continental faults elsewhere.
not exceed that of the minimum horizontal principal
stress or natural hydraulic fracturing would occur,
7.4.2 Transpression and transtension
and so fault slip would not occur. Zoback et al. thus
propose that some other mechanism must be active Since ancient fractures may control the locations of
in reducing fault strength. fault planes in many transform fault zones, the strike
TRANSFORM AND TRANSCURRENT FAULTS 131
'^'/--Restraining
honri
_^;:;^Releasing
hpnri
(a)
y-yy-'y/
(b) (c)
Fig.7,9 Structures developed along continental wrench The combination of strike-slip motion and exten
faults:(a) at a double bend a pull-apart forms at a releasing sion (divergent strike-slip) is known as transtension^
bend and deformation and uplift at a restraining bend;(b) and the combination of strike-slip motion and
sharp pull-apart;(c) severe deformation at a sharp compression (convergent strike-slip) is known as
restraining bend causing folding and thrusting (redrawn
transpression. Transtensile regions exhibit normal
from Crowell, 1974b, with permission from the Society of
faulting, basin extension and volcanicity, whereas
Economic Palaeontologists & Mineralogists).
transpressive regions exhibit thrust faulting, folding
and uplift. Gentle strike-slip fault curvature can thus
of the faults may depart from a simple linear trend give rise to both extension and compression. Various
following a small circle on the Earth's surface. The forms of basin develop under extension, while the
curvature of the faults imposed by existing litho- elevated areas resulting from compression may pro
spheric weaknesses gives rise to alternate zones of vide sediment for the basin filling.
convergence and divergence, where the fault shape
causes blocks on either side of the fault to be com
7.4.3 Pull-apart basins
pressed or extended (Fig. 7.9a). An example ofsuch
behaviour is illustrated by the Transverse Ranges of Where the curvature of a strike-slip fault is pro
California (Fig. 7.8), which have developed along a nounced,or where one fault terminates and sidesteps
portion ofthe San Andreas Fault which strikes signi to an adjacent, parallel fault, the curved zone or area
ficantly more westerly than the general northwest separating the ends of the faults is thrown into
trend of the fault system. A consequence of the tension (Fig. 7.9b)or compression (Fig. 7.9c). Com
compression experienced between southwest and pression results in the development of a region ele
northeast California across this part of the fault is vated by crustal shortening and typified by folds and
the formation of the Transverse Ranges by crustal thrust faulting. Tension gives rise to an extensional
shortening. trough known as pull-apart basin.
132 CHAPTER?
Well documented examples of pull-apart basins ation, so only relatively small extensions are needed
include the Dead Sea (Quennell, 1958) and the to initiate the formation of pull-apart basins. A
Salton Trough of southern California (Crowell, sketch of an idealized pull-apart basin is presented in
1974a). The Salton Trough is located in the San Fig. 7.11, although it is unlikely that any one basin
Andreas fault zone to the north of the Gulf of would exhibit all the features shown.
California and is growing due to continued dextral
motion along two parallel traces of the fault. It
7.4.4 Fault wedge basins
contains about 6000 m of young sediments
(Crowell, 1974b). Pull-apart basins are of economic Strike-slip faults may diverge and converge to form
importance as they represent excellent targets for an anastomosing pattern. Duplexes (Fig. 7.12) may
hydrocarbon exploration. form at bends (Woodcock & Fischer, 1986), analo
Pull-apart basins progressively grow in the same gous to the duplexes formed on the ramps ofdip-slip
direction as the fault movements so that the oldest faults but differing in that vertical movements are not
sediments occupy the margins only (Fig. 7.10). The constrained at the upper (ground) surface. In this
strike-slip margins are initially straight and parallel, environment, pull-apart basins are frequently associ
but may sag with time so that continued fault move ated with lens-shaped basins and high-standing
ment may slice off segments and provide a complex ridges and banks ofsimilar shape. These features owe
braided zone. The pull-apart margins can exhibit any their origin to the stress regime resulting from the
shape. They are initially identical in form, but extend confluence of two sub-parallel strike-slip faults. The
and sag during development of the basin so that in individual blocks defined by the strike-slip faults are
time their similarity is lost. compressed and uplifted when the faults converge
As the pull-apart basement grows, its floor and stretched and downthrown where the faults
stretches and becomes attenuated, so that eventually diverge (Fig. 7.13).
igneous material may be emplaced in the centre of In an anastomosing pattern of sub-parallel strike-
the basin. The presence of basic igneous material is slip faults, those most closely following the direction
suggested by the presence of positive gravity anoma of plate movements predominate, grow longer and
lies and the location of geothermal areas within assume near vertical dips. Other faults at an angle to
basins such as the Salton Trough. However, many the overall direction of movement may then rotate
basins,ofthe order of50-100 km in length,have not farther out of alignment and develop dips signifi
attenuated to the point of lithospheric rupture, and cantly less than vertical, so that the fault involves a
contain no igneous rocks. Calculations on several component of dip-slip motion. If the fault's curva
basins indicate that only about 5-10% extension is ture carries it to a region of extension, a normal
required to produce the observed lithospheric attenu oblique-slip fault develops;ifto a region ofcompres-
Time
Complex
Pull-apart basin
intersection
Complex corner
Oldest intact basin
Remnants of marginal
rocks within volcanics
Oblique-slip
faults
Talus breccias and rapid facies
changes basinward
sion, a reversed oblique-slip fault forms. In braided 30-35 km thick, and about 5-10 km of the upper
fault systems of this type, the faults may converge at mantle. The lower lithosphere is not displaced. One
depth to produce a flower structure (see Fig. 7.14) possible interpretation ofthese data is that the upper
(Harding, 1985). lithosphere is decoupled along a horizontal plane,
and has reached its present position by westerly
shearing. The plate boundary at depths in excess of
7.4.5 Deep structure of transform zones
40 km is, then, situated farther to the east than at the
Important data on the deep structure of continental surface (Fig. 7.15). This idea is supported by earth
strike-slip faults have been provided by teleseismic quake data from southern California (Nicholson et
studies of the crust and upper mantle in southern fit/., 1986b), which suggest that much deformation
California by Hadley & Kanamori(1977). They have takes the form oftectonic rotations ofrelatively small
detected a high velocity body beneath the Transverse blocks. If this is the case, a detachment surface is
Ranges which extends from a depth of40 km down required at depth.
to about 100 km. This body extends 150 km east of A lack of coherence between the upper crust and
the San Andreas Fault before terminating, and does mantle in the region of strike-slip faults has been
not appear to be oflfset by the fault. The San Andreas catalogued by Molnar (1988), who notes that the
fault system only appears to affect the crust, which is deformation of near-surface rocks can be of such
134 CHAPTER?
Separation
Straight Overlap Straight
Contractional duplex
Leading contractional imbricate fan
Trailing contractional imbricate fan
Fig.7.12 Map view ofidealized dextrai strike-slip systems San Andreas Fault do not continue down into the
(redrawn from Woodcock & Fischer, 19S6,Journal of underlying upper mantle. Elsewhere, however,
Structural Geolqgy^ with permission from Pergamon Press, Lemiszki & Brown (1988) have shown that the San
Ltd). Andreas Fault appears to pass directly into the mantle
as a narrow vertical feature. Indeed, Lemiszki &
Brown suggest, from available deep reflection data,
magnitude as to cause the amount of horizontal that past or present plate boundary transform faults
displacement to vary along the fault. Ifthis deforma in continents are characterized by a near vertical fault
tion affected the whole crust, variations in crustal zone penetrating the whole crust, while intraplate
thickness would be expected. These have not been wrench faults sole out at depth along a subhorizontal
noted by deep seismic reflection surveys, implying detachment, allowing the rotation and deformation
that,in at least some fault zones, the fault penetrates of blocks of upper crust.
no deeper than 8-10 km. The upper crust thus The possibility of the occurrence of decoupling
appears to be detached from the lower continental within the continental Iithosphere along horizontal
iithosphere, and to act as a suite of small blocks shear planes is clearly most important as it suggests
separated by minor faults. This suggests that the that, in addition to transpression and transtension,
crustal blocks are driven by flow in the underlying the upper brittle Iithosphere can respond to distrib
lower crust and mantle. Thus,some portions of the uted shear in the more ductile Iithosphere at depth.
Upper llthosphere
Divergence
Transform motion between
oontinental crust
(a) \
Transform motion between
oceanic crust
Oceanic crust
Fig. 7.16 Simplified model of the formation of a sheared sheared margin. This is probably recorded in a de
continental margin,(a) Continental crust in transform crease in the subsidence rate of the margin, as shown
contact,(b) Continent and continental margin in by a succession ofsediment onlaps, and results in the
transform contact,(c) Progressive drift of the continental uplift of the marginal ridges. In phase 4(Fig. 7.16d)
margin past hot ocean crust and the ocean ridge,(d) the transform is only active between blocks of
Formation of an oceanic fracture zone as separation is oceanic crust and thus appears as a fracture zone
completed (redrawn from Mascle & Blarez, 1987).
(Section 7.3). The sheared margin is then in contact
with cooling oceanic lithosphere and its subsidence
should evolve in a manner similar to other rifted
crustai thinning (Section 10.3.5). The sediments are margins.
folded and faulted by the transform motion and a
marginal ridge is formed. This tectonism is also
recorded in unconformities in the sedimentary se Further reading
quence. In phase 3 (Fig. 7.16c) there is contact Bonatti, E. (1978) Vertical tectonism in oceanic fracture
between continental margins and oceanic crust. Dur zones. Earth planet. Set. Lett. 37, 369-79.
ing this phase the sheared margin passes adjacent to Bonatti, E. 8c Crane, K.(1984) Oceanic fracture zones. Sci.
the hot oceanic crust of the spreading centre and the Am. 250(5), 36-47.
thermal exchanges it experiences must affect the Crowell, J.C. (1974) Origin of late Cenozoic basins in
TRANSFORM AND TRANSCURRENT FAULTS 137
southern California. In Dickinson, W.R. (ed.) Tectonics H.G. (eds) Sedimentation in Oblique Slip Mobile Zones.
and Sedimentation. Soc. econ. Paleont. Mineral. Spec. Spec. Publ. int. Ass. Sediment. 4,7-26.
Publ. 22,190-78. Woodcock, N.H. & Fischer, M. (1986) Strike-slip
Hadley,D.& Kanamori,H.(1977)Seismic structure ofthe duplexes./, struct. Geol. 8,725-35.
Transverse Ranges, California. Bull. £feol. Soc. Am. 88, Zoback, M.D. et al. (1987) New evidence on the state of
1469-78. stress of the San Andreas fault system. Science 238,
Reading, H.G. (1980) Characteristics and recognition of 1105-11.
strike-slip fault systems. In Ballance, P.P. & Reading,
8/ Subduction zones
Fig. 8.2 Generalized schematic morphology of an oceanic some island arc systems there may be up to three
subduction zone. generations of marginal seas developed on the land
ward side of the island arc.
amori, 1979). Proceeding from the oceanward side
ofthe system, a bulge about 500 m high occurs about
120-150 km from the trench. The forearc region 8.2 Gravity anomalies of
comprises the trench itself, the subduction complex subduction zones
(the 'first arc' or accretionary wedge or prism) and Figure 8.3 shows a free air gravity anomaly profile
the forearc basin. The subduction complex is con across the Aleutian arc which is typical of most
structed of thrust slices of trench fill (flysch) sedi subduction zones. The bulge of the downgoing
ments and also possibly oceanic crust, which have lithosphere to seaward of the trench is marked by a
been scraped off the downgoing slab by the leading positive gravity anomaly of about 500 g.u. (Talwani
edge ofthe overriding plate. The contact between the & Watts, 1974). The trench and accretionary prism
accretionary wedge and forearc basin is often a region are typified by a large negative anomaly of some
of backthrusting (Silver & Reed, 1988). The forearc 2000 g.u. amplitude which results from the displace
basin is a region of tranquil, flat-bedded sedimenta ment ofcrustal materials by sea water and sediments.
tion between the forearc ridge and island arc. The Conversely, the island arc is marked by a large posi
island arc ('second arc') is made up of an outer tive anomaly. Isostatic anomalies over the trench and
sedimentary arc and an inner volcanic arc. The sedi arc are large and exhibit the same polarity as the free
mentary arc comprises coralline and volcaniclastic air anomalies. These large anomalies result from the
sediments underlain by volcanic rocks older than dynamic equilibrium imposed on the system by com
those found in the volcanic arc. This volcanic sub pression, so that the trench is forced down and the
strate may represent the initial site of volcanism as arc held up out of isostatic equilibrium by the forces
the relatively cool oceanic plate began its descent. As driving the plates.
the 'cold' plate extended further into the astheno-
sphere the position of extrusive igneous activity
moved backwards to its steady state location now 8.3 Structure of island arc systems
represented by the volcanic arc. The island arc and from earthquakes
remnant arc (back-arc ridge or'third arc'), first recog Island arc systems exhibit intense seismic activity. A
nized by Vening Meinesz(1951),enclose a marginal large number of events occur on a plane which dips
sea (back-arc basin) behind the island arc. Such mar on average at an angle of about 45° away from the
ginal seas are generally 200-600 km in width. In underthrusting oceanic plate. The plane is known as
140 CHAPTERS
South
2000 r
Volcanic arc
Bering Sea
Lithosphere
Fig.8.3 Gravity anomalies of an oceanic subduction zone The BeniofFzone had originally been interpreted as a
(after Grow, 1973, with permission from the Geological large thrust fault between different crustal provinces.
Society of America). The seismic data allowed a new interpretation to be
made in terms of a high belt ofPacific lithosphere
a BeniofF (or Benioff-Wadati) zone, after its dis- underthrust into the mantle. This interpretation has
coverer(s), and earthquakes on it extend from the been refined by Barazangi St Isacks(1971) by the use
surface, at the trench, down to a maximum depth of of a local seismometer network (Fig. 8.7) in the
about 680 km. Figure 8.4 illustrates how shallow, region of the Tonga arc. In addition to the pret ious
intermediate and deep focus earthquakes in the results, a zone of very high attenuation (extremely
southwestern Pacific occur at progressively greater low of about 50) has been defined in the upper
distances away from the site of underthrusting at the most mantle above the downgoing slab in a region
Tonga Trench. Figure 8.5 shows a section through about 300 km wide stretching between the active
the Tonga-Kermadec island arc system with earth island arc (Tonga) and back-arc ridge (Lau Ridge).
quake foci projected on a vertical plane parallel to the This implies that the mantle beneath the marginal sea
direction of underthrusting. Further information on (Lau Basin) is much weaker than elsewhere or that
the nature of the BeniofF zone was obtained from a the lithosphere is considerably thinner. The data have
study of the body wave amplitudes from deep earth important ramifications for the origin of back-arc
quakes (Fig. 8.6). Seismic arrivals at the volcanic basins and will be considered in more detail in
islands ofthe arc, such as Tonga, were found to be of Section 8.9.
far greater amplitude than those recorded to the Detailed investigations of the region above
front or rear of the arc at stations such as Raratonga the subducting lithosphere have also been carried
and Fiji. The differences in amplitude are usually out using seismic tomography (Section 2.1.8).
described quantitatively in terms of the ^-factor, the Figure 8.8, from Hasemi et al. (1984), shows a
inverse of the specific attenuation factor, and in section through the Japan arc at 39.8°N with con
general the higher the ^.-factor the stronger the rock. tours in percentages of the relative velocity variation
High travel paths give rise to little attenuation, and in the region ofthe subducting Pacific Plate. The top
vice versa. Seismic waves travelling up the length of ofthe downgoing slab is clearly indicated by a region
the seismic zone appear to pass through a region of of high velocity. Also present is a low velocity anom
high 2,(about 1000), while those travelling to lateral aly in the wedge above the subducting plate which
recorders pass through a more normal region of low extends from the surface to a depth of about 65 km.
Q (about 150). The BeniofF zone thus appears to This underlies the volcanic arc and is interpreted as
define the top of a high Qzone about 100 km thick. an upwelling diapir of partially molten material. Also
SUBDUCTION ZONES 141
shown is the double BeniofF zone in this region, occurring on a vertical plane perpendicular to the
which is discussed later in this section. island chain.
The earthquake activity of the downgoing slab The earthquakes occurring in the Benioff zone in
occurs as a result of three distinct processes zone *c' (Fig. 8.9) at depths greater than the thick
(Fig. 8.9). In region 'a' earthquakes are generated in ness of the lithosphere at the surface are not gener
response to the bending of the lithosphere as it ated by thrusting at the top of the descending plate,
begins its descent. Downward flexure of the litho because the asthenosphere in contact with the plate is
sphere throws the upper surface of the plate into too weak to support the stresses necessary for exten
tension, and the normal faulting associated with this sive faulting. At these depths earthquakes occur as a
stress regime gives rise to the observed earthquakes, result of the internal deformation of the strong
which occur to depths of up to 25 km. descending slab of lithosphere, so that the majority
Flexural bending of the lithosphere also gives rise of events lie about 30-40 km beneath the top of the
to the topographic bulge present in the subducting slab. Hasegawa etal.(1978), by making use ofa local
plate on the oceanward side of the island arc. This array of seismographs, have identified what appears
regional rise of sea bed topography is located about to be a double Benioff zone beneath the Japan arc
120-150 km from the trench axis and has an ampli (Fig. 8.13; see also Fig. 8.8). The arrival times of
tude ofseveral hundred metres. It is associated with a different seismic phases indicate that the upper of
regional positive gravity anomaly of about 500 g.u. these corresponds to the top of the descending slab
(Watts & Talwani, 1974). Simple beam theory pre and that the lower corresponds to the main Benioff
dicts that the presence ofthis bulge is a consequence zone where earthquakes originate by internal deform
of the downward deflection of the subducting plate ation. It is probable that the upper plane earthquakes
(Fig. 8.10). However, closer investigation of litho- result from the stresses caused by the unbending of
spheric behaviour in this environment (Chappie & the plate which had suffered a certain amount of
Forsyth, 1979)indicates that the flexure is not com permanent, plastic deformation during its initial de
pletely elastic and must involve considerable plastic scent(Isacks & Barazangi, 1977). Fujita & Kanamori
(permanent)deformation. A two-layer elastic-plastic (1981) and Spence (1987) attribute the lower seis
lithospheric model would deform in a very similar mic zone to tension in the slab resulting from the
way to that observed (Fig. 8.11). deeper sinking plate, and the upper seismic zone to
Region 'b' (Fig. 8.9) is characterized by earth localized compression of the plate as its motion is
quakes generated from thrust faulting along the con resisted by the surrounding mantle. Wiens et aL
tact between the overriding and underthrusting (1993) have suggested that the double Benioffzone
plates. Indeed, the overriding plate suffers compres- results from transformational faulting (see below) at
sional deformation for several tens of kilometres to the edges of the upper part of the descending slab.
the landward side of the trench. Focal mechanism Kawakatsu (1986) and Wiens et al. (1993) have
solutions for earthquakes associated with regions 'a' described a similar double seismic zone beneath the
and 'b' of Fig. 8.9 are shown in Fig. 8.12, which Tonga arc and Gorbatov et al. (1994) a double
represents the distribution of earthquake types seismic zone beneath Kamehatka. With the imple
around the Aleutian island arc (Stauder, 1968). The mentation of more refined location techniques, it is
belt ofearthquakes to the south ofthe islands is caused probable that more will be discovered in the future.
by normal faulting associated with the flexure of the The presence ofearthquakes at depths in excess of
top part of the Pacific Plate which is underthrusting 70 km is paradoxical in that below this level the high
the Bering Sea in a northwesterly direction. The pressure causes materials to flow rather than fracture.
groups ofearthquakes lying under orjust to the south From 70 km to 300 km it appears that faulting
of the island chain are indicative of thrust faulting. occurs during the rapid dehydration of serpentinite
The nodal planes dip steeply to the south and gently formed higher in the subduction zone by the hydra-
to the north. It is probable that the latter planes tion of olivine. Below 300 km the earthquake mech
represent the faults so that these earthquakes are anism is believed to be the sudden phase change
generated by the relative movements between the from olivine to spinel, known as transformational or
Pacific and Bering Sea lithosphere. The single focal anticrack faultin£i., which takes place by rapid shear
mechanism solution indicative of strike-slip move ing of the crystal lattice along planes on which
ment is probably an arc-arc dextral transform fault minute spinel crystals have grown (Green, 1994).
IS)
Samoa Is.
Samoa Is.
15°S —
Tonga Is.
Tonga-fiji Tonga-Fiji
earthquakes
Focal depth earthquakes
o 0-100 km
Kermadec A ! X 200-300 km
A 100-200 km
Trench < 300-400 km
Water depth
contours in km o 400-500 km
Kermadec rnO-6 ■ >6 km
Trench
L_^ fJmfltSJBif
175°W 175°W
SUBDUCTION ZONES 143
Samoa Is.'
15°S
Fiji Is.
Tonga
20°
± wv
#f-v
Tonga
Trench
V
V
V
•V V
25° - w
Tonga-Fiji
earthquakes
V 500-600 km
in kilometres;(b)intermediate;(c) deep
(redrawn from Sykes, 1966^ Journal of
Geophysical Research 71, with permission 180° 175°W
from the American Geophysical Union). (c)
The change to spinel is probably complete by a depth plate is thrown into either downdip compression or
of about 700 km,explaining the termination of sub- extension. Isacks & Molnar (1969) have suggested
duction zone seismicity at this depth. It is also that the distribution ofstress type in the seismic zone
probable that a similar transformation from enstatite may result from the degree of resistance experienced
to ilmenite contributes to subduction zone seismic by the plate during its descent, and Spence (1987)
ity(Hogrefe et al,^ 1994). has described this resistance in terms of the nett
Unlike the shallow events of regions 'a' and 'b' in eflPect of ridge push and slab pull forces (Section
Fig. 8.9 where the principal stress directions are hor 11.6). In Fig. 8.15a the plate is sinking through the
izontal or vertical, the deep events of region 'c' are asthenosphere because of its negative buoyancy and
characterized by principal stress directions which are is thrown into downdip tension as its descent is
either parallel or orthogonal to the dip of the de unimpeded. In Fig. 8.15b the bottom of the plate
scending plate (Isacks et al.^ 1969)(Fig. 8.14).Con approaches the mesosphere, which resists descent
sequently, the nodal planes determined by focal and throws the leading tip into compression. As the
mechanism solutions do not correspond to the dip of plate sinks farther (Fig. 8.15c), the mesosphere pre
the Benioff zone or a plane perpendicular to it. The vents further descent and supports the lower margin
principal stress directions show that the descending of the plate so that the majority of the seismic zone
144 CHAPTERS
Volcanic arc
Coralline arc
1-
A;vy;.vLN
Water depth - >>v Tonga Trench
(km) 5
+200
200
500
A 8o
ZO
Depth
(km)
550
400 -1r A
A
600
km
600 A
Tonga Tonga
Fiji Islands Lau Ridge Lau Basin Ridge Trench
Hioh Q
Extremely low O
LowQ LowQ
Depth(km)
400
Fig.8.7 Schematic section across the
Tonga arc showing the zone of very
high seismic attenuation beneath the
Lau marginal basin (redrawn from
Barazangi & Isacks, 1971^Journal of
Geophysical Research 76, with
permission from the American 1000 900 800 700 600 500 400 300 200 100
Geophysical Union). Distance from trench(km)
Volcanic arc
3 O O O
50 —
100 - 8f%
Q.
(D
o
150 -
O O
200 —Qo-
142°E
146 CHAPTERS
Lithosphere
KAM. ^^^A^eriosph^ei
w«a:
'•V«%V.V#V.V»*.W.V.V.V»*«%%*A
v.%v.v.%v >v.w.;
.v.v.w.
•.%•••«•••••
vwlm
{.VAV.
K'MwS
i«iS55!
itt r--'?:•:•:%■
%wfSft .%%vK%;
5S5S^g
w.w.ss m v,w>>
'.W.K
>X.X<<<«»555m
Fig. 8.9 Plate model of subduction
zones; a, b and c indicate regions of
distinctive focal mechanism
solutions. See text.
determined from focal mechanism solutions (Isacks of the trench adjacent to Alaska. The unconsolidated
& Molnar, 1971) is shown in Fig. 8.16. nature of these sediments probably prevents any
The stress distributions shown in Fig. 8.15b,d pro build-up of the strain energy necessary to initiate
vide one possible explanation for the seismic gaps earthquakes, and their high positive buoyancy may
observed along the middle parts of the Benioff zone force the subducting plate to descend at an anom
at certain trenches, such as the Peru-Chile Trench alously shallow angle.
(Figs 8.16 and 9.30) (Section 9.5) where it is known Information on the depth reached by the down-
that the slab is continuous (James & Snoke, 1990). A going slab has been provided by seismic tomography
further type of seismic gap appears to be present in (Section 2.1.8), showing it to be dependent on the
some island arcs at shallow depths. Figure 8.17 past history of the subduction zone (van der Hilst,
shows sections through the BeniofF zone at the 1995). In the western Pacific, where the lithosphere
Aleutian-Alaska arc (Jacob et aL^ 1977). There is a is descending at a rapid rate, the slab reaches depths
prominent gap in seismicity between the trench and a of at least 800 km and passes directly through the
point about half-way towards the volcanic arc which base of the mantle transition zone at 670 km (Sec-
becomes progressively greater from west to east. The
angle of underthrusting is very shallow in this region.
The probable cause of this seismic gap and shallow
underthrusting is the presence of copious quantities
of terrigenous sediments within the trench which
Mariana profile
become increasingly abundant towards that section Elastic
Sea level
Tonga profile
Elastic
Elastic-plastic
-100 300 km
Bering Sea
+ 50°
Pacific Ocean
Fig.8.12 {Top) Focal mechanism solutions of Fig.8.13 (Bottom) Distribution ofearthquakes beneath
earthquakes in the Aleutian arc, compressional quadrant northeastern Japan arc. Shaded line is probably the top
shaded (redrawn from Stauder, 196%^ Journal of ofthe descending lithosphere (redrawn from Hasegawa
Geophysical Research 73, with permission from the etal.^ 1978, with permission from the Royal Astronomical
American Geophysical Union). Society).
NE Japan
E 100
Q 150
250
140°E 142°E 144°E
148 CHAPTERS
Trench axis
200 km
Depth(km)
North South
100
500
South/
100
300
km
500
Flores
100
300
/ :
km
500
Fig.8.16 Summary ofthe distribution of downdip neither P- nor T-axis parallel to zone; solid lines,
stresses in Benioffzones. Open circles, events with approximate form ofseismic zone (redrawn from Isacks
compressional axis parallel to dip ofzone; solid circles, & Molnar, 1971, with permission from the American
events with tensional axis parallel to dip ofzone; crosses. Geophysical Union).
transition at about 670 km which may be exo- or have deduced that the downward deflection of iso
cndothermic. therms, and hence the length ofthe seismic zone, is
Different solutions ofthe temperature distribution proportional to both the rate of subduction and the
have been derived by various workers depending square of the thickness of the lithosphere. Litho-
upon the relative contributions of the above phe sphere thickness is proportional to the square root of
nomena. Two such models derived by Toksoz et al, its age (Section 6.9) so that the length of seismic
(1971) and Schubert et al. (1975) are presented in zones should be proportional to the product of
Fig. 8.19. Although differing in detail, all such mod convergence rate and age. That this is generally so is
els indicate that the downgoing slab maintains its illustrated by Fig. 8.20, and although there is con
thermal identity to great depths. Temperature con siderable scatter the data appear to fit the relationship
trasts of 700® C are shown to exist between the slab length (km)= rate (mma"^)xage (Ma)/10. Other
and mantle at a depth of700 km. controls of the geometry of the subducted plate are
The depths to which the subducting oceanic litho- discussed by Cross & Pilger (1982). An alternative
sphere maintains its separate thermal identity control explanation for the absence ofseismicity below about
the length of the Benioff zone. Molnar et at.(1979) 700 km is that the phase changes suffered by the slab
150 CHAPTERS
Alaska
170'
~
- 100
f Shumagin Islands
View N59°E
iNW
. . . . 1 '• ■ . ... 1 . .. 1 . .. . 1 . 200
may produce fine-grained materials which behave in a An abnormally low geothermal gradient of about
superplastic manner and thus cannot generate earth 10®Ckm~^ results from the rapid descent of rela
quakes (Ito & Sato, 1991). tively cool oceanic lithosphere at trenches (Section
8.4) to depths of about 30 km. The high pressures
and low temperatures in this environment give rise to
8.5 Metamorphism at convergent a metamorphic complex characterized by the pres
margins
ence ofglaucophane and jadeite, which are indicative
The anomalous thermal and pressure conditions as of the blueschist facies. Blueschist metamorphism
sociated with subduction zones give rise to distinctive has been observed in the Mariana subduction zone
suites of metamorphic rocks whose disposition de (Maekawa et 1993) and blueschist terrains are
pends on the direction ofunderthrusting(Fig. 8.21). associated almost everywhere with ophiolitic suites
SUBDUCTION ZONES 151
Fig. 8.18 The vertical extent of subduction revealed by suggested that their common margin, the Median
seismic tomography,(a) With no back-arc spreading, the Tectonic Line, has been the locus ofsome 400 km of
slab can directly penetrate to the lower mantle, e.g. strike-slip movement responsible for bringing the
Kermadec arc.(b) Back-arc spreading causes a horizontal Sanbagawa and Ryoke belts into juxtaposition. This
slab deflection at the 670 km discontinuity, e.g. Tonga arc
transcurrent movement has been amply confirmed
(after van der Hilst, 1995).
by detailed mapping (Takagi, 1986). The Hidaka/
Kamuikotu paired belt shows the opposite metamor
(Ernst, 1973)(Section 2.5). The ascent of magmas phic polarity, and must have been formed during a
generated by dewatering-induced mantle melting different phase of plate movements when the direc
(Section 8.8) gives rise to anomalously high geother- tion of subduction was from the west ofJapan.
mal gradients of more than 25°Ckm"'^ up to about Paired metamorphic belts have now been identi
50°Ckm~^ A second metamorphic complex conse fied from many of the subduction zones, of both
quently arises associated with surface volcanism, island ate and Andean type, around the margin ofthe
characterized by minerals such as andalusite, which Pacific, as shown in Fig. 8.24. Full descriptions of
are generated at high temperatures and low pressures many of these belts are given in Frnst (1975). In
(Fig. 8.21). Figure 8.22, from Miyashiro (1973), other subduction zones, however, such as the Phan-
shows the minerals expected to be formed at the erozoic orogenic belts of the Atlantic region, un
temperature and pressure conditions experienced in paired metamorphic belts occur in which one of the
these two metamorphic environments. two belts is not developed or the contrast between
Subduction zones consequently contain paired belts is unclear. Miyashiro (1973) has noted that a
metamorphic belts (Miyashiro, 1972); an outer high very low geothermal gradient is required for blue-
pressure/low temperature type on the oceanward schist grade metamorphism, and if this is not
side and a parallel low pressure/high temperature achieved only medium pressure metamorphism will
belt of similar age associated with the island arc, result (Fig. 8.22). Thereafter, if the rate of plate
which are typically about 100-250 km apart. descent is slow, or if young, hot oceanic lithosphere
Paired metamorphic belts were first recognized in is being subducted, the very low thermal gradient
Japan (Fig. 8.23) where three pairs of different ages necessary for blueschist grade metamorphism will
have been identified. At present Pacific lithosphere is not be achieved and the high pressure belt will be
being subducted in a northwesterly direction be missing. Investigations in California and Japan have
neath the Japan arc, and the metamorphic polarity of shown that in past subduction systems, blueschist
the Sangun/Hida and Ryoke/Sanbagawa paired metamorphism ceased 30-40 Ma before the under-
belts indicates that they were formed by underthrust- thrusting ofan ocean ridge as young, hot lithosphere
ing in this same direction.The Sanbagawa/Ryoke started subducting. The rapid plate motion and rate
belts however, are much closer together than sug of underthrusting of the Pacific lithosphere thus
gested by the plate tectonic model,and so it has been account for the common occurrence of paired meta-
152 CHAPTERS
1200°C -200 km
• 600X
Olivine
1600"C -400 km
Spinel
800°C
— 600 km
Spinel
1700°C 1700°C
Oxides
(a)
- 800 km
300
km
-0
400°C
800°C
1200°C
1600°C
- 200 km
- 400 km
2000°C
- 600 km
2400X
800 km
Japan Sea Japanese Islands Shikoku Shelf Nankai Trough Pacific Ocean
Volcanicity
' Alkaline Calc-alkaline Tholeiitic
Marginal sea Shelf Magmatic arc Forearc Forearc Trench Ocean floor
basins ridge /
Partial melting of
100-1 mantle peridotite
(100-200 km)
r^Zoneof partial
melting of basaltic
Z = zeolite fades magma
PP = pumpellyite-prehnite Asthenosphere
gl = glaucophane-iawsonite
eel = eclogite A
154 CHAPTERS
High pressure
metamorphism
1 Albite + H2O = analcime + quartz
1000
2 Lawsonite +2quartz + HgO = laumontite
3 Aragonite = calclte
800 4 Jadelte + quartz = albite
Pressure 5 Kyanite = andalusite
(MPa) 6 Kyanite = sillimanite
600
7 Sillimanite = andalusite
8 Muscovite + quartz= Kfeldspar + AI2Si05+ H2O
400 9 Onset of granite melting
10 Onset of olivine tholeiite melting
400 km
I I
ijggilHonshu
Sangun belt (High pressure) I Permian -
Hida belt (LOW pressure)
Shikoku
NW Canada^
X\ ^ ^
Kamchatka Washington
^California j
^/A'Taiwan
/7 Celebes
were originally situated in equatorial latitudes, and reaches depths of 2-4 km below the surrounding
that the Japanese islands have undergone a clockwise ocean floor so that its base is 7-8 km below sea level.
rotation since Palaeogene times, when the major axis Trenches are generally 50-100 km in width and in
of the islands was parallel to the Asian coast. Barber section form an asymmetric V-shape with the steep
suggests that northward subduction at an oblique est slope of 8-20® on the side opposite the under-
angle at this time was accompanied by sinistral strike- thrusting ocean. The sediment fill of trenches can
slip movements along the trench. The various Japan vary greatly from virtually nothing as in the Tonga-
ese high and low pressure belts were accreted and Kermadec Trench to almost complete as in the Lesser
brought together during these phases ofmovements, Antilles Trench. All are associated with very large
that is, they represent suspect terrains (Section 9.5). negative free air and isostatic gravity anomalies which
Although the paired metamorphic belt model result from the downward displacement ofthe litho
appears reasonable, this new synthesis for the type sphere so that mantle material is displaced by either
example in Japan questions its validity (Windley, sea water or low density sediments.
1984). It is necessary to await detailed investigations In section the morphology of the sub-sedimentary
of other paired belts before a decision can be made basement of all trenches is very similar (Fig. 8.25),
about its general applicability. and is independent of the amount of sediment
present. Consequently, it is inferred that the sedi
ment loading plays no role in the formation of
8.6 Ocean trenches
trenches. This sediment, however, may be deformed
Oceanic trenches are the direct manifestation of by the underthrusting to form the accretionary
underthrusting oceanic lithosphere and are devel prism.
oped on the oceanward side of both island arcs and
Andean-type mountain ranges. They represent the
largest linear depressed features of the Earth's sur
8.7 Accretionary prism
face, and are remarkable for their depth and continu The accretionary prism (first arc, subduction com
ity. The Peru-Chile Trench is 4500 km long and plex, accretionary wedge)forms on the inner wall of
(Jl
O)
O
X
>
"0
3J
00
—I 0
10 10
100
km
Vertical exaggeration: 50
Fig.8.25 Stack of 35 projected topographic profiles ofPacific subduction zones(redrawn from Hayes & Ewing, 1970, with permission from John Wiley).
SUBDUCTION ZONES 157
an ocean trench (Moore & Silver, 1987). The inter rocks encountered at structurally lower levels. Seeley
nal structure and construction have been deduced (1979) and Cowan & Silling (1978) have suggested
from seismic reflection profiles and sampling at that, in addition to underplating, mass flow may
active subduction zones and by the study of ancient occur within the prism. Material may be transported
subduction complexes now exposed on land. downward by the underthrusting and subsequently
Karig (1974) and Karig & Sharman (1975) have flow upwards on reaching the lowest part of the
proposed that accretionary prisms develop when trench (Fig. 8.27). Cowan & Silling (1978) have
trench-fill turbidites (flysch), and sometimes also constructed a scale model demonstrating that this
pellagic sediments and underlying oceanic crust, are form of uplift is feasible, and have also modelled the
scraped from the descending oceanic plate by the mechanics theoretically in terms of the two-
leading edge of the overriding plate to which they dimensional flow of a viscous fluid round an acute
become accreted. The internal structure ofthe accre corner in response to sliding a rigid plate beneath
tionary prism is consequently made up of a series of another fixed plate. Davis et al.(1983) have likened
imbricate listric thrusts dipping towards the arc the overall mechanics in an accretionary prism to
which define a series ofwedge-shaped pockets within those of a wedge of soil scraped in front of a bull
which are developed complex folds verging towards dozer. The material in the prism deforms internally
the trench (Dickinson,1977)(Fig. 8.26). As subduc and steepens its slope until a certain critical taper
tion continues, older thrust wedges are gradually between the top and base of the wedge is attained,
moved upwards and rotated arcwards by the addition whereupon it slides stably and continues to grow at a
ofnew wedges to the base ofthe prism (Platt, 1986). constant taper as new material is added to its leading
The older thrusts consequently become more steeply edge. Indeed, with the addition of new sedimentary
dipping with time. This mechanism has been termed material at its edges, the prism will deform internally
'offscraping', 'plastering', 'underplating' and 'sub- until it reaches a stable configuration in which the
cretion'and is responsible for the outward growth of internal gravitational forces ofthe wedge balance the
the prism. This model requires that the most intense traction on its base from the subducting slab (Platt,
deformation occurs at the oceanward base of the 1986).
sedimentary pile, although the older thrusts may Karig (1982) has suggested that the trajectory of
remain active during their rotation. During this de materials within the subduction complex is influ
formation a forearc basin may develop between enced by the thickness and mechanical properties of
trench and island arc in a trough resulting from the the sedimentary cover on the subducting plate, the
inability of sediment sources to keep pace with basement relief on this plate and the convergence
forearc development. The forearc basin covers the rate. Thick sedimentary cover gives rise to large
oldest members of the accretionary wedge. In addi subduction complexes, although the proportion ac
tion, small sedimentary pockets may develop on top tually accreted to the overriding plate varies signifi
ofthe thrust wedges. cantly and is controlled by rheological and porosity
This model is in accord with observations made on variations within the sediments. Accretion depends
ancient subduction complexes (Ernst, 1975) in upon the nature and geometries of the arriving sedi
which sediment age and metamorphic grade increase ment (Fig. 8.28). Relatively thick trench turbidites
from trench to arc and major thrust faults dip to overlying a thin pelagic cover,such as in the Cascade
wards the arc. The oldest, high grade, rocks are arc, allow thrusting to take place at the base of the
structurally higher and have been uplifted with re turbiditic sediments (Fig. 8.28a). At the Makran arc,
spect to younger deposits. These older rocks, which however, where the sediment carried to the trench
may have formed at depths as great as 30 km, have may be in excess of 6 km thick, only the upper
been translated upwards and may be juxtaposed with 2-2.5 km are accreted (Fig. 8.28b). Where the sedi
much younger, lower grade rocks. ment cover is thin, such as the Marianas arc, the
The uplift of high grade rocks cannot be fully situation at the trench is less certain. An increased
explained by continued movements along thrust proportion ofoceanic lithosphere may participate in
faults. It is possible that the high grade rocks rise the accretion, and this would be enhanced by relief
buoyantly (Ernst, 1975), but this seems unlikely on the oceanic basement surface (Fig. 8.28c) due to
as their density would be higher than the younger the normal faulting resulting from downbending of
158 CHAPTERS
Volcanic arc
Trench
\1
Oceanic crust -
!w.v.?.v.v
»»!W555553
•Krw.v.%x»x*:<.:
Subduction complex
»:*K^K?S
Lithosphere
Asthenosphere
<»M%w5
IwvXvIvI
.%%%%%%" >>>»!•;
WM
5>>>S'
.•.w.v
wtwl
100 km
Oceanic crust
Fig.8.26 (a-c) Idealized sequence ofthe evolution ofan lines, inactive thrusts,(d)Schematic, vertically
accretionary prism (redrawn from Dickinson, 1977, with exaggerated section ofan accretionary wedge (redrawn
permission from the American Geophysical Union):(a) from Cowan & Silling, 1978, with permission from the
incipient stage;(b)forearc basin;(c)full forearc basin. American Geophysical Union).
Solid lines in subduction complex, active thrusts; broken
SUBDUCTION ZONES 159
sediments at depth, and this requires the decolle- Pichon et al.^ 1987) has revealed the presence of
ment zone to possess a lower fluid pressure than its unique benthic communities which exist by chemo-
surroundings. This apparent conflict can be recon synthetic processes along the active tectonic zone at
ciled by a model in which the fluid pressure in the the leading edge of the prism and whose shells
decollement zone varies cyclically. During periods of contain carbonate derived from methane.
strain accumulation there is an interval in which fluid Knowledge of the structure and models of evolu
flows towards fractures developed in the decollement tion of subduction complexes have allowed new
zone. Eventually the fractures join up as fluid pres interpretations to be made of ancient complex sedi
sure increases, and failure and slip take place on the mentary terrains in terms of subduction-related pro
decollement zone. As this takes place fluid is expelled cesses. One such terrain is the Southern Uplands of
along the decollement zone as the connected frac Scotland (Fig. 8.29a)(McKerrow et al.^ 1977; Leg-
tures offer a higher permeability route than the gett et«/., 1982; Leggett, 1987). This area has been
source sediments. There is then a period of recovery interpreted in terms of an accretionary wedge devel
as deformation of the source sediments increases oped at a subduction zone. In late Ordovician times,
their pore pressure to the point at which fractures are Laurentia (early Palaeozoic North America, includ
again developed and the cycle recommences. Fluid ing Scotland) was separated from Europe by the
expulsion thus takes place by a pumping mechanism lapetus Ocean (Fig. 8.30). lapetus subsequently be
(Moore, 1989), and is manifest in the development gan to close as the result of underthrusting at the
of mud volcanoes and diapirism (Westbrook et al.^ southeast margin of Laurentia. A subduction com
1984). plex grew here by sediment accretion which was
The outward flow of water from accretionary partly overlain by a foreland basin on its landward
prisms can explain many of the phenomena associ side. Laurentia eventually collided with Europe in
ated with them. Reck (1987) has reported thermal late Silurian times, giving rise to a continent-
gradients from DSDP holes in the prism ofnortheast continent collision represented in the stratigraphic
Japan of24-36°C km"^.Ifextrapolated downwards, record as the Caledonian Orogeny, and welding
temperatures in excess of400®C would be expected Scotland to England. The subduction complex is
at 15 km depth, taking into account the increasing now preserved as the Southern Uplands and the
thermal conductivity with depth associated with de forearc basin as the Midland Valley of Scotland
creasing porosity. These temperatures are in conflict (Fig. 8.29a). Figure 8.29c shows a reconstructed
with the temperatures of about 150®C predicted section ofthis area in Wenlock times, and its similar
from conductive thermal models and studies of the ity to models of mdoern subduction complexes is
metamorphic rocks in ancient subduction com readily appreciated.
plexes. These data can be reconciled by the upward
flow of water in the prism decreasing the thermal
gradients to much lower than those at the surface
8.8 Volcanic and piutonic activity
below a depth ofa few kilometres. Dewatering ofthe Where subducting oceanic lithosphere reaches a
Barbados prism at the decollement zone creates a depth in excess of 80 km, an island arc is created at
unique chemical environment in which clay minerals the surface by volcanic and piutonic activity 150-
of distinctive character are formed (Schoonmaker, 200 km from the trench axis. The majority ofisland
1986). Dewatering also produces characteristic au- arcs are situated along the western and northern
thigenic carbonates derived from the methane-rich margins of the Pacific, but are also located in the
pore waters ofthe Washington-Oregon accretionary western Atlantic at the Caribbean and Scotia arcs. At
prism (Ritger et al., 1987). The methane, which is the eastern edge ofthe Aleutians an island arc gradu
dominantly of biogenic origin,is carried to the upper ally transforms into an Andean-type orogenic belt off
sediments of the prism by fluids and oxidized by Alaska (Fig. 8.17).
sulphate reducers before being incorporated into a Relatively young island arcs are structurally simple
carbonate cement of magnesian calcite, dolomite and are underlain by a crust of less than 20 km
and aragonite at the sites of pore water expulsion. thickness. These include the arcs of Tonga-
Direct observation by submersible ofthe accretionary Kermadec, the New Hebrides, Aleutians and Lesser
prism of the Nankai Trough off southern Japan (Le Antilles. Older, mature island arcs are more complex
SUBDUCTION ZONES 161
Devonian granites
Upper Palaeozoic
Caradoc
Site of Southern
Upland Fault
NW SE
Sea level
(b)
Wenlock
Site of Midland
Valley inliers Site of Southern Site of
Upland Fault Solway Firth
Cockburnland
Fig. 8.29 (a) Map of the Southern v:-:: Sea level
Uplands of Scotland showing the
locations of the seven sequences of Present
(b) and (c); (b) Southern Uplands level of
accretionary prism in Caradoc erosion
times; (c) Southern Uplands
accretionary prism in Wenlock (c)
times (redrawn from McKerrow et
V.E.X2
al.^ 1977, with permission from 10 km
Nature 267, 237-9. Copyright © Northern belt Central belt Southern belt
1977 Macmillan Magazines Ltd).
162 CHAPTERS
Carbonates
-z'zi CJastics
E3 Volcanics
Land
Trench
VVV V
vvvvvvvv
1000 km
I
lapetus Ocean
Fig.8.30 Possible configuration ofthe lapetus Ocean tholeiites have been interpreted as being derived by
and bordering continents in late Ordovician times the fractional crystallization of olivine from a pri
(redrawn from Leggett etal., 1982,in Trench-Forearc mary olivine tholeiite magma originating at relatively
Geolqgy^ pp. 495-520, with permission from the shallow levels of 80-120 km. The calc-alkaline and
Geological Society).
alkaline series are encountered in more mature arcs,
and are believed to be derived from magmas gener
ated at increasingly greater depths. Indeed, in some
as they were constructed upon previous generations mature arcs a compositional trend is apparent, with
of subducting plate margin products. They are gen increasing distance from the trench, of tholeiite-
erally underlain by thicker crust of 20-35 km thick calc-alkine-alkaline volcanic rocks, which represent
ness, and include the Japanese and Indonesian arcs. magmas derived from increasingly greater depths
Three series of volcanic rocks occur in the island (Fig. 8.21). There is also a systematic variation in
arc environment(Baker, 1982): abundance, and in elemental and isotope ratios
1 The low potassium tholeiitic series, which is dom (Fig. 8.31), which correlates with the depth to the
inated by basaltic lavas associated with lesser volumes Benioflfzone.In mature island arcs plutonic rocks are
ofiron-rich basaltic andesites and andesites. exposed which represent the residua of magma cham
2 The calc-alkali series, dominated by andesites bers which have crystallized at depth. They are gener
(Thorpe, 1982) which are more enriched in potas ally of granodioritic composition and exhibit similar
sium, other incompatible elements and light rare variations to the volcanic rocks.
earth elements. In Andean-type belts dacites and In the past there has been controversy over the
rhyolites are more abundant. origin of the magmas supplying island arc volcapic-
3 The alkali series, which includes the subgroups of ity. Certainly the generation ofthe magmas must be
alkaline basalts and shoshonitic lavas. linked to the BeniofF zone, as there is a very strong
The tholeiitic series is found in young arcs. The correlation between its depth and the systematic
variation in volcanic rock composition and elemental at depths of 100-200 km producing acidic magmas.
abundances. It was originally believed that the mag These react with the overlying pyrolite of the asthe
mas were derived from the melting of the top of the nosphere forming diapirs which rise and undergo
descending oceanic slab. However, the volcanic partial fusion. During their ascent they fractionate to
suites of island arcs show little variation in spite of produce the calc-alkaline series of andesite, dacite
great differences in geographic location and tectonic and rhyolite typical of mature island arcs. For a more
environment, so it appears that these suites must detailed account of island arc petrogenesis, see
develop from a parent with a ubiquitous, common Wilson (1989).
composition. This, petrological and mineralogical Ringwood's models thus imply that island arc
evidence (Arculus & Curran, 1972) and helium magmas originate from both the top ofthe descend
isotope ratios (Hilton & Craig, 1989)strongly indi ing plate and the overlying asthenosphere. The resid
cate that the parental magmas originate largely by ual asthenosphere has been irreversibly differentiated
partial melting of the asthenospheric mantle imme and will never again be able to participate in basaltic
diately overlying the descending plate. However magma generation at ocean ridges. Ringwood esti
Karig & Kay(1981),Davidson(1983)and Hilton & mates that 30-60% of the asthenosphere has been
Craig(1989),among others, have demonstrated that affected in this way.
certain isotopic ratios require a large contribution
from continent-derived sediments. Consequently,
sediments from the trench must be carried down the
8.9 Marginal seas
subduction zone and incorporated into the asthen Marginal seas (back-arc basins) are small ocean basins
ospheric melt(Plank & Langmuir, 1993). lying on the inner, concave, sides of island arcs,
A major problem is the source ofthe heat required bounded on the side opposite the arc by a back-arc
for melting the asthenosphere above the descending ridge ('third arc', remnant arc). They are most com
slab. It was originally believed that this was derived mon behind the island arcs ofthe western Pacific, but
by shear heating at the top of the slab. This is are also found in the Atlantic behind the Caribbean
unlikely, however, as the viscosity of the astheno and Scotia arcs. Tarney & Windley (1981) have
sphere decreases with increasing temperature, and at presented evidence for their existence throughout
the temperatures required for partial fusion the asthe the geological record from Archaean times.
nosphere would have such a low viscosity that shear The crust beneath marginal basins is similar in
melting could not occur. Ringwood (1974; 1977) composition to that ofthe ocean basins, although in
has therefore suggested that partial melting takes some cases layer 1 is somewhat thicker as would be
place at a relatively low temperature because of the expected in a small, enclosed basin. Several genera
high water vapour pressure resulting from the dehyd tions of marginal basin may occur behind any given
ration of various mineral phases. Indeed, the greater island arc, and Karig (1971) has classified them
the amount of water present, the more the melting according to their heat flow (Fig. 8.33) in terms of
temperature of the mantle is reduced (Stolper & active marginal basins with high heat flow, inactive
Newman, 1994). basins with high heat flow and inactive basins with
At depths shallower than 100 km amphibolite normal heat flow. Active basins always occur imme
transforms to eclogite and releases water (Fig. diately adjacent to island arcs. The high heat flow
8.32a). The water rises into the asthenosphere wedge experienced in these basins (Watanabe et aL^ 1977)
overlying the Benioffzone causing partial melting of correlates with the region of very low (g,in the upper
the pyrolite and the production of tholeiitic basalt mantle (Fig. 8.7) and supports the interpretation of
magmas at depths of 80-100 km. These magmas this region as a zone of partial melting.
fractionate as they ascend and give rise to the basaltic Karig (1970) has interpreted marginal basins as
andesite and andesite magmas typical of the tholeit- developing in response to tensional tectonics,
tic suite present in young island arcs. whereby an existing island arc was rifted along its
At depths in excess of 100 km water is released by length and the two halves, corresponding to the
the dehydration ofserpentinite bodies in the oceanic volcanic and third arcs, separated to give rise to the
crust (Fig. 8.32b). The resulting high water vapour marginal basin. This model was deduced from
pressure causes partial melting of the quartz eclogite the Tonga-Kermadec arc(Fig. 8.34)which is flanked
164 CHAPTERS
Island arc
tholeiite series
crust
\ '\ / / //
Partial \\-\20'.,a
melting of}'///'
pyrolite
^;s ^/
< ^ ^£Sgy -- - -/:=«
Eclogite / ^
Calc-alkaline
volcanics
Peridotitic
lithosphere
/ ^ ^
Serpentinite^
/
— —' Y';/>-SegregatiQn and /
/ / / ^
— — T JZ/VA-differentiation of
— — —^ . calc-alkaline
-magmas — Amphibolite ^
— 7 / / / / /
Diapiric uprise \ / / / / /
and partial
^
melting of .
pyroxenite ^\\
'\
\
/ ^ ^ ^ ^ ^^
/
High PhjO
Wet garnet 1
pyroxenite'
/>
Fig.8.32 (a) Early phase of development of an island arc. London^ 1974, with permission from the Geological
(b)Later phase ofisland arc development(redrawn from Society).
Ringwood, 1974,in Journal ofthe Geological Society of
SUBDUCTION ZONES 165
- 60°N
Aleutian Basin
f ' •«.
"Kamchatka Basin
Bowers Basin
Okhotsk Basin
Shikoku Basin
Sulu Basin
Celebes Basin
Bismarck Basin
Solomon Basin
^Coral Basin
^i. j.
Fiji Plateau
New Hebrides
South Fiji
Basin
- 30°S
7 Lau-Havre
^ Trough
i
Tasman Basirf
Pacific
3 Contour interval
2000 m Fig.8.34 Bathymetry of the Tonga-
Kermadec region (after Karig, 1970,
with permission from the American
170" 175" New Zealand 180' Geophysical Union).
on its landward side by the Lau-Havre Trough and Tamaki(1985) has noted that the initial rifting of
Lau-Colville Ridge. Evidence for extension comes a back-arc basin takes place at the island arc, as this is
from: the asymmetric cross-section of both arc and the most ductile part of the system. The dip of the
ridge, which are mirrored across the centre of the subduction zone controls the nature ofthe split, with
marginal basin, and the fact that the basin's topo a single rift forming within the narrow volcanic zone
graphy is aligned parallel to both; the considerable associated with a steeply dipping Benioffzone, and a
sediment thickness present to the seaward side ofthe multirift system forming in the wider zone associated
arc and landward side ofthe ridge and the absence of with shallow angle subduction.
sediment within the basin; and the continuation of The oceanic nature of the crust underlying mar
the arc-basin-ridge system to the south which corre ginal basins implies that it is generated in a manner
lates with the Taupo volcanic region of the north similar to that occurring at the ocean ridges, and
island of New Zealand, a zone of active extension. older theories involving the thinning of continental
Further support for this model comes from the crust have been abandoned. Indeed, a magma cham
subsidence generally experienced by remnant arcs ber beneath the Lau Basin has been identified on
after the development of marginal basins, as their seismic reflection projects (Collier & Sinha, 1992).
dynamic support is removed. However, structures corresponding to an accretive
SUBDUCTION ZONES 167
margin are not commonly present within marginal pressional events are rare. Figure 8.35d shows the
seas. Moreover, magnetic lineations are not devel formation of a marginal basin by a step-back in the
oped in all marginal basins (Weissel, 1981). Those location of underthrusting. The cause of this behav
lineations that are present can usually be correlated iour is unknown,but its effect is to isolate a fragment
with the magnetic polarity timescale (Section 4.6), ofunderthrusting oceanic lithosphere and to accrete
although they tend to be oflower amplitude and less it to the leading edge ofthe overriding plate. Dvorkin
clearly defined than oceanic anomalies. These data etal.(1993)have suggested that back-arc basins only
support the proposal that marginal crust formation form behind trenches where narrow slabs are sub
takes place over a relatively broad, diffuse zone of ducted, that is, slabs which are restricted in width in
volcanic and plutonic activity. the direction parallel to the trench. Such slabs allow
Although some workers have suggested that cer the ingress of asthenosphere above the slab from
tain marginal basins have been constructed from both the end and sides ofthe slab, so overcoming the
leaky transform fault systems, four possible mecha hydrodynamic suction believed to support wide
nisms (Karig, 1974) have been put forward to ex slabs. An alternative model,unrelated to subduction,
plain the evolution of the majority of marginal seas has been proposed by Miyashiro(1986)for marginal
(Fig. 8.35). Karig (1971) has suggested that mar seas in the western Pacific. He suggested that the
ginal sea extension has resulted from the forcible spreading of these back-arc basins was related to the
intrusion of basaltic mantle diapirs generated by northward migration of a hot region in the mantle
partial melting at the top of the descending slab from Australia in the Late Cretaceous to its present
(Fig. 8.35a) in the same manner as magmas which location around north China. In this way Miyashiro
erupt at the volcanic arc (Section 8.8). Other work explains the general northerly younging of the mar
ers, such as Packham & Falvey (1971), have sug ginal seas, their location and short duration of back-
gested that this magma upwelling is passive rather arc spreading. He cannot explain, however, the
than active (Fig. 8.35b), and occurs in response to presence of up to three generations of basins behind
other factors which have caused regional extensional any given island arc.
stresses in the back-arc region. These factors include The mechanism shown in Fig. 8.35d can be iden
the trench suction force (Section 11.6) which some tified as having occurred behind island arcs by dating
workers, such as Chase (1978) and Fein & Jurdy the remnant arc or marginal basin crust. If these
(1986), believe originates from a 'roll-back' of the features are older than the frontal arc, a step-back in
trench. Such a roll-back is believed to occur in sub- underthrusting has probably occurred. If, however,
duction systems where the 'absolute' direction of the back-arc structures are contemporaneous with
movement of the overriding plate is away from the the arc,some continuous process in this region must
trench. Carlson & Melia (1984) have calculated the be active. Both the Bering Sea and eastern Caribbean
rates of roll-back in a hotspot reference frame, and Sea in all probability formed by a step-back in under
have found values of 100 mm a"^ at the Tonga arc, thrusting. The Bering Sea crust exhibits magnetic
50-70 mm a"^ at the South Sandwich arc,zero at the lineations typical ofocean basins, but these strike at a
New Hebrides arc and negative values, that is, an high angle to the trend ofthe Aleutian arc(Cooper et
advancing trench, at the Izu-Bonim Trench. Hsui & 1977). The lineations can be dated and related to
Toksoz (1981) and Jurdy & Stefanick (1983) have anomalies in the Pacific to reconstruct the events
suggested that marginal sea extension is a result of preceding the formation of the Bering Sea. This
secondary convection cells in the mantle wedge over technique and other geological evidence suggest that
lying the Benioff zone which have been induced by the present plate subduction at the Aleutian Trench
the descent of the underthrusting slab (Fig. 8.35c). occurs 400-1000 km farther south than in Mesozoic
Such convection is feasible provided that the upper times, and that the Bering Sea was formed by the
mantle viscosity is low and/or the superadiabatic initiation of a new subduction zone in Late Creta
temperature gradient high (Section 11.5.2). How ceous times(Section 9.5). The Caribbean case will be
ever, there is little geological evidence of the crustal discussed in Section 8.10.
compression that would be produced either side of More commonly, marginal basins appear to have
the median line of marginal seas by this mechanism, formed contemporaneously with underthrusting at
and focal mechanism solutions indicative of com- their parental island arc, and the correlation of active
168 CHAPTERS
Lithosphere
Mesosphere
^r-S-:
'
== w
Remnant arc
A
Pacific Plate
Indo-Australian Plate
Fig. 8.36 Schematic model of deformation in the region known active deformation; diagonal ornament, stable
between the New Hebrides and Tonga arcs. Heavy arrows, areas (redrawn from Hamburger & Isacks, 1988, with
relative motion between Pacific and Indo-Australian permission from Nature 332, 599-604. Copyright ©
plates; open arrows,'pull-apart basin'; shading, areas of 1988 Macmillian Magazines Ltd).
marginal basins with underlying regions of very high southwestern Pacific in the region behind the island
attenuation of seismic waves (Section 8.3) provides arcs of the New Hebrides and the Tonga Islands, and
support for models which invoke either the active or have found that the deformation is diffuse and dom
passive intrusion of diapirs into the back-arc region. inated by strike-slip mechanisms. This is very dif
However, the criticism of such models is that the ferent from the linear earthquake belt with normal
zone ofextension apparently remains in the centre of faulting mechanisms that would be expected if the
the marginal basin and migrates away from the volca back-arc basin developed in a manner similar to the
nic arc. The cessation of activity in a marginal basin, ocean basins at accretive plate margins. Volcanism is
and the development of successive generations of similarly diffuse in this region, and the concept of a
marginal basins, may be related to times when the single spteading centre is not supported by the sea
whole of the mantle wedge above the Benioff zone floot bathymetry and magnetic anomaly data. Ham
had undergone partial melting and the supply of burger 8c Isacks consider that this back-arc region has
materials for the creation ofnew back-arc lithosphere developed in response to the opposite directions of
had been exhausted. underthrusting at the Tonga and New Hebrides
It is probable that certain marginal seas have trenches (Fig. 8.36), in which deformation takes
evolved in response to specific configurations ofplate place by shearing and localized crustal generation at
boundaries in their vicinity. Hamburger 8c Isacks irregular spreading centres of short duration in the
(1988) have compiled earthquake data from the whole region between the island arcs. The entire
interarc region is viewed as a large, internally de
formed pull-apart basin (Section 7.4.3)formed along
Fig. 8.35 (Facingpage) Models for the formation of
a 700 km long sinistral offset of the transform fault
marginal seas: (a) active diapirism generated by Benioff
zone;(b) passive diapirism from regional tensional
boundary between the Pacific and Indo-Australian
stresses;(c) subsidiary convective circulation;(d) step-back Plates. It is possible that other back-arc basins which
in underthrusting. No vertical exaggeration (after Karig, show no evidence of a discrete spreading centre and
1974,Annual Review ofEarth and Planetary Sciences 2, lie in a region of complex plate boundaries, such as
Copyright © 1974 by Annual Reviews Inc.). the Tyrrhenian Basin and Sea of Japan, developed by
170 CHAPTERS
diffuse spreading of a similar nature. A major prob East Pacific rise (a)
lem is the reason why marginal seas have developed Pacific West CoastTrench
preferentially in the western Pacific and not in the Plate Farallon Plate \ American Plate
east. The answer may lie in the rates at which the \ \-/' / N-,'.' t
Marginal basin
Island arc
Remnant arc
Chile' Patagonia
Rise Soi^th Georgia
\ Dyk^orientation Tierra del Fuego
Drake
Andesitic volcanoclastid' Passage
100 km s '/! South Sandwich
arc
Antarctic Peninsula
Fig.8.38 Reconstruction ofthe Rocas Verdes complex of merge at the island of Guadeloupe, and to the south
southern Chile in Early Cretaceous times (redrawn from the islands consist of volcanics on the western flank
de Wit & Stern, 1981, with permission from Elsevier). and sediments on the eastern. In the south the oldest
rocks are ofEocene age, while in the north the oldest
rocks on the volcanic arc are Oligocene and the
transform faults along its northern margin and to the sedimentary islands are underlain by rocks ofEocene
east ofthe Lesser Antilles. age. There is no evidence anywhere ofa substructure
The topographic expression ofthe trench is largely older than Eocene in age.
obscured by a thick sediment fill derived from the A back-arc basin, the Grenada Trough, flanks the
Orinoco River of Venezuela, but its presence is inner side of the island arc, and is bounded to the
indicated by a belt of negative free air and isostatic west by the Aves Ridge (Kearey, 1974), an almost
gravity anomalies (Kearey a /»/., 1975). These sedi totally submarine bathymetric prominence. The east
ments have been thrust and deformed into a large ern edge ofthe Aves Ridge is noticeably arcuate, with
accretionary wedge over 20 km thick known as the a radius ofcurvature very similar to that ofthe island
Barbados Ridge(Westbrook,1982;Westbrook et al., arc. The ridge is of volcanic origin, and samples of
1984), which emerges at the island of Barbados igneous rocks recovered by dredging and drilling
(Fig. 8.40). Between this ridge and the island arc lies provide ages no younger than Cretaceous and Palae-
a forearc basin of tranquil deposition known as the ocene. The Aves Ridge consequently has many ofthe
Tobago Trough. The island arc stretches from Som characteristics of a remnant island arc. Moreover, in
brero in the north to Grenada in the south, and has a order to explain the sequential nature of volcanism
history ofcalc-alkaline volcanism since the Eocene. It from the Aves Ridge to Lesser Antilles, it probably
exhibits the very small radius ofcurvature of400 km. originated by a step-back in underthrusting
In the north the arc is diflferentiated into an outer (Fig. 8.3Sd) rather than by any process of back-arc
sedimentary arc and an inner volcanic arc. These spreading.
172 CHAPTERS
Volcanic islands
Sombrero
I I Sedimentary islands
18°N
100 km
St Kitts
Guadeloupe
16°N
6
Dominica
Martinique
l!)
14°N
St Lucia
St Vincent Barbados
Grenada 12°N
100 200
I I
km
Fig.8.40 Section through the eastern Caribbean (after Hasemi,A.H.,Ishii, H.& Takagi, A.(1984)Fine structure
Westbrook, 1982,in Trench-Forearc Geolq0y, pp. 275-90, beneath the Tohoku district, northeastern Japan arc, as
with permission from the Geological Society). derived by the inversion ofP-wave arrival times from local
earthquakes. Tectonophys. 101, 245-65.
Karig, D.E. (1971) Origin and development of marginal
basins in the western V2X\^c.J.£feophys. Res. 76,2542-61.
Karig, D.E.(1974)Evolution of arc systems in the western
Further reading Pacific. Ann. Rev. Earth planet Sci. 2, 51-75.
Baker,P.E.(1982)Evolution and classification oforogenic McKerrow, W.S., Leggett, J.K. 8c Eales, M.H. (1977)
volcanic rocks. In Thorpe, R.S. (ed.) Andesites^ pp. 11- Imbricate thrust model of the Southern Uplands of
23, Wiley, New York. Scotland. Nature 267, 237-9.
Barazangi, M. & Isacks, B. (1971) Lateral variations of Miyashiro, A. (1973) Paired and unpaired metamorphic
seismic wave attenuation in the upper mantle above the belts. Tectonophys. 17, 241-54.
inclined earthquake zone of the Tonga island arc: deep Molnar,P., Freedman, D.& Shih, J.S.F.(1979)Lengths of
anomaly in the upper mzntit.].£feophys. Res. 76,8493- intermediate and deep seismic zones and temperatures in
516. downgoing slabs oflithosphere. Geophys.J. Roy. astr. Soc.
Barber, A.J.(1982)Interpretations ofthe tectonic environ 56,41-54.
ment ofsoutheast Japan. Proc.£ieol. Ass. 93,131-45. Moore, J.C.(1989) Tectonics and hydrogeology of accre-
Hasegawa, A., Umino, N. & Takagi, A. (1978) Double- tionary prisms: role of decollement zone.J. struct. Geol.
planed seismic zone and upper mantle structure in the 11,95-106.
northeastern Japan arc. Geophys. J. Roy. astr. Sac. 54, Ringwood, A.E. (1974) The petrological evolution of
281-96. island arc systems.J.^eol. Soc. Lond. 130,183-204.
9/ Mountain ranges
174
MOUNTAIN RANGES 175
New
Australia TasmanSea Hebrides Pacific Ocean Ar^s South America Atlantic Ocean
Continental
(a)
crust
Oceanic
crust
Subcrustal
lithosphere
Indian Ocean India Tibet
Asthenosphere
(b)
Fig.9.1 (a) Section from Australia through the Pacific to gravity data, which show that the base of the crust is
the Atlantic Ocean illustrating island arc and Andean plate at 12 km below sea level beneath the ocean, at 76 km
margins,(b)Section from the Indian Ocean to Tibet under the Altiplano and probably at 40 km beneath
across the Himalayan mountain range (redrawn from the Brazilian Shield.
Dewey & Horsfield, 1970, with permission from Nature
225, 521-5. Copyright © 1970 Macmillan Magazines
Ltd). 9.2.3 Seismicity of the Andes
The general pattern of seismicity and focal mechan
The volcanosedimentary trough of the Western ism solutions of earthquakes are in accord with the
Cordillera includes the massive Coastal Batholith. easterly subduction of the Nazca plate beneath the
This is a multiple intrusion of granite, tonalite and Andes(Molnar & Chen, 1982). However,a detailed
gabbro (Cobbing & Pitcher, 1972) extending for assessment of the most accurately located events
some 1600 km. It consists of more than a thousand (Barazangi & Isacks, 1979) indicates that the dip of
interlocking plutons intruded into the andesitic vol the BeniofFzone is not constant(Fig. 9.4). In central
canic rocks(Cobbing, 1982) via a persistent series of and northern Peru the seismic zone dips shallowly at
deep-seated fractures. Drake et aL (1982) have esti about 10" while in southern Peru the dip is 30". The
mated that the Peruvian continental margin has transition between these two areas must conse
increased in width by up to 200 km over the last quently be represented by a tear zone or distorted
200 Ma by the emplacement ofsuccessive plutons. downgoing plate which would allow subduction to
The Andean Cordilleras are flanked to the east by take place at different angles. Barazangi & Isacks
the sub-Andean zone, which consists of continental (1979) and Jordan et aL (1983) have noted that the
sediments deposited on the Brazilian Shield, and to dip of the BeniofF zone appears to control the topo
the west by the Precambrian Arequipa Massif, which graphy and geology ofthe overlying areas. Above the
may make up a large part ofthe Peruvian continental shallow dipping plate there is no significant Quater
shelf. The boundaries with the older rocks are prob nary and Neogene volcanism, there is a steady in
ably deep faults originating during periods of exten crease in topography from the coast to the crest of
sion (Myers, 1975) which were probably caused by the Andes and shallow seismicity occurs over a broad
the trench suction force (Section 11.6). region. Above the more steeply dipping plate there is
Figure 9.3 illustrates the generalized structure of Neogene and Holocene volcanism, a longitudinal
the region. Controls on the deep structure come valley is developed between the Eastern and Western
from the interpretation of seismic refraction and Cordilleras which contains the deposits of the
176 CHAPTER 9
Eugeosynclinal
volcaniclastics
Miogeosynclinal
elastics
Intrusives
Altiplano molasse
Ordovician-Devonian
Dm flysch trough
Pre-Ordovician
schist belt
Precambrian craton
"- Moho
Tertiary volcanic
81°W
S60W 13030's
\ 550 ^ ^ _ 1100km _
S 200-
300 4 A
Dominant surface
■*■ compression •*-
+5j
- -sc
(D E o4
VT
^ 51
I 100-
g- 200-
Q
1100 km
E
100
Q.
a>
200
Q
300
Dominant surface
— extension — Fig. 9.4 Cross-sections along profiles
A and B, shown in Fig. 9.2, showing
earthquake foci and interpretative
models. In the models, the dashed lines
— 100-
under the Moho (M) show the
-c 200-
approximate position of the base of the
o. South American Plate (redrawn from
0>
Q 300- Megard & Philip, 1976, with
permission from Elsevier).
stress generated by magma intrusion caused folding resulted from the intrusion of magmatic products
and thrusting of the Altiplano and Eastern Cordillera from this region. In addition, Molnar & Chen
(Fig. 9.5d). During Cenozic times the Altiplano re (1982) suggest that thickening has also arisen from
ceived some 15 km of molasse from Eastern and crustal shortening by thrusts, and infer that the
Western Cordilleras. During late Quaternary, the present seismic activity indicates that this process is
igneous activity diminished, compressive stresses
relaxed and graben formed within the Altiplano,
probably in response to the trench suction force
Fig. 9.5 (Facingpa^e) Schematic sequence of the
(Section 11.6). evolution of the Peruvian Andes. M, Moho; I, presumed
The crustal rocks of the Andes are thus interpreted isotherm at which melting occurs in the lithosphere
as deriving from the melting of the downgoing slab (redrawn from James 1971, with permission from the
and overlying mantle, and the crustal thickening has Geological Society of America).
MOUNTAIN RANGES 179
Distance (km)
200 400 ^0 800 1000
—I— —I—
Late Palaeozoic 50
Lithosphere
(a) 100
Triassic-
^
,/
+ + + + + + + + + + + + + —W-—
+ + + + + + + + + + + + + + + + + + + + + + + + +
t * * * * * + + + + + + + + + + + + + + + +*
++++++ J -
Jurassic
50
1
100
(b) -"150
^ ✓ N'
\ T'
' + + .
Cretaceous- - +
+
+ + + + + +
+ + + + + + + + + +
+ + + + +
+
+
+
early Tertiary + + + + + + + + + + +
• i + + + + + +. ^
^
M 50
100
(c)
150
Western Eastern
cord
Pliocene-
Pleistocene
- 100
Asthenosphere
Lithosphere -150
2 X Vert,exag
(d)
Pre-Palaeozoic Sedimentary
still continuing. The present rate of subduction is the new continental margin. This process has been
70-120 mm a"^. documented in the Indian Ocean, where a new plate
There are objections to James's plate tectonic boundary is developing in response to the collision
model (Helwig, 1973) which Zeil (1979) has exam ofIndia and Asia(Van Orman 1995). It will be
ined in detail. These include the absence of blue- shown later (Section 9.3.6), however,that the behav
schist facies rocks except in the northern Andes, the iour of continents after collision is rather more
variation in time between compressional and ten- complex.
sional episodes, the presence of seismic gaps and The surface marking the locus ofcollision is known
variable dips in the Benioff zone and the question as a suture. The suture zone may preserve slivers of
able validity ofextrapolating recent seismic data back the oceanic lithosphere which formerly separated the
in time. James's model consequently represents a continents, known as ophiolites (Section 2.5).
working hypothesis which must be critically tested The Himalaya represent the youngest collisional
against continuing field and laboratory investiga mountain range, in which continental collision
tions. Certainly, the Andes cannot be considered to began in Tertiary times. Older mountain ranges are
have formed by a continuum process, as there have represented by the Appalachians, the Caledonides,
been considerable changes in the orogen with time. the Alps and the Urals. Several of the mobile belts
The history ofthe Andes cannot be related to uniform present within Precambrian shields may also have
subduction, but rather to a variety of processes which originated in this way.
gave rise to different characteristic features in dif It is probable that mechanisms operating in the
ferent sections of the orpgen. Ramos (1988), for Himalaya today are analogous to those which af
example, has interpreted the history of the central fected these older mobile belts and mountain ranges.
Andes in terms of a model which involves the west Consequently, this section will concentrate on the
wards migration ofthe trench since Palaeozoic times evolution of the Himalaya. It has only recently be
and the accretion ofexotic terrains (Section 9.5). come possible to investigate the geology and geo
physics of this region, as it has only been in the last
few years that it has been opened to Western investi
9.3 Collisional mountain ranges gators. Before considering the Himalaya in detail,
however, a number of theoretical models of colli
9.3.1 Introduction
sional mountain building will be examined.
When an oceanic plate is subducted beneath conti
nental lithosphere, an Andean mountain range de
9.3.2 Models of continental collision
velops on the edge of the continent (Section 9.1). If
the subducting plate also contains continental litho Several models ofthe evolution of collisional moun
sphere, continued underthrusting causes conver tain ranges have been proposed, and with the accu
gence of the continents, and eventually brings them mulation of increasing knowledge of these features,
into juxtaposition (Fig. 9.1b). Whereas oceanic greater geological and geophysical constraints can be
lithosphere is relatively dense and sinks into the used to discriminate between them.
asthenosphere,the greater sialic content ofcontinen A review of models for the evolution ofthe Hima
tal lithosphere causes it to possess a positive buoy laya and Tibetan plateau has been given by Powell &
ancy in the asthenosphere which prevents its being Conaghan (1975), and these are shown in Fig. 9.6.
subducted any great distance (McKenzie, 1969). Figure 9.6b shows a model in which both converg
Consequently, the arrival of continental lithosphere ing plates are equally affected by folding and thrust
at a trench results in collision with the overriding ing,so that the resulting orogenic belt is symmetrical
continent; rapid relative motion is halted, and a about the suture. This model is discounted because
collisional mountain range is formed by crustal the main Himalayan suture does not occupy a central
shortening (Fig.9.1b). According to classic plate position within the mountain belt.
tectonic theory, no further relative movements be Figure 9.6c shows a model first proposed by Ar-
tween the continents are possible, and the forces gand (1924), and suggests that the suture between
responsible for collision, if they remain active, may India and Eurasia developed beneath a cover of
cause the initiation ofan oceanic subduction zone at Tethyan sediments. Deformation was thought to
MOUNTAIN RANGES 181
Lower and
higher
S Himalaya . Tibetan Plateau
(a)
Sub
Himalaya
^"u1o^TangWa_^ChangThangPlatformH- p'^oldB^elt-H- Stable^ri sban^
Basin l^oldBelt
- Tibetan Dzuhgarian
Brahmaputra Himalaya Stable Block
Sea level
js ^ J-
Base of crust
Central
Thrust
ndia
Tethys
Indian Plate
ndia
Aiianl31iBi-r^
f^r'^'atpr Indj
(km)
No vertical exaggeration
Fig.9.6 Different models suggested for the formation of (after Dewey & Burke, 1973, with permission from the
the Himalaya. S, postulated suture between India and Asia Geological Society of America);(e) multiple block
(after Powell & Conaghan, 1975, with permission from collision model (after Chang & Cheng, 1973);
the Geological Society of America),(a)Surface geology (f) underplating model (after Powell & Conaghan, 1973
and nomenclature;(b) orogenic belt symmetrical about and Mattauer, 1986);(g) limited underplating model
suture;(c) plastic deformation beneath Tethyan sediments (after Molnar, 1984).
(after Argand, 1924);(d) basement reactivation model
182 CHAPTERS
occur by plastic deformation and thrusting of the elevation of the Himalaya is then attributed to the
upper crust, which resulted from the underthrusting isostatic response ofa thickened,low density crust. A
ofIndia beneath Asia. problem ofthis type of model is that some process of
Figure 9.6d is based on the basement reactivation 'mantle peeling' is required beneath the Asian block
model of Dewey & Burke (1973). They envisage a so that Indian crust can be emplaced directly beneath
thickening of the Asian crust following collision by Asian crust rather than the base of the Asian litho
thrust faulting in the upper crust and ductile creep sphere. This problem is obviated, however, by the
in the lower crust. The increased heat flow of the model of Molnar (1984) in which only a limited
thickened crust then causes partial melting and amount of Indian underthrusting is postulated. No
the buoyant ascent ofa melt rich in potassium and the crust-mantle decollement is required, and crustal
lithophile elements. The enriched upper crust would thickening takes place in response to sequential
be underlain by a depleted lower crust ofanorthositic thrusting along the Main Central Thrust and Main
composition, with the boundary at mid-crustal Boundary Fault (Fig. 9.6g). This model is also con
depths being represented by a zone of migmatites. sistent with the seismic studies ofHirn (1995),
This boundary may well represent the Conrad seis which suggest that no Indian lithosphere is present
mic discontinuity (Section 2.4.1), and it is interest at depth north of the Indus-Zangbo Suture Zone
ing to speculate that the Conrad discontinuity is only (Section 9.3.3) and that crustal thickening is caused
developed in continental crust which has been af by imbrication. A major criticism ofall models ofthis
fected by collision, thus explaining its absence else type which involve the underthrusting ofIndia is that
where. This model is inadequate to explain the they appear to contravene the assertion of McKenzie
evolution of the Himalaya, however, as most of the (1969) that continental lithosphere is too buoyant
crustal thickening has occurred within the Indian to subduct. Subduction might be achieved if the
Plate and not beneath Tibet. Also, if the increased continental crust were greatly attenuated, but this
heat flow beneath Tibet is caused by the presence of would probably only affect relatively small areas.
subducting oceanic lithosphere, even a very shallow Subduction would be facilitated, however, if the
dip of 10® for the subduction zone would only affect crust were delaminated from the lithospheric mantle
the southern half ofTibet (Fig. 9.6d). along the brittle-ductile transition at the level ofthe
Figure 9.6e, based on the model of Chang & Moho predicted by some theories of the rheologic
Cheng (1973), attempts to explain the presence of structure of the continental lithosphere (Section
many ophiolite belts within the Himalaya by the 2.10.3). Butler (1986) has suggested that if crustal
successive collision of several small crustal blocks thickening is accomplished by thrusting along a
after the progressive closure of several small oceans detachment climbing from the Moho (Fig. 9.7), the
during the Phanerozoic. The model is based on base of the footwall section of the crust would be
radiometric dates which span Phanerozoic time and depressed below the level at which the granulites of
which young southwards. This mechanism success the lower crust undergo a phase change to eclogite.
fully explains the thickening of the underthrusting The increased density of the basal eclogite would
southern plates, but cannot explain why uplift ofthe then decrease the buoyancy of the crust and allow
Himalaya did not occur until Tertiary times. This limited subduction to take place, although long
model is also highly relevant to the accretion of distance underthrusting does not appear to have
suspect terrains to mainland Asia (Section 9.5). occurred (Change?At/., 1986).
Figure 9.6f is based on the models of several
workers (e.g. Powell & Conaghan, 1973; Mattauer,
9.3.3 Himalayan geology
1986) who believe that the thickened crust beneath
the Himalaya has resulted from the underthrusting The Himalayan mountains are 250-350 km wide
ofIndian crust beneath the crust of Asia rather than and extend for about 3000 km from Afghanistan to
by penetrative deformation ofthe crust. This appears Burma. They comprise a series of lithologic and
to be supported by the seismic tomographic analysis tectonic units which run parallel to the mountain
(Section 2.1.8) of Roecker (1982) and the thrusts belt and maintain a constant character for long
and double thickness crust revealed by a seismic distances (Fig. 9.8). Summaries of the known geol
reflection profile (Zhao et 1993). The great ogy are to be found in Molnar(1984) and Windley
MOUNTAIN RANGES 183
Karakorum Batholith
Trans-Himalayan Batholiths
and volcanics
Ophlollte klippe
_ Higher
Tethyan sediments ,
Himalaya
Miocene granites J
Central crystallines ]
Sediment, crystallines I Lower
and granites i Himalaya
Outer crystalline klippe
Siwaliks — Sub-Himalaya
Fig.9.8 Geological map of the Himalaya (redrawn from 9.3.4 Deep structure of the Himalaya
VVindley, 1985 with focal mechanism solutions of
earthquakes from Molnar, 1984). IZS,Indus-Zangbo Bouguer anomalies become increasingly negative
Suture; MCT, Main Central Thrust; MET, Main northwards from India. This indicates that the crust
Boundary Thrust; MFT, Main Frontal Thrust. thickens in this direction and that the Himalayan
mountains are at least partly compensated by an Airy
type of compensation (Section 2.11.2).
Himalaya. Thrust faulting occurs beneath the Lower The most detailed imaging of crustal structure has
Himalaya, with the fault planes probably dipping come from a 500 km-long seismic traverse across the
north at shallow angles. Thrust faulting gives way to Himalaya and Tibet in which emphasis was placed on
normal faulting on north-south planes in the Higher the recording of critical reflections from the Moho
and Trans-Himalaya, and probably reflects the in (Hirn et a-L, 1984a; 1984b). A line drawing of the
creased loading of the lithosphere, causing the verti interpretation of these events is shown in Fig. 9.10.
cal stress to pass from being the least to the most The data show that the crustal thickness ofthe Indian
compressive stress. The overall pattern of fault plane Shield is some 35 km, and that this increases to
solutions appears to be compatible with a model in 55 km beneath the Himalaya and to 70 km under
which the Indian Plate bends downwards in front of Tibet. The crust-mantle boundary, rather than being
the Himalaya and descends beneath them. a sharp contact, occupies a transition zone over a
MOUNTAIN RANGES 185
Sediments
Tethyan
Sediments
Upper crus
Melting
Upper mantle zone
50 km
Fig.9.9 Diagrammatic section across the central intracontinental thrusting affecting both the crust
Himalaya. Symbols as for Fig.9.8 (redrawn from Windley, and upper mantle.
1983,in the Journal ofthe Geological Society ofLondon^ Lyon-Caen & Molnar(1983) have interpreted the
with permission ofthe Geological Society). gravity field of the Himalaya by assuming that India
slides beneath Eurasia and that the Indian Plate flexes
elastically (Section 2.11.4) under the weight of the
vertical interval of some 12 km. The phase has a overthrust Himalayan mountain range (Fig. 9.11).
velocity of8.5 km s"\ indicating that normal mantle The only unknowns in this interpretation are the
material is present with no temperature anomalies so flexural rigidity of the lithosphere, density contrasts
that no isostatic support originates from hot, low between sediments,crust and mantle,the location of
density material in the upper mantle. the end of the elastic plate and the bending moment
The Moho topography is not smooth (Fig. 9.10), applied to its end. Figure 9.11 shows an interpreta
and exhibits a number of steps, some of which are tion which has optimized the correspondence of
vertical and some in which there is an overlapping of observed and calculated gravity anomalies by varying
a Moho at different depths. All steps produce a Moho these parameters. In the light ofthe seismic reflection
displacement of less than 20 km. The vertical dis data described above, this model is incorrect in that
continuities are associated with the Indus-Zangbo India does not descend smoothly and coherently
Suture and the Danqiao Ophiolite, and probably beneath Eurasia. However, the model successfully
indicate that these features are now strike-slip faults simulates the observed gravity field, and it is possible
associated with indentation tectonics (Section that modification of the model might well produce
9.3.6). The stacked Moho segments probably repre an interpretation consistent with the seismic reflec
sent the thrust faults which were responsible for the tion results.
thickening of the Himalaya-Tibet crust. The thick
ness ofthese thrust slices is less than 20 km,which is
9.3.5 Evolution of the Himalaya
considerably less than the normal crustal thickness.
This implies that crustal thickening does not take The Himalaya formed in response to the collision of
place by simple underthrusting of Indian crust be India with Eurasia. This collision was brought about
neath Eurasia (i.e. underplating) in which the Moho by the northward migration ofthe Indian Plate as the
is smoothly depressed northwards. Rather, it sug intervening Tethyan Ocean was subducted beneath
gests that thickening has taken place in response to the southern margin ofthe Eurasian Plate. Magnetic
186 CHAPTER 9
50 km j g - 50 km
70km -i 1- 70km
Fig.9.10 Moho reflections beneath Tibet (after Hirn et geological observations (Allegre et al., 1984) can
«/., 1984b, with permission from Nature 307, 25-7, provide more detail to the sequential movements.
Copyright © Macmillan Magazines Ltd). These indicate that the main continental conver
gence was preceded by the collision of two smaller
anomalies in the Indian Ocean and palaeomagnetic plates (Fig. 9.13). The North Tibetan Plate became
measurements on the subcontinent confirm the welded to Eurasia by about 140 Ma and was followed
northerly drift of India and allow the reconstruction by a step-back in the subduction zone to the south of
of its path (Fig. 9.12). The initial continental colli North Tibet. The underthrusting brought the South
sion took place about 50 Ma ago and marks the Tibetan Plate into juxtaposition with North Tibet
cessation of marine sedimentation and the com and eventually welded North and South Tibet to
mencement of terrestrial deposition in the suture gether. A further step-back in underthrusting pre
zone. All oceanic lithosphere had disappeared by ceded the main collisional event. India collided with
45 Ma, and at about 36 Ma there was a decrease in southern Tibet, which then represented the southern
the velocity of India's northward drift from over margin of the Eurasian Plate, 50 Ma ago, with the
ICQ mm a"^ to about 50mma~^. This probably formation ofthe Indus-Zangbo Suture.
marked the last stage of true continental collision Continued convergence at a slower rate led to the
and the initiation of a regime of indentation tecton progressive formation of the Main Central Thrust
ics (Section 9.3.6). and Main Boundary Thrust, accompanied by some
Magnetic lineations can only provide a relatively 2000 km of crustal shortening.
coarse model of the India-Eurasia convergence, and It is thus apparent that the Himalaya contain at
Computed
Observed
50°S 40 30 20 10 10 20 30 40 ?in°N
S.Tibet N. Tibet
140 Ma
Deccan Traps
IZS^x t
mot\x V
MBTx^,
L/'y.i'!: Present
Fig.9.13 Possible sequence of events in the evolution of and Altyn Tagh faults run approximately parallel to
the Himalaya. KS, Kokoxili Suture; BNS,Bangong the a and P lines, respectively, generated by this form
Nujiang Suture; IZS,Indus-Zangbo Suture; MCT,Main of indenter. The asymmetry of the indenter, whose
Central Thrust; MBT,Main Boundary Thrust(redrawn northeastern edge impinges more obliquely than the
from Allegre etal., 1984, with permission from Nature
northwestern/edge, may explain why deformation
307,11-22. Copyright © 1984 Macmillan Magazines
occurs over a very large area to the northeast of
Ltd).
the Himalaya. However, a triangular indenter
(Fig.9.15c) may represent a more appropriate model
a and P lines, which correspond to dextral and to explain this phenomenon.In this case the indenter
sinistral strike-slip motion, respectively. The pattern has penetrated a significant distance into the plastic
ofslip lines is controlled by the shape ofthe indenter medium. The Kangting Fault, a remarkably curved
and the lateral constraints placed on the plastic feature on satellite images,follows a p line, while the
medium. Figure 9.15 shows the patterns developed linear Red River Fault follows an a line. This model
when the plastic medium is semi-infinite. A flat can also explain how underthrusting can simulta
indenter (Fig. 9.15a) produces a symmetrical pattern neously take place in an easterly direction at the
of slip lines extending into the plastic medium a Burman arc (Fig. 9,14). The subduction results from
distance approximately equal to the width of the the material shouldered aside to the east by the
die. Wedge-shaped indenters (Figs9.15b,c) produce northerly motion ofIndia.
more complex arrangements of slip lines. Figure A further class of models ofthis type is obtained if
9.15b appears to simulate the structure in the west plastic material of finite width is indented, as shown
ern Himalaya and Pakistan where the Indian Plate's in Fig. 9.16. In this case the region ofdeformation is
boundary has a wedge-shaped geometry. The Herat bounded by the pair of slip lines which leave the
MOUNTAIN RANGES 189
TWl
Strike-slip fault
[)s Motions with respect
Siberia
Normal fault
1^ Direction of extension
Oceanic crust
Thrust fault in SE Asia
60°N
Siberia
V 150°E
WI*. 30°N
50°N
Mongolia
f
140°E
20°N
40°N
China
30°N
thrust
China•§:
Sea i 130°E
20°N
Borneo
10°N
120°E
80°E 100°E
cx lines
p lines
a lines
Mongolia
20 mm
ation of numerous, short-lived, dextrai and sinistrai tectonic features ofthe Himalayan region (Fig.9.14)
faults near the apex of the triangle. The unilateral can be explained by this experiment,for example:
case provides asymmetrical deformation in which 1 sinistrai oflfsets in central China are much greater
faults allowing displacement towards the free edge than dextrai;
predominate, such as F^. The block translated side 2 pull-apart structures along the sinistrai faults may
ways rotates about 25 ® clockwise and is followed by be analogous to the extensional regimes in Shansi,
the extrusion of a second block along a further Mongolia and Baikal;
sinistrai fault, F2, which allows a continued rotation 3 the Altyn Tagh Fault may correlate with the major
of the first block by up to about 40°. Numerous dislocation F2, and its bending in the west may be
pull-apart troughs (Section 7.4.3) develop along the caused by the effects ofthe indenting Indian Plate;
sinistrai faults because oftheir irregular geometry. As 4 the Red River Fault maycorrelate with Fi,and the
these movements progress, a gap grows between the succession of activity between the Altyn Tagh and
indenter and extruded Plasticine. Red River faults might explain many ofthe complex
Tapponier et al. (1982) suggest that many of the ities in the evolution ofsoutheast Asia.
192 CHAPTERS
The concept of indentation tectonics is clearly in which the deformation of Asia is viewed as the
successful in explaining the spatial distribution and response of a continuous viscous medium to both
relationship of the various styles of faulting and the edge forces of continental collision and the
deformation in the Himalayan region. In applying internal forces generated by differences in crustal
the results of indentation studies to deformation of thicknesses. This continuum model takes no account
the lithosphere, however,it should be borne in mind of faulting, but it is argued that fault-induced de
that several simplifying assumptions are made which formation is small and restricted to the upper brittle
are not completely true for natural Earth materials. layer, so that over long intervals of time the ductile
The deformation resulting from indentation is as lower crust is the major control of deformation. The
sumed to occur in a homogeneous plastic material. lithosphere is not regarded as rigid, but simply as a
The lithosphere does not exhibit such a simple theol layer of much greater strength than its substrate.
ogy,and the upper crust certainly stores strain energy England (1982) has shown that horizontal forces
elastically, as evidenced by the presence of earth are generated by regions of elevated and thickened
quakes. In mitigation, at greater depths creep is crust. He has combined these with the edge forces
probably a dominant mechanism of deformation. arising from continental collision and, in order to
A major criticism of the indentation model is that account for the observed deformation of a region
it does not predict, or take account of, the variable such as central Asia, has determined the required
thickness of the deformed plastic layer, in which rheological properties of the viscous sheet that is
thickening occurs in response to thrusting, and thin believed to represent this region. Solution of the
ning in response to extension. England (1982) and response of this sheet to the applied edge and inter
England & McKenzie (1982) have adopted a com nal forces then provides the evolutionary trends in
pletely different approach to the India-Asia collision the distribution ofdifferent styles ofdeformation for
(a)
140°
Fig.9.19 Distribution of tectonic styles in Asia (redrawn the main thrust front, is explained here by the ten
from Tapponier & Moinar, 1976, in Nature 264,319- dency of thickened crust to spread laterally.
24, with permission. Copyright © 1976 Macmillan In contrast to the indentation model, in which a
Magazines Ltd). plastic layer of constant thickness progressively de
forms because of the steady advance of an indenter,
the continuum model places greater emphasis on the
a model in which collision has occurred along part of role of forces arising from crustal thickening. The
one edge (Fig. 9.18). The black areas correspond to initial continental convergence results in compres-
regions ofcompression in which thrusting would be sional deformation and thickening ofthe crust in the
dominant;the vertically and diagonally hatched areas region of the suture. Eventually the crust reaches
are regions of compression and extension, respec such a thickness that its buoyancy exerts lateral forces
tively,in which deformation has a significant compo large enough to deform the hinterland. The thick
nent of strike-slip; the dotted areas are regions of ened crust is then thrown into tension and the
predominantly strike-slip with components of com collisional forces deform the region outside this area
pression(+)and extension(-). by predominantly strike-slip movements.
Such a geometrically simple model would not be Neither the indentation nor the continuum model
expected to simulate exactly the distribution of can be completely successful in simulating the pat
deformation styles in such a complex region. The tern of deformation styles of a collisional orogen,
model can explain in broad terms, however, the although the presence of large strike-slip displace
northerly progression from thrust faulting in the ments has been proved (Tapponier et al.^ 1990).
Himalaya through a region ofstrike-slip faulting to a Moreover, work by England & Houseman (1986)
region of tension in northern China and eastern on the continuum model predicts that deformation
Siberia (Fig. 9.19). The region of normal faulting in is restricted to the region about 1000 km north and
Tibet, which has no analogue in the indentation northeast of the Himalaya, and argues against the
model and which occurs immediately to the north of model of Tapponier et al. (1982) in which Indo-
194 CHAPTERS
(a)
Crustal low
velocity zone'
Foreland basin
Peripheral bulge
Foreland/hinterland crust
Fig.9.22 Schematic sections comparing (a) Alpine-type of crust was thrust into the mantle and metamor
and (b) Himalayan-type orogens (redrawn from Dewey et phosed to eclogite, causing it to create its own
al., 1986). 'slab-pull' force. These differences in thrust style
might owe their origin to the forces acting on the
plates immediately prior to collision. The Indian
it into separate Indian and Australian plates. A fur Plate, because ofits larger northerly flanking compo
ther contributor to the continued northward move nent of oceanic lithosphere, would possess a greater
ment of India is the ridge-push force exerted on the horizontal stress component which might have led
southern margin of the Indian Plate. to lower angles of thrusting.
The Himalaya are the youngest mountains on Such suggestions are tentative, but strongly suggest
Earth, and must represent our analogue for ancient that if plate tectonic interpretations are to be made of
continental collisions. However, an examination ancient collisional mountains, a full study must be
of schematic sections (Dewey et«/., 1986) through made of the plate regime extant for some consider
the Himalaya and Alps reveals several differences able time before collision.
(Fig.9.22), the most important of which is the lack
of a continental plateau, and correspondingly thick
9.4 Arc-continent coliision
ened crust equivalent to Tibet, in the hinterland of
the Alps. It is not easy to provide an explanation for Collisional mountain ranges can also develop as the
the differences between Alpine- and Himalayan-type result of a collision between an island arc and a
collisions. It is possible that the differences arise continent, although the resulting mountains are
because of the initial presence, between India and smaller than those resulting from continent-
Eurasia, of a large intervening ocean whose subduc- continent collision, such as the Tertiary fold belt of
tion caused extensive relative movements. This large northern New Guinea(Dewey & Bird, 1970).
oceanic area was missing from the northern margin The sequence of events occurring in an arc-
of Africa prior to the Alpine collision (England, continent collision is shown in Fig. 9.23. In
1982). Butler (1986) has suggested that the differ Fig. 9.23a a continent approaches an island arc by
ences between the Alps and Himalaya depend upon the consumption of an intervening ocean. When
the initial geometry of the thrust faulting experi the continental margin is driven into the trench
enced on collision (Fig. 9.7). Low angle thrusting (Fig. 9.23b), significant underthrusting of con
would have occurred in the Himalaya and much tinental lithosphere is inhibited by its buoyancy
steeper thrusting in the Alps, so that a larger volume (McKenzie, 1969). Thrust slices moving towards the
196 CHAPTERS
(a)
Flysch
oceanic crust
I I lithospheric
mantle
mantle
Molasse
iT'r \'w r
Fig.9.23 Schematic sequence ofsections illustrating the Timor. The inner volcanic arc stretches from Java to
collision ofan island arc with an Andean-type plate the east ofFlores. In the west the arc is active and the
margin, followed by a change in the direction of volcanism results from the northwards underthrust
underthrusting (redrawn from Dewey & Bird, 1970, with ing ofthe oceanic Wharton Basin at the Java Trench.
permission from the American Geophysical Union).
In the east, however, north of Timor, volcanic activ
ity ceased in the late Tertiary. Strontium isotope
continent develop while flysch and slices of oceanic studies of the volcanic arc in this region are consis
crust are driven on to the continental margin tent with the involvement of continental material in
(Fig. 9.23c). No further lithosphere convergence is the magma genesis. Here the outer arc is underlain
possible, and so if the two plates continue to con by continental crust which is continuous with that of
verge a new trench may develop on the oceanward Australia.
side ofthe island arc (Fig. 9.23d). The structural and tectonic units of this region
Arc-continent collision zones are relatively rare, as originated by the underthrusting of oceanic litho
they usually represent an intermediate step in the sphere northwards at both the Java Trench and
closure of an ocean and so are relatively short-lived. Timor Trough. About 3 Ma ago this consumption of
The most recent site of an arc-continent collision is oceanic material brought the Australian continental
represented by the Banda arc (Fig. 9.24) (von der margin to the Timor Trough (Fig. 9.25). Volcanic
Borch,1979;Bowin^f/i/., 1980;Johnston & Bowin, activity in this eastern part of the arc then ceased
1981). The Banda arc is a double arc enclosing the when the buoyant continental lithosphere choked
oceanic crust ofthe Banda Sea.The outer arc(or first the subduction zone, creating a discontinuity be
arc—see Section 8.1)is made up ofa submarine ridge tween the Timor Trough and Java Trench. No fur
ofdeformed sediment which emerges at the island of ther convergence of the Australian shelf and Banda
MOUNTAIN RANGES
ulnearoiiiiis^
Ceram
pFlores thrust
Sumb
Q Australian-New Guinea shelf Ms* 10°S
Arc discontinuity J
Wharton Basin
500 km Australia
15°S
110°E 120°E 130°E
Fig.9.24 Structural elements of the Banda Sea region. and may represent the precursors of a new subduc-
Diagonal shading, continental or quasi-continental crust; tion zone caused by a polarity reversal ofthe arc.
S-S-S,sedimentary arc; V-V-V,volcanic arc (redrawn A similar arc-continent collision zone has been
from von der Borch, 1979, with permission from Elsevier).
described in Taiwan by Hsu (1990).
Charlton (1991) has noted that extension of the
Sea is possible along the northward dipping trench at Banda arc collision zone commenced within 5 Ma of
the Timor Trough. However, it appears that the the collision, and that such extension appears to be
forces driving this convergence are still active, and common in the orogenic belts of eastern Indonesia.
SiWtretal.(1983) have discovered two large thrusts This extension may arise following decoupling ofthe
in the back-arc region. These Wetar and Flores subducted oceanic lithosphere as continental litho-
thrusts are northerly directed (Figs 9.23 and 9.24), sphere chokes the subduetion zone.
50 km
Fig.9.25 True scale section across I
as suspect, exotic or displaced terrains. They are de 2 Disrupted terrains, which contain a heterogeneous
scribed in these terms as they are allochthonous and assembly of flysch, serpentinite, shallow water lime
have become accreted to the continent over a consid stones and graywacke, with occasional exotic blocks
erable period of time, in many cases after travelling of blueschist metamorphic rocks. Some of these
great distances. The process of amalgamation has terrains may have originated as subduction zone
been referred to as accretionary, colla£fe or mosaic assemblages (Section 8.7).
tectonics. 3 Metamorphic terrains, in which a metamorphic
A suspect terrain is a mappable unit characterized overprint has destroyed the original stratigraphy.
by a geological history different from that ofadjacent 4 Composite terrains, which comprise two or more
units (Jones et at.., 1983). Terrains are bounded by sub-terrains which had amalgamated prior to accre
normal,reverse or wrench faults, and the boundaries tion to the continent.
occasionally preserve thin ophiolites, rocks of blue- The chronological sequence of accretion of
schist metamorphic grade or highly deformed flysch. terrains to the continent can be determined from
The recognition of suspect terrains should ideally geological events which postdate accretion and
be based on contrasts in detailed stratigraphic his link adjacent terrains (Fig.9.26). These include the
tory, although in many cases this has been destroyed deposition of sediments across terrain boundaries
by later tectonism and the bounding faults obscured (Fig. 9.26a), the appearance of sediments derived
by later processes. Consequently, in order to deter from an adjacent terrain (Fig.9.26b) and the 'stitch
mine whether the geological histories of adjacent ing' together of terrains by plutonic activity
terrains are compatible with their present spatial (Fig.9.26c).
relationships, extremely detailed and comprehensive The most fully investigated region of suspect
investigations are necessary (e.g., Yorath & Chase, terrains is the North American Western Cordillera
1981). In practice several criteria are used to distin (CoTity et at., 1980). Figure 9.27 shows the distribu
guish the identity ofseparate terrains(Shermer et at., tion ofthe major terrains in this region, which form a
1984), based on contrasts in: zone some 500 km wide and make up about 30% of
1 the stratigraphic and sedimentary history; continental North America. Many of the known
2 petrogenetic affinity and magmatic history; terrains are too small to be illustrated at the scale of
MOUNTAIN RANGES 199
Sp I Principal terrains
NS North Slope
Kv Kagvik
En Endicott
R Ruby
Sp Seaward Peninsula
xP mPM
j
/
I
NF
Innoko
Nixon Fork
PM Pingston and McKinley
YT Yukon-Tanana
01 Chuiitna
P Peninsular
W Wrangellia
Cg Chugach and Prince William
TA Tracy Arm
T Taku
Ax Alexander
G Goodnews
Ch Cache Creek
St Stikine
BR Bridge River
E Eastern assemblages
Washington, Oregon and Cali'ornia
Ca Northern Cascades
SJ San Juan
O Olympic
S Si letzia
BL Blue Mountains
Trp Western Triassic and Palaeozoic
of Klamath Mountains
KL Klamath Mountains
Fh Foothills Belt
"■V
San Francisco Ti ,
Si S F Franciscan and Great Valley
C Calaveras
(_RM^ Si Northern Sierra
SG San Gabriel
Mo Mohave
Sa Salinia
Or Orocopia
?5^0r S Sonomia
RM Roberts Mountains
X USA GL Golconda
^MEXICOl
B Baja
V Vizcaino
Fig. 9.27 Generalized map of suspect terrains in western ornament, Wrangellia; diagonal ornament. Cache Creek
North America. Stippled ornament, North American terrain (redrawn from Coney rt a/., 1980, with permission
cratonic basement; barbed line, eastern limit of from Nature 288, 329-33. Copyright © 1980 Macmillan
Cordilleran Mesozoic-Cenozoic deformation; solid Magazines Ltd).
'Chukchi '
Rise
Voering
Plateau
JS'i.' -Faeroe
' . ^^-/^Plateau
Shirshov
Ridge
Rockall
Bowers
Plateau
Umnak
Porcupine^
Alaska Q Bank
Plateau Seamounts
Flemish ^
Shatsky Cap Madiera'
Rise Hess Rise
§ Marcus
Bermuda Rise ^ Plate^
ratosthenes East ^
Plateau Chagos-Laccadive Mid .. Blake-Bahama Canary
Ridge Pacific Mts £>• Plateau Islands
Medina Aves Ridge
PI . Palau-<s^ ^ (Plateau
Kyushu Ridge Barbados Ridge
Manikiki
ySeychelles Cocos Ridge
Plateau
Bank Ontong-Jav
g^Plateau Tuamotu
\ ^^ Ridge Galapagos
Mozambique Exmouth Rise Sao Paulo
Plateau
Fiji Plateau Walvis
Nazca
Q Plateau
Ridge Ridge
Lau Ridge
Madagascar Tuba! Is Juan
Plateau ^ Broken Lord Ridge Fernandez Ridge Rio Grande
\ Ridge Norfolk Plateau Rise
Cape Agulhas South ^Rise Chatham
\ Falkland
Rise Plateau Tasman
Rise South Georgia
Rise
Campbell
® «p. Plateau ^ North Scotia Ridge
Thirty-East Maud
(
Spur Balleny Island South Scotia Bank
Peter Ridge
Island
1 ♦
20° 40° 60° 80° 100° 120° 140° 160° 180° 160° 140° 120° 100° 80° 60° 40° 20° 0°
Fig.9.28 Distribution of present day oceanic plateaux (redrawn from Ben-Avraham ^a/., 1981,Science 213,47-54, with permission from the AAAS. Copyright © 1981 by the AAAS).
202 CHAPTER 9
of volcanic rock and uplifted oceanic crust. Indeed, Plateau moved northwards as a result ofsubduction
Silver & Smith (1983) have suggested that the island at a now extinct trench within the Bering Sea
arc, melange zones and continental fragments found (Fig. 9.31a). The arrival of these features at the
in the Indo-Pacific region of modern southeast Asia trench may have locked the subduction zone and
may resemble the configuration ofsuspect terrains of caused a new trench to develop at the presently active
the Western Cordillera prior to accretion. Coffin & Aleutian arc (Fig. 9.31b). The Shirshov Ridge may
Eldholm (1994) have shown that many of the frag have been emplaced at the same time by motion
ments are ofigneous origin and must have originated along a now extinct transform fault. A fuller account
from a complex mantle circulation coexisting with of the evolution of this region is given by Cooper et
the convection associated with plate movements. al.(1992).
Geophysical data and sampling have shown that Similar processes of accretion have been proposed
some of the oceanic topographic highs have strong for other regions of suspect terrains (Shermer et al.^
continental affinities, possessing, for example, a gra 1984), and such models have prompted Nur &
nitic substrate. Others are of oceanic origin. Most, Ben-Avraham (1982) to construct a world map
however, lie on a crust two to five times thicker than showing possible present and future suspect terrains
normal oceanic values. These oceanic highs situated (Fig. 9.32). While it is accepted that the terrains
within shrinking oceans will ultimately be brought to shown have accreted, or will accrete, to major conti
a trench. Their thickened crust implies a relatively nental plates following large horizontal movements,
high buoyancy, which may be sufficient to prevent it is important that the nomenclature does not get
their subduction. If not, they should alter the geo out of hand. Most workers would describe the re
metry ofthe Benioffzone and consequently affect the gions of Somalia, Arabia and India as plates or
tectonic and magmatic process associated with the sub-plates, and it would seem more reasonable to
subduction. restrict the term 'suspect terrain' to considerably
At present, no oceanic high of continental affin smaller entities with a rather more obscure history of
ities, that is, too buoyant to be subducted, is in the movement.
neighbourhood of a trench. There are several loca It is probable that many of the suspect terrains in
tions, however, where oceanic rises are in the process the Alpine-Himalayan chain represent fragments
ofsubduction (Nur & Ben-Avraham, 1982), such as rifted from the northern margin of Gondwana and
at the Peru-Chile Trench (Fig. 9.29) and the Nankai transported across the Tethys Ocean before accretion
Trough (Hirahara, 1981). The subducting rises to the southern margin of Laurasia (Fig. 9.33).
strike obliquely to the direction of underthrusting, Similarly, the circum-Pacific terrains may have
so that the locus ofthe point ofdescent progressively originated from the eastern margin of Gondwana.
moves along the trench. The cpntinental region The fragments ultimately travelling to the Western
behind the segment of trench swept by this locus is Cordillera of North America are believed to have
marked by a gap in volcanic activity. Figure 9.30 come from a region of Gondwana named Pacifica
shows the vertical distribution of earthquakes across (Fig. 9.33)(Kamp, 1980), which was probably situ
these zones of anomalous subduction and compares ated over a geoid high (Anderson, 1982). Since the
them with regions of normal subduction and conti Pacific Ocean occupied two-thirds of the Earth's
nental collision. The buoyancy ofthe descending rise surface at the start of the Mesozoic, the Pacifica
causes the Benioff zone to dip at a shallower angle terrains would have become widely dispersed during
than normal, and a distinct seismic gap is present. their long voyage. The mechanism of rifting, how
A conceptual model for the accretion of a suspect ever, is unknown. It is possibly related to the hotspot
terrain is shown in Fig. 9.31,illustrating the possible and volcanic activity associated with the geoid high
evolution of the Bering Sea. Within this region the (Section 10.4)(Anderson, 1982). Wmketal.(1984)
Umnak Plateau and Bowers and Shirshov ridges form have demonstrated the weakness ofcontinental litho-
positive topographic features, all of which are under sphere compared to oceanic, so that if rifting occurs,
lain by thickened crust. Bowers Ridge is probably it will preferably follow a continental path and create
an extinct island arc. Ben-Avraham et al. (1981) the microplates which will subsequently become
propose that the present structural configuration suspect terrains.
evolved as the proto-Bowers Ridge and proto-Umnak The mechanism of accretion of suspect terrains is
MOUNTAIN RANGES 203
Cocos Ridge
Galapagos
Carnegie R. (KJ
Nazca Ridge^
Juan Fernandez R
Submarine ridges
and islands
Cenozoic voicanism
Cenozoic graben
Fig.9.29 Map showing the faults and
relationship between subducting intra-Andean depressions
oceanic rises, Cenozoic voicanism
and graben in western South
America.
poorly understood. Where the terrain is of a buoy way. In other cases, terrains are accreted directly to
ancy sufficient to prevent its subduction, the trench the continent as coherent masses of significant
may become locked and a new trench may develop to crustal thickness which do riot, however, behave as
the oceanward side of the old one, perhaps entrap rigid sub-plates and undergo strong internal deform
ping a marginal sea. Several of the present marginal ation (Burchfiel, 1980). Telescoping and thrusting
seas for which there is no convincing evidence for are dominant and there is no evidence ofsubduction-
back-arc spreading (Section 8.9), such as the eastern related igneous activity. It is possible that the process
Caribbean and Bering Sea, may have formed in this resembles an obduction mechanism similar to that
204 CHAPTER 9
E 200 200
200
400
B 400 Gap
400
600
Q 600
Izu Trench 600 X
/ '-N
Tibet(85°E)
Juan Fernandez gap(~30°S)
S N
100 200 300 (km)
I . r T"
1100 - /Gap/...
#200
Q.
0
Q Kos(SW-NE)
Fig.9.30 Comparison offocal distributions beneath occur at terrain margins. It is also possible that
subduction zones and Himalayan mountain ranges: terrain accretion was a dominant mechanism in the
(a) normal subduction-related seismicity;(b) disrupted growth of the continents during Precambrian times
subduction zone seismicity;(c) collision-related seismicity (Section 12.2).
(redrawn from Nur & Ben-Avraham, 1982, with
The post-accretionary fragmentation of suspect
permission from the American Geophysical Union).
terrains is probably a consequence of plates under
going oblique convergence (Barber, 1985),as such a
regime frequently gives rise to transcurrent faulting
proposed by Oxburgh (1972). In this model an in the associated mobile belt (Section 8.5). Oblique
irregular continental mass approaches a subduction subduction is common in many present-day
zone (Fig. 9.34a). When the promontory X reaches trenches, and it would be expected that such faulting
the trench and resists subduction, plate movement is a ubiquitous feature ofancient mountain belts and
continues in response to the subduction of embay- responsible for the fragmentation of accreted ter
ments Y, Z, and the leading edge of the overriding rains.
plate acts as a wedge, splitting the lithosphere hori As the Western Cordillera of North America
zontally (Fig. 9.34b) and obducting the upper part formed by accretion tectonics, it seems possible that
on to the adjacent continent for distances perhaps in the Andes formed in a similar way. Although detailed
excess of 100 km. The split propagates progressively analyses are not yet available, there are several pieces
downwards so that the whole of the crust may be ofevidence which suggest that this might be the case.
come detached. Confirmation of such a mechanism The Andes comprise several distinct tectonic/
must await more detailed investigations. However,it stratigraphic provinces, often separated by major
appears that terrain accretion is a viable mechanism faults; many Palaeozoic and early Mesozoic struc
of ophiolite emplacement, as ophiolites frequently tures strike at an oblique angle to the mountain belt;
MOUNTAIN RANGES 205
Siberia Alaska
Shirshov
-Ridge umnak
Plateau
/ V .h. Bowers
Ridge
during the Palaeozoic and Mesozoic sediment margin has been integral during the Phanerozoic.
sources were located to the west. These geological If both the Western Cordillera and Andes devel
observations (e.g. Ramos, 1988) and palaeomag- oped by accretionary tectonics,it could be argued that
netic results (Beck et al.^ 1986)suggest the presence all orogenic activity is related to the single process of
ofsuspect terrains. Other workers, however, such as collision, the complexity of mountain belts being
Cobbing,in Pitcher (1985), question the pres explicable in terms of multiple deformational epi
ence of exotic terrains and contend that the Andean sodes associated with the successive accretions of
North Slope
w Endicot
o
c
o
20° 0° 20° 40° 60° 80° 100° 120° 140° 160° 1P0° 160° 140° 120° 100° 80° 60° 40°
Fig.9.32 Distribution of past, present and future suspect terrains (redrawn from Nur & Ben-Avraham, 1982, with permission from the American Geophysical Union),
MOUNTAIN RANGES 207
^''^gnients
Seyche//
Antarctica ^^stralia
Fig.9.33 Possible original
configuration of many ofthe suspect
J^smania
terrains in the Alpine and Pacific
Mesozoic-Cenozoic orogenic belts
(redrawn from Nur & Ben-Avraham,
1982, with permission from the
American Geophysical Union).
terrains. True continent-continent collision then ending orogenic activity. The concept of a unique
represents but one end of a spectrum of possible suture marking the ancient juxtaposition of con
accretionary events.It is not essential to the formation verged plates may thus be illusory for many orogens.
of a mountain belt, but may have the function of Palaeogeographic reconstructions have previously
been attempted from the fragments of plates pre
Fig.9.34 (a) Schematic representation ofan irregular served in mountain belts. The recognition ofsuspect
continental margin approaching an Andean-type terrains implies that such attempts cannot be success
subduction zone,(b)Formation of a crustal flake after ful. Terrains now adjacent may have originated from
collision (redrawn from Oxburgh, 1972, with permission completely different areas, and even after accretion
from Nature 239, 202-4. Copyright © Macmillan may have been translated large distances along
Magazines Ltd). wrench faults.
Trench
Continent Continent
Moho
i-
(a) (b)
208 CHAPTER 9
209
210 CHAPTER 10
(a)
AAAAAAA
AAAA A A
AAAAAAA A A A A A A A A A A A A A A A A 0 AA
Fig. 10.1 Diagrammatic maps showing the development Such collision-generated graben have been termed
of a rift following collision oftwo continents A & B. See impactqgens.
text for explanation (redrawn from §eng6r, 1976, with The distinction between rifts generated by conti
permission from the Geological Society of America). nental rupture or collision is made by some authors
by using the terms'mantle activated'(or 'active') for
the former and 'lithosphere activated' (or 'passive')
indentation tectonics (Section 9.3.6). In this way for the latter (Burchfiel, 1980). These terms refer to
relatively small, isolated rift valleys can form such as the role of upwelling asthenosphere in creating the
the Rhine Graben in the hinterland of the Alpine 'mantle activated' rift. Mantle activated rifts are char
Orogeny and the Baikal and Shansi Graben in the acterized by large volumes of extrusive rocks, while
distal region of the Himalaya. Such 'collisional' rifts lithosphere activated rifts usually contain a greater
can be distinguished from aulacogens, which also abundance of immature clastic sediments than vol
strike into collisional fold belts, as they have no canic rocks. Passive rifts may evolve into active rifts
pre-collisional history in their geological record. by the subsequent upwelling and spreading of the
§eng6r (1976) has also proposed a model for asthenosphere.
certain collision-related rifts whereby graben develop
in the foreland region ofa Himalayan-type orogen as
10.1.4 Petrology and petrogenesis of
a result of the collision of irregular continental
rift rocks
margins (Fig. 10.1). The localized compression re
sulting from the collision of a promontory The volcanic rocks extruded at continental rifts are
(Fig. 10.1a) also causes tensional stresses parallel to highly enriched in alkalis, large ion lithophile ele
the collision front. These give rise to dilational fea ments (LILE) such as K, Ba, Rb and the light rare
tures,including graben striking at right angles to the earths, and volatiles, in particular CO2 and the halo
orogen (Fig. 10.1b). Continued convergence length gens.
ens the suture so that collision occurs along an Bailey (1983) has suggested that the distinctive
extensive front, and the foreland may then be cut by nature of these rocks arises from their formation by
large scale wrench faults (Fig. 10.1c). The Upper metasomatic processes in which the volatile content
Rhine Graben may be an example ofa rift ofthis type. of a large volume of mantle material is focused and
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 211
concentrated. Volatiles escaping upward from the mantle or the sub-crustal lithosphere. High level
deep mantle probably carry LILE, and these would enrichment may result from the trapping of en
cause metasomatism at higher levels where amphi- riched, primitive asthenosphere at the base of the
bole, biotite and carbonate are stable. Bailey suggests lithosphere before depletion, underplating of the
that the volatiles originate in a large mantle reservoir lithosphere by undepleted mantle diapirs, or diffu
and migrate upwards through the narrow rift zone sion ofLILE-rich volatiles from the asthenosphere or
(Fig. 10.2). LILE build up in the conduits in the deeper mantle into the lithosphere.
overlying mantle and crust by metasomatism. The
volatiles modify the melting behaviour of their sur
10.1.5 Structure of continental rifts
roundings and eventually depress the solidus to the
point at which melting occurs, producing volatile Earthquake focal mechanism determinations, geo
and LILE-rich magmas. logical investigations and gravity and seismic reflec
The alkaline magmas probably originate at depths tion surveys have demonstrated that the bounding
of approximately 60-100 km. Tholeiitic and calc- faults of rift valleys are of normal type. Rifts con
alkaline volcanics may also be present, which equi sequently develop in response to tensional stresses,
librated at much higher levels. In addition to these and older theories which attributed their formation
systematic variations, which are caused by depth and to compression, in which a central block was de
degree of partial melting, there also exist unsystem pressed beneath bounding reverse faults, have been
atic variations related to mantle inhomogeneities and discounted. Seismic reflection data (Brown et al.^
complex combinations of processes which affect the 1980) have shown that the faults bounding the Rio
magmas during and after melting (Neumann & Grande Rift do not decrease in dip with depth,and so
Ramberg, 1978). are not of listric type. Whether or not this observa
Volcanic rocks occurring within continental rifts tion applies to all rifts must await similar surveys of
and oceanic islands are enriched in incompatible other areas.
elements and alkalis. Ocean floor basalts, however, Seismic refraction experiments have shown that
are relatively depleted in these components. It is not the crust underlying many rifts is significantly thin
possible to explain these differences in terms of the ner than beneath adjacent regions, and indeed
conditions ofmagma genesis and evolution,and so it gravity anomalies indicate that the whole of the
appears that the mantle from which the magmas are lithosphere is thinned (Searle, 1970). Seismological
derived is heterogeneous. Volcanic rocks in conti data using both natural and artificial sources have
nental rifts and oceanic islands must be derived from shown that a zone of anomalously low seismic velo
a mantle source enriched in incompatible elements cities occurs in the upper mantle beneath most rifts
and radiogenic strontium and lead isotopes. Ocean (Fig. 10.3) and is commonly two to three times
ridge basalts must be derived from a source depleted wider than the rift zone. The extent of this zone in
in these components. the region of the Baikal Rift (Puzyrev et al. 1978) is
It is generally recognized that the depleted mantle shown in Fig. 10.4 and Gao et al (1994) have
source is the asthenosphere, but opinions are divided demonstrated a shallowing of the asthenosphere in
as to whether the enriched source originates above or this region. The vertical extent of the zone has been
below the asthenosphere, that is, from the deep studied beneath the Gregory Rift in Kenya by Long
& Backhouse (1976). They have shown that it
widens downwards into an ellipsoidal structure at a
depth of 150 km. The regional gravity field of the
Gregory Rift is negative (Baker & Wohlenberg,
1971),indicating that the low velocity mantle mater
Mobile elements
ial is also of low density. Heat flow measurements,
the presence of volcanic rocks and geothermal phe
nomena show that this low density results from
Fig. 10.2 Fissuring ofthe lithosphere to allow the escape thermal expansion. Local positive gravity anomalies
of volatiles from a mantle reservoir (stippled)(after Bailey, over the rift axis indicate the presence of a high
1983, with permission from Elsevicr). density igneous body about 20 km wide.
212 CHAPTER 10
China
LVZ
Crust
Moho
100° El 20
Low velocity
50 \ Hi
upper mantle
Normal mantle
km
100-
I
s Basni Province Low velocity
mantle Irkutsk
150
V(km 8"')
200 km
I
Fig. 10.3 P wave velocity structure of continental rift
zones (redrawn from ^uzyrcv et al., 1978, with y Contours of
depth to mantle (km)
permission from Elsevier).
Limit of
•••*' low velocity region
X Geological boundaries
Gravity surveys have shown that the uplifted re of rift
Contours of the
, Dulsburg
/ crust-mantle
boundary
' (depth In km)
Border of the
Alpine fold belt
Miocene volcano
of the Kalserstuhl
iLuxembourg, V
Amount of uplift:
Stuttgart
1000-1500
M 1500-2000
2000 - 2500
> 2500 m
Fig. 10.6 Mantle bulge beneath
the Rhine Graben with post
middle Eocene uplift (redrawn
from lilies, 1977, with
permission from Kluwer
Academic Publishers).
opment of rim uplift by elastic upbending of the lower level. This is easily understood ifone considers
crust on the outer, upthrown, sides of the rift that a wedge-shaped block of wood would float
(Fig. 10.7a). There would be a corresponding out somewhat lower in water than a block of the same
ward flow of ductile material in the upper mantle. thickness with parallel edges (Fig. 10.7b).
This hypothesis clearly explains the dimensions The theory of Heiskanen & Vening Meinesz
and morphology of rifts. Moreover, it accounts for (1958) breaks down, however, in that the crust is
the presence of negative isostatic anomalies. This is usually thinned beneath rifts and there is no evidence
not because the downfaulted wedge is actually out of for a fault wedge penetrating into the mantle (Artem-
isostatic equilibrium, but rather that it deviates from jev & Artyushkov, 1971). It is readily modified ifit is
the assumptions inherent in theories of isostatic assumed that the faulting and subsidence are re
compensation. The Airy model (Section 2.11.2) as stricted to the upper, brittle, part of the crust and
sumes that equilibrium is attained by differential compensating lateral flow at depth takes place in the
movements of adjacent blocks of crust along vertica.1 ductile lower crust (Section 2.10), whose viscosity is
fault planes until their upthrust is sufficient to bal decreased by the elevated temperatures experienced
ance the mass of the overlying material. In the rift in this region. There is geochemical evidence that the
model described above, however, the subsiding fault lithosphere thins beneath rifts (Perry a/. 1988).
wedge is bounded by inclined rather than vertical A more realistic sequence of events in rift forma
faults so that equilibrium is attained at a somewhat tion (Fig. 10.8) is thus as follows (Bott, 1981):
214 CHAPTER 10
(a)
?^65 km-^
Mantle —i Ductile
< lower
crust
Moho
Upper
mantle
0$k
Sea level
Aberdare detachment Elgeyo detachment
Lithosphere
20 km
216 CHAPTER 10
1/1 -'I
I- WN-.- ~\y-yl^. - /
■l*i
Mantle contraction
Arc-trench system
Volcanic arc
Arc-derived Trench
^ sediments
aM;ri 'n:t;,r"XT^
' ^ / y\
/C' ^ 1 f 'T 'w \<
¥:¥:%¥x*x*x«x*x
Axial Dyke
rrr junction
Volcano—
(a)
Sediment transport
Riftyalley crust rrrjunction
Miogeocline
Foreland basin
occurs as a result of the changing pattern of convec times would represent the initiation of a phase of
tion in the mantle associated with the growth of the splitting of continents sited over the rising limbs of
core (Fig. 10.18).The growth ofthe core would have adjacent convection cells. This mechanism is no
progressively decreased the volume ofthe mantle and longer valid as it is now known that the core stabi
so controlled the extent of the region in which lized early in the history ofthe Earth. It also suggests
convection could take place. The width of convec that plate motion is via the mantle drag mechanism
tion cells is of the same order as their thickness. (Section 11.7.1), which is unlikely.
Hence, early in the history of the Earth the mantle Oxburgh & Turcotte (1974) have proposed that
would have been considerably thicker than today and the lithospheric stresses which give rise to continen
so would have been populated by a few large convec tal splitting are only indirectly related to the forces
tion cells. As the core grew and the mantle became which cause plate motion. They suggested that these
smaller there would have been a series of times at stresses originate because of the Earth's oblate ellip
which the convection cells became smaller and more soidal form, whose radius ofcurvature at the poles is
numerous. Runcorn suggested that each of these greater than that at the equator. When a plate moves
220 CHAPTER 10
xvivX::
Alkaline
magmatism
5%:xSixSxSSxix
V.V.V.V.V.V.V.•••••.V.V.'.V
X-X-XvX*X*X'X'X*X*XvX
:-Xy:;::Xv:::::::x
XWX'X'X-X'SWW^^
wr¥:yx*xW:';":
xwr-x-wy:-:- X'I'X'X-X •x*x-x*:X'X
River flowing down
failed arm
'•*,*»**w***»\'»\*»w»\***0*»*»*»***»
X'I'X'X'
:'X'x:::x:x>:::::
xwxWx
"^A~ Continent Ocean il
x«
SSi:::!®:!;
- Ridge
- :wx<Wx::
Hotspot
Fig. 10.14 Evolution of a continental rift to continental more commonly interpreted in terms of the passage
margins by the concatenation of a series of rift-rift-rift of the lithosphere over nearly stationary hotspots
triple junctions (redrawn from Dewey 8c Burke, 1974). (Section 5.4), and a membrane tectonic origin does
not appear likely as the linear volcanic zones are,
longitudinally across the Earth's surface it must ac within a single plate, all elongated in the direction of
commodate to the changing radius ofcurvature, and motion of that plate rather than forming a radial
this results in intraplate stresses known as membrane pattern. Oxburgh & Turcotte suggested that the East
stresses^ which originate as a result of membrane African rift system was formed as a result of Africa's
tectonics (Fig. 10.19). When a plate moves from northward motion over the past 100 Ma,as revealed
equator to pole,compressional stresses are generated by palaeomagnetic investigations. Tensile stresses
in the centre of the plate as it attempts to adjust to a were generated as Africa moved towards the equator
greater radius of curvature, and radial graben are so that the rift propagated southwards with time.
formed at the edges ofthe plate. When a plate moves There is no evidence, however, of any other radial
towards the equator, however,the stresses generated cracks expected to be formed by this mechanism.
in the centre of the plate are tensional and radial Membrane tectonics consequently does not re
compression ridges form at the edges ofthe plate. It quire the active participation of the mantle in the
was proposed that many types of propagating frac rifting process, the only contribution being passive
ture could form in this way, such as rift valleys and as mantle material wells up into the propagating rift.
intraplate volcanic zones. The latter phenomena are It would seem that membrane stresses are neither
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 221
Mesozoic took place approximately along the suture A deviatoric tension of 100-200 MPa is required
ofthe proto-Atlantic that closed in Palaeozoic times. to rift the continental lithosphere. The horizontal
Wilson (1965) also drew attention to lineaments drag on the base ofthe lithosphere resulting from the
222 CHAPTER 10
Late Devonian-Mississippian
Pennsylvanian-Permian
Granitic basement
Fold Belt
Southern
Oklahoma
\ w Ouachlta
Aulacogen
Fold Belt
500 km
Fig. 10.17 Major geological structures ofsouthern North response to the uplift associated with the buoyant
America. ascent ofthe mantle plume. This would be in accord
with the observation that crustal doming precedes
lateral spreading of material welling up from the rifting. Tensional stresses would then be created as
hotspot is unlikely to be large enough to cause such the lithosphere was squeezed between the weight
tension as the material is oflow viscosity and so would of the uplift and the upthrust of the low density
not effectively couple with the overlying lithosphere. material associated with the mantle plume. If the
0
0
224 CHAPTER 10
whole lithosphere behaved in an elastic manner, an ation ofregional stress whereby a whole continent is
uplift of 2km would produce a tensional stress of placed under tension. Also, previous episodes of
only SOMPa. Bott & Kusznir (1984), however, rifting were not always associated with uplift, such as
recognized that the crustal component of the litho the rifting ofthe North Sea in Mesozoic times.
sphere was subdivided into an upper, brittle layer and The notion that hotspots are responsible for conti
a lower plastic layer. Consequently, only the upper nental rupture has been expanded by Burke (1976;
crustal component ofthe lithosphere would respond 1977). He overcomes the objection that rifting is
elasticaily to stress. This would have the effect of confined to the region of uplift associated with a
concentrating the tensional stress in the upper crust hotspot by postulating that continental splitting
as the lower crust and lithospheric mantle deformed takes place in response to the tensional forces gener
by creep(Fig. 10.20).The amplification ofthe tensile ated by a series ofhotspots which form a line marking
stress by its concentration in the upper crust would the location of the initial rift (Fig. 10.14). Each
increase the stress to a value ofsome 400 MPa,which hotspot would initially give rise to a rift-rift-rift
is more than enough to cause rifting. triple junction (Burke & Dewey, 1973). A modern
This mechanism of stress concentration is clearly example of such evolution is the Afar rift system of
effective in initiating the rifting process. The stress Djibouti (Barberi et 1974), at which sea floor
would, however, be confined to the area of plateau spreading is taking place at two arms in the Red Sea
uplift, and no explanation is provided for the gener and Gulf of Aden and rifting on the third—the
Tertiary northeast Atlantic igneous province extends low density compensation then increase the tension
from west Greenland to Scotland and is character about the hotspot, leading to rifting and the intru
ized by volcanism which correlates in time with the sion of dykes into the rupture (Fig. 10.22b).
initiation of opening of the North Atlantic (White 3 Once commenced,the split propagates away from
et al.^ 1987). This region, then, was most probably the hotspot as the tensional forces are concentrated
underlain by a hotspot which became active some at the tips of the rupture (Fig. 10.22c). It eventually
60 Ma ago and had moved away from it after a period reaches the margins ofthe plate and splits it into two.
of some 3 Ma or less. This hotspot now underlies The tensile stress then gives over to compression near
Iceland, and was responsible for the creation of the the rupture as the ridge push force is initiated.
Greenland-Scotland Ridge (Bott, 1983). The differ It is thus apparent that this model predicts the
ence between volcanic and non-volcanic margins development of two types of rift: those associated
thus arises because ofthe difference in temperature of with doming, extensive volcanism and late stage
the underlying asthenosphere. Mutter (1988b) rifting, and those characterized by initial graben
suggest that the copious igneous material emplaced formation with little or no volcanism. These have
at volcanic margins arises from convective partial been termed 'mantle active' and 'mantle passive'
melting ofthe upper mantle. rifts, respectively, and the model predicts that the
The evidence from volcanic margins argues against former are related to the phase ofinitial rifting under
the adoption of the East African rift system as a the influence of hotspot activity and the latter to
model of continental rupture. In East Africa, for propagation of the rift through a continent placed
example, rifting commenced 25 Ma ago, but separa under tension by plate driving forces. This is not the
tion is still not complete;the North Atlantic splitting only explanation, however, and Dunbar & Sawyer
implies that rifting should have been completed (1988) have proposed that both styles of rifting can
within a timespan ofonly a few million years. Pavoni occur in response to regional horizontal stresses. The
(1993) has suggested that the horizontal extensional different rift types are proposed to be related to
regime of Africa originates from a large mantle different kinds of pre-existing weaknesses in the con
plume;it is probable that the tractional forces on the tinental lithosphere, with volcanic rifts formed over
base of the lithosphere arising in this way are too areas of mantle weakness and non-volcanic rifts over
small to initiate full continental break-up. weakness in the crust. Storey(1995)has summarized
the information available about the role of mantle
10.3.5 Models of continental splitting plumes in the break-up of Gondwana and has
demonstrated that not all episodes of continental
The most reasonable model for continental splitting splitting are associated with plume activity.
seems to require a regional tension to exist, affecting A criticism of the model of continental splitting
the whole plate, and on this is superimposed a described above is that it assumes the rifting process
localized region of hotspot-related uplift which is to be infinitely rapid, so that sea floor spreading
responsible for the initial rupture of the lithosphere. commences simultaneously at every point along the
Once ruptured at a single point, the regional tension rift. This cannot be the case, and it appears likely that
would cause the rupture to propagate rapidly at right the rift would propagate through the continent over
angles to its direction, with the location of the rift a finite period oftime. The process would be akin to
controlled to some extent by regional lines of weak that proposed for propagating ocean ridges (Section
ness in the plate which trend in the appropriate 6.11), but differs in that the rift would propagate
directions. independently through a continent, whereas in the
Bott(1982b; 1992)has recognized three principal propagating ridge model one ridge propagates at the
stages in the rifting process (Fig. 10.22): expense of a nearby ridge receding at the same
1 The continental lithosphere is thrown into a state velocity.
of tension by the trench suction force acting on Courtillot (1982) and Courtillot & Vink (1983)
opposite edges ofthe plate (Fig. 10.22a). have described a simple model of rift propagation.
2 A hotspot develops in the upper mantle by local They envisage that a rift would probably form along
ized convection. This thins, and consequently weak existing lines of weakness in a continent, such as old
ens, the lithosphere. The resulting uplift and its hot. sutures or hotspot tracks. Under extensional forces
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 227
(a)
Enhanced tension
the lithosphere deforms and thins by normal faulting continues to stretch (Fig. 10.23d) until it too be
at shallow levels and plastic necking at depth, to comes invaded by the rifts (Fig. 10.23e). The process
produce an extension ofa few tens ofkilometres over continues by incremental invasion of the rifts until
a few tens of millions of years. Thinning reduces the the tips of the rifts meet and the locked zone is
strength of large tracts of the future plate boundary ruptured to produce two distorted zones on either
and rifting begins. However, the likely nonunifor- side of the ridge (Fig. I0.23f). The model is easily
mity of the lithosphere would cause the plates to modified to encompass continuous, rather than dis
remain attached at a number of locked zones. Even crete, step-wise penetration of the locked zone. The
tually the extensional forces increase to the point at magnetic character of the deformed zones will be
which the tips of the rift begin to enter the locked different from that of typical oceanic crust, and may
zones. The sequence of events during the rupture of well produce a magnetic quiet zone at the new
a locked zone is shown in Fig. 10.23. In Fig. 10.23a continental margin.
the tip of a rift has reached both ends of a locked Evidence for this model is provided from a study of
zone. Oceanic crust is produced along the rift, lead the evolution of the Gulf of Aden (Fig. 10.24),
ing to stretching ofthe locked zone (Fig. 10.23b). In whose history can be deduced by the fitting of mag
Fig. 10.23c the rifts have penetrated the locked zone, netic anomalies on either side of the central ocean
so that its outer parts are carried away by each plate. ridge (Section 4.6). The ridge can be seen to propa
The part ofthe locked zone not penetrated by the rift gate at a rate of 30 mm a"^ in a westerly direction as
228 CHAPTER 10
Continental crust
Locked zone
(c)
(d)
(e)
(f)
Fig. 10.23 Schematic illustration of the incremental the seismicity of the region, in which earthquakes
splitting of a Mocked zone'in a continent by a propagating cluster along the rift and extend into the Afar region
rift (redrawn from Courtillot & Vink, 1983). (Fig. 10.21) into which the rift is penetrating. Fur
ther evidence is provided by tholeiitic basalts in Afar,
which show strong affinities with ocean ridge basalts.
the Gulf opens. Moreover, the magnetic anomalies Another example ofa propagating rift may be present
are interrupted at the coastal magnetic quiet zone so in the GulfofCalifornia, but the intense interruption
that the oceanic margins become younger to the of the ridge in this area by transform faulting makes
west. The model ofa propagating rift is supported by the magnetic anomalies difficult to interpret.
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 229
(a)10 Ma
(c)Present
It is significant that the palaeogeographic recon tion (Fig. 10.26). The overlap would then be equiva
structions ofFig. 10.24a,b show an area ofoverlap of lent to the amount of extension of the continental
the coastlines of Arabia and Africa to the NW of the edges.
propagating rift tip. This slight misfit is partly a Models of continental rifting have largely been
consequence of the model, as the continental mar derived from studies of old passive margins or conti
gins would have formed diachronously as the rift nental regions ofextension where sea floor spreading
propagated westwards. The reconstruction thus at has not yet commenced. There is, however, a re
tempts to fit features of slightly differing age. The stricted number of areas where the transition from
overlaps represent the stretching of the continental rifting to spreading can be observed in action. One
crust which occurred as the crust was ruptured by the such area, for which geophysical data have recently
propagating rift. Similar overlaps are present in the become available, is the Woodlark Basin off north
reconstruction of the continents around the Atlantic eastern Papua New Guinea (Taylor et a/., 1995). In
of Bullard et al. (1965) shown in Fig. 3.2, and are many respects,the scenario there is in accord with the
apparent along the juxtaposed margins of eastern model of continental splittirig described above, that
South America and western Africa. It is possible that is, the progressive truncation of magnetic lineations
these result from the opening of the South Atlantic and uplift above the propagating rift. Other observa
by rift propagation from south to north (Fig. 10.25), tions, however, indicate greater complexity than im
accompanied by southerly rift propagation in the plied by the model. Rather than the rift tip's
North Atlantic (Bonatti & Crane, 1984). This would propagating continuously, it appears that short
be in accord with the observation that the oldest spreading segments can develop instantaneously,
magnetic anomalies are found only in the southern probably reflecting the contrasting strengths of dif
most Atlantic. Courtillot (1982) has calculated that ferent parts of the continental lithosphere. Fiarth-
the mean separation between these overlaps is some quake activity along the developing passive margin is
700 km,and they tend to be about 400 km long and not immediately localized at the plate boundary as
150 km wide. Perhaps the best way of depicting spreading commences. During its evolution, there is
continental reconstructions taking continental dis evidence that the developing ocean ridge reoriented
tortion on rifting into account is, according to Vink instantaneously by about 5° by synchronous jump
(1982), to represent the pre-rift configuration by ing during the recent past, possibly in response to
matching the point ofinitial rifting and allowing the rapidly reorienting stresses. The data even cast some
overlap to increase in the direction of rift propaga doubt on the postulate that oceanic transform faults
Amount of
extension
(a) (C)
Reconstruction
Amount of
overlap <
Initial
rifting
h Rifting
complete
Closure
(b) <d)i
Fig. 10.26 Propagating rift model for continental break after rifting complete,(f) Reconstruction of pre-rift
up (in plan),(a) Continent under tension,(b)Initial configuration. Extension due to rifting gives rise to an
rifting, parallel lines show amount of extension, apparent overlap (redrawn from Vink, 1982, with
(c) Mid-rifting, with sea floor spreading in lower half and permission from the American Geophysical Union).
extension in upper half,(d) Rifting complete,(e) 3 Ma
always reflect the original shape of the rift (Section the more normal areas of continental crust across
7.2). Clearly,then,our models ofcontinental rifting, which the rift subsequently propagated. The former
although generally reasonable, are not perfect in have been termed volcanic margins.
detail. It is unlikely that perfection will be achieved An elegant comparison ofthe two types of margin
even with the availability of these new data, because has been provided by the work of White & McKenzie
the continental lithosphere is so variable in character (1989; see also White, 1993), who have described
that it is doubtful that a single model could satisfy all sections across the rifted continental margin of the
cases. northeastern Atlantic (Fig. 10.27a) derived from
seismic surveys. The profile across Goban Spur
(Fig. 10.27c), situated at the mouth of the English
10.3.6 Volcanic margins
Channel, shows that the continent-oceanic transi
It follows from the theory of continental splitting tion is marked by very little volcanic activity and that
(Bott, 1982b) described in Section 10.3.5 that the continental rifting was dominated by extensional
continental margins ofan ocean formed by continen faulting and the tilting of blocks ofcrust. The profile
tal rifting should be of two types: those associated at Hatton Bank (Fig. 10.27b),at the Atlantic margin
with volcanic products, situated close to the thermal between Scotland and Greenland, shows the pres
anomalies responsible for the initial rifting, and ence ofextensive lava flows,5 km in thickness, which
those with no associated volcanism, which represent dip gently out to sea. At the base ofthe crust there is
232 CHAPTER 10
■ Sediment
Sediment
; Fault
Tilted ^ -- : ; , ' r, : ■
fault X z'' ~''i
', block ^ ^ - X ■ 'X
\\ 'j^ ' L~, ^ Continental
- ,x- ' -' ; crust
Fig. 10.27 (a) Location map. Profiles across (b) the deed, it has been estimated that some 10^ km^ of
volcanic margin at Hatton Bank; (c) the non-volcanic igneous rocks were generated when the Atlantic first
margin at Goban Spur (after White, 1993). rifted. It is thus the presence of the causative hotspot
that is responsible for the differing nature of the
an even larger body of igneous rock some four times continental margins formed by rifting. Although the
greater in volume than the volcanics. Subsequent hotspot, or mantle plume, was probably quite small,
work has shown that very large tracts of the north of the order of a few tens of kilometres in horizontal
eastern Atlantic continental margin are associated section, as it ascended in the mantle, modelling of its
with these supra- and infracrustal igneous bodies behaviour as it reached the cold boundary layer at the
(Fig. 10.28). base of the lithosphere indicates that it would spread
The presence of extensive bodies of igneous rocks considerable distances horizontally (Courtney 8c
at volcanic margins is explained by their proximity to White, 1986), producing a layer of hot material at
the hotspot responsible for the initial breaching of the top of the asthenosphere with a diameter of some
the continental lithosphere. McKenzie & Bickle 2000 km; in the North Atlantic the plume would
(1988) have demonstrated that the amount of mol have been centred on Iceland. This behaviour ex
ten rock generated as mantle material ascends and plains why the igneous products associated with the
melts due to the reduced pressure is strongly depen plume occupy a broad area around it. The partition
dent upon the initial temperature of the mantle ing of these products into volcanic and intrusive
material. Only very small increased temperatures, of fractions depends on their relative density; the major
the order of 100-200 °C, are necessary for the ity of the igneous rocks are relatively dense and only a
amount of molten material produced by this decom small proportion are of sufficiently low density to rise
pression melting to increase very dramatically. In buoyantly thtough the crust to the surface. The
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 233
\^\roU_of_zoneof
^oornilousi^T^
Hotspot
Labradbr
Massive
layered
basalts
Abundant
volcanic
rocks
Vpring Plateau
j ^—v.^Hotspot
Fig. 10.28 (a) Reconstruction of
the North Atlantic region at 57 Ma.
(h)The present North Atlantic.The
locations of massive layered basalts, Hatton
Bank
abundant volcanic rocks,the
hotspot and its associated region of
anomalously hot mantle are shown
(after White & McKenzie, 1989).
thickening of the crust also results in less subsidence ic margins are present at many continental margins
than seen at non-volcanic margins because of isos- world-wide at the location of mantle plumes,and the
tatic effects (Section 2.11). concept of mantle plume-related volcanic margins
Subsequent investigations have shown that volcan has been widely embraced. However, there are some
234 CHAPTER 10
New
^o^vwacting Collision expanding
BC ocean
B C Incipient rift
Incipient
subduction
^^'Oanding oc®^^
(a) (b) (c)
Fig.10.29 The Wilsonian cycle,(a) Continents AB and the predominantly eastern flow of rivers in India is
CD separated by expanding and contracting oceans AD related to drainage down the eastern flank of a dome
and BC,respectively,(b) Collision of AB and CD as BC associated with the plume responsible for the Deccan
closes completely; initiation of subduction beneath A and Traps.
D.(c) Splitting ofsupercontinent ABCD into two new
continents AB' and CD and the growth of a new
expanding ocean B'C. Ocean AD starts to contract. 10.4 The Wilsonian cycle
The Pacific Ocean opened about 300 Ma ago and will
inconsistencies to this simple idea. The volcanic close completely in a further 200 Ma, so that the
sequences do not thin with distance from their re lifetime of an ocean basin is about 500 Ma. Plate
lated hotspot as might be expected. The effects ofthe tectonic processes, however, have probably been
hotspots appear to diminish very rapidly with time active at least since Early Proterozoic times, some
after rifting and normal oceanic crust is found imme 2500 Ma ago; consequently, there must have been
diately adjacent to volcanic margins; it might be several cycles of ocean creation and destruction dur
expected that the effects of the hotspot would have ing the Earth's history. Very little remains of these
been longer-lived. There are also locations where ancient oceans, although fragments are preserved in
volcanic sequences appear to be present with no ophiolite assemblages (Section 2.5), but their exist
accompanying hotspot, for example the passive mar ence is implied by continental reconstructions.
gin of the eastern USA (Holbrook & Kelemen, The periodicity of ocean formation and closure
1993). Indeed, in this location the generation of the is known as the Wilsonian cycle (Dewey & Burke,
observed lower crustal seismic velocities would seem 1974), named after J. Tuzo Wilson in recognition
to require a hotspot at 550°C above normal. It thus of his major contribution to the theory of plate
appears that the division of passive continental mar tectonics.
gins into volcanic and non-volcanic types is too Figure 10.29a shows a section through an Earth
simplistic. on whose surface exist two continents AB and CD.
Also associated with mantle plumes are thick, Ocean AD is expanding and BC contracting. Some
extensive flood basalts (Fig. 5.9), such as the Deccan time later (Fig. 10.29b) ocean BC closes completely,
Traps of western India and the Siberian Traps of and this phase of plate movements ceases with the
northern Russia. These represent the volcanic expres collision of margins B and C, the construction of a
sion of plumes which were often associated with Himalayan-type mountain range at the suture BC
continental rifting (Section 10.3.4). Cox (1989) has and the formation of a supercontinent AD. Ocean
demonstrated that the topography associated with AD must develop subduction zones at its margins to
such plumes is still in evidence today, and is respon maintain constant global surface area, although the
sible for characteristic drainage patterns,for example reasons for the initiation of subduction are un-
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 235
known; sediment loading does not appear to be the controlled, to some extent, by the formation of
cause (Cloetingh et al.^ 1984). The trench suction supercontinental assemblages.
force (Section 11.6) places the supercontinent AD
under tension.Ifother conditions are extant(Section
10.3.2), the supercontinent will rift (Fig. 10.29c) at 10.5 Vertical movements
B'C, not necessarily along the previous suture BC. associated with plate tectonics
Two continents AB' and CD now separate across a
10.5.1 Introduction
new expanding ocean B'C, and to compensate for
the newly formed oceanic lithosphere, ocean AD The study of plate tectonic processes is frequently
must start contracting by the continued develop concerned with the horizontal movements of plates
ment of subduction zones at continental margins A over the Earth's surface. It must be recognized,
and D. This phase will continue until A and D however, that associated with these horizontal mo
collide. Vink et al. (1984) have demonstrated that tions are vertical movements, which play a very im
the strength of the continental lithosphere decreases portant role in the evolution of the Earth. Many of
as the thickness of the crust increases. Conse these vertical movements are implicit in the phenom
quently, the large crustal thicknesses associated with ena described in previous chapters of this book, and
suture zones make them likely locations for future will be briefly listed. Others have not been previously
rifting. There will thus be a tendency for new oceans considered, in particular the predominantly down
to open along the suture where a former ocean ward motion of the continental lithosphere associ
closed. ated with sedimentary basin formation, and these
Present day analogues of the oceans involved in will be discussed in more detail.
the first and last phases ofFig. 10.29 are: Fig. 10.29a, Hotspots (Section 5.4) produce vertical move
AD = Atlantic Ocean, BC = Pacific Ocean; Fig. ments of both continental and oceanic lithosphere
10.29c, AD = Pacific Ocean, B'C = Red Sea. because of thermal expansion and because they iso-
The Wilsonian cycle will continue to operate until statically support positive, topography by a form of
the concentration of radioactive elements within the Pratt-type compensation (Section 2.11.3). In the
Earth becomes too low to provide sufficient heat ocean basins, passage of the lithosphere over a hot-
energy for the driving of plate motions. spot creates an aseismic ridge, which progressively
Gurnis(1988) has made numerical simulations of subsides as it moves away from the hotspot. Beneath
continental aggregation and dispersal, and has postu continents, hotspots create pleateau uplifts which
lated that plate motion during the Wilsonian cycle is may subsequently rupture to produce rift valley
episodic rather than continuous. He suggests that systems(Section 10.1.6) and, ultimately, aulacogens
continents tend to aggregate over cold downwellings (Section 10.2)or complete lithospheric rupture(Sec
in the mantle, where they act as an insulating blanket tion 10.3).
(Anderson, 1982). The mantle consequently heats In the oceans, the elevation of ocean ridges is
up, altering the convection pattern, and the super- similarly supported isostatically by a hot,low density
continent splits in response to the resulting tension. region in the upper mantle (Section 6.3), and the
The continental fragments then move towards the gross topography ofthe sea floor is controlled by the
new cold downwellings resulting from the changed progressive cooling of the lithosphere as it moves
convective regime. Gurnis points to the fact that the away from the ridge (Section 6.9). Similarly, the
continents, except Africa, are currently moving to juxtaposition of oceanic lithosphere ofdifferent ages
cold regions of the mantle, characterized by geoid at fracture zones produces abrupt scarps on the sea
lows, few hotspots and high seismic velocities. Pan- bed (Section 7.3). The morphology of subduction
gaea was probably positioned, 200 Ma ago, over zones (Section 8.1) is controlled by plate motions.
what is presently the Atlantic-African geoid high. The accretionary wedge(Section 8.7)is progressively
Since Africa has moved only slowly with respect to built up by the stacking of thrust slices of sediment
the hotspot reference frame, it appears that Pangaea scraped off the top of the downgoing slab by the
may have been situated over a geoid high prior to overriding plate. Island arcs maintain their positive
break-up, in accord with the model. It would thus topography by the continuous construction of
appear that the pattern of mantle convection may be the volcanic edifice by the addition of magmatic
236 CHAPTER 10
products produced in the upper mantle immediately would be expected. More sediment can be accom
above the BeniofF zone. Marginal seas and back-arc modated in the flexural model, as the margins ofthe
ridges owe their negative and positive topographies basin are able to prograde seawards(Walcott, 1972).
to tension in the rear-arc area generated as a result of However, both models suggest that gravity loading
plate movements (Section 8.9). causes a basin subsidence which is far smaller than
On the continents, vertical movements can occur the observed sediment thickness.
in a number of different environments. At Andean- Watts & Ryan (1976) applied this method to the
type plate margins (Section 9.2) a mountain range Gulf of Lion, and progressively backstripped the
develops as the result of the upward migration of sediment layers to determine the basement depth
magmas from the Benioffzone in a fashion similar to throughout the history ofthe basin. When compared
island arcs but accompanied by large volumes of with known basement depths (Fig. 10.30), they
acidic material. In collisional orogens (Section 9.3) found that the subsidence had been much greater
Himalayan-type mountain ranges grow vertically by than would have arisen from the sedimentary load
the stacking of thrust sheets in the upper crust and ing. This additional loading was attributed to some
ductile movements at depth. In the distal regions form oftectonic subsidence or'driving force',so that
of such orogens, impactogens (Section 10.1.3) may the main role of the sediment loading is to amplify
be developed in response to tensional stresses associ the subsidence produced by this primary mechanism.
ated with collision. Transform and transcurrent In many cases this driving force originates directly or
faults which have a sinuous trend may be associated indirectly from plate motions.
with vertical movements resulting from transpres- Sleep (1971) and Sleep & Snell (1976) attributed
sion and transtension (Section 7.4.2) giving rise to the subsidence of basins situated on continental
small compressional uplifts and pull-apart basins, margins to simple thermal contraction following
respectively. continental rupture. Thermal expansion ofthe litho
sphere would have taken place during rifting as a
10.5.2 Sedimentary basin formation thermal anomaly develops under the continent. This
expansion causes a plateau uplift which is isostatically
A category of structures not covered in the previous supported by the hot,low density region in the upper
section concerns those sedimentary basins which mantle. As rifting is completed and an ocean grows,
occur in intraplate environments out of the direct the continental margin becomes removed from the
influence of plate margin processes. The problem of hot mantle, and subsidence takes place. Some thin
their formaLtion was first recognized in the 19th ning of the crust is necessary for subsidence to
century when it was observed that more than 12 km continue below sea level, as otherwise the crust
shallow water sediments had been deposited in the would only sink back to its pre-rifting level. This, in
Appalachian geosyncline. It can be shown, however, fact, is a major criticism of this mechanism of basin
that the sediment loading itself cannot cause the subsidence as there is often no evidence of the
basin subsidence, as described below. extensive supra-crustal erosion necessary to produce
In xhc gravity loading hypothesis it is assumed that the observed subsidence, and sub-crustal erosion
the lithosphere reacts by faulting or flexing solely in seems an unlikely process.
response to the sediment load. By using either an Bott (1971) has presented a different model for
Airy-type model, in which isostatic equilibrium is the development ofsedimentary basins on the conti
maintained by movement along vertical faults, or a nental shelf. Their development is viewed as the
flexural model, it is possible to predict, if the densi response of the crust after rifting to the progressive
ties of sediment, crust and mantle are known, the loss ofgravitational energy associated with the juxta
depth to basement at any stage in the filling of the position of thick continental crust and thin oceanic
basin (Steckler & Watts, 1978). It has been found crust. At that time, geophysical observations had
that the Airy-type model predicts a sediment fill of suggested that the transition from continental to
only about 2.4 times the initial water depth before all oceanic crust, which was quite abrupt immediately
the available water is displaced by sediment. Since the after rifting, actually takes place over a distance of
shallow water sediments were deposited in water less 50-100 km,although more recent data indicate that
than 150 m deep, a thickness of only about 350 m this is only the case for continental margins not
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 237
Local
loading
E 1
Flexural ^
loading g
Subsidence due to
Ocean ridge
sediment loading
Fig. 10.30 Summary curves for the subsidence history moves away from the thermal anomaly. The high
recorded in boreholes in the Gulf of Lion. Gaps in the temperatures existing at this time permit a broaden
data correspond to the Mediterranean salinity crisis ing of the continent-ocean transition by hot creep
between 7 and 5 Ma ago. Local and flexural loading from the base of the continental crust towards the
models are shown (redrawn from Watts & Ryan, 1976, ocean (Fig. 10.31). The juxtaposition of thick, low
with permission from Elsevier). density continental crust and thin, dense oceanic
crust causes the continental crust to be thrown into
affected by extensive volcanism (Mutter et al.^ tension. To maintain isostatic equilibrium as lower
1988b). Bott suggested that the continent-ocean crustal material moves oceanwards, the continental
transition becomes gradational with time, and the shelfsubsides and the continental slope and adjacent
readjustment allows the development ofsedimentary oceanic crust rise. These crustal movements are bal
basins parallel to the continental margin. When anced by the flow of asthenosphere inwards beneath
rifting takes place, the whole region is uplifted in the continent. The brittle upper crust responds to
response to the hot,low density material in the upper these movements by faulting and the development of
mantle. Regional subsidence then occurs as the crust basins on the continental shelf by wedge subsidence.
238 CHAPTER 10
Brittle layer
The development of these basins is thus associated and the intrusion of granites during the extensional
with the maintenance of isostatic equilibrium as the episoide associated with continental rupture. The
potential gravitational energy associated with the intrusion of igneous material would have weakened
margin is gradually dissipated. the continental lithosphere and provided a zone of
The above mechanism cannot provide an explana regional stretching in which cratonic basins formed
tion for the formation of basins far from plate mar almost simultaneously over a wide region.
gins. Haxbyef a/. (1976) have proposed a model for
these intraplate basins with reference to the Michigan
Basin. Basin subsidence is attributed to the flexural
response of the continental lithosphere to a load
arising from the effects of a thermal anomaly in the
upper mantle. The process is initiated by the action
Hotspot,
of a hotspot which thins the lithosphere. so that
asthenospheric material ascends to the base of the
crust (Fig. 10.32a). The hotspot then causes thermal
metamorphism of the lower crust in which gabbroic
material is converted to garnet eclogite. Garnet Asthenospnere/
eclogite is denser than normal lower crust, and the
additional mass is initially supported by the negative
buoyancy of the hotspot. When the hotspot decays,
the low density support is removed and the litho
sphere subsides by flexing (Fig. 10.32b). Thus, basin
subsidence in this model is caused by the excess
weight of eclogite in the lower crust, and no major
deformation should be visible at the surface.
Klein & Hsui (1987) have noted that cratonic
basins of North America, Europe, Africa and South
America all formed during the period 550-500 Ma
and have similar sedimentation histories and dates of
regional unconformities. They suggest that their
Fig. 10.32 Model for the formation ofthe Michigan
common dates offormation may have arisen by their Basin:(a) development of a hotspot in the upper mantle
originating during the break-up of a Late Precam- and its ascent to the base of the crust which it
brian supercontinent (Section 12.2). The supercon- metamorphoses to eclogite;(b) removal ofthe hotspot
tinent may have acted like a heat lens which caused and sinking of the basin in flexural response to the load of
partial melting of the lower crust and upper mantle. eclogite (after Haxby era/., 1976).
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 239
Sediment .Sediment
~ Crust"'~
, 2,8 Mg m~3
Mnhn \"/r O ^ 'i""'V \
Upper mantle
0 50 km ij
3.3 Mg m~^
Fig. 10.33 Sedimentary basins paired with an eroding the subsidence as the response to stretching of the
mountain range (redrawn from Bott, 1964, with continental lithosphere and subsequent cooling as it
permission from Nature 201, pp. 1082-4. Copyright © relaxes to its predeformation state. On the initiation
1964 Macmillan Magazines Ltd). of basin formation a unit length of lithosphere is
abruptly stretched to a length P (Fig. 10.35). Thin
An alternative mechanism of intracratonic basin ning of the lithosphere gives rise to normal faulting
subsidence has been proposed by Bott (1964) in and rapid subsidence, and is accompanied by an
which sedimentary basins are paired with mountain upwelling of the asthenosphere. As the thermal per
ranges(Fig. 10.33). The uplift ofa mountain range is turbation associated with the upwelling of the asthe
followed by rapid erosion, with the clastic sediments nosphere decays,the lithosphere thickens and further
transported into adjacent basins. Erosion causes the slow subsidence takes place which is not accompa
mountain range to become overcompensated and nied by faulting. The crust remains permanently
out of isostatic equilibrium. To regain equilibrium, thinned even after the decay of the thermal perturba
the mountain range rises and mantle material flows tion. Calculations based on an isostatically compens
beneath it. Renewed uplift promotes further erosion, ated system show that subsidence only takes place for
and so the cycle continues until the mountain range crustal thicknesses which were initially in excess of
disappears or the stress differences become too small 18 km. The amount of subsidence and the heat flow
to cause further uplift. This behaviour differs slightly depend only on the amount of stretching.
from classic Airy isostasy in that the mantle flows in Figure 10.36 shows the relationship between the
from localized regions only, rather than from the thermal component of subsidence and the square
whole region beneath the depth of compensation,
which may be due to the mantle's possessing a small Sediments
finite strength, making unrestricted flow impossible.
Brittle layer
The model was subsequently modified by Bott(1976)
in the recognition that only the upper crust can
behave in a brittle manner. Rather than occurring
within the mantle, the inflow of material at depth is
jMotio
believed to occur in the ductile lower crust 1 -»■ Mantle
(Fig. 10.34), which stretches and thins, producing
tension in the upper crust which responds by faulting. Fig. 10.34 Graben formation by faulting in the brittle
One of the most widely accepted models of basin upper crust and ductile flow in the lower crust (redrawn
formation is that ofMcKenzie (1978), which views from Bott, 1976, with permission from Elsevier).
240 CHAPTER 10
r(°c)
(a) 0 1000
hJTime(Ma)
r\
(b)
.o[
pa-
a/p
I
•VSlli
Fig. 10.36 Thermal subsidence as a function of the
square root oftime for the model shown in Fig. 10.35,
t-^co shown in terms of both water depth and sediment
thickness (redrawn from McKenzie, 1978, with
(c) permission from Elsevier).
100-
E
Fig. 10.35 Subsidence model combining stress and
§
thermally based mechanisms,(a) At time r = 0 a block of E
continental crust is extended by a factor p, causing
subsidence,(b) Hot asthenosphere wells up to replace the (0
thinned lithosphere.(c) Further subsidence as the thermal X
a>
those basins which are likely to contain accumula formation of marginal seas (Section 8.9); and sec
tions ofoil and gas(MacKenzie & McKenzie,1983). ond,in the neighbourhood ofan incipient continen
The model does not take account ofthe radioactiv tal rift, by the mechanisms described in Section 10.3.
ity of the continental crust and adopts a simplified Consequently, the extensional stresses required in
version of the variation of temperature with depth. McKenzie's model would only be expected to occur
However,the inclusion of these effects is unlikely to within a few hundred kilometres of destructive or
alter the general form of the results. The model has incipient accretionary plate margins, or perhaps in
been refined by White & McKenzie (1988), who the vicinity of passive continental margins by the
postulate a stretching ofthe lithospheric mantle over mechanism proposed by Bott(1971). It is difficult to
a wider area than the crust in order to explain the envisage, however, how the mechanism could apply
post-rift sedimentary onlap observed in many basins. to basins within the interior of continental plates
A major objection to the model is that observa unaffected by indentation tectonics, especially since
tions of sedimentary basins do not reveal the large focal mechanism solutions of earthquakes in this
amounts of extension required. McKenzie (1978) environment indicate that they are subject to com-
has presented evidence from several regions which pressional stresses.
strongly suggests that the dip ofthe high angle faults McKenzie's model of basin formation involves
seen at the surface decreases with depth. In particu brittle failure of the upper crust and ductile (pure
lar, seismic reflection sections reveal that rotation of shear)stretching ofthe lower crust and mantle litho
the sedimentary reflectors has occurred, indicating sphere. Wernicke (1985) has proposed a different
that the faults are listric. The curved nature of these model which entails simple shear only, possibly af
faults prevents accurate estimates being made of fecting the whole ofthe lithosphere, along low angle
their associated extension from surface observations normal faults. The model was derived in explanation
and it is proposed that listric faults conceal large of the structure of the Basin and Range Province of
extensions. the western USA (Wernicke, 1981), but it may also
Barton & Wood (1984) have tested McKenzie's be applicable to rifting in general (Section 10.1.6)
model in the evolution of the Cfentral Graben of the and the evolution of passive continental margins
North Sea (see also White, 1989), which is probably (Lister et 1986). The model arose from the
an aulacogen (Section 10.2) associated with the recognition in the Basin and Range area of a region
opening of the Atlantic. Stretching estimates were ally extensive low angle detachment beneath normal
obtained by seismic refraction experiments over fault systems, flanked by an adjacent zone of ductile
stretched and unstretched crust, and independently deformation whose maximum elongation is in the
from the history ofsubsidence revealed by well data. same direction as that implied by the normal faults.
Both methods suggest a P value of about 1.6 for the The majority of ductile deformation was observed to
last extensional event. Seismic refraction experi pre-date the episode of extension.
ments in the Aegean Sea, a back-arc basin (Section The model for the development of the simple
8.9) of the Hellenic Trench, suggest a P value of shear model is shown in Fig. 10.38. The undeformed
about 2. These data and model computations sug section of Fig. 10.38a is affected by a major, shal-
gest that large extensions, of the order of p = 2, are lowly dipping (10-30") fault penetrating the crust
required to produce a basin fill of 4.5 km thickness. and, perhaps, the whole lithosphere (Fig. 10.39).
The lack of a proposed mechanism for the origin of Within the sedimentary basin so formed, coarse
the regional stresses responsible for the extension is clastic sediments accumulate. After continued fault
the least satisfactory aspect of McKenzie's model. displacement (Fig. 10.38c), the sedimentary basin
Consideration ofthe forces acting on plates (Section and its basement undergo penetrative brittle defor
11.6) and the mechanism of continental splitting mation with the formation of a normal fault duplex,
suggests that there are only two environments in having the internal geometry of a number of highly
which regional tensile stresses may occur in the rotated faults and thin fault blocks. Extension is not
continental lithosphere: first, in the area on the sufficient to expose the basal detachment. The thin
landward side of an Andean-type mountain range, ning ofthe hanging wall ofthe detachment resulting
where the trench suction force creates regional ten from this extension causes the tectonics of the shear
sion by a similar mechanism to that proposed for the zone to be raised into the zone of brittle behaviour. A
242 CHAPTER 10
Time at_
)mm a~^ o
Future upper plate
of core complex
K-15 km-H
Incipient faults
3 Ma
Imbricately 25km^-
extended range
Future "core-complex
8 Ma upper plate
32 km- H
14Ma
Denudation = 20 km
Depth = 25 km
Preorogenic datuhri
^
Orogenic c.astics: Firie
hoarseclastic-lacustrine
clastic
Total extension = 72 km (100%)
irrmn Highly attenuated'
rocks
Ductile shear zone
Fig. 10.38 Model of an extensionai simple shear system Wernicke, 1985, with permission from the National
in the upper and middle continental crust (redrawn from Research Council of Canada).
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 243
Limit of
Upper significant
crustal Thinnest upper crustal 'Moho Topographic
breakaway Extensional crust extension hinge' Culmination
allochthons ^DiscrepantO Ju
zone' 0 I
Proximal Distal
Mantle
80
Uthos
Asthenosph
Tectonic:
Thermal:
+ +
Zone A ZoneB ZoneC Zone D ZoneE
Fig. 10.39 Model of an extensional simple shear system Wernicke has indicated the possibility that the
involving the entire lithosphere (redrawn from Wenicke, whole lithosphere may have failed by simple shear
1985, with permission from the National Research (Fig.10.39). Five different zones can be postulated
Council of Canada). on the basis of their subsidence/uplift history.
Because mantle lithosphere is denser than the asthe-
part of the upper plate may begin to extend beyond nosphere, its thinning should result in uplift. Con
adjacent areas (Fig. lO.SSd) and, if the extension is versely, thinning of the lower density crust should
sufficiently large, folds may form in the detachment result in subsidence. The subsidence resulting from
as a result of the isostatic rebound of the unloaded extensional strain is referred to as tectonic subsidence^
footwall. Old rocks then become exposed in the and that related to subsequent cooling ofthe mantle
cores of anticlines. These 'core-complex' ranges are lithosphere as thermal subsidence. In zone A the
highly extended regions >vhich passed through the lithosphere is unstrained, but a small amount of
stage of imbricate normal faulting before becoming uplift may occur from the isostatic edge effect result
thinned by low angle normal faulting to expose the ing from the unloading ofthe crust by the breakaway
footwall of the detachment. Large fault-block ranges fault, to which zone A is coupled.In zone B the crust
occur adjacent to the core complex, which exhibit is extended but not the mantle lithosphere, and so
little internal distortion and ride passively on the subsidence takes place. The distal boundary of zone
hanging wall. B is defined by the point at which the shear zone cuts
Examples ofall the stages ofdevelopment of Basin the mantle. In zone C both crust and mantle are
and Range structures are found in the western USA. extended, producing a gradual change from subsid
The model is supported by deep seismic reflection ence to uplift. In zone D only the mantle lithosphere
profiling, which has identified a fault which may is thinned, producing uplift and broad doming of
extend to a depth of 20 km with a displacement of the surface. Erosion may thin the crust so that the
30-60 km. This model, unlike the pure shear model subsequent thermal subsidence produces an overall
ofMcKenzie(1978),can account for the close juxta subsidence of the zone. Zone E is unstrained, but
position of brittle deformation and ductile deforma may be uplifted by its coupling with zone D and may
tion and also for the temporal separation of these eventually undergo overall subsidence because of
different types of deformation. erosion. Magmatism may play an important role in
244 CHAPTER 10
the evolution of such a region, as the extension in heterogeneous and patchy. In this case there would
proximal areas may induce melting ofthe middle and be localized uplift and subsequent thermal subsid
lower crust and the rest of the lithosphere. ence within any particular faulted basin. This may
Wernicke has applied this model to Arizona and explain several problems associated with the end
Utah and also to the Red Sea, where the uplift of the member mechanisms such as the differing stretching
Arabian plateau to the northeast may be the result of estimates for the same basin provided by different
an extensional fault system dipping in that direction. techniques, the variations in dip of detachment
Two rather different models of basin formation zones in large basins, the regional uplift associated
have been described which entail both an initial with parts ofsome basins which has led to the erosion
tectonic subsidence and a later thermal subsidence. ofearly fault blocks and the presence ofvolcanism in
The model of McKenzie (1978) thins the lower basins with only small upper crust extensions. Cow
lithosphere homogeneously by ductile extension be ard proposes that concentrated stretching of the
neath an upper crust thinned by faulting, and super lower lithosphere below an extensive zone of upper
imposes a thermal subsidence on an early fault- crustal extension may develop instantaneously or via
bounded basin. The model of Wernicke (1985) off an initial stage of uniform extension throughout the
sets the regions of lower lithosphere stretching and lithosphere,followed by extension ofthe lithospheric
thermal subsidence. Coward (1986) considers that mantle and lower crust over a relatively narrow area
these models represent end members in a range of while the upper crust extends over a much wider
possible mechanisms. Coward suggests that there region. The model appears to explain satisfactorily
will often be an overlap between the zones of upper structures in the Basin and Range Province and the
and lower lithosphere stretching, which is likely to be northeastern European continental shelf.
Mountain
(a) Forebulge Foreland
range
basin
'7/////////(/77-^
OrCist' ^;
Mantle
Lithosphere
Asthenosphere
Forebulge
Orogenic
Fig. 10.40 (a) The formation of a
wedge
foreland basin by lithospheric
About flexure associated with the weight of
2-4 km a mountain range (after Coward,
1994a).(b)Sketch of the
stratigraphic pattern of a foreland
basin, showing the onlap at
About 50-100 km
positions Xi to corresponding
to chronostratigraphic time lines Tj
to (redrawn from Allen & Alien,
1990).
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 245
Fig. 10.41 Foreland basins associated with Alpine the collision transports material laterally. The fore
mountain ranges (redrawn from Allen etal., 1986). land basin extends much further into the surround
ing craton than the margin of the load and forms in
advance of the thrust front of the mountain range.
Foreland basins associated with the Alpine chain are
10.5.3 Foreland basins
shown in Fig. 10.41.
Continental collision results in a Himalayan-type The shape of the basin depends on the flexural
mountain range with a thickened crust(Section 9.3). rigidity (Section 2.11.4) of the lithosphere beneath
This additional mass represents a load which causes and surrounding the mountain belt; a low value,
flexure (Section 2.11.4) of the adjacent lithosphere arising in young, hot, weak lithosphere, gives rise to a
and creates a depression peripheral to the mountain narrow, deep basin, whereas a high value, arising in
range, bounded on the side distant from the moun old, cool, strong lithosphere, produces a wide basin
tains by a forebulge (Fig. 10.40a). An example ofthis with a better developed forebulge (Coward, 1994a).
is the development of the Indo-Gangetic Plain Basin Varying flexural rigidity of the lithosphere can thus
adjacent to the Himalaya. This flexural depression is cause the character of the foreland basin to change
termed a foreland basin or foredeep, and it is the along strike.
depository for the sedimentary material (molasse) Unlike basins formed by lithospheric stretching
that pours off the newly elevated mountain range as (Section 10.5.2), foreland basins do not experience
it experiences erosion and as the active thrust belt of a temporal diminution in the rate of subsidence.
246 CHAPTER 10
\
^ \ D
CD
h5 "S
-10
Pin line
5 km
inversion in the Upper Cretaceous
H Upper Cretaceous Jurassic and
(redrawn from Hayward & Graham,
Lower Cretaceous
HI Trlasslc V=H 1989).
and subsidence persists over a long period of time moving thrust sheets in addition to the foreland
(Coward,1994a).Ifthe lithosphere rheology is visco- region. Such basins are referred to asplayback basins.
elastic rather than elastic, the flexural rigidity would
decrease with time. This would cause the basin
10.5.4 Inversion
to narrow and deepen and the forebulge to become
higher (Allen & Allen, 1990). Many regions have undergone reversal in the sense of
As a result of the formation mechanism, the vertical movement at different stages in their evolu
sediment thickness in a foreland basin decreases tion. This reversal is known as inversion., and is
away from the mountain front to a feather edge on referred to as positive, the most common type, when
the forebulge. Close to the mountain range the a region of subsidence becomes one of uplift and
sediments are coarse grained and deposited in a negative when uplift is followed by subsidence. In
shallow water or continental environment; at the version can arise from several mechanisms(Coward,
feather edge they are fine grained and often tur- 1994b). Plate motions can cause inversion as a result
biditic. The sediments thus form a wedge-shaped of the tectonism associated with continent-
unit whose stratigraphy reflects the subsidence his continent collision, resulting in the uplift and de
tory of the basin as it grows and migrates outwards formation ofa sedimentary basin. At major strike-slip
as convergence continues (Allen & Allen, 1990). faults, inversion can occur as a result oftranspression
Figure 10.40b shows the progressive onlap of sedi and transtension (Section 7.4.2) at bends in the
ments onto the foreland. The stratigraphy is thus fault. Inversion can also occur as blocks rotate,
characterized by units which thin laterally, overstep narrow and shorten. Isostatic mechanisms can also
older members or may be truncated by erosion. result in inversion. Isostatic rebound (Section
As convergence continues in the mountain range, 2.11.5) can cause uplift as a surface load, such as an
small sedimentary basins may form on the top of icecap or mountain range, is removed by melting or
CONTINENTAL RIFTS, CONTINENTAL MARGINS AND SEDIMENTARY BASINS 247
erosion, causing the structure to be overcompen- Beaumont, C. & Tankard, A.J. (eds) (1987) Sedimentary
sated and followed by progressive uplift in an at Basins and Basin-Forminpf Mechanisms. CSPG Mem.12/
tempt to regain isostatic equilibrium. On a smaller AGS Spec. Publ. 5. Can. Soc. Petrol. Geol.
scale, the diapirism of evaporites can cause uplift as Bosworth, W. (1987) Off-axis volcanism in the Gregory
the evaporites mobilize and ascend because of their Rift, east Africa: implications for models of continental
rifting. Geolopfy 15,397-400.
low density. Lithospheric heating can cause uplift by
Bott, M.H.P.(1982) The mechanism of continental split
doming above a hotspot, which is followed by sub ting. Tectonophys. 81,301-9.
sidence as the effects of the hotspot are removed. Burke, K. (1977) Aulacogens and continental breakup.
Extensional faulting can cause uplift as the litho- Ann. Bsv. Earth planet. Sci. 5, 371-96.
sphere is thinned and rises isostatically (Section Burke, K. & Dewey, J.F. (1973) Plume generated triple
10.5.2). junctions: key indicators in applying plate tectonics to
In orogenic belts,inversion is frequently caused by old rocks.J.pieol. 81,406-33.
the reversal of motion on major dip-slip fault sys Courtillot, V.& Vink, G.E.(1983) How continents break
tems. Figure 10.42 shows sections through the up. Sci. Am. 249(1),40-7.
Broad Fourteens Basin (offshore northwest Holland) Coward, M.P. (1986) Heterogeneous stretching, simple
shear and basin dc\c\opmcnt.Earth planet. Sci. Lett. 80,
before and after inversion. The bounding normal
325-36.
faults of the graben present in Upper Cretaceous McKenzie,D.P.(1978)Some remarks on the development
times have been reactivated as thrusts in response to ofsedimentary basins. Earth planet. Sci. Lett. 40,25-32.
subsequent compression, leading to uplift and fold Oxburgh, E.R. & Turcotte, D.L. (1974) Membrane tec
ing of the graben fill. It is clear that such inversions tonics and the East African Rift. Earth planet. Sci. Lett.
are indicated by breaks in the stratigraphic record. 22,133-40.
Similarly, thrust systems may be reactivated as nor Watts, A.B. 8c Ryan, W.B.F.(1976) Flexure of the litho-
mal faults to produce negative inversion. sphere and continental margin basins. Tectonophys. 36,
25-44.
Wernicke, B. (1985) Uniform sense simple shear of the
continental lithosphere..C/»w./. Earth Sci. 22,108-25.
White, R.S. 8c McKenzie, D.P.(1989) Magmatism at rift
Further reading zones; the generation of volcanic continental margins
Allen, P.A. & Allen, J.R. (1990) Basin Analysis, Principles and flood bzszdts.J.pfeophys. Res. 94,7685-730.
and Applications. Blackwell Scientific Publications, Ox White, R.S., Spence, G.D., Fowler, S.R., McKenzie, D.P.,
ford. Westbrook, G.K. 8c Bowen, A.N.(1987) Magmatism at
Bailey, D.K.(1983)The chemical and thermal evolution of rifted continental margins. Nature 330,439-44.
rifts. Tectonophys. 94,585-97. Whitmarsh, R.B., Avedick, F. 8c Saunders, M.R. (1986)
Barberi, P., Bonatti, E., Marinelli, G. & Varet, J. (1974) The seismic structure of thinned continental crust in the
Transverse tectonics during the split of a continent; data northern Bay of Biscay. Geophys. J. Roy. astr. Soc. 86,
from the Afar Rift. Tectonophys. 23,17-29. 589-602.
11/ The mechanism of plate tectonics
Complete shell of Continental Oceanic small element of the body and r the distance of the
continental lithosphere lithosphere element from the axis of rotation. The moment of
lithosphere
inertia of a uniform sphere is given by IMR^S,
where M is the mass of the sphere and K its radius.
Newton's laws of mechanics for linear motion state
that the momentum (mass x velocity) of a system is
conserved unless an external force acts upon it.
Radius 3500 km Radius 6370 km These laws apply equally to angular (rotational)
motion, in that angular momentum (moment of
inertia x angular velocity) is conserved unless the
system is acted upon by an external torque.
Since the Earth's mass remains constant,any deter
mination of the ancient moment of inertia of the
Fig. 11.2 Expanding Earth model. Earth would allow a calculation of its ancient radius
and so demonstrate if any expansion had occurred.
The theory behind any such determination is com
and dispersed, while mantle material welled up into plicated by the fact that momentum is conserved
the consequent gaps to form the oceans (Fig, 11,2). within a system comprising the Earth and the Moon.
Independent evidence for the expanding Earth Throughout geological time the angular momentum
hypothesis appeared to be provided by certain theor inherited from the fragments that accreted to form
etical physicists, who suggested that the universal the Earth and Moon has been progressively parti
gravitational constant was decreasing with time as tioned between the two bodies, by a mechanism
the universe expanded and its constituent matter known as tidal interaction, in such a way as to reduce
became more widely dispersed. Gravitational forces the rotational energy of the system. At present the
are responsible for binding the Earth into a spherical stage has been reached in which the Moon spins very
form, and since the gravitational constant directly slowly and must consequently lie at a greater distance
controls the magnitude of the force of attraction from the Earth than in the past,so that momentum is
between masses, its decrease would imply a progres conserved in its orbital motion. The tidal interaction
sive relaxation of the binding forces and an increase of the Moon on the Earth is similarly causing the
in the Earth's radius. latter's angular rotation to decelerate. Since the
The most recent versions of the expanding Earth number of rotations in a complete orbit of the Sun
hypothesis correlate the period of rapid expansion determines the number of days in a year, the year in
with the break-up and fragmentation of Pangaea the past would have consisted of more days than at
since 200-300 Ma. These argue that continental present. This also implies that the length of the day
reconstructions can be arranged more accurately on a has progressively increased.
globe ofsmaller radius, and propose that during this Transfer of angular momentum from Earth to
period the surface area of the Earth increased by a Moon thus causes an increase in the length of the
factor of three while the radius increased from an day. A further contributor to this phenomenon
initial value of70% ofits present value. This calcula would be an increase in the Earth's moment of
tion was based on the surface area of the present inertia, which would allow angular momentum to be
continents, and provides a mean radial expansion conserved by a slower rate of rotation. Knowledge of
rate of about 6 mm a"^. the length of the lunar month would allow an esti
There are two methods available which can be used mate to be made of the lunar contribution to the
to test the expanding Earth hypothesis directly. Earth's rotational deceleration and allow any change
in its moment ofinertia to be isolated.
Information on the rotational history of the
11.3.1 Calculation of the ancient
Earth-Moon system has been provided from a
moment of inertia of the Earth
detailed examination of fossil organisms whose
The moment of inertia of a rigid body about a given patterns of growth are strongly affected by diurnal
axis is defined as where m is the mass of each effects. In particular, certain rugose corals of Middle
250 CHAPTER 11
100km would be ISOO^C. This temperature is in have overcome the problem of limited observations
excess of the melting temperature of mantle rocks at in some areas by predicting the heat flow in those
this depth, and so a fluid layer is implied. Such a areas on the basis ofthe correlation of heat flow with
molten layer does not exist as S waves are known to the age of the oceanic lithosphere and the age of the
propagate through this region (Section 2.1.3). Two last tectonic event to affect continental crust. In
possibilities exist in explanation of this pheno Fig. 11.4 their results are presented by a spherical
menon: first, that heat sources are concentrated harmonic analysis of the heat flow measured or
above a depth of 100 km; and second, that a more predicted in 5"x5® grid areas of the globe. This
efficient mechanism than conduction operates below procedure imparts a certain smoothing of the true
this depth whereby heat is transferred at a much pattern, so that variations with wavelengths of less
lower thermal gradient. Distinction between these than about 3300 km are not represented. Figure 11.4
processes can be made by considering the variation in illustrates the high heat flow associated with the
heat flow over the Earth's surface in conjunction ocean ridge system and the youngest marginal basins
with the variation in content of radioactive minerals of the western Pacific. Low heat flow values are
of different crustal types. associated with old oceanic crust and with Precam
Heat flow generally decreases with the age of the brian shields.
crust (Sclater et 1980). Within the oceans heat Histograms of heat flow measurements from
flow decreases from the ocean ridges to the deeper oceans and continents (Pollack & Chapman, 1977a;
water flanking basins and it has been shown (Section VoWzcket 1993) are presented in Fig. 11.5. The
6.9) that this cooling correlates with a progressive significantly greater number of heat flow values mea
thickening ofthe oceanic lithosphere and an increase sured in the oceans reflects the relative ease oftaking
in water depth. Similarly, the heat flow of marginal measurements at sea (Section 2.13). The greater
seas (Section 8.9) decreases with age, with the pres dispersion of the oceanic values, and their positive
ently active basins exhibiting the greatest heat flow. skewness, reflects variability arising from localized
Within continental regions the heat flow generally extreme values at the crests of ocean ridges. By
decreases with increasing time since the last tectonic contrast, there are fewer extreme high or low values
event. Consequently, Precambrian shields are char present in the continental values. The mean of oce
acterized by the lowest heat flow and young fold anic heat flow measurements is 65 mW m~^. How
mountain belts by the highest. ever,this only represents the heat loss by conduction,
The representation of the global pattern of heat and ignores the heat reaching the surface by the
flow is difficult because the density of the obser discharge of hot fluids such as water and lava. It is
vations is highly variable so that the location of now recognized that the hydrothermal contribution
contours can be greatly biased by only a small num accounts for about a quarter of the global heat loss,
ber of measurements. Chapman & Pollack (1975) and the average oceanic heat flow is increased to
c 1200
.2
> 1000
Continents
101 mW m"^ (Davies, 1980; Sclater et al.^ 1980). originate at sub-crustal levels. Because of the melting
The mean continental heat flow is 57 mW m~^, in problems discussed above, this heat must be trans
cluding the small contribution from lavas. The mean ported under the influence ofa low thermal gradient.
continental heat flow is thus only about 60% The mechanism of heat transfer by convection is the
of the oceanic mean. The global average heat flow is only feasible process conforming to these con
87 mW m-2. straints. Therefore, although heat transfer by con
The majority of the heat escaping at the Earth's duction takes place within the rigid lithosphere, heat
surface originates from the decay oflong-lived radio transfer by convection must predominate in the
active isotopes of uranium, thorium and potassium sublithospheric mantle. Indeed, conduction cannot
(Section 2.13) which have half-lives of the same occur to any great depth as the rate of heat transfer by
order as the age of the Earth. These isotopes are this mechanism is much slower than required. The
relatively enriched in the upper continental crust, feasibility and form ofsuch convection is discussed in
and it has been estimated that their decay contributes the following sections.
18-38 mW m~^ to the observed heat flow (Pollack 8c
Chapman, 1977b). Consequently up to about 70%
ofthe heat flow in continental regions may be gener 11.5 Convection in the mantle
ated within the upper 10-20 km of the crust. The
11.5.1 The convection process
oceanic crust, however, is virtually barren of radio
active isotopes, and only about 4 mW m"^ can be The nature of convective flow in the mantle is prob
attributed to this source. Over 96% of the oceanic lematic. Analytical solution is difficult because of the
heat flow must originate from beneath the crust, and complex theological structure, including the pres
so different processes of heat supply must act beneath ence of a transition zone (Section 2.8.5), because of
continents and oceans (Sclater 8c Erancheteau, the presence of heat sources within the convecting
1970). layer as well as beneath it, because of the unknown
Thus, a large proportion of the continental heat influence of an overlying rigid lithosphere on the
flow is from sources concentrated at a shallow depth, pattern of convection, and because the convecting
and only a small sub-crustal component is required. layer has the form of a spherical shell. However, by
Conversely, the majority of oceanic heat flow must making use ofsimple hydrodynamic theory and con-
THE MECHANISM OF PLATE TECTONICS 253
Table 11.1 Computed Rayleigh numbers for the mantle and its subdivisions assuming a superadiabatic temperature
gradient p = 0.1 Km"^ and^ = 10 m s"^(from Bott, 1982a, with permission from Edward Arnold).
Thermal expansion Layer thickness Thermal diflusivity Kinematic viscosity Rayleigh number
(a(K-')) (<<(m)) (4(m^s->)) (vKs-i)) m
Upper mantle 2.5x10-® 3.0x10® 1.1x10"® 3x10^® 6x10®
Lower mantle 1.5x10"® 2.0 X 10® 1.3x10"® 2x10'^ 9x10®
Whole mantle 1.9x10"® 2.7x10® 1.3 X 10"® 2x10'^ 4x10®
The nature of the flow ofthe convecting fluid can convection of 10®-10^, Nu is about 10, indicating
be judged by the magnitude ofthe Reynolds number the predominance of heat transfer by convection.
(i^), which allows discrimination between laminar Exact solution ofthe form of mantle convection is
and turbulent flow. Re is defined. not possible, although approximate solutions (Par-
mentier et al., 1976)reveal that the pattern offlow is
Re=vd/v
as shown in Fig. 11.7a. Heat is transferred to the
where v is the velocity of flow. Taking v = 200 mm convection cell by conduction at the base of the cell
a"^ = 6 X 10~^ms"^, //= 3000 km = 3 x lO^m and and is removed by conduction at its top. Within the
V = 2 X 10^^ m^ s"^, jR^ = 9 X 10"^°. This very low fluid most of the heat is transported around the
value indicates that the flow is laminar, and that the margins of the cell, with an isothermal core rotating
amount of heat transferred by convection never in in sympathy with these movements but not contrib
creases to the level at which turbulent flow occurs uting to the heat flow. Parmentier et al, (1976) have
and the regular pattern of convection cells disap also demonstrated that the form ofconvection is not
pears. radically diflferent if the mantle acts as a non-
The effect of the Earth's rotation on convection Newtonian fluid (Fig. 11.7b), the only major effect
can be judged by the magnitude of the Taylor num being to increase the critical Rayleigh number. Con
ber(T), which is defined: sequently, for non-marginal convection, the nature
ofthe creep mechanism in the mantle is not of great
r= {2wd^/v)^
significance in determining the flow pattern of con
where w is the angular velocity of rotation. Putting vection.
= 7.27 X 10"® rads"^ and other values as above, The mantle transition zone (Section 2.8.5) must
T«4xl0"^^.A value ofTless than unity implies no influence the vertical extent of convection in the
significant effect ofrotation on convection and so the mantle. If this zone represents a change in chemical
Earth's rotation should have no effect on the pattern composition, convection currents could not cross it.
of mantle convection. However, at the present time In this case convective circulation would be in the
the majority of plates are moving approximately
east-west (Solomon et 1975), and during the
Mesozoic the predominant movements were north-
south. It is possible, therefore, that there is some
poorly understood mechanism responsible for this
phenomenon (Doglioni, 1993).
The efficiency of convection is measured by the
Nusselt number (N«), which is the ratio of the total
heat transferred to that transferred by thermal con
duction. Elder(1965) has computed experimentally
the relationship between Nu and R, and has found Fig.11.7 Streamlines (dashed) and isotherms (solid
that at values of R appropriate to marginal convec lines) for(a) Newtonian and (b) non-Newtonian
tion Nu is unity and very little heat can be transferred convection (redrawn from Parmentier 1976, with
by convection. At R values appropriate to vigorous permission from the American Geophysical Union).
THE MECHANISM OF PLATE TECTONICS 255
Lithosphere takes place in paired cells above and below the mantle
transition zone (Fig. 11.8b).
The nature of the mantle transition zone is equi
vocal, but the majority view appears to be that it
represents a region in which solid state phase changes
take place whereby the mineralogy ofmantle material
changes to higher pressure forms with depth rather
than representing a change in chemical composition
(Section 2.8.5). For example. Watt & Shankland
(1975) have shown, from an inversion of velocity-
density data, that the mean atomic weight of the
mantle shows no change across the transition zone.
(a) Core-mantle boundary Ifthis is the case, convection currents could cross the
transition zone as long as the phase changes take
Lithosphere place very rapidly, and convection cells would then
be mantle wide (Fig. 11.8a). The phase changes
would have two important effects on convection.
700 km The phase changes,such as from olivine to spinel, are
temperature dependent. The change from low pres
sure to high pressure forms takes place at shallower
than average depths in the cold descending currents
and at greater than average depths in the hot ascend
ing currents. Consequently,low density minerals are
created deeper on ascent than high pressure forms on
descent, and their negative and positive buoyancies
then help to drive the convection cells. Also, the
(b) Core-mantle boundary phase changes are associated with a release or absorp
Fig. 11.8 Possible influence of the mantle transition zone tion oflatent heat, the high to low pressure reaction
on the pattern ofconvection,(a) The transition zone being endothermic and the low to high pressure
represents a phase boundary which can be traversed by reaction being exothermic. This causes steepening of
convection currents,(b) The transition zone represents a the thermal gradient across the transition zone, so
chemical boundary prohibiting its crossing by convection that the temperature in the lower mantle would need
currents (redrawn from Bott, 1982a, with permission to be 100-150"C higher than if the zone did not
from Edward Arnold). exist.
Ringwood & Irifune (1988) have examined the
form of paired cells (Fig. 11.8b), with heat trans physical and chemical processes occurring in the re
ported across the separating transition zone by con gion of the mantle transition zone, in particular the
duction. The mechanism seems unlikely, as density relationships of the mantle and components
convection in the upper mantle is not far from of the oceanic lithosphere at the pressures corre
marginal so that it could not cope with the amount sponding to this region. It appears that when young,
of heat transported from below by much more vigor relatively thin oceanic lithosphere is subducted, the
ous convection. This idea is discussed more fully in various phase changes associated with increasing
the following section. pressure and temperature cause both the oceanic
crust and lithospheric mantle to be trapped gravita-
tionally. Growth of this boundary layer with time
11.5.3 The vertical extent of convection
would lead to its thickening until it occupied a shell
The vertical extent of mantle convection is still between about 600 and 700 km depth. If no other
controversial, and there is no general agreement phenomena operated,this layer would cause a barrier
whether convection is mantle wide (Fig. 11.8a) or to the descent of young lithospheric slabs and isolate
256 CHAPTER 11
convection systems in the upper and lower mantle. etrate the layer and allow the exchange of material
However, the interiors of thick, mature, oceanic between the upper and lower mantle. Thus,although
plates have a high thermal inertia and viscosity so that whole mantle convection is suggested, the process
when the tip of the slab meets the boundary layer at need not be continuous and depends upon the age of
a depth of about 650 km it buckles and thickens the oceanic lithosphere being subducted. This model
plastically to form what Ringwood & Irifune term a has many important ramifications for mantle pro
megalith, which grows with time. The megalith is cesses, including the petrogenesis of ocean ridge
neutrally buoyant and in isostatic equilibrium at this basalts and intraplate volcanic associations and the
level, so that it 'floats' straddling the boundary layer nature ofthe convective regime in the Archaean; it is
(Fig. 11.9). If the megalith lies in the region of a clearly testable by geophysical methods, and the
descending convection current in the lower mantle, outcome of such investigations will be awaited with
it would become entrained and gradually mix into interest.
the lower mantle; this must be accompanied by the Tackley et al. (1993) have numerically modelled
movement of an equal volume of lower mantle into mantle convection in three dimensions with an en-
the upper mantle. The model envisages a dynamic dothermic phase change at the base of the transition
equilibrium between the subduction of young litho- zone. They suggest that, periodically, cold down-
sphere,that creates and maintains a boundary layer at welling material accumulates above 670 km and then
about 650 km depth, and the subduction of mature rapidly flushes into the lower mantle. This sugges
lithosphere which gives rise to megaliths which pen tion is supported by seismic tomographic imaging
Island arc
Calc-alkaline volcanism Basaltic crust
jI /'^^^^^^eso''btion
W Resorbtion ofof
y/depleted pyrolite
'IM, 400 km discontinuity
'1^//1 V
'Buckling and plastic
thickening of crust
550 km discontinuity
"^Ancient oceanic
- Former harzburgite
vCpol highly viscous megalith;^^ - Former basaltic crust
lithosphere
(Section 2.1.8) of the Tonga subduction zone (van the transition zone, and is therefore further evidence
der Hilst, 1995)in which the descending iithosphere in favour ofdeep convection. The existence ofearth
apparently penetrates the transition zone (see Fig. quakes with compressive focal mechanism solutions
8.18). at great depth in BeniofF zones (Section 8.3) was
Thus, the transition zone may not be a barrier to once cited as evidence that a barrier to slab penetra
mantle-wide convection, and a number of workers tion existed at a depth of about 700 km which
have presented evidence in accord with this premise. limited convection to the upper mantle. This argu
Kanasewich (1976) has noted an organized distribu ment no longer holds, as Isacks & Barazangi(1977)
tion of plates, in which the Pacific and African plates have explained the compressive nature of these
are approximately circular with the smaller plates events in terms of an unbending, rather than a
having an approximately elliptical form and arranged stopping, of the slab. The cessation of earthquake
systematically between these two large plates. Kana activity probably results from the loss of identity of
sewich attributes this organization to convection that the downgoing slab as it heats up to the same
is mantle-wide. Davies (1977) has conducted model temperature as the surrounding mantle.
experiments and has concluded that only extreme It is possible that the controversy between whole
viscosity contrasts would restrict convection to the mantle and layered convection is artificial and that
upper mantle, and maintains that such contrasts do both modes may be possible. Stein & Hofmann
not exist. Elsasser et al. (1979) have employed a (1994) have described a model in which the different
scaling analysis in which the depth of convection is types of convection alternate during the Earth's his
derived as a function ofknown parameters, and have tory. During the present phase of plate motions, the
concluded that this depth is consistent with convec majority of new continental crust is formed by volca
tion throughout the entire mantle. Grand(1994)has nic processes at subduction zones. However, in the
demonstrated, by seismic tomographic analysis past, geochemical studies suggest that some large
(Section 2.1.8), that oceanic Iithosphere descends igneous terrains appear to have formed in environ
beneath western America into the lower mantle, so ments different from subduction zones, and at a
that at least one convection cell is mantle-wide. much higher rate than at present. It has been pro
Morgan &Shearer(1993)have derived the buoyancy posed that these ancient terrains formed when the
distribution in the mantle from seismic tomographic bulbous head of a mantle plume approached the
maps and concluded that there must be significant surface, partially melted and gave rise to copious
flow between the lower and upper mantle. Other basaltic volcanism. The Cretaceous, when there were
workers, however, such as O'Nions (1987), suggest many events of flood basalt volcanism, may have
that the geochemical and isotopic pattern of trace been one of these episodes. The ascent of material
elements found in oceanic volcanic rocks supports a from the lower mantle in the plume would have
model in which portions of the mantle have been replenished the incompatible elements in the upper
chemically isolated for much of Earth history. This mantle removed during continental crust formation.
would suggest that the mixing implied by whole Replenishment by the subduction of continental
mantle convection has not occurred,and that layered crust would be inconsistent with the known isotopic
convection is more likely. Loper (1985), however, composition of the upper mantle. The new model
has been able to produce a model of whole mantle suggests that periods of layered convection, during
convection in which the preservation of chemically which incompatible elements were depleted in the
distinct reservoirs is possible. The topography on the upper mantle by crust formation, were followed by
base of the mantle transition zone has an amplitude whole mantle convection, when the incompatible
of about 30 km (Shearer & Masters, 1992), which is elements were replenished. This would explain both
an order of magnitude lower than predicted for a the episodic nature ofcontinental crust formation in
chemical, rather than a phase, change at this depth. the geological record and the isotopic composition
It would seem, therefore, that convective circula of the upper mantle. The model is still speculative
tion is most likely to be mantle-wide and not con and a true understanding of mantle convection will
strained by the transition zone. The presence of only come when the nature of the transition zone is
earthquakes to depths of680 km certainly indicates fully characterized.
that the Iithosphere circulates almost to the base of The motion ofthe plates must be compensated by
258 CHAPTER 11
return flow at depth in the mantle,so that the mantle Earth's surface. The asthenosphere's role in this
material displaced by downgoing slabs is balanced by scenario was considered to be essentially passive.
material flowing towards spreading centres. Several A similar set of computations based on a similar
workers have discussed the nature ofthis return flow method and providing similar results was made by
(e.g. Garfimkel, 1975; Harper, 1978; Chase, 1979; Chappie & Tullis (1977). The following description
Alvarez, 1982), and consider that generation and of forces is based on the extensions of the work of
consumption of plates is not balanced around single Forsyth & Uyeda (1975) made by Bott (1982a).
plates, but globally. For example, Garfimkel (1975) At ocean ridges the ridge push force JFpj,
and Alvarez (1982) suggest that material at present (Fig. 11.10) acts on the edges of the separating
displaced by the circum-Pacific subduction zones plates. This derives from the buoyancy of the hot
streams towards other oceans which are increasing in inflowing material causing the elevation of the ridge
size. The form of the return flow must change con and hence an additional hydrostatic head at shallow
siderably with time as the plate configuration changes, depths which acts on the thinner lithosphere at the
and so its geometry must be complex. These studies ridge crest. It may also arise from the cooling and
implythatflow in the mantle does not take the form of thickening of the oceanic lithosphere away from the
regular circuits linked to the overlying plates. ridge (Section 6.9), which exerts a pull on the ridge
region. The ridge-push force may be two or three
times greater if a mantle plume (Section 5.4) under
11.6 The forces acting on plates lies the ridge (Bott, 1993), because of the increased
In order to understand the structural styles and pressure in the asthenosphere at the ridge crest. The
tectonic development of plate margins and interiors, separation of plates at ocean ridges is opposed by a
it is necessary to consider the nature and magnitude minor ridge resistance which originates in the
ofall the forces which act on plates. Forsyth & Uyeda brittle upper crust, and whose existence is demon
(1975) have solved the inverse problem ofdetermin strated by earthquake activity at ridge crests. The
ing the relative magnitude of plate forces from the resisting forces are small so that the net effect is the
observed motions and geometries of plates. Since the presence of a driving force.
present velocities of plates appear to be constant, Beneath plate interiors a mantle drag force acts
each plate must be in dynamic equilibrium, with the on the base of both the oceanic and continental
driving forces being balanced by inhibiting forces. lithospheres ifthe velocity ofthe underlying astheno
Forsyth & Uyeda (1975) used the corollary of this, sphere differs from that of the plate. If the asthenos
that the sum of the torques on each plate must be phere velocity exceeds that of the plate, mantle drag
zero, to determine the relative size of the forces on enhances the plate motion (Pdo^ but if the
the 12 plates which they assumed make up the asthenosphere velocity is lower, as shown in
Continent Ocean
Frp- Ridge push I?r -Ridge resistance Fig. 11.10 Some ofthe forces
Fnb- Negative buoyancy He - Bending resistance
Fsp -Slab pull Rs -Slab resistance
acting on plates (developed
-Trench suction Ro -Overriding plate resistance from Forsyth & Uyeda(1975)
Hdo- Mantle drag under ocean by Bott, 1982a, with permission
Hoc- Mantle drag under continent from Edward Arnold).
THE MECHANISM OF PLATE TECTONICS 259
subduction zones (Fig. 11.1 Id) as described in Sec thermal energy is employed in driving the litho-
tion 8.9. However, once back-arc spreading com spheric plates. The proposal by Morgan (1971;
mences the landward plate becomes decoupled from 1972b) that plates are driven by the horizontal flow
the trench system (Fig. ll.lle). of material brought to the base ofthe lithosphere by
When two plates of continental lithosphere are hotspots is discounted (Chappie & Tullis, 1977), as
brought into contact after the complete consump the lateral flow would probably be equal in all hori
tion of an intervening ocean at a subduction zone, zontal directions and thus would not apply a direc
the resistance to any further motion is known as tional force to the plates. Two models have been
collision resistance. The mechanism ofthis resistance proposed. The classical, or mantle drag, model con
is complex because it takes place both at the suture siders that the upper, cool, boundary layer of the
between the plates and within the overriding plate convecting system is represented by the upper part of
(Section 9.3). Finally, transform fault resistance af the asthenosphere, and that plates are driven by the
fects conservative plate margins in both continental viscous drag of the asthenosphere on their bases. In
and oceanic areas. The resistance acts parallel to the contrast, the edge-force model recognizes the litho
faults and gives rise to earthquakes with a stick-slip sphere itself as the upper,cool, boundary layer ofthe
mechanism (Section 2.1.5) confined to a shallow convection cells and proposes that the plates are
depth. More complex resistance is encountered when driven by forces applied to their margins. The two
the fault trend is sinuous so that motion is not purely models thus differ in the importance placed on
strike-slip (Section 7.4.2). the various forces acting on plates (Section 11.6)
The relative magnitude of the forces acting on described by Forsyth & Uyeda (1975).
plates and their relevance to the driving mechanism
of plate tectonics will be discussed in Section 11.7.
11.7.1 Mantle drag mechanism
Mantle drag was the first driving mechanism to be
11.7 Driving mechanism of
proposed, and envisages plate motion in response to
plate tectonics the viscous drag exerted on the base of the litho
The energy available to drive plate motions is the sphere by the lateral motion of the top of mantle
heat generated in the core and mantle which is convection cells in the asthenosphere (Fig. 11.12a).
brought to the surface by convection in the mantle. The convection cells would consequently rise be
It now remains to consider the manner in which this neath oceanic ridges and descend beneath trenches.
Ocean Ocean
ridge Mantle drag Continent ► Mantle drag ridge
Tension-^Compression Comoression
Tension
the satellites, the height ofthe sea level surface can be which might give rise to the observed bathymetry,
determined from the two-way travel time of a radar gravity and geoid anomalies.
pulse. The geoid and bathymetric data were subse Examination of Fig. 11.14a shows that the sea
quently filtered to remove variations with wave floor around Hawaii is shallower than suggested by
lengths of less than 300 km, which originate within its age and is associated with a positive geoid anom
the lithosphere, in order to isolate anomalies caused aly. This suggests that the Hawaiian Islands are
by more regional phenomena(Fig. 11.14). Although underlain by a hot ascending plume, which is in ac
it might be expected that the low density region cord with other observations suggesting this mode of
associated with a rising plume would give rise to a origin (Section 5.4). Other regions are too deep for
negative gravity anomaly, this effect is swamped their age, and must thus be underlain by cool, desc
by the bathymetric bulge over the plume. Con ending mantle material. There is no correspondence
sequently, hot plumes are associated with positive between positive geoid anomalies and ocean ridges;
geoid anomalies and cold descending plumes with indeed, some ridges are associated with negative
negatives. The correlation between Figs 11.14a and anomalies. The study suggests that accretive plate
11.14b is further demonstrated by the profiles shown margins are not necessarily associated with the ascend
in Fig. 11.15, which also show the temperature dis ing limbs ofconvection cells, and so sea floor spread
tribution and flow pattern in a two-dimensional layer ing in some areas is merely a passive phenomenon
30''N 20" N
10" N
10"S
20"S
30" N 20" N
10" N
Hawaii
10"S
ously accepted models. It must be stated, however, 300 km new velocity features develop which become
that there is no universal agreement that the plan- fully established in the transition zone. Their pattern
form of mantle convection is evident in the geoid is completely different from that near the surface. For
(Sandwell & Renkin, 1988). We must now await example, there are two broad, high velocity zones,
further studies of this kind in other ocean basins to one extending from the western Pacific under south
see if results from the Pacific Ocean are generally ern Australia to the Antarctic coast, the other be
applicable. neath South America,the South Atlantic and parts of
Much important information on the three- Africa. At this depth the localized low velocity zones
dimensional structure of the mantle has been associated with hotspots near the surface change to
supplied by seismic tomography (Section 2.1.8). high velocities in the transition zone. This suggests
Convection is driven by lateral differences in that hotspots originate at shallow depths. However,
temperature and density. These variables affect seis they might be fed through conduits from much
mic velocity, which decreases with decreasing density greater depths which are too narrow to be resolved by
and increasing temperature (Dziewonski & Ander the model.
son, 1984). By mapping velocities in the mantle it is A velocity model of the lower mantle has been
possible to infer the differences in temperature and derived from P wave arrival times by Dziewonski
density which are a consequence ofconvection. Also, (1984). In general, velocity anomalies in this region
by mapping seismic anisotropy both vertically and are much smaller than in the upper mantle. The
laterally it is possible to obtain estimates of the Pacific is surrounded by a ring ofhigh velocities, and
direction of mantle flow. there is some correlation between hotspots and ve
Woodhouse & Dziewonski(1984) have derived a locity lows.
three-dimensional velocity model of the upper Figure 11.17 shows four global cross-sections
mantle by the inversion of a large dataset of shear illustrating the velocity models for both the upper
wave arrivals. Because of the wavelengths involved and lower mantle(note that the vertical exaggeration
the resolution is about 2500 km horizontally and ofthe upper mantle model is twice that ofthe lower
200 km vertically. They found that the velocity struc mantle). Unfortunately the resolution at a depth of
ture is only related to surface tectonic features down 670 km is quite low, and this makes it difficult to
to a depth ofabout 200 km. Ocean ridges are under trace features between the two models.Figure 11.17a
lain by low velocity mantle, shields by very fast shows an equatorial section. The high velocity
velocities and old ocean basins by high velocities. regions beneath the Brazilian and African shields are
High velocities probably reflect the presence of ma separated by the low velocities of the Mid-Atlantic
ture mantle materials which are colder and denser Ridge near the surface. At about 400 km the highs
than average. Between 200 km and 400 km there are merge to form a broad high velocity region which
some remnants of the surface features. Most shield appears to continue into the lower mantle at a lower
areas are still underlain by positive velocity anoma angle. The low velocities associated with the Carls-
lies, but the velocity contrasts are lower than near the berg Ridge extend into the uppermost lower mantle,
surface. The origin of this high velocity material while the Pacific Basin is underlain by low velocity
beneath shields is not known;it could reflect material material extending to the core-mantle boundary.
that is translating with the overlying shield,implying The marginal seas of Indonesia are underlain by low
that shields maintain a continuous identity to this velocities to a depth of200 km.The underlying high
depth; it might equally well represent cold, sub velocity zone, which appears to penetrate the transi
ducted material overridden by the continent. Within tion zone, may be related to subducted material.
this depth range only a small proportion ofthe ocean Figure 11.17b shows a section across the Pacific
ridge system retains anomalous velocities. The fast approximately parallel to the spreading direction. As
spreading East Pacific Rise is underlain by nearly expected, lithospheric velocities increase with in
normal mantle material below 200 km. However, creasing age. At the western margin ofthe Pacific the
isolated regions beneath ocean ridges are underlain high velocity material is probably related to subduc-
by low velocities extending to 400 km, and there is tion, and merges with the high velocity material
evidence that some ridge segments are fed by lateral under Eurasia to form an extensive positive region in
transport of material at shallow depth. Below about the lower mantle extending to the core-mantle
266 CHAPTER 11
670
(km)
I + 3.0%
I + 0.75%
boundary. The profile crosses the velocity low be prisingly, associated with high velocities in the upper
neath the Atlantic which, as in Fig. 11.17a, divides mantle, but is underlain by a high velocity column in
the high velocity material beneath South America the lower mantle.
and Africa which converge at depth. The low velocity Unfortunately, because of the lack of resolutiiin at
anomaly beneath the Red Sea may continue into the the depth of the transition zone, these models can
lower mantle. not provide any information on the vertical extent of
Figure 11.17c runs along the ridges ofthe east and convection.
south Pacific. It illustrates the variability ofthe depth More recent tomographic studies of the mantle,
extent of the low velocity material underlying the which have been summarized by Montagner (1994),
ridges, and the persistence of the high velocities of have greater resolution. All investigations sho^v ve
the North American and Eurasian plates to about locity anomalies in the upper 400 km which correlate
400 km. Figure 11.17d crosses the low velocity re with surface tectonic features. The upper 200 km
gion south of Hawaii, and shows it to be continuous have slow velocities below plate boundaries. Only
with an anomaly extending to the core-mantle fast spreading ridges are visible between 200 and
boundary. The Tonga-Kermadec Trench is not, sur 400 km and the corresponding low velocity zones are
THE MECHANISM OF PLATE TECTONICS 267
I + 3.0%
I + 0.75%
r...r^
I + 3.0%
Fig.11.17 {Continued). I + 0.75%
often displaced laterally. High velocities are only causing it have ceased. The preferred orientation of
present beneath large continents. Below 400 km ve crystals causes seismic waves to propagate faster in
locity anomalies decrease in size, with the largest in the direction of elongation than transverse to it, that
the western Pacific, below the Alpine-Himalayan is, it imparts an azimuthal anisotropy to the mantle.
mountains and below the Atlantic; these may be Also, transverse, or polarization, anisotropy can be
related to subducting lithosphere. The anomalous studied in which S waves propagate horizontally
velocity zones of spreading centres are no longer (SH)with a different velocity from vertically(SV). By
visible, but large lows are present beneath the central making use ofthe very accurate seismic data available
Pacific and eastern Africa. from digital recording systems it is now possible to
Information on the actual directions offlow in the map past and present flow lines using azimuthal
mantle can be obtained by tomographic analyses of anisotropy, whereas transverse anisotropy can be
anisotropy in the mantle. Most mantle minerals are used to determine the relative importance of hori
elongate and so, if subjected to recrystallization, zontal and vertical flow and to map the boundaries of
shearing or flow they can assume a predominant convection cells, as SH is greater than SV for waves
orientation which is retained long after the processes travelling in the direction of flow. Results are still
268 CHAPTER 11
269
270 CHAPTER 12
600 Ma 500 Ma
450 Ma 400 Ma
375 Ma 350 Ma
Laurasia
Acadian
FB
\ /\
Gondwana
IMPLICATIONS OF PLATE TECTONICS 271
325 Ma
Appalachian
Hercynian
LA-Laurentia
SN-Chlnes'eShieid
SB-Siberian Shield
BA-Baltica
B-Bohemian FB
Fig. 12.1 {Facin^f pa^e and above) Palaeomagnetically- (e) Early-Middle Devonian;(f) Late Devonian;
derivcd continental reconstructions for the Palaeozoic: (g) Middle Carboniferous. FB,fold belt(redrawn from
(a) end-Precambrian;(b) Late Cambrian-Early Morel & Irving, 1978, with permission from the
Ordovician;(c) Late Ordovician;(d) Late Silurian; University ofChicago Press).
20-25° km"^ (Bickle, 1978). The conduction mech tirely of basaltic crust. It also seems unlikely that the
anism of heat transfer in the continental crust is lithosphere could have been as thin as 20 km. More
inefficient; in contrast, the convection process opera over, the high temperature of the asthenosphere
tive beneath the oceanic lithosphere is far more implies a low viscosity, so that the viscous coupling
effective in dissipating heat from the interior of the between lithosphere and asthenosphere would be
Earth. Since the oceanic lithosphere probably repre expected to be too low to create significant mantle
sents the upper layer ofthe mantle convecting system drag forces. If the Earth were not to have expanded
(Section 11.7.2), the increased heat supply in the considerably during its early history, some sort of
mantle could be dissipated by increasing the length subduction must have taken place. If plate motions
ofthe ocean ridge system or by increasing the rate of were not powered by mantle drag forces it is probable
plate production by a factor of at least 6 with respect that the edge-force mechanism (Section 11.7.2) was
to the present (Bickle, 1978). Assuming a non- active, as it is believed to be at present. Certainly
expanding Earth (Section 11.3),this implies a similar increased ocean ridge length/more rapid spreading
increase in plate subduction rate and suggests that rates imply a powerful ridge-push force and it must
mantle convection was very vigorous. Certainly, be concluded that the density of the lithosphere,
these computations suggest that some form of plate even though lower than at present, must have been
tectonics was taking place during the Precambrian at sufficient to allow the initiation ofsubduction. Once
a much greater rate than today. the subducted plate reached depths of 70-80 km,
The increased heat flow in the Precambrian also phase changes affecting the lithosphere would have
has implications bearing on the nature of the plate increased its density and facilitated further subduc
driving forces. The increased length of ocean ridges tion. The magmatic products of the underthrusting
or increased rate of plate accretion implies that either would have increased the crustal thickness in island
the mean distance from ridge to trench, or the mean arc environments, and the collisions of different arcs
age of oceanic lithosphere, decreased. The thickness would have created regions of high negative buoy
and density of oceanic lithosphere are a function of ancy which may have evolved into protocontinents.
age (Section 6.9), and it has been argued (Hargraves, As the radioactive heat sources decayed,the heat flow
1981)that subduction only occurs when the oceanic diminished and the lithosphere would have thick
lithosphere has attained some critical density at ened and become more dense. The thermal necessity
which it becomes negatively buoyant. The conclu of extensive ocean ridges and/or rapid spreading
sion is that, in Precambrian times, this density was would gradually be removed, and the plate tectonic
not achieved so that the principal plate driving force regime would have gradually changed over to that
ofslab pull could not have operated. This conclusion typical of Phanerozoic times. Given the differing
is questionable, however, as at the present time there tectonic environments of the older and younger
is little correlation between lithospheric age and Precambrian, it seems reasonable to postulate that
subduction. In the Pacific, oceanic lithosphere as the changeover to modern-type plate tectonics took
young as 5 Ma is being subducted while in the place around the time of the Archaean-Proterozoic
Atlantic there are extensive regions of lithosphere boundary. It must be appreciated, however, that
200 Ma old which are not being subducted. conditions in the Early Precambrian are quite conjec
Hargraves (1981) has also argued that the mantle tural, so that these conclusions are speculative.
drag force beneath Precambrian oceanic lithosphere
would have been stronger than at present because of
12.2.3 Archaean tectonics
the more rapid convective overturn, and could have
been of sufficient magnitude to drag down the posi The Archaean is dominated by two types of terrain
tively buoyant lithosphere. Hargraves has concluded which must have formed in different tectonic
that the gravitational forces on a crust of density environments: the greenstone or greenstone-granite
contrast 0.25 Mgm"^ would, if less than 20 km belts and granulite gneiss belts (Windley, 1981).
thick, exceed the critical value of 5xlO^Nm"^ Greenstone belts are typified by a synformal struc
(McKenzie & Weiss, 1975) at which it would be ture, 40-250 km wide and 120-800 km long. They
dragged down into the mantle. This calculation, exhibit a low pressure regional metamorphism of
however, assumes that the lithosphere consisted en the greenschist facies and are intruded by plutons
IMPLICATIONS OF PLATE TECTONICS 273
ranging in composition from tonalite through developed from a continuous global scum of inter
granodiorite to potassic granite. They are divided mediate composition which was continuously re
stratigraphically into three groups: a lower group of cycled as a result of surface volcanism and remelting
komatiitic ultramafic and mafic volcanic rocks with at its base (Hargraves, 1981). Windley (1984) has
pillow structures which have a bulk composition summarized the mechanisms proposed for the two
similar to ocean ridge basalts; a central group of main crustal members preserved from this time. A
andesites and calc-alkaline volcanic rocks whose trace typical non-uniformitarian model for the formation
and rare earth elements are similar to rocks found on of greenstone belts is shown in Fig. 12.3. A down-
island arcs; and an upper group of clastic sediments, warp in the primitive crust was formed in response to
such as greywackes, sandstones, conglomerates and the load of injected ultramafic rocks, and subse
banded ironstones with cherts and limestones. quently filled by sediments. Later the crust was be
Granulite-gneiss belts exhibit a high grade of lieved to have thickened and intruded by diapiric
metamorphism and comprise three components: granites which lead to compression of the belt.
quartzofeldspathic gneiss oflargely tonalitic to gran- Most workers who suggest a plate tectonic model
odioritic composition; layered peridotite-gabbro- for the formation of greenstone belts propose that
anorthosite or leucogabbro-anorthosite complexes, they originated in a subduction-related setting. A
and metavolcanic amphibolites and metasediments. particularly convincing model has been proposed by
Most of the regions in which granulite-gneiss belts Tarney et al, (1976). Their studies of the Rocas
are situated were subjected to a major accretionary Verdes complex in southern Chile (Section 8.9 and
event, indicated by isotope abundances, during the Fig. 8.38) lead them to postulate that greenstone
period 2700-3100 Ma. belts originated in ensialic marginal basins of this
Workers adopting a non-uniformitarian view of type. Their model is shown in Fig. 12.4. It envisages
the Archaean propose that the crust of this time was the greenstone belts as developing on thinned
Sediments
Calc-alkaline
Ultramafic volcanics
group
Mantle tapping
fracture
(a)
(b)
Volcanic
arc
Extension
.5555%%%?.%?;?;
® Deep source
»»»»&»»»»^
komatiite '»M*W
Volcanic phase
5!*%5!W.swX*••••*******^^^^
Deformation 4-
synorogenic tonalite
Basin closure
•»;«X*Xw.
I%%%%wXv.v
5:S!%»»:oK*;w5Wi%%x*>>
w.vKv.w.?.!
Late K-granites
Fig. 12.4 Development of an Archaean
greenstone belt from a marginal basin, based
on a comparison with the Rocas Verdes
complex in southern Chile (redrawn from
WK«5!W
Tarney flj/., 1976, with permission from
v.%VA%v.v.v.v.v*:.;.:.v.v
John Wiley).
continental crust in a back-arc environment, covered Far less attempt has been made to propose tectonic
by sediments originating from the flanking continent models for the formation of granulite-gneiss belts.
and volcanic arc, and subsequently becoming de Clearly, though, any reasonable model must include
formed and intruded by granites as the basin was a mechanism for the production ofthe great quanti
subjected to compression.The structure,volcanosedi- ties of tonalite which make up the major part of
mentary rock associations,geochemistry and igneous these belts. Windley(1981) has noted the geological
history of the Rocas Verdes complex are remarkably and geochemical similarities between Archaean
similar to Archaean greenstone belts (Windley, granulite-gneiss belts and the deeply eroded granitic
1984). The back-arc environment also explains how rocks associated with Andean-type subduction
Archaean oceanic crust came to be preserved;ifit had zones. He considers this the only environment in
originated in a normal ocean basin it would probably which voluminous quantities of tonalite can be pro
have been completely destroyed by subduction. duced, and concludes that this represents a reason-
IMPLICATIONS OF PLATE TECTONICS
Sediments and lavas in Siaiic 'basement' of folded gneisses Tonaiites as horizontal Sheif-type
extensional backarc (Amitsoq-type) with relic sheet intrusions in main arc sediments
basin metasupracrustais (Isua-type)
Fig. 12.5 Plate tectonic model for the development and continents producing an extensive continental plate
growth of continents in the Archaean,(a) Widespread consisting of greenstone belts and granulite gneiss belts
small continental plates with ensialic back-arc basins. The (redrawn from Windley, 1984, with permission from John
Andean-type margins are invaded by mantle-derived Wiley).
tonaiites in batholithic quantities,(b) Aggregation of the
able analogue for the formation of granulite-gneiss Archaean (Fig. 12.5b), these crustal fragments had
belts. coalesced into composite siaiic blocks characterized
Windley (1984) has argued that some form of by alternate belts of greenstone and granulite gneiss
plate tectonics is the only mechanism capable of formed, respectively, from the oceanic crust and
continuously replenishing, over hundreds of mil eroded tonaiitic batholiths. This phase gave rise to
lions of years, the mantle-derived material which the larger, thicker plates which dominated litho-
enabled a large proportion ofthe present continental spheric evolution in the Proterozoic.
crust to have been created during Archaean times.
The mechanism was probably not identical to
12.2.4 Precambrian palaeomagnetic
present day plate tectonics because the increased heat
studies
flow in Archaean times would have led to a thinner
lithosphere and greater plate motions. It can, how Potentially the most rigorous method of investigat
ever, provide a mechanism for the complementary ing the presence of relative plate motions in Precam
origin of both greenstone and granulite-gneiss belts brian times is that of palaeomagnetism (Section 3.6).
whereby the marginal basin and subduction-related The application of this method is rather less straight
batholith models are combined (Fig. 12.5). This forward than for Phanerozoic times for three main
mechanism views the Archaean crust as being made reasons (Dunlop, 1981). First, the error limits of
up of a large number of small, relatively thin, siaiic radiometric age determinations are larger. Second,
crustal blocks. Subduction of oceanic crust between isotopic and magnetic records may be partially reset
these gave rise to back-arc spreading and the forma during metamorphism to different degrees, and the
tion of oceanic-type crust (Fig. 12.5a). Subduction distinction between pre- and post-orogenic isotopic
beneath these blocks also gave rise to calc-alkaline and magnetic overprints is extremely difficult. Third,
volcanics and the intrusion of large bodies of overprints occur during post-orogenic cooling and
tonaiitic magma (Fig. 12.5a). By the end of the uplift, and the temperatures at which isotopic
276 CHAPTER 12
Fig. 12.6 Summary apparent polar wander path for the that the APW paths for the best studied continents
proposed Proterozoic supercontinent between 2200 and are similar in shape, while the poles of the more
800 Ma. Dated points: approximate locations of Eulerian sparsely sampled continents are not inconsistent
poles of rotation between 2700 and 1000 Ma(redrawn with APW paths of the same form. The APW paths
from Piper, 1987, with permission from the Open
for the individual cratons can be brought into coin
University Press).
cidence by a sequence of angular rotations about
Eulerian poles to provide the composite APW curve
systems close and magnetizations stabilize are dif shown in Fig. 12.6. From this Piper concluded that
ferent,so that the dates may be younger or older than in late Archaean times the individual cratons coa
the magnetizations by intervals of tens of millions of lesced into a single supercontinent which remained
years. However, even given these uncertainties and coherent, within the limits of palaeomagnetic detec
the gaps in the palaeomagnetic record arising from tion, throughout the whole of Proterozoic times.
the lack of suitable samples of certain ages, the data The composite APW curve of the supercontinent is
should be adequate to determine the major features illustrated as a broad swathe whose width reflects the
of apparent polar movement relative to each craton effects of a number of factors, including incomplete
(Briden, 1976). subtraction of non-primary magnetizations, non-
Piper (1987) has synthesized all available palaeo dipole components of the geomagnetic field and
magnetic data for the Precambrian. Although the warping of the continents not evident in their geol
data are patchy both in space and time, he has shown ogy (Piper, 1982). The APW path exhibits eight
IMPLICATIONS OF PLATE TECTONICS 277
88
+ (BaitiM
traversion of the
an. 1991h
IMPLICATIONS OF PLATE TECTONICS 279
fragments. The age of the oldest sediments associ bile belts in which deformation was wholly ensialic
ated with break-up, and the provinciality of certain (Sutton & Watson, 1974) with no rock associations
animal groups across the split,suggest that the super- that could be equated with ancient ocean basins.
continent fragmented at about 750 Ma (Storey, Similarly, mineralization in the Proterozoic indicates
1993), with the blocks now making up eastern a lack of the exhalative volcanic processes related to
Gondwana (east Antarctica, Australia and India) calc-alkaline volcanism which are so abundant in the
moving anticlockwise, opening the proto-Pacific Phanerozoic. However, ophiolite complexes have
Ocean, and colliding with the blocks of western been recognized in the Proterozoic record, and these
Gondwanaland(Congo,west Africa and Amazonia). Piper attributes to processes associated with subduc-
The intervening Mozambique Ocean closed by the tion zones around the periphery of the supercon
pincer-like movements of these blocks and Gond tinent, a well-studied example being the Wopmay
wana was created when they collided to form the Orogen at the northwestern margin of the Lauren-
Mozambique belt of east Africa and Madagascar. tian Shield (Hoflfman, 1980). This model thus views
Eastern Gondwana then rotated clockwise away from Proterozoic tectonism as occurring in two distinct
Laurentia about 200 Ma later. Southern Africa was environments, interior mobile belts and marginal
located at the pivot of these movements and Baltica subduction, which must have been of Andean-type
moved independently away from Laurentia, opening and similar to processes occurring in the Phanero
the lapetus Ocean, which was subsequently to close zoic. Indeed, Nakajima^f At/. (1990)have recognized
during the assembly ofPangaea. Figure 12.8b shows blueschists typical ofa subduction zone(Section 8.5)
the configuration at 500 Ma. It appears to be con in the Proterozoic of China.
firmed by the similarity in the structural grains of An example of the type of ensialic orogeny envis
northwestern Canada,southern Australia and Adelie aged by this scenario is provided by the Labrador
land in Antarctica. Trough of northeast Quebec and Labrador, whose
It is claimed that palaeomagnetic data are ofinsuf geology is particularly well known. The Labrador
ficient quality to disprove the proposed reconstruc Trough is a linear belt of sedimentary and volcanic
tion, and more rigorous testing will have to await rocks of Early Proterozoic age some 800 km long
better data, should this become available, and an and 100 km wide. To the west it is flanked by
improved geochronological database. Archaean rocks of the Superior Province which were
The process of fragmentation of the supercon- unaffected by the Hudsonian Orogeny (1800-
tinent took only 200 Ma, implying rapid rates of 1600 Ma) which deformed and metamorphosed the
movement.It is possible that the process affected the Labrador Trough. To the east lies the Archaean
sudden changes to metazoan life experienced at the terrain of the Churchill Province which was sub
beginning of the Phanerozoic, perhaps by the in jected to high grade metamorphism during the Hud
creased number of shelf seas surrounding the new sonian Orogeny.
continents and climate changes associated with the Dimroth (1981) has summarized the geological
new tectonic regime. history of the Labrador Trough as shown in
Fig. 12.9. Sedimentation was initiated by the deposi
tion of a continental red-bed sequence in a normal
12.2.5. Proterozoic tectonics
fault-bounded elongate basin (Fig. 12.9a) which
The recognition of a single supercontinent during was followed by slow subsidence and the deposition
the Proterozoic places important constraints on the of shelf sediments (Fig. 12.9b). There followed a
types of tectonism which could have been active period of tectonic instability with faulting and ero
during this time. It must be noted, however,that this sion and the deposition of flanking conglomerates
supercontinental assembly was probably not com (Fig. 12.9c). Subsidence then increased greatly with
pletely rigid. The APW data are insensitive to rela the outpouring of voluminous basalts and the depo
tively small movements and can accommodate sition of shale and greywacke (Fig. 12.9d). The re
internal relative motions of up to about 1500 km gion then stabilized and a further cycle ofthe events
(McElhinny & McWilliams, 1977)and angular rota shown in Fig. 12.9b-d took place. The development
tions of up to about 10®. Piper (1987)suggests that ofthe trough culminated in the Hudsonian Orogeny
the stable Archaean cratons were subdivided by mo during which 100-200 km of crustal shprtening
280 CHAPTER 12
occurred as the volcanosedimentary sequence was formed by a process of vertical accretion and thus
folded and thrust on to the Superior block differ markedly from ophiolite complexes. He has
(Fig. 12.9e). suggested an intracratonic origin based on the crustal
Dimroth (1981) has demonstrated that the basalts delamination model of Bird (1979), shown in
of the Labrador Trough are autochthonous and Fig. 12.10. The trough was initiated when a crack in
IMPLICATIONS OF PLATE TECTONICS 281
the sub-crustal lithosphere developed and astheno- vection currents, then caused the final orogenic
spheric material rose and spread laterally at the crust- phase (Fig. 12.10c), induced by weakening of the
mantle interface (Fig. 12.10a)(Section 2.10.3). The crust by the intrusive activity or changes in the
flow of the asthenosphere was believed to be under velocity of the convection cells.
the influence of convection currents, which caused There are several points which should be consid
lateral spreading of the asthenosphere to only one ered concerning this model. The main driving force
side of the trough and dragged down the subcrustal oftrough formation and deformation is proposed to
lithosphere. The low density asthenosphere invasion be convectional motion in the mantle. At the present
caused uplift and the deposition of the continental day, in order to explain observed heat flow values
red beds. The uplift decayed after about 100 Ma and (Section 11.4), it is believed that convection cells are
shelf sediments were deposited. This phase was only present beneath the oceanic lithosphere. Given
followed by a massive intrusion of asthenosphere- that the radioactive content of the Proterozoic crust
derived mafic and ultramafic magmas into the crust, would have been greater than today,the temperature
whose high density caused tectonic instability and of the crust would have been higher and the contri
rapid subsidence as large quantities of basalts were bution of additional heat from sub-lithospheric con
erupted on to the sea floor (Fig. 12.10b). Continued vection cells would have raised the temperature of
sinking ofthe sub-crustal lithosphere, driven by con- the lower crust even more. It is probable that this
increased heat flow would have led to anatexis ofthe
lower crust and the ascent of considerable quantities
(a) Crust of acidic magma. There is no geological evidence for
the presence of such material within the Labrador
Mantle Littiosphere
Trough. The powering of trough evolution by man
tle convection implies that lithospheric motions were
Asthenosphere ■
driven by mantle drag forces. This appears unlikely
because, at the present time, mantle drag tends to
inhibit rather than enhance the motion of continen
tal plates (Section 11.7). Finally, the production of
magmas at the relatively shallow levels proposed for
the model requires that the melting temperature of
Mantle Lithosphere the mantle be considerably lowered in some way. In
modern subduction zones this achieved by the deri
vation of water from hydrous phases in the crustal
Asthenosphere^
portion of the subducting lithosphere (Section 8.8).
Such a mechanism is not possible for a downgoing
slab which has been delaminated from its crustal
component, and no other mechanism is proposed.
(c) Clearly, our understanding of mechanisms operative
in these so-called ensialic belts is incomplete. An
alternative mechanism is presented in Section 12.2.6.
Mantle Lithosphere
The concept of a Proterozoic supercontinent, al
though accepted by many workers, is not without its
Asthenosphere critics. For example, Dunlop (1981) accepts that
palaeomagnetic poles derived from cratons of indi
vidual continents tend to fall on a single APW
swathe, and thus that the mobile belts between these
cratons are the product of ensialic processes. He
Fig. 12.10 Intracratonic delamination model for the suggests, however, that palaeopoles from different
formation of the Labrador Trough (redrawn from continents do not follow the same APW path and
Dimroth, 1981, with permission from Elsevier). record large relative rotations and latitudinal shifts.
282 CHAPTER 12
For example, he argues that no match is possible of are many examples in the literature in which the
the APW paths ofLaurentia and Africa for the period origin of Precambrian orogenic belts has been ex
2300-1750 Ma unless the former is more complex plained in terms of modern plate tectonic processes.
or the latter less complex than indicated by available Gibb et aL (1978; 1983) have proposed that the
data, and that, throughout most of the Proterozoic, structural province boundaries of the Canadian
Laurentia, Africa and Australia drifted indepen Shield can be explained in terms of plate margin
dently. Confirmation ofthis model or ofthe Protero interaction during Proterozoic times. Using a combi
zoic supercontinent must await the accumulation of nation of geological information and gravity inter
high quality palaeomagnetic data in both space and pretation, they argued that the structures marking
time. This would provide information on those era- these suture zones, which are all characterized by a
tons at present only poorly sampled and fill in the remarkably similar gravity signature, were formed by
gaps in the records of better sampled cratons by the the continent-continent collision of the juxtaposed
identification of suitable rocks of the desired ages. provinces following the consumption of oceanic
The very major differences between the grouping of lithosphere beneath the member of the pair whose
the continents in the Late Precambrian proposed by structure and radiometric age were more recently
Morel & Irving (1978)(Fig. 12.1) and Piper(1987) overprinted at the time of the collision. Similar
(Fig. 12.7) suggest that it is probably premature to gravity signatures have been described from crustal
accept palaeogeographic reconstructions for the Pre province boundaries in the Precambrian of Australia
cambrian based on palaeomagnetic results. (Wellman, 1978; 1988), southern Africa (Reeves,
1976) and the Brazilian Shield (Lesquer et al.^
1981).
12.2.6. Proterozoic plate tectonics
The nature of these province boundaries is exem
Given the uncertainties inherent in the palaeomag plified by studies of the Labrador Trough. Kearey
netic investigation of Precambrian rocks, it is un (1976) has proposed a model for the evolution of
likely that it will ever be possible to reduce the width this structure (Fig. 12.11) which differs markedly
of the APW swathe which represents a form of from that of Dimroth (1981) described above. Be
confidence limit on the pole determinations. The fore the Hudsonian Orogeny the Superior and
uncertainty means that the presence of narrow ocean Churchill provinces were separated by a small ocean
basins, of up to 1500 km width, between cratonic basin (Fig. 12.11a) whose initiation by rifting may
areas cannot be discounted. Also, it must be remem have created the environment of the early red-bed
bered that palaeomagnetic methods give no control sediments in the trough. Subsequently, a subduction
on palaeolongitude (Section 3.6),so that linear inter- zone formed at the western margin of the Churchill
cratonic regions which strike in the same direction as block (Fig. 12.11b). Shales and greywackes filled the
the Eulerian pole used to bring the cratons into trench and forearc basin, while the eruption of volu
juxtaposition are not constrained to have had any minous quantities of tholeiitic basalt, derived from
particular width. The method of continental reas the upper mantle above the subducted lithosphere,
sembly based on palaeomagnetic poles assumes that built up a volcanic edifice on the continental margin,
these intracratonic regions would have the minimum A briefcessation of movement would have permitted
width possible when, in fact, they might have been shallowing of the forearc region and the deposition
represented by wide ocean basins. of shelf sediments before subduction recommenced
Palaeomagnetic data for the Precambrian thus can with the deposition of further turbiditic sediments
not preclude the presence ofat least relatively narrow and the further emplacement of basaltic rocks. Even
ocean basins between the Archaean cratons. Conse tually the ocean closed completely during the Hud
quently, rather than accepting that both intracra sonian Orogeny (Fig. 12.11c) and the Churchill
tonic and plate marginal tectonic processes were block and Labrador Trough were thrust over the
active in the Proterozoic, it is possible to postulate Superior craton. The Churchill block was then thick
that the mobile belts of this age were related to the ened and metamorphosed by the basement reactiva
Wilsonian cycle of present day plate tectonics (Sec tion mechanism of Dewey & Burke (1973)(Section
tion 10.4), involving the opening and closure of 9.3.2). The orogeny would have ended as the sub
ocean basins up to 1500 km in width. Indeed, there ducting oceanic lithosphere decoupled, and the sys-
IMPLICATIONS OF PLATE TECTONICS 283
Superior Churchill
t^UTT] Continental
■.\-/".^M crust
Subcrustal
lithosphere
n;; Potassic
^ \ upper crust
V/ '/\ Refractory
Superior Churchill I/- 'A lower crust
\jrZ\ Migmatite
LT - Labrador Trough
S - Sediments
V - Basic igneous rocks
G-Granitic rocks
C - Conrad discontinuity
M-Moho
iiiiiiiiiliiilllliii
Fig. 12.11 Plate tectonic model for the formation of the oped in response to indentation tectonics (Section
Labrador Trough: (a) Late Archaean (pre-2560 Ma); 9.3.6). If this parallel is correct, it implies that plate
(b) Lower Aphebian (2560-1800 Ma); (c) Hudsonian tectonic forces remained active after Precambrian
Orogeny (1800 Ma) (redrawn from Kearey, 1976, with collisions, and may have arisen from ridge-push
permission from Elsevier).
forces acting upon the opposite edge of the under-
thrusting plate.
tem stabilized. Thomas & Kearey (1980) have There are many other regions where plate margin
produced a further model for the Labrador Trough processes appear to have been active in Proterozoic
which recognizes its similarity to an Andean-type times which do not necessarily coincide with the
plate margin at an early stage in its development margins of the Proterozoic supercontinent. Hoffman
(Fig. 12.12). Elsewhere in the Canadian Shield, Gibb (1980) has described a complete Wilsonian cycle
(1978) and Hoffman (1987) have recognized, from (Section 10.4) in early Proterozoic rocks in the
the displacement of magnetic anomaly patterns, a northwest Canadian Shield. Stauffer et al. (1975)
series of wrench faults associated with province have described what they interpret to have been an
boundaries which bear a strong geometrical similar island arc sequence of andesitic lavas in the Amisk
ity to those apparent in the Himalaya which devel Group of Manitoba. Krogh (1977) has presented
284 CHAPTER 12
□ Metasediments
m Basic igneous rocks
|+U[ Granitic rocks
□ Gneisses & migmatites
(a) Granulites
-200 n
Bouguer ^
anomaly ■
(g-u.) Zone of
Granite Dorset
-1200 compression
Site of Hudsonian belt Fold Belt
Trench
500^
Height (m) 0-
Main
20-
Depth (km)
40- 4tQ.30|
200 km
West Peruvian Trough Zone of
(b) Peru-Chile Coastal Batholith
East Peruvian compression
Present Trough
Trench
coastline
//
0
40 H
Basic intrusions
I+ + +I Coastal batholith
Fig. 12.12 (a) Gravity interpretation of a profile across interpretation of a crustal section across the central Andes
the Labrador Trough in which the Churchill Province is at 100 Ma (redrawn from Thomas & Kearey, 1980, with
made up of several fault-bounded blocks, each in isostatic permission from Nature 283, 61-3. Copyright © 1980
equilibrium. Density contrasts in Mgm"^, (b) Schematic Macmillian Magazines Ltd).
evidence for subduction and continent-continent Shanti & Roobol, 1979). It has also been proposed
collision in the Proterozoic of western Norway. that the Midcontinent Rift of North America repre
Black et al. (1979) have proposed a cycle of ocean sents a continental rift developed in middle Protero
opening and closure in the Late Precambrian of West zoic times but which failed to develop significant
Africa and Abdelsalam & Stern (1993) a similar crustal extension (Chase & Gilmer, 1973; Green
Wilsonian cycle (Section 10.4) in the Proterozoic of 1983). For further examples see Windley (1993).
the Sudan. Watters (1976) has identified a possible It therefore appears that there is a large body of
late Precambrian subduction zone in southwest Af evidence to suggest that true plate tectonic processes
rica. Matthews (1972) has described an episode of were active in Proterozoic times, and that rather than
Precambrian obduction in southeastern Africa. Prot developing in ensialic environments, Proterozoic
erozoic subduction and ophiolite obduction have mobile belts evolved in plate margin settings. The
been suggested for several regions of the Arabian major difference from Phanerozoic times was that
Shield (Leblanc, 1976; Al-Shanti & Mitchell, 1976; the ocean basins are palaeomagnetically constrained
IMPLICATIONS OF PLATE TECTONICS 285
to have been somewhat narrower than at the present alkaline igneous rocks, and ore minerals are gener
day. The imaging, by seismic reflection,ofa structure ated from this magmatism. Carbonatites, which form
similar to a suture zone in the 2700 Ma Superior ring complexes within the alkaline rocks, are fre
Province of the Canadian Shield (Calvert et quently associated with extensive mineralization.
1995)suggests that modern-type plate tectonics may The important elements found in this environment
also have been extant in the Archaean. are phosphorus (as apatite), niobium (pyrochlore),
rare earths(monazite, bastnaesite), copper, uranium,
thorium and zircon. Also found are magnetite, fiuo-
12.3 Plate tectonics and
rite, barite, strontianite and vermiculite. Carbon
economic geoiogy atites may also contribute to the sodium carbonate,
chloride and fluoride found in vast quantities in the
12.3.1 Introduction
lakes of the East African rift system, although it is
The application of plate tectonic theory has not yet possible that these derive from weathering of the
led to the discovery ofeconomically viable mineral or alkaline rocks. Directly related to the magmatism are
hydrocarbon deposits. This is probably because it is porphyry and vein-type molybdenum deposits associ
only relatively recently that attempts have been made ated with subalkaline granites, copper-nickel depos
to relate the evolution ofeconomic deposits to plate- its associated with mafic intrusions and hydrothermal
controlled environments. Plate tectonics, however, copper deposits. Within the sediments related to
is important in this context as it provides exploration rifting, stratiform copper deposits of great volume
geologists with a framework to which they can relate are associated with specific shale or sandstone hori
the habitats and spatial relationships of economic zons. These disseminated ores are believed to form
deposits and to trace their structural development during the first marine transgression into the conti
(Rona, 1977). Studies of this kind are expected to nental interior and overlie red-bed horizons, and are
become increasingly important in the future as probably derived from the copper-rich rift basalts in
the search concentrates on smaller, more obscure response to the elevated heat flow ofthe rift. Carbon
deposits. ate hosted lead-zinc-barite ores are also found in the
Several publications in this field are available and intracratonic rifts and passive continental margins,
the following account will attempt to summarize the typified by the deposits of the Upper Mississippi
most important findings. For more detail the reader Valley of North America.
should refer to Mitchell & Garson (1976; 1981), With the development of a narrow ocean basin
Rona (1977), Tarling (1981), Hutchinson (1983), between the rifted continental fragments, new min
Sawkins(1984) and Evans (1987). eral deposits are created at the mid-ocean ridge. The
There are several types of economic deposit which present day example of this environment is the Red
can be related to plate tectonic processes: Sea. Here 13 pools of hot brines (Fig. 12.14) have
1 autochthonous deposits directly related to mag- been located along the central ridge where it is
matism at plate margins and interiors; intersected by transform faults. These contain zinc-
2 allochthonous deposits related to plate margin copper-lead sediments of possible economic value,
magmatism; for example the Atlantis II Deep, which contains
3 deposits related to sedimentary basins formed by sulphide layers with zinc contents of up to 20%
plate motions; which are 20 m thick and cover an area of over
4 deposits related to climate and hence north-south 50 km^. It is generally agreed that the metals are of
plate motions. volcanic origin and have been concentrated into
brines by the thermally-induced circulation of sea
water through the volcanic rocks and thick evaporite
12.3.2 Autochthonous and
sequences found in this region (Cowan & Cann,
allochthonous deposits
1988). As the ocean basin evolves, these deposits
The various plate tectonic environments in which may become buried by sediment and reappear in
metalliferous deposits are found are shown in collisional orogens where the tectonism obscures
Fig. 12.13. The initial rifting of a continent (Section their original setting. Also associated with this ad
10.3) includes the emplacement of alkaline and per- vanced phase of rifting are sediment-hosted massive
N)
00
O)
o
X
>
-0
(a) Intracontinental hotspot Sn, F, Nb (b) Aborted rift zone Carbonatite Nb, PoO-,,
REO, Ba, F
?Sullivan-type Pb-Zn-Ag j
Sn, F. Nb
}'\ iTv--1'I// J
-'r/x> ''r '"/»' V^
(c) Intercontinental rift zone Evaporitic layer (d) Ocean rise and Hawaii-type chain
. LIL-rich Potassic
Red Sea Mn '^^gnd'^m^ ds Mississippi-type LIL-depleted
I II ^ I . .basalt basalt basalt
^ ^' l\ Pb-Zn-Ba ?Fe, Mn,As,B,Zn,Pb,Hg seds
Cyprus-type Cu-F^ '
Alkali-Depleted
?i*A*i*i%*»*i'PodIfnrrri Cr Ni^
(e) Andean-type magmatic belt (f) Island arc and inter-arc basin
Mn nodules with Porphyry Sn, W , ^ Incipient Inactive Cyprus-type
Cu, Nl, Co Cu, Mo, Hg / f Kuroko Zn, Cu ^ riff inter-arc Cu-Fe
basin
I ^ 1.,^,V
(g) Magmatic and outer arc with marginal basin
D ~ Ophiolite Sn, W, Bi, Mo, F
Cyprus-type Cu-Fe~,^°"''°''7^
I ■ or mantle diapir C,p™.-W, J s.
Porphyry
^u-Au —7*^ .
^;vtVTW«
ts)
00
vj
288 CHAPTER 12
• Brine pool
A Metalliferous sediments
Oceanographer
Kebnt deep
Gypsum deep
Vema deep
Nereus deep
Thetis deep
Atlantis II deep
Valdivia deep Albatross deep
Discovery deep Erba deep
Sudan deep
Shagara deep
Suakin deep
;P
Saudi
Arabia
sulphide deposits which occur in thick continentaiiy- which allows fluids to percolate with relative ease.
derived clastic sediments on passive continental mar Hydrothermal processes of low intensity lead to the
gins. They comprise single or multiple lenses of formation offerromanganese nodules, and encrusta
pyrite, galena and sphalerite ores with minor silver tions of iron and manganese on pillow basalts at the
and copper. These are not common and probably layer 1-layer 2 interface. Higher intensity hydrother
reflect the influence of metal-rich formation waters mal activity has been observed at some locations,
powered by long-lived geothermal systems such as on the East Pacific Rise where discharge is of
As an ocean basin continues to grow,contempora two types. Black smokers are vents where pyrrhotite
neous mineralization takes place at the mid-ocean particles are discharged, producing ores which may
ridge, and has been observed at certain locations be zinc- or iron-rich and containing lesser amounts of
along the Pacific (Corliss et al.^ 1979), Atlantic cobalt, lead, silver and cadmium. At white smokers
(Scott et 1973) and Indian Ocean ridges. The little sulphide material is discharged, and the main
mineralization is of hydrothermal origin and its loca precipitate is barite.
tion depends upon the availability ofoceanic crust of In open ocean conditions ferromanganese nodules
high permeability overlying the magma chamber and encrustations form on top of basalt or sediments
IMPLICATIONS OF PLATE TECTONICS 289
where strong ocean currents prevent the accumula lar masses within the harzburgites and dunites ofthe
tion of clastic sediments. These deposits are of upper lithospheric mantle. These deposits may have
hydrous origin and accumulate slowly, sometimes formed by partial melting of primitive mantle ma
forming extensive pavements. As well as iron and terial or by crystal fractionation within the magma
manganese, they also contain smaller amounts of chamber underlying ocean ridges (Section 6.10).
copper, nickel and cobalt. Similarly associated with the magma chamber are
In addition to the exhalative processes of mineral nickel and platinum sulphides. The model for miner
ization described above, ore bodies may form within alization of the oceanic lithosphere derived from
the oceanic lithosphere as it is created. Much of our ophiolite studies is shown in Fig. 12.16 (Rona,
knowledge of these deposits is derived from studies 1984).
of ophiolites (Section 2.5), which are interpreted as Several forms of mineralization are present in
allochthonous slices of oceanic or back-arc basin subduction zone environments, their types depend
lithosphere tectonically emplaced within continental ing upon whether the overriding lithosphere is con
crust during collisional orogenesis. One of the most tinental or oceanic. Hedenquist & Lowenstern
intensively studied bodies ofthis type is the Troodos (1994) have reviewed the role of magmas in the
complex of Cyprus(Fig. 12.15), and other examples formation of hydrothermal ores in such environ
include ophiolites of northeastern Newfoundland, ments. The most important mineralizations are the
the Semail Ophiolite of Oman and Ergani Maden in porphyry coppers. These are very large, low grade
Turkey. At high levels in the lithosphere, massive deposits which are gold-rich and molybdenum-poor
sulphide deposits (marcasite, chalcopyrite and when associated with island arcs and gold-poor and
sphalerite) occur on top ofor within the pillow lavas molybdenum-rich in Andean-type mountain ranges,
oflayer 2. It has been suggested that these sulphides being found, for example, in the Andes themselves,
formed either in a manner similar to the brines ofthe the Philippines, Taiwan, Puerto Rico and the Ryuku
Red Sea or by precipitation from hydrothermal solu and Burma arcs. The controls on the location of
tions which became enriched in metals by circulating porphyry coppers are poorly understood, but they
within the volcanic rocks. The deeper, plutonic por appear to be related to the primary calc-alkaline
tions of ophiolites contain economic deposits of magmatism and do not require the presence of any
chromite, which occur as podiform bodies and tabu specific crustal type. Solomon(1990)has noted that.
Umber
o
(/) Upper pillow lavas A AA
T3 Massive Fe-Cu sulphides
O Lower pillow lavas
2'Q.i5
Basal group
0
-D 0
^ X
"S Q- Diabase
B E
0 o
0 o 1 km
CO
- Plaglogranlte
Gabbro +
[ 0
Q.
E Layered gabbro
8 Pyroxenltes -
O
Fig. 12.15 Mineralization in the 'c Wehrllte Cu-Ni-Co-Fe sulphides
o
Troodos Ophiolite (redrawn from 3 □unite £- Podiform chromite
Searie & Panayiotou, 1980, with Q.
Layer 1:
Sediment
sediment
Metalliferous sediment ^ Layer 2:
(Cu,Fe, Mn,Pb,Zn,Ba,Co,Ag,Au) mafic rocks
Stratiform oxides, hydroxides, _ E
(basalt)
silicates(Fe, Mn) >
Layer 3:
Massive,disseminated,and
mafic and
stockwork sulphides(Cu,Fe,Zn,Ag,Au)
ultramafic rocks
Dykes -5
Gabbro and serpentinite Upper mantle:
ultramafic rocks
Cumulates —
Chromite-
Lherzolite and harzburgite —
Nickel and platinum sulphides -
Fig. 12.16 Schematic block diagram showing the ment from Cyprus- or Kuroko-type. They are associ
potential distribution of mineral deposits in the oceanic ated with intermediate to basic volcanic rocks with
lithosphere (redrawn from Rona, 1984, with permission carbonaceous mudstones, clastic limestones or
from Elsevier).
quartzites, all of which suggest a deep water environ
ment unlike ocean ridges, ocean basins or island arcs.
in the southwestern Pacific rim, porphyry copper- They have been termed Besshi-type deposits. They
gold deposits mostly form after a reversal of arc may have formed in a trench or a tensional environ
polarity following a collisional event. Sometimes ment, but their origin remains, as yet, obscure.
associated with porphyry coppers are mercury depos There are several types ofdeposit which are specific
its (as cinnabar or quicksilver), which may have to Andean-type subduction. These include strata-
originated in a similar manner. bound copper sulphide deposits,such as are found in
Another important class of deposits found associ Chile, which are closely related to episodic calc-
ated with oceanic subduction zones (Fig. 12.17) is alkaline volcanism and occur within porphyritic
stratiform massive sulphides ofzinc,lead and copper andesite lavas. The principal minerals are chalcosite,
known,after their type area ofoccurrence in Japan,as bornite and chalcopyrite, and they contain signifi
Kuroko-type ores. They were deposited in a shallow cant amounts of silver. The intercalation of these
marine environment and occur interbedded with deposits with shallow marine and terrestrial deposits
pyroclastics and silicic calc-alkaline lavas. It is be suggests their formation in small lagoons. Tin and
lieved that they are ofsyngenetic origin, and occur at tungsten mineralization occurs in the eastern Andes
a late stage of arc evolution. Halbach et al. (1989) of Peru and Bolivia on the landward side of the
suggest that they formed in a back-arc basin, and cite porphyry copper belt. It appears to be derived from
the Okinawa Trough as a modern analogue. They the same Benioff zone region as the magmas, and
may have been deposited by saline submarine hot may owe its existence to the anomalously shallow
springs arising from the separation of aqueous ore dip of the subduction zone in this region (Section
fluids during the final stages of magmatic fraction- 9.2.3).
ation or from the leaching of older volcanic rocks. In the back-arc environment of Andean-type sub
Kuroko-type ores may be incorporated into conti duction zones, such as in east and southeast Asia,
nents during continent-island arc collisions, such as there occur granite belts with which are associated tin
at Rio Tinto in Spain, Umm Samiuki in Egypt and and tungsten with lesser molybdenum, bismuth and
the Buchans mine in Newfoundland. fluorite. The origin ofthe tin, in particular, is contro
There also exist other forms of stratiform massive versial. Tin is present in only minute quantities in the
sulphides which differ in their depositional environ oceanic crust, and is similarly absent in island arcs.
IMPLICATIONS OF PLATE TECTONICS 291
Outer arc
Batholiths
7^^^~^^'Continentai
Oceanic crusT™®^ "beanie crust'
Descending lithospheric
plate
Ascending
calc-alkaline
magma
Asthenosphere
Fig. 12.17 Development and emplacement of mineral major episode offaulting. They probably originate in
deposits in a subduction-related setting (redrawn from magmatically associated brines.
Evans, 1987, using data from Sillitoe, 1972a; 1972b, with In the marginal seas of oceanic subduction zones
permission from the Economic Geology Publishing Co. (Section 8.9) the crust is similar to oceanic, although
and the Institute of Mining and Metallurgy).
generated in a different fashion, and so mineral
deposits would not be expected to differ greatly from
An oceanic origin of the tin appears unlikely. One those in oceanic crust. The mineralization would be
hypothesis is that tin is derived from deep in a Benioff similar to that formed during the early development
zone which is migrating away from a continent of a spreading ridge, and would thus be related to
during back-arc spreading. The fluorine originating magmatic and exhalative volcanic processes. If the
at these levels would extract tin from deep levels of greenstone belts of Precambrian terrains are the an
still hot granite plutons and deposit it at the surface cient analogues of back-arc basins (Section 12.2.3),
in their vicinity. Another hypothesis is that the gen then the copper-zine sulphides and gold deposits
eration oftin requires the presence ofthick continen found in them must have originated in this environ
tal crust, such as is present in the tin belts of the ment.
Andes, Alaska and upper Burma, and a shallow Collisional environments may contain allochtho-
dipping Benioff zone which acts as a source of heat nous members which contain deposits typical of the
and volatiles. In this case the tin may originate from earlier stages of crustal accretion, such as ophiolites,
pre-existing concentrations in the lower continental ferromanganese nodules, subduction-related de
crust. posits and mineralization related to early stages of
Within ensialic back-arc basins vein-type gold and rifting. Additionally, deposits which originate in
silver deposits are common,such as are found in the continental collision are also present. Commonly
Great Basin of Nevada. These are associated with granite bodies are emplaced during and after the
andesites, dacites and rhyolites and pre-date the collisional event, such as in the Variscan belt of
292 CHAPTER 12
northwest Europe. Associated with these granites are tions are necessary for the development and preserva
tin-tungsten deposits of cassiterite and wolframite. tion of these resources.
It is believed that the mineralization, like the gran There are four principal criteria which must be met
ites, is derived from anatexis ofthe lower continental for the development of petroleum and gas, herein
crust, either by magmatic concentration or the melt after referred to as hydrocarbons: layers rich in or
ing ofexisting mineral accumulation. Also associated ganic matter within the sedimentary succession; a
with these granites are vein-type deposits ofuranium. source of heat applied for a time sufficient for the
There are no economic deposits associated with gran maturation of organic materials into hydrocarbons;
ites in the Himalaya,all known examples being older. permeable pathways which allow movement of the
Some models of the evolution of collisional oro- hydrocarbons; and a porous reservoir whose top is
gens require the formation of an anorthositic lower sealed by impermeable capping beds.
crust after the upward migration of potassium and The main source ofthe disseminated organic mat
the lithophile elements(Section 9.3.2). The majority ter, or kerogens,in sediments is plankton. The abun
of massive anorthosites are of Precambrian age and dance of plankton is controlled by climate, the
so this hypothesis cannot be tested by modern ana quantity of nutrients available and water body geo
logues. Associated with anorthosite bodies are iron- metry. The first two factors are latitude dependent,
titanium deposits of magnetite and ilmenite, which and the majority of oil basins originate at low lati
probably originated in the same partial melting tudes. The latitude is obviously affected by the
episode. north-south component of plate motion, while the
Oceanic transform faults are favourable environ plate configuration at any given time determines
ments for mineralization because they may be associ water body geometry. Organic material is especially
ated with high heat flow and provide highly fractured abundant along continental margins where there is
and permeable conduits for both the downward major river runoff into large deltas.
percolation ofsea water and the upward migration of The preservation of kerogens requires conditions
mineralizing fluids. Iron sulphide concretions have which are not oxidizing. These are achieved along
been reported from the Romanche Fracture Zone of continental slopes where the production of organic
the equatorial Atlantic which may have originated by matter exceeds the availability of free oxygen to
this mechanism. The brine pools of the Red Sea convert them to carbon dioxide, and in closed anoxic
appear to be located where transform faults intersect basins. It follows that the shales and mudstones
the central ridge, and it is possible that the metals produced in such environments are the most com
ascend along these faults. Indeed, base metal depos mon source rocks as they have the ability to absorb
its are found along the continental continuation of kerogens and remove them from the effects of free
the faults. A similar mechanism has been postulated oxygen.
for the brines of the Salton Sea, California. It is The temperature experienced by the kerogens after
probable that the ultramafic intrusions occurring in burial is critical, and depends on the local geothermal
fracture zones (Section 7.3) contain high propor gradient. Temperatures of 70-85 are required to
tions of nickel, cobalt and copper. develop liquids and 150-175 ®C for dry gas. It is
Very little mineralization appears to be associated also important that a critical exposure time to these
with continental transform faults at the present level temperatures is exceeded, so the basin must be free
oferosion. from tectonism and uplift during this period.
Precambrian mineral deposits can be widely inter After formation, the hydrocarbons undergo pri
preted as forming in plate tectonic environments, mary migration from the fine-grained source rocks
particularly those related to divergent plate margins and secondary migration as they concentrate and
and subduction zones (Gaal & Schulz, 1992). accumulate in a reservoir of high porosity. Migration
occurs because ofthe buoyancy ofthe hydrocarbons,
and it follows that all hydrocarbon accumulations are
12.3.3 Deposits of sedimentary basins
allochthonous. There are several types of oil trap,
The majority of fossil fuels are found within sedi including anticline, fault, stratigraphic, unconfor
mentary basins whose formation can be related di mity and lithological traps, which have the effect of
rectly or indirectly to plate motions. In addition to providing a capping to the reservoir with an imperme
the sedimentary environment,quite stringent condi able cover which prevents further upward movement.
IMPLICATIONS OF PLATE TECTONICS 293
mmmm
Trade Winds
Ocean current
Major oil basin
Mid-cretaceous
Fig. 12.18 Possible relation of oil basins to the Tethys- type mountain ranges, e.g., the Aquitaine Basin,
Atlantic seaway in middle Cretaceous times (redrawn from southwest France,
Irving et al., 1974, with permission from the National 11 tensional basins associated with indentation tec
Research Council of Canada). tonics (impactogens), e.g. the Rhine Graben.
Not only can plate tectonics create the habitat of
hydrocarbon deposits, it can also explain why certain
Plate tectonics contrors the locations of reservoirs regions are particularly rich in these deposits. A large
in that it is responsible for the formation and preser proportion of the Earth's hydrocarbon reserves are
vation of the sedimentary basins in which hydro located in the Middle East, and the evolution and
carbons are generated and trapped. These include: preservation of these deposits has been a conse
1 intracratonic basins formed by hotspot activity, quence of a specific pattern of plate interactions
Paris and Michigan basins (Section 10.5.2); (Irving 1974).
2 basins associated with continental rifting, e.g. During Mesozoic and Early Cenozoic times two
the Gulf of Suez, Red Sea; large embayments existed on the continental shelf of
3 aulacogens (Section 10.2), e.g., the North Sea; the Afro-Arabian continent on the southern side of
4 passive continental margin basins, e.g., the Ga the Tethys Ocean (Fig. 12.18). Such embayments
bon Basin; around the Tethys, which also included the Gulf of
5 ensialic back-arc basins, e.g. the Oriente Basin of Mexico and the Persian Gulf, may have been con
Ecuador and Peru; nected via the proto-Mediterranean Sea, or Tethys-
6 marginal seas, e.g., the Andaman Sea; Atlantic seaway, which was situated at low latitudes.
7 accretionary prisms, e.g., the coastal oil-fields of At about 100 Ma the rate ofspreading of the seaway
Ecuador and Peru; increased, maximizing the development of hydro
8 forearc basins, e.g., the Cook Inlet of southern carbon source rocks because of the formation of
Alaska; extensive, warm,shallow seas to which were supplied
9 pull-apart basins associated with wrench faults large quantities of nutrients from the spreading
(Section 7.4.3), e.g. the Los Angeles Basin, western centre. When the Seaway subsequently began to
USA (Moody, 1973); close following the development of a subduction
10 foreland basins (Section 10.5.3) of Himalayan- zone at its north margin, the geometry of the plate
294 CHAPTER 12
movements was such as to protect the Persian Gulf are likely to be preserved due to their stable sur
from major tectonism. This arose because the rapid roundings. Coal deposits are also found in aulaco-
northerly motion of the Indian Plate absorbed most gens and ensialic back-arc basins. The tectonism
of the energy associated with the collision with the associated with collisional orogens provides an envi
Eurasian Plate. The Gulf of Mexico was similarly ronment whereby coals increase in grade by high
protected by northeastward motion of the Greater pressure metamorphism.
Antilles.
Coal is a combustible sedimentary rock containing
12.3.4 Deposits related to climate
in excess of50% by weight ofcarbonaceous material.
It is formed by the decomposition, compaction and It has already been stated that the formation of
diagenesis ofan accumulation ofterrestrial and fresh hydrocarbon and coal deposits is dependent upon
water plant debris. Coals thus appear in the geologi both climate and special conditions of sedimenta
cal record from Devonian times when the first plants tion. There are certain deposits, however, which
appeared. appear to be related solely to climate. Since climate is
In order to prevent the total destruction of the largely dependent upon latitude, north-south plate
vegetable matter by biochemical decomposition, motion can be considered as controlling the forma
very wet conditions are required to stop the decay by tion of such deposits. They include laterites and
the accumulation of toxic waste products. The con evaporites.
ditions under which this process occurs are con The most important lateritic deposit is nickel later-
trolled by climate and topography. Normally a warm, ite, which results from the extreme weathering ofthe
wet climate is required to promote luxuriant growth, ultramafic parts of ophiolite bodies under tropical
and this should be under the condition of constant conditions. The original nickel content offresh peri-
standing water. Although,in regions of high rainfall, dotite becomes enriched by a factor of about seven
peat forms in upland areas, it is rarely preserved due under the influence of such weathering by percolat
to the erosion experienced in this environment. The ing ground water. These deposits are becoming in
prime conditions for coal formation are those offlat, creasingly important sources of nickel, and are
low-lying ground invaded by swamps with stagnant exploited in the southwestern Pacific and the north
water. The slow sinking ofthese regions preserves the ern Caribbean.
organic layers by progressive burial. A similar deposit is bauxite, a residual deposit
The process of coalification refers to the physical enriched in aluminium hydroxide, which provides
and chemical changes experienced by the organic the vast majority of the world's aluminium. This
matter after burial in response to rising temperature forms by the in situ weathering of aluminosilicate
and pressure. On compression, water and volatiles minerals on stable peneplaned topography in a wet
are expelled and the deposit becomes enriched in tropical climate by the intense leaching of alkalis and
carbon. The degree ofcoalification is reflected in coal silica. Bauxite only forms within 30® of the equator
of different ranks, varying from the low rank lignite and requires high rainfall and high ambient tempera
to high rank anthracite. tures. It is mined in Jamaica, northern Australia and
Plate tectonics affects coal formation in that it China.
controls the latitude of a region, since most coal Evaporites form in an arid climate by the evapora
deposits originated at low latitudes. Plate tectonics tion ofsea water in semi-isolated basins which receive
also creates the environments necessary for the pre periodic marine influxes. They cannot develop by the
servation of organic matter, of which the most im evaporation ofa single isolated body of water, as this
portant are passive continental margins. Deltas could not explain the vast observed thicknesses of
formed on such margins produce the most favour evaporite deposits. The sequence of minerals pre
able conditions for coal formation, and swamps can cipitated is calcium carbonate and sulphate, sodium
develop on a regional scale. Present day examples chloride and finally magnesium and potassium min
include the Niger, Amazon and Mississippi deltas, erals. Evaporites are important commercially in the
and ancient examples the Carboniferous coals of chemical industry, particularly for the potash salts.
North America and northwest Europe. Intracratonic They are also important in the generation of hydro
deltas,such as the Rhine,are similarly productive and carbon traps because, being of low density, they
IMPLICATIONS OF PLATE TECTONICS 295
Paris Basin, where space heating for over 20 000 Windley, B.F.(1981)Precambrian rocks in the light ofthe
plate-tectonic concept. In KrOner, A.(ed.) Precambrian
dwellings is provided by deep fluid circulation.
Plate Tectonics., pp. 1-20. Elsevier, New York.
Windley, B.F. (1984) The Evolving Continents (2nd edn).
Wiley, New York.
Windley, B.F. (1993) Uniformitarianism today: plate tec
tonics is the key to the past./.geol. Soc. Land. 150,7-19.
Review questions
1 Summarize the subdivisions of the Earth's sili latitude and orientation of a continental area.
cate mantle and metallic core which have been Outline the various stages involved in carrying
deduced from seismological studies. Indicate out such a palaeomagnetic study.
briefly the type ofseismological study or observa Describe two ways in which studies based on
tion which provides the evidence for each layer the remanent magnetization of rocks have con
that you mention. tributed to the verification of continental drift
Suggest a plausible temperature distribution and the formulation of plate tectonics.
within the Earth that is consistent with the phys 6 (a) Define the term transformfault and illustrate
ical state and theological properties of these the six theoretical possibilities for active,
different layers. right-lateral transform faults which do not
2 Describe the way in which the seismic 'first mo terminate in triple junctions. Indicate the
tion' at a number of seismological observatories way in which the geometry of each type
across the world can be used to obtain a focal evolves with time and, where possible, cite a
mechanism solution for an earthquake of suffi recent or present day example.
ciently large magnitude. (b)Explain the significance of transform faults
Explain how, for near-surface earthquakes, when deducing the relative motion between
such solutions can be related to a simple theory adjacent lithospheric plates.
of faulting and hence to the principal stress (c) Explain the origin and nature of pull-apart
directions in the vicinity of the earthquake and basins and strike-slip/sheared continental
the sense of motion on the fracture. margins.
3 Review critically the various lines of geophysical 7 Explain the principles upon which the direction
evidence suggesting that there is a layer of and rate of relative motion between lithospheric
low strength (the asthenosphere) in the upper plates are deduced for the different types of plate
mantle. How,ifat all, does the depth ofthis layer boundary developed at the Earth's surface. Re
vary beneath different geological provinces? view briefly the types of argument used to con
4 In the late 1960s four aspects of seismology vert these 'relative' motions into 'absolute' plate
provided the basis for the formulation of the motions with respect to the Earth's deep interior.
concept of plate tectonics: To what extent do the resulting models for
(a) more precise epicentral and focal depth deter absolute plate motions differ?
minations; 8 Briefly outline the basic assumptions involved in
(b)the study of surface waves and free oscilla making a palaeomagnetic study.
tions; Explain qualitatively how the global palaeo
(c) attenuation studies; and magnetic dataset for rocks formed during the
(d)the determination of focal mechanism solu past 200 Ma may be analysed to test:
tions. (a) the geocentric dipole model for the Earth's
Explain briefly the nature and significance of magnetic field;
each of these contributions. Hence show how a (b)the hotspot model for absolute plate mo
nearly complete statement of plate tectonics can tions; and
be derived from seismology alone. (c) the suggestion that much of the Western
5 Explain the underlying assumptions ofthe palaeo- Cordillera of North America consists of'dis
magnetic method for determining the palaeo- placed' or 'suspect' terrains.
297
298 REVIEW QUESTIONS
In each case take care to state any assumptions your answer by outlining the calculations in
and limitations of the method. volved in the simple predictive models of:
9 Explain the following plate tectonic concepts, (a) the maximum sedimentary loading of oce
illustrating your answer with specific examples: anic crust;
(a) failed arms or rifts; (b)the subsidence history of extensional basins;
(b)triple junctions; and
(c) back-arc spreading; and (c) the evolution of passive continental margins.
(d)Benioff plane reversal. 16 Formulate a possible 4ife cycle' for an ocean
10 Explain current ideas and models for the evolu basin within the framework of continental drift,
tion of passive continental margins. Pay particu sea floor spreading and plate tectonics. Give
lar attention to those aspects of the models that present-day examples to illustrate various stages
help to explain the observed diversity of such in your proposed cycle.
margins at the present day. 17 Compare and contrast island arcs and Andean-
11 Cite the empirical, i.e. geological and geophysi type mountain ranges.
cal, evidence for brittle fracture in the outer 18 Discuss the statement by Wyllie that 'Plate tec
layers of the Earth's crust. Explain the way in tonics has given us a framework of tectonic
which the nature and geometry of the fractures environments within which various suites of
reflect the state of stress in these layers at the igneous rocks can be located'.
present day or in the past. 19 Continental rift zones are thought to form in a
Focal mechanism solutions for certain earth number of different ways. Explain the different
quakes can be used to give information on the models for the formation of such rifts, paying
state of stress at the present day. Outline the particular attention to similarities and differences
principles and the limitations of this method for regarding crustal structure and geological his
determining the principal stress directions. tory.
12 Outline the McKenzie (1978) model for the 20 Some geologists and geophysicists claim that
development ofsedimentary basins. studies of the structural setting, internal struc
Describe some of the ways in which this ture and composition of ophiolites are highly
model has been modified subsequently to ac relevant to a better understanding of the pro
count for discrepancies between the observed cesses occurring in subduction zones and at
subsidence history and thermal evolution of ocean ridge crests. Review critically the basis and
sedimentary basins, and the predictions of the implications of these claims.
simple model. 21 Compare and contrast continental and ocean
Indicate briefly the potential importance of lithosphere.
such variations in relation to the formation of 22 Explain why studies ofisostasy are important in
oil and/or gas in a particular sedimentary basin. considering the Earth's dynamic behaviour.
13 Explain how earthquake focal mechanism solu 23 Write an essay on the evolution of the theory of
tions can give an indication of the stress field in plate tectonics.
the vicinity of the earthquake focus. Cite exam 24 Describe the data and techniques used in making
ples of focal mechanism solutions and inferred continental reconstructions.
maximum and minimum principal compressive 25 In what ways have studies ofterrestrial heat flow
stress directions for earthquakes occurring: contributed to plate tectonic theory?
(a) at plate boundaries; and 26 Explain the strengths and limitations of palaeo-
(b)within plates. climatological techniques for determining past
14 Extension ofthe Earth's crust may be accommo plate motions and geographies.
dated in a number of different ways. Write a 27 Describe the evidence for plate tectonic processes
short essay outlining the different possibilities, being operative in Palaeozoic times.
distinguishing between brittle and ductile re 28 Relate metamorphism to plate tectonic environ
sponses, and pointing out any thermal and/or ment.
isostatic consequences of different models. 29 Discuss the role of supercontinents within the
15 Explain the principle of isostasy and illustrate basic framework of plate tectonics.
REVIEW QUESTIONS 299
30 How and why does the Archaean tectonic regime tances from the source to place detectors if
differ from that ofthe present? critical-angle reflections from the interfaces are
31 Write an essay on energy resources and plate to be observed.
tectonics. 47 The following table lists data obtained from two
32 Discuss the evidence for the concept that suspect tectonic provinces in North America. Calculate
terrains played an important role in the tectonic the heat flow at each site and, using graphs or
development ofthe Western Cordillera of North calculations as appropriate, discuss the signifi
America. cance of the results:
33 Compare and contrast the evolution of the
Andes and the Himalaya.
34 Describe the scope for the development of thick Heat production Thermal Temperature
of surface rocks conductivity gradient
wedges of sediment within the context of plate
Site (lO-^Wm-3) (Kkm-^)
tectonic environments.
35 Discuss the evidence for convection in the man Cenozoic province (last thermal event at 20Ma)
tle. A1 4.28 6.6 15.0
36 Outline and discuss the data which have led to A2 5.50 4.9 22.5
A3 5.90 3.8 30.0
current theories of the causes of plate motions.
A4 3.78 3.9 23.9
37 To what extent can the occurrence of mineral
deposits be related to plate tectonic processes?
Caledonian province (last thermal event at 400Ma)
38 Explain the evidence for and role of suspect B1 2.14 2.0 25.0
terrains in the geological development of the B2 2.90 3.0 18.5
Himalaya. B3 3.57 3.0 20.2
39 What has been the impact of plate tectonics on B4 4.83 4.2 16.0
geological studies of continental areas?
40 Give an account of the principal mechanisms of
crustal extension and relate them to specific 48 Assuming that Airy's mechanism for isostatic
tectonic settings. compensation is perfectly obeyed, show that the
41 Describe the structural features of the upper thickness of continental crust at any point is a
crust which might be associated with transform linear function of the topographic height at that
and transcurrent faults. point.
42 Describe how polarity changes of the Earth's The Bouguer gravity anomaly map for a
magnetic field can be used for dating purposes. mountainous area of Utopia was corrected for
43 Discuss the statement 'All orogenesis is related isostatic compensation using an Airy model. In
to continental collision'. general the anomalies on the resulting isostatic
44 What geophysical measurements might assist in anomaly map were negligible, with the exception
elucidating the vertical extent of convection in of one area where a broad band of negative
the mantle? anomalies persisted. This was interpreted as
45 Describe the generalized morphology of a sub- arising from compensation for topography
duction zone, explaining all the features de which has since been removed by erosion. If this
scribed in terms of a subduction model. residual isostatic anomaly had an amplitude of
46 In a particular continental area, the crust is -1000 g.u. and the area is broad, estimate the
believed to consist ofan upper layer, ofthickness excess thickness ofthe root. To what thickness of
25 km and P wave velocity 6.0 km s"^, overlying eroded crustal material does this correspond?
a lower layer of thickness 15 km and P wave (G = 6.67 X 10"^^ m^ kg"^ s"^; 1 g.u. =
velocity 7.0kms~^. The upper mantle P wave 10"^ms~^;
velocity is about 8.2 km s"^ By calculating ar Mean density ofcrust = 2.85 Mg m~^;
rival time curves for refracted arrivals from a Density of upper mantle = 3.35 Mg m"^).
surface explosion,show whether or not the head 49 The thermoremanent magnetization of the Pali
wave from the lower layer of the crust can be sade Sill, which overlooks the Hudson River
expected as a first arrival. Deduce the best dis and the city of New York,is directed towards the
300 REVIEW QUESTIONS
north and has an inclination of 25.6° tion at T isy mm a"^, describe the way in which
downwards. Calculate the implied palaeomag- the situation evolves with time for each of the
netic pole position (in present day coordinates), following five cases:
and the geographic latitude ofthe sill at the time (a)y less than x\
ofits intrusion, approximately 190 Ma. (b)y equal tox;
Assuming that the difference between the (c)y between x and 2x;
present and the palaeolatitude ofthe sill is due to (d)y equal to 2x; and
the drift of North America with respect to the (e)y greater than 2x.
geographic pole, calculate the implied average
rate of drift during the past 190 Ma. Is this an
absolute, maximum or minimum average rate.>
Explain.(Present latitude and longitude ofsill =
40.5°N,75°W; 1 degree oflatitude = 111 km.) R
a:
50 In the plate tectonic situation illustrated below
three plates (A, B and C) are separated by a
symmetrically spreading ridge at R and a subduc- Hint: Consider plate C to be fixed and assign
tion zone at T. If the rate of spreading at R is velocities to A,R and B,relative to C,in terms of
X mm a"^ per ridge flank, and the rate ofsubduc- X andy.
Appendix / The geological timescale and
stratigraphic column
Epoch
Quaternary Pleistocene
Miocene
Oligocene
Cenozoic Tertiary
Paiaeocene
Cretaceous
Jurassic
Mesozoic
Triassic
Permian
Carboniferous
Devonian
Palaeozoic
Ordovician
Cambrian
(Eon)
Proterozoic
Precambrian
Archaean
Abdelsalam, M.G.& Stern, R.J.(1993)Tectonic evolution Artemjev, M.E. & Artyushkov, E.V.(1971) Structure and
ofthe Nakasib suture, Red Sea Hills, Sudan: evidence for history of the Baikal rift and the mechanism of rifting.J.
a late Precambrian Wilson cycle. J.£feol. Soc. Lond. 150, geophys. Res. 76,1197-211.
393-404. Ashby, M.F. & Verrall, R.A.(1977) Micromechanisms of
Aki, K., ChristofFerson, A.& Husebye, E.(1977) Determi flow and fracture,and their relations to the theology ofthe
nation of three-dimensional seismic structures of the upper mantle. Phil. Trans. Roy. Soc. Lond. 288A,59-95.
lithosphere.J.^eophys. Res. 82, 277-96. Atherton, M.P. (1984) The coastal bathshift of Peru.
Allegre, C.J. et al. (1984) Structure and evolution of the In Harmon, KS. & Barreiro, B.A. (eds) Andean
Himalaya-Tibet orogenic belt. Nature 307,17-22. Magmatism: Chemical and Isotopic Constraints, pp. 152-
Allen, C.R. (1981) The modern San Andreas Fault. In 67. Shiva, Cheshire.
Ernst, G.(ed.) The Geotectonic Development of California., Atwater, T.(1981) Propagating rifts in seafloor spreading
pp. 511-34. Prentice Hall, Englewood Cliffs, NJ. patterns. Nature 290,185-6.
Allen, P.A. & Allen, J.R. (1990) Basin Analysis Principles Auzende,J.-M. et al.(1989)Direct observation ofa section
andApplications.Blackwell Scentific Publications,Oxford. through slow-spreading oceanic crust. Nature 337,
Allen, P.A., Homewood, P. & Williams, G.D. (1986) 726-9.
Foreland basins: an introduction. In Allen, P.A. & Bailey, D.K.(1983)The chemical and thermal evolution of
Homewood, P. (eds) Poreland Basins. Spec. Pub. Inter- rifts. Tectonophysics 94, 585-97.
nat. Assoc. Sedimentologists, 8, 3-12. Blackwell Scien Baker,B.H. & Wohlenberg, J.(1971)Structure and evolu
tific Publications, Oxford. tion of the Kenya Rift Valley. Nature 229,538-42.
Al-Shanti,A.M.S.& Mitchell, A.M.G.(1976)Late Precam Baker,P.E.(1982)Evolution and classification oforogenic
brian subduction and collision in the Al Amar-ldsas volcanic rocks. In Thorpe, R.S. (ed.) Andesites, pp. 11-
region, Arabian Shield, Kingdom of Saudi Arabia. Tec- 23. Wiley, Chichester.
tonophysics 30,T41-7. Ballard, R.D. & van Andel, T.H.(1977) Morphology and
Alvarez, W.(1982) Geological evidence for the geograph tectonics of the inner rift valley at lat 36®50'N on the
ical pattern of mantle return flow and the driving mecha mid-Atlantic ridge. Bull.geol. Soc. Am. 88,507-30.
nism of plate lectomcs.J.pfeophys. Res. S7^ 6697-710. Banks, R.J., Parker, R.L. & Huestis, S.P. (1977) Isostatic
Anderson,D.L.(1982)Hot spots, polar wander, Mesozoic compensation on a continental scale: Local versus re-
convection and the geoid. Nature 297,391-3. . gional mechanisms.Geophys.J. Roy. astr. Soc. 51,431-52.
Anderson,D.L.(1989)Theory ofthe Earth. Blackwell Scien Barazangi, M.& Dorman,J.(1969)World seismicity maps
tific Publications, Oxford. compiled from ESSA, Coast and Geodetic Survey, epi
Anderson, D.L. (1994) Superplumes or supercontinents? center data, 1961-1967. Bull, seism. Soc. Am. 59, 369-
Geology 22, 39-42. 80.
Anderson, D.L., Tanimoto, T. & Zhang, Y. (1992) Plate Barazangi, M. & Isacks, B. (1971) Lateral variations of
tectonics and hotspots: the third dimension. Science 256, seismic wave attenuation in the upper mantle above the
1645-51. inclined earthquake zone of the Tonga island arc: deep
Andrews, J.A. (1985) True polar wander: an analysis of anomaly in the upper mdinxXc. J.geophys. Res. 76,8493-
Cenozoic and Mesozoic palaeomagnetic poles./.geophys. 516.
Res. 90,7737-50. Barazangi, M.& Isacks, B.(1979)Subduction ofthe Nazca
Arculus, R.J. & Curran E.B. (1972) The genesis of the plate beneath Peru: evidence from spatial distribution of
calc-alkaline rock suite. Earth planet. Sci. Lett. 15, 255- earthquakes. Geophys. J. Roy. astr. Soc. 57,537-55.
62. Barber, A.J.(1982)Interpretations ofthe tectonic environ
Argand, E.(1924) La tectonique de PAsie. 13th Intemat. ment ofsouthwest Japan. Proc.geol. Assoc. 93,131-45.
Geol. Cong. Brussels 1924, Comptes Rendus No. 5, pp. Barber, A. (1985) A new concept of mountain building.
171-372. Geology Today 1,116-21.
303
304 REFERENCES
Barberi, F., Bonatti, E., Marinelli, G. & Varet, J. (1974) Bird, P. (1979) Continental delamination and the Colo
Transverse tectonics during the split of a continent; data rado Vbxtzu.J.geophys. Res. 84,7561-71.
from the Afar rift. Tectonophysics 23,17-29. Black, R., Caby, R., Moussine-Pouchkine, A. et al.(1979)
Barker, P.P.(1979) The history of ridge crest offset at the Evidence for late Precambrian plate tectonics in West
Faiklands-Aguihas Fracture Zone from a small circle Africa. Nature 278,223-7.
geophysical profile. Geophys,]. 59, 131-45. Bloxham,J. 8c Gubbins,D.(1985)The secular variation of
Barton, P. & Wood, K (1984) Tectonic evolution of the Earth's magnetic field. Nature 317,777-81.
North Sea Basin: crustal stretching and subsidence. Geo Bonatti, E. (1978) Vertical tectonism in oceanic fracture
phys. J. Roy. astr. Soc. 79,987-1022. zones. Earth planet. Sci. Lett. 37, 369-79.
Bath, M.(1979)Introduction to Seismology. Birkhauser Ver- Bonatti, E. (1987) Rifting or drifting in the Red Sea?
lag, Basel. Nature 330,692-3.
Batiza, R., Melson, W.G. & O'Hearn, T. (1988) Simple Bonatti, E. 8c Crane,K.(1984)Oceanic fracture zones. Sci.
magma supply geometry inferred beneath a segment of Am. 250(5), 36-47.
the Mid-Atlantic Ridge. Nature 335,428-31. Bonatti, E. 8c Honnorez, J.(1976)Sections of the Earth's
Beaumont, C. & Tankard, A.J. (eds)(1987) Sedimentary crust in the equatorial h.i\zni\Q.J.geophys. Res. 81,4104-
basins and basin-forming mechanisms. C.S.P.G. 16.
Mem.l2/A.G.S. Spec. Pub. 5. Can. Soc. Pet. Geols. Bonatti, E., Harrison, C.G.A., Fisher, D.E. et al. (1977)
Bechtel, T.D., Forsyth, D.W., Sharpton, V.L. & Grieve, Easter volcanic chain (southeast Pacific): a mantle hot
RA.F.(1990) Variations in effective elastic thickness of line.J.geophys. Res. 82, 2457-78.
the North American lithosphere. Nature 343,636-8. Bond, G.C., Nickeson, P.A. 8c Kominz, M.A. (1984)
Beck, M.E. Jr (1980) The palaeomagnetic record of plate- Breakup of a supercontinent between 625 Ma and
margin tectonic processes along the >vestern edge of 555 Ma: new evidence and implications for continetal
North America. Res. 85,7115-31. histories. Earth planet. Sci. Lett. 70, 325-45.
Beck, M.E. Jr (1986) Model for late Mesozoic-early Ter Borch, R.S. 8c Green,H.W.11(1987)Dependence ofcreep
tiary tectonics of coastal California and western Mexico in olivine on homologous temperature and its implica
and speculations on the origin of the San Andreas fault. tion for flow in the mantle. Nature 330, 345-8.
Tectonics 5,49-64. Bosworth, W.(1987) Off-axis volcanism in the Gregory
Beck, M.E., Drake, R.E. & Butler, R.F.(1986) Palaeomag- Rift, east Africa: implications for models of continental
netism of Cretaceous volcanic rocks from central Chile rifting. Geology 15, 397-400.
and implications for the tectonics of the Andes. Geology Bott, M.H.P.(1964) Formation of sedimentary basins by
14,132-6. ductile flow of isostatic origin in the upper mantle.
Ben-Avraham, Z. 8c von Herzen, R.P. (1987) Heat flow Nature 201,1082-4.
and continental brczk-up.J.geophys. Res. 92, 1407-16. Bott, M.H.P.(1967)Solution ofthe linear inverse problem
Ben-Avraham, Z., Nur, A., Jones, D. 8c Cox, P. (1981) in magnetic interpretation with application to oceanic
Continental accretion and orogeny. From oceanic pla magnetic anomalies. Geophys. J. Roy astr. Soc. 13, 313-
teaus to allochthonous terranes. Science 213,47-54. 23.
Ben-Avraham,Z.,Nur,A.8c Jones,D.(1982)The emplace Bott, M.H.P. (1971) Evolution of young continental
ment of ophiolites by collision. ]. geophys. Res. 87, margins and formation ofshelf basins. Tectonophysics 11,
3861-7. 319-27.
Berggren, W.A., Kent, D.V., Flynn, J.J. 8c van Couvering, Bott, M.H.P.(1976) Formation of sedimentary basins of
J.A.(1985)Cenozoic geochronology. Bull.geol. Soc. Am. graben type by extension of the continental crust. Tec
96, 1407-18. tonophysics 36, 77-86.
Bickle, M.J.(1978) Heat loss from the Earth: a constraint Bott, M.H.P.(1981) Crustal doming and the mechanism
on Archaean tectonics from the relation between geo- ofcontinental rifting. Tectonophysics 73,1-8.
thermal gradients and the rate of plate production. Earth Bott, M.H.P.(1982a)The Interior ofthe Earthy its Structurey
planet. Sci. Lett. 40,301-15. Constitution and Evolution (2nd edn). Edward Arnold,
Bicknell, J.D.,Sempere,J.-C., Macdonald, K.C. 8c Fox,P.J. London.
(1988) Tectonics of a fast spreading center—a deep-tow Bott, M.H.P.(1982b)The mechanism ofcontinental split
and seabeam survey at EPR 19"'30'S.Marinegeophys. Res. ting. Tectonophysics 81, 301-9.
9, 25-45. Bott, M.H.P.(1983) Deep structure and geodynamics of
Bignell, R.D., Cronan,D.S. 8c Tooms,J.S.(1976) Red Sea the Greenland-Scotland Ridge: an introductory review.
metalliferous brine precipitates. In Strong, D.F. (ed.) In Bott, M.H.P., Saxov, S., Talwani. M. 8c Theide, J.
Metallogeny and Plate Tectonics. Geol. Assoc. Can. Spec. (eds) Structure and Development of the Greenland-
Paper 14, 147-79. Scotland Ridge, pp. 3-9. Plenum Press, New York.
REFERENCES 305
Bott, M.H.P. (1992) The stress regime associated with old rocks./. Geol. 81,406-33.
continental break-up. In Storey, B.C., Alabaster, T. & Burke,K., Kidd, W.S.F. 8c Wilson,J.T.(1973)Relative and
Pankhurst, R.J.(eds)Ma^fmatism and the causes ofConti latitudinal motion of Atlantic hotspots. Nature 245,
nental Break-up. Geol. Soc. Spec. Pub. 68,125-36. 133-7.
Bott, M.H.P.(1993) Modelling the plate driving mecha Burnett, M.S., Caress, D.W. 8c Orcull, J.A.(1989) Tomo-
nism./, Soc. Lond. 150,941-51. graphic image ofthe magma chamber at 12®50'N on the
Bott, M.H.P. & Kusznir, N.J. (1979) Stress distribution East Pacific Rise. Nature 339, 206-8.
associated with compensated plateau uplift structures Butler, R.W.H. (1986) Thrust tectonics, deep structure
with application to the continental splitting mechanism. and crustal subduction in the Alps and Himalayas.
Geophys.J. Roy. astr. Soc. 56,451-9. Soc. Lond. 143,857-73.
Bott, M.H.P. & Kusznir, N.J. (1984) Origins of tectonic Calvert, A.J.(1995)Seismic evidence for a magma chamber
stress in the lithosphere. Tectonophysics 105,1-13. beneath the slow-spreading Mid-Atlantic Ridge. Nature
Bowin, C., Purdy, G.M., Johnston, C. et at. (1980) Arc- 377,410-14.
continent collision in the Banda Sea region. Bull. Am. Calvert, A.J., Sawyer, E.W., Davis, W.J. 8c Ludden, J.N.
Assoc. Petroleum Geols. 64,868-915. (1995) Archaean subduction inferred from seismic im
Brace, W.F. & Kohlstedt, D.L. (1980) Limits on litho- ages of a mantle suture in the Superior Province. Nature
spheric stress imposed by laboratory experiments. /. 375,670-4.
pfeophys. Res. 85,6248-52. Campbell, A.C. et al. (1988) Chemistry of hot springs on
Brett, R.(1976) The current status of speculations on the the Mid-Atlantic Ridge. Nature 335, 514-19.
composition ofthe core ofthe Earth. Rev. Geophys. Space. Cande, S.C. 8c Kent, D.V. (1992) A new geomagnetic
Phys. 14, 375-83. polarity time scale for the late Cretaceous and Cenozoic.
Briden, J.C. (1976) Application of palaeomagnetism to J.geophys. Res. 97,13 917-51.
Proterozoic tectonics. Phil. Trans. Roy. Soc. Lond. 208A, Cann,J.R (1970)New model for the structure ofthe ocean
405-16. crust. Nature 226,928-30.
Briggs, J.C.(1987)Biogeopfraphy and Plate Tectonics, Devel Cann, J.R. (1974) A model for oceanic crustal structure
opments in Palaeontology and Stratigraphy 10. Elsevier, developed. Geophys. J. Roy. astr. Soc. 39, 169-87.
Amsterdam. Cannat, M., Mevel, C., Maia, M.et al. (1995) Thin crust,
Brown, G.C. & Mussett, A.E. (1981) The Inaccessible ultramafic exposures, and rugged faulting patterns at the
Earth. Allen & Unwin, London. Mid-Atlantic Ridge (22'-24'N). Geolop^ 23,49-52.
Brown, L.D., Chapin, C.E., Sanford, A.R., Kaufman,S. 8c Carbotte,S.M. 8c MacDonald,K.C.(1994)The axial topo
Oliver, J. (1980) Deep structure of the Rio Grande Rift graphic high at intermediate and fast spreading ridges.
from seismic reflection profiling. /. geophys. Res. 85, Earth planet. Sci. Lett. 128,85-97.
4773-800. Caress, D.W., Menard, H.W. 8c Hey, RN.(1988) Eocene
Brozena, J.M. 8c White, RS. (1990) Ridge jumps and reorganization of the Pacific-Farallon spreading center
propagations in the South Atlantic Ocean. Nature 348, north ofthe Mendocino fracture zone,.],geophys. Res. 93,
149-52. 2813-38.
Bullard, E.G., Everett, J.E. 8c Smith,A.G.(1965)The fit of Carey,S.W.(1976)The Expanding Earth. Elsevier, Amster
the continents around the Atlantic. Phil. Trans. Roy. Soc. dam.
Lond. 258A,41-51. Carey, S.W. (1988) Theories of the Earth and Universe.
Bullen, K.E. 8c Bolt, B.A. (1985) An Introduction to the Stanford University Press, Stanford, CA.
Theory of Seismology (4th edn). Cambridge University Carlson, RL. 8c Melia, P.J. (1984) Subduction hinge mi
Press, Cambridge. gration. Tectonophysics 102, 399-411.
Burchfiel,B.C.(1980)Tectonics of non-collisional regimes Carter, N.L. 8c Ave'lallemant, H.G.(1970) High tempera
—the modern Andes and the Mesozoic Cordilleran oro- ture flow ofdunite and peridotite. Bull.geol. Soc. Am.81,
gen ofthe western United States. In Burchfiel, B.C.(ed.) 2181-202.
Continental Tectonics, pp.65-72. National Academy of Carter, W.E. 8c Robertson, D.S.(1986)Studying the earth
Sciences, Washington D.C. by very-long-baseline interferometry. Sci. Am. 255(5),
Burke, K.(1976) Development of graben associated with 44-52.
the initial ruptures of the Atlantic Ocean. Tectonophysics Cathles, L.M. (1975) The Viscosity of the Earth's Mantle.
36,93-112. Princeton University Press, Princeton NJ.
Burke, K. (1977) Aulacogens and continental breakup. Cawood, P.A. 8c Suhr, G.(1992) Generation and obduc-
Ann. Rev. Earth planet. Sci. 5, 371-96. tion of ophiolites: constraints from the Bay of Islands
Burke, K. 8c Dewey, J.F. (1973) Plume generated triple complex, western Newfoundland. Tectonics 11,884-97.
junctions: key indicators in applying plate tectonics to Chang, C. 8c Cheng, H.(1973)Some tectonic features of
306 REFERENCES
the Mt. Jolmo Lungma area, southern Tibet. Sci. Sinica during a reversal. Nature 374,687-92.
16, 257-65. Coffin, M.F. 8c Eldholm, O. (1994) Large igneous
Chang, C. et al, (1986) Preliminary conclusions of the provinces: crustal structure, dimensions, and external
Royal Society and Academica Sinica 1985 geotraverse of consequences. Geophys. 32,1-36.
Tibet. Nature 323, 501-7. Cohen, S.C. 8c Smith, P.E. (1985) LAGEOS scientific
Chapman,D.S. 8c Pollack, H.N.(1975)Global heat flow: a results: introduction./.pfeophys. Res. 90,9217-20.
new look. Earth planet. Sci. Lett. 28, 23-32. Collette, B.J.(1979)Thermal contraction joints in spread
Chappie, W.M. 8c Forsyth, D.W.(1979) Earthquakes and ing seafloor as origin of fracture zones. Nature 251,
bending of plates at tvtnchts. J.pfeophys. Res. 84,6729- 299-300.
49. Collier,J. 8cSinha, M.(1992)Seismic mapping ofa magma
Chappie, W.M. 8c Tullis, T.E. (1977) Evaluation of the chamber beneath the Valu Fa Ridge,Lau Bzsm.J.£ieophys.
forces that drive phtcs.J.^feophys. Res. 82,1967-84. Res. 97,14 031-54.
Charlton T.R (1991) Postcollision extension in arc- Condie, K.C.(1982)Plate Tectonics and Crustal Evolution.
continent collision zones, eastern Indonesia. Geolqgy 19, Pergamon Press, Oxford.
28-31. Coney, P.J. (1989) Structural aspects of suspect terranes
Chase, C.G. (1978) Plate kinematics: the Americas, east and accretionary tectonics in western North America. J.
Africa, and the rest of the world. Earth planet. Sci. Lett. struct. Geol. 11,107-25.
37, 355-68. Coney, P.J., Jones, P.L. 8c Monger,J.W.H.(1980) Cordil-
Chase, C.(1979) Asthenospheric counterflow: a kinematic leran suspect terranes. Nature 288,329-33.
model. Geophys.J. Roy. astr. Sac. 56,1-18. Cooper, A.K., Marlow, M.S. 8c Scholl, D.W.(1977) The
Chase, C.G. 8c Gilmer, T.H. (1973) Precambrian plate Bering Sea—a multifarious marginal basin. In Talwani,
tectonics: the Midcontinent Gravity High. Earth planet. M. 8c Pitman, W.C. Ill (eds) Island Arcs, Deep Sea
Sci. Lett. 21, 70-8. Trenches and Back-arc Basins., pp.437-50. Maurice
Choukroune,P., Francheteau,J. 8c Le Pichon,X.(1978)In Ewing Series 1. Am. Geophys. Union.
situ structural observations along Transform Fault A in Cooper, A.K., Marlow, M.S., Scholl, D.W. 8c Stevenson,
the FAMOUS area. Mid-Atlantic Ridge. Bull. £feol. Sac. A.J. (1992) Evidence for Cenozoic crustal extension in
Am. 89,1013-29. the Bering Sea region. Tectonics 11,719-31.
Christensen, N.l. 8c Fountain, D.M.(1975) Constitution Corliss, J.B. etal.(1979)Submarine thermal springs on the
of the lower continental crust based on experimental Galapagos Rift. Science 203,1073-83.
studies of seismic velocities in granulite. Bull, pfeol. Sac. Courtillot, V. (1982) Propagating rifts and continental
Am. 86, 227-36. breakup. Tectonics 1, 239-56.
Christensen, N.I. 8c Salisbury, M.H. (1972) Seafloor Courtillot, V. 8c Besse, J. (1987) Magnetic field reversals,
spreading, progressive alteration of layer 2 basalts, and polar wander and core-mantle coupling. Science 237,
associated changes in seismic velocities. Earth planet. Sci. 1140-7.
Lett. 15, 367-75. Courtillot, V. 8c Vink, G.E.(1983) How continents break
Christodoulidis, D.C., Smith, D.E., Kolenkiewicz, R, up. Sci. Am. 249(1),40-7.
Klosko, S.M., Torrence, M.H. 8c Dunn, P.J. (1985) Courtney,RC.8c White,RS.(1986)Anomalous heat flow
Observing tectonic plate motions and deformations from and geoid across the Cape Verde Rse: evidence for
satellite laser V2in^mg.J.^eophys. Res. 90,9249-63. dynamic support from a thermal plume in the mantle.
Clark, T.A., Gordon, D., Himwich, W.E., Ma, C., Earth planet Sci. Lett. 79, 373-84.
Mallama, A. 8c Ryan, J.W. (1987) Determination of Cowan, J. 8c Cann, J.(1988) Supercritical two-phase sepa
relative site motions in the western United States using ration of hydrothermal fluids in the Troodos ophiolite.
Mark 111 Very Lx)ng Baseline Interferometry. J.£feophys. Nature 333,259-61.
Res. 92, 12 741-50. Cowan, D.S. 8c Silling, R.(1978) A dynamic scaled model
Cloetingh, S., Wortel, R 8c Vlaar, N.J. (1984) Passive of accretion at trenches and its implications for the
margin evolution, initiation of subduction and the Wil tectonic evolutions of subduction complexes. J.£feophys.
son cycle. Tectonophysics 109, 147-63. Res. 83,5389-96.
Cobbing, E.J. (1982) The segmented coastal batholith of Coward, M.P. (1986) Heterogeneous stretching, simple
Peru: its relationship to volcanicity and metallogenesis. shear and basin development. Earth planet. Sci. Lett. 80,
Earth Sci. Rev. 18, 241-51. 325-36.
Cobbing, E.J. 8c Pitcher, W.S.(1972) Plate tectonics and Coward, M.P.(1994a) Continental collision. In Hancock,
the Peruvian Andes. Nature Rhys. Sci. 240, 51-3. P.L. (ed) Continental Deformation., pp. 264-88. Perga
Coe, RS., Prevot, M. 8c Camps, P.(1995) New evidence mon Press, Oxford.
for extraordinarily rapid change of the geomagnetic field Coward, M.P. (1994b) Inversion tectonics. In Hancock,
REFERENCES 307
P.L. (ed) Continental Deformation; pp. 289-304. Perga- tectonics. Geophys.J. Roy. astr. Soc. 49,459-86.
mon Press, Oxford. Davies, G.F.(1980)Review ofoceanic and global heat flow
Cox,A.(1975)The frequency ofgeomagnetic reversals and estimates. Rev. Geophys. Space Phys. 18,718-22.
the symmetry of the nondipole field. Rev. Geophys. Space Davies, G.F. & Richards, M.A.(1992) Mantle convection.
Rhys. 13(3), 35-51. J. Geol. 100,151-206.
Cox, A.(1980) Rotation of microplates in western North Davis, D.,Suppe, J. & Dahlen, F.A.(1983)The mechanics
America.In Strangway,D.W.(ed.)The Continental Crust offold-and-thrust beks.J.geophys. Res. 88,1153-72.
and its Mineral Deposits. Geol. Assoc. Canada. Spec. Davis, P.M.(1984) The median valley, a result of magma
Paper 20, pp. 305-21. fracture beneath mid-ocean ridges. Nature 308,53-4.
Cox,A.& Hart,RB.(1986)Plate Tectonics. How It Works. Debiche, M.G., Cox, A. & Engebretson, D. (1987) The
Blackwell Scientific Publications, Oxford. motion of allochthonous terranes across the North
Cox, A., Doell, R.R & Dalrymple, G.B.(1964) Geomag Pacific Basin. Spec. Pap. 207, Geol. Soc. Am.
netic polarity epochs. Science 143, 351-2. DeLong, S.E., Dewey, J.F. & Fox, P.J. (1977) Displace
Cox,A.v.,Dalrymple,G.B.& Doell,R.R (1967)Reversals ment history of oceanic fracture zones. Geolopfy 5, 199-
ofthe Earth's magnetic field. Sci. Am. 216(2),44-54. 201.
Cox, K.G.(1989) The role of mantle plumes in the devel DeMets, C., Gordon, R.G.,Stein,S.& Argus, D.F.(1987)
opment of continental drainage patterns. Nature 342, A revised estimate of Pacific-North America motion and
873-7. implications for western North America plate boundary
Creer, K.M. (1965) Palaeomagnetic data from the zone tectonics. Geophys. Res. Lett. 14,911-14.
Gondwanic continents. In Blackett, P.M.S., Bullard, E. DeMets, C., Gordon, RG.,Argus, D.F. & Stein, S.(1990)
& Runcorn,S.K.(eds)A Symposium on Continental Drift. Current plate motions. Geophys. J. Int. 101,425-78.
Phil. Trans. Roy. Soc. Lond. 258A,27-40. Detrick, R.S., Buhl,P., Vera, E.etal.(1987)Multi-channel
Cross, T.A. & Pilger, R.H.(1982) Controls ofsubduction seismic imaging of a crustal magma chamber along the
geometry,location of magmatic arcs, and tectonics ofarc East Pacific Rise. Nature 326, 35-41.
and back-arc regions. Bull.^eol. Soc. Am. 93,545-62. Detrick, R.S., White R.S. & Purdy, G.M.(1993) Crustal
Crough, S.T. (1979) Hotspot epeirogeny. Tectonophysics structure of North Atlantic fracture zones. Rev. Geophys.
61,321-33. 31,439-57.
Crough, S.T., Morgan, W.J. & Hargraves, R.B. (1980) Detrick, R., Collins, J., Stephen, R. & Swift, S.(1994). In
Kimberlites: their relation to mantle hotspots. Earth situ evidence for the nature of the seismic layer 2/3
planet. Sci. Lett. 50, 260-74. boundary in oceanic crust. Nature 370, 288-90.
Crowell, J.C.(1962) Displacement along the San Andreas Dewey, J.F. (1969) Evolution of the Appalachian/
Fault, California. Spec. Paper 71, Geol. Soc. Am. Caledonian orogen. Nature 222, 124-9.
Crowell, J.C. (1974a) Sedimentation along the San Dewey,J.F.(1972)Plate tectonics. Sci. Am.226(5),56-68.
Andreas fault, California. In Dott, R.H. Jr & Shaver, Dewey,J.F.(1976) Ophiolite obduction. Tectonophysics 31,
R.H. (eds) Modem and Ancient Geosynchrial Sedimenta 93-120.
tion. Soc. econ. Pal. Mineral. Spec. Pub. 19, 292-303. Dewey, J.F. & Bird, J.M. (1970) Mountain belts and the
Crowell, J.C., (1974b) Origin of late Cenozoic basins in new global itQtomcs.J.pfeophys. Res. 75, 2625-47.
southern California. In Dickinson, W.R (ed.) Tectonics Dewey,J.F. & Burke,K.C.A.(1973)Tibetan, Variscan,and
and Sedimentation. Soc. econ. Pal. Mineral. Spec. Pub. 22, Precambrian basement reactivation: products of conti
190-204. nental collision./. Geol. 81,683-92.
Dalziel, I.W.D. (1991) Pacific margins of Laurentia and Dewey, J.F. & Burke, K.C.A.(1974) Hot spots and conti
East Antarctica-Australia as a conjugate rift pair: evi nental break-up: implications for collisional orogeny.
dence and implications for an Eocambrian superconti- Geology 2, 57-60.
nent. Geology 19, 598-601. Dewey,J.F. & Horsfield, B.(1970)Plate tectonics,orogeny
Dalziel, I.W.D. (1995) Earth before Pangea. Sci. Am. and continental growth. Nature 225, 521-5.
272(1), 38-43. Dewey, J.F., Hempton, M.R,Kidd, W.S.F., Saroglu, F. &
Davidson, A. & Rex, D.C. (1980) Age of volcanism and §eng6r, A.M.C. (1986) Shortening of continental
rifting in south western Ethiopia. Nature 283,657-8. lithosphere: the neotectonics of eastern Anatolia: a
Davidson, J.P. (1983) Lesser Antilles isotopic evidence of young collision zone. In Coward, M.P. & Ries, A.C.
the role of subducted sediment in island arc magma (eds) Collision Tectonics. Geol. Soc. Spec. Pub. 19,3-36.
genesis. Nature 306, 253-5. de Wit, M.J. & Stern, C.R.(1981) Variations in the degree
Davies, D.(1968) A comprehensive test ban. Sci. J. Lond. ofcrustal extension during formation ofa back arc basin.
Nov. 1968, 78-84. Tectonophysics 72, 229-60.
Davies, G.F. (1977) Whole mantle convection and plate Dickinson, W.R.(1971) Plate tectonic models of geosyn-
308 REFERENCES
dines. Earth planet. Sci. Lett. 10,165-74. 39. Am. Geophys. Union, Washington DC.
Dickinson, W.R.(1974)Plate tectonics and sedimentation. Elsasser, W.M.(1969) Convection and stress propagation
In Dickinson, W.R. (ed.) Tectonics and Sedimentation. in the upper mantle. In Runcorn,S.K.(ed.) The Applica
Soc. econ. Pal. Mineral. Spec. Pub. 22, 1-27. tion ofModem Physics to the Earth and Planetary Interiors,
Dickinson, W.R (1977)Tectonostratigraphic evolution of pp. 223-46. Wiley-Interscience, London.
subduction-controlled sedimentary assemblages. In Tal- Elsasser, W.M.(1971) Sea-floor spreading as thermal con
wani, M.& Pitman, W.C. Ill (eds)Island Arcs, Deep Sea vection./. Geophys. Res. 76,1101-12.
Trenches and Back-arc Basins, pp. 33-40. Maurice Ewing Elsasser, W.M.,Olsen,P.& Marsh,B.D.(1979)The depth
Series 1. Am. Geophys. Union. of mantle convection./,geophys. Res. 84,147-55.
Dietz, R.S.(1961) Continental and ocean basin evolution Elthon, D. (1981) Metamorphism in oceanic spreading
by spreading ofthe sea floor. Nature 190,854-7. centres.In Emiliani, C.(ed.) The Oceanic Lithosphere, The
Dietz, R.S. & Holden, C.(1970) The breakup of Pangaea. Sea 7, pp. 285-303. Wiley, New York.
Sci. Am. 223(4), 30-41. Elthon, D.(1991) Geochemical evidence for formation of
Dimroth,E.(1981)Labrador Geosyncline: type example of the Bay of Islands ophiolite above a subduction zone.
early Proterozoic cratonic reactivation. In Kroner, A. Nature 354,140-3.
(ed.) Precambrian Plate Tectonics, pp. 331-52. Elsevier, Engebretson, D.C., Cox, A. & Gordon, R.G.(1985) Rela
Oxford. tive motion between oceanic and continental plates in
Dixon, T.H. (1991) An introduction to the Global Posi the Pacific Basin. Geol. Soc. Am. Spec. Paper 206.
tioning System and some geological applications. Rev. Engeln, J.F., Stein, S., Werner, J. 8c Gordon, R.G.(1988)
Geophys. 29, 249-76. Microplate and shear zone models for oceanic spreading
Dixon,T.H.,Stern,R.J.& Hussein,I.H.(1987)Control of center reorganisations. Res. 93,2839-56.
Red Sea Rift geometry by Precambrian structures. Tecton England, P.C. (1982) Some numerical investigations of
ics 6, SSl-72. large scale continental deformation. In Hsu, K.G. (ed.)
Doglioni, C.(1993) Geological evidence for a global tec Mountain Buildin£f Processes, pp. 129-41. Academic
tonic polarity. Soc. Lond. 150,991-1002. Press, London.
Drake, R. Vergar, M., Munizaga, F. & Vincente, J.C. England, P.C. 8c Houseman, G.A. (1986) Finite strain
(1982) Geochronology of Mesozoic-Cenozoic magma- calculations of continental deformation 2. Comparison
tism in central Chile, lat 31®-36°S. Earth Sci. Rev. 18, with the India-Asia collision zone./. geophys. Res. 91,
353-63. 3664-76.
Dunbar, J.A. & Sawyer, D.S.(1988) Continental rifting at England, P.C. 8c McKenzie, D.P. (1982) A thin viscous
pre-existing lithospheric weaknesses. Nature 333, sheet model for continental deformation. Geophys.J. Roy.
450-2. astr. Soc. 70, 295-321.
Duncan, RA. & Richards, M.A.(1991) Hotspots, mantle England, P. 8c Molnar, P. (1990) Right-lateral shear and
plumes, flood basalts, and true polar wander. Rev. rotation as the explanation for strike-slip faulting in
Geophys. 29, 31-50. eastern Tibet. Nature 344, 140-2.
Dunlop, D.J.(1981) Palaeomagnetic evidence for Protero Ernst, W.G. (1973) Blueschist metamorphism and P-T
zoic continental development. Geophys. J. Roy astr. Soc. regimes in active subduction zones. Tectonophysics 17,
301A,265-77. 255-72.
du Toit, A.L.(1937) Our Wandering Continents. Oliver & Ernst, W.G. (ed.)(1975) Subduction Zone Metamorphism.
Boyd, Edinburgh. Benchmark Papers in Geology, Vol. 19. Dowden,
Dvorkin,J., Nur,A., Mavko,G.& Ben-Avraham,Z.(1993) Hutchinson 8c Ross,Stroudsberg, PA.
Narrow subducting slabs and the origin ofbackarc basins. Evans, A.M. (1987) An Introduction to Ore Geology (2nd
Tectonophysics 227,63-79. edn). Blackwell Scientific Publications, Oxford.
Dziewonski, A.M. (1984) Mapping the lower mantle: Faure, G. (1986) Principles of Isotope Geology (2nd edn).
determination of lateral heterogeneity up to degree and Wiley, Chichester.
order 6.J.pfeophys. Res. 89, 5929-52. Fein, J.B. 8c Jurdy, D.M.(1986) Plate motion controls on
Dziewonski, A.M. & Anderson, D.L. (1984) Seismic to back-arc spreading. Geophys. Res. Lett. 12, 545-8.
mography of the Earth's interior. Am. Sci. 72,483-94. Flower, M.F.J. (1981) Thermal and kinematic control on
Edgar, N.T. (1974) Acoustic stratigraphy in the deep ocean ridge magma fractionation: contrasts between At
oceans. In Bur, C.A. & Drake, C.L.(eds) The GeoUtgy of lantic and Pacific spreading axes./.geol. Soc. Lond. 138,
Continental Margins, pp. 243-6. Springer-Verlag, Ber 695-712.
lin. Force, E.R. (1984) A relation among geomagnetic rever
Elder, J.W.(1965)Physical processes in geothermal areas. sals, sea floor spreading rate, paleoclimate, and black
In Lee, W.M.K.(ed.) Geophys. Mono^r. No. 8, pp. 211- shales. Eos (Trans. A.G.U.) 65,18-19.
REFERENCES 309
Forsyth, D.W.(1975) The early structural evolution and Geol. Surv. Cyprus.
anisotropy of the oceanic upper mantle. Geophys, J. Roy. Gente,P., Pockalny, RA.,Durand,C.etal.(1995)Charac
astr. Sac. 43,103-62. teristics and evolution of the segmentation of the Mid-
Forsyth, D.W. & Uyeda, S.(1975) On the relative impor Atlantic Ridge between 20"N and 24°N during the last
tance of the driving forces of plate motion. Geophys. J. 10 million years. Earth planet. Sci. Lett. 129, 55-71.
Roy. astr. Sac. 43,163-200. Gibb, RA. (1978) Slave-Churchill collision tectonics.
Fowler, C.M.R. (1976) Crustal structure of the mid- Nature 271,50-2.
Atlantic ridge crest at 37°N. Geophys.J. Roy. astr. Soc. 47, Gibb, RA., Thomas, M.D., Kearey, P. & Tanner, J.G.
459-91. (1978) Gravity anomalies at structural province bound
Fowler,C.M.R.(1990)The Solid Earth. Cambridge Univer aries of the Canadian Shield: products of Proterozoic
sity Press, Cambridge. plate interaction. GeoSkrifter 10, 21-57.
Fox,P.J., Schreiber, E., Rowlett, M.& McCamy,K.(1976) Gibb, RA., Thomas, M.D., Lapointe, P. & Mukhopad-
The geology of the Oceanographe Fracture Zone: a hyay, M. (1983) Geophysics of proposed Proterozoic
model for fracture zones.J.geophys. Res. 81,4117-28. sutures in Canada. Precambrian Res. 19, 349-84.
Fox,P.J. & Stroup,J.B.(1981)The plutonic foundation of Glatzmaier, G.A. & Roberts, P.H. (1995) A three-
the oceanic crust. In Emliani, C. (ed.) The Oceanic dimensional self-consistent computer simulation of a
Lithosphercy The Sea 7, pp. 119-218. Wiley, New York. geomagnetic field reversal. Nature 377, 203-9.
Frakes, L.A. (1979) Climates throughout Geologic Time. Glikson, A.Y. (1981) Uniformitarian assumptions, plate
Elsevier, Amsterdam. tectonics and the Precambrian Earth. In Kroner, A.(ed.)
Francheteau, J.(1983)The oceanic crust. Sci. Am. 249(3), Precambrian Plate Tectonics.^ pp.91-104. Elsevier,
68-84. Amsterdam.
Francis, T.J.G. & Porter, l.T. (1973) Median valley Goetze,C.(1978)The mechanism ofcreep in olivine. Phil.
seismology: the mid-Atlantic ridge near 45°N. Geophys.]. Trans. Roy. Soc. Lond. 288A,99-119.
Roy. astr. Soc. 34, 279-311. Gorbatov, A., Suarez, G., Kostoglodov, V. & Gordeev, E.
Frank, F.C. (1968) Curvature of island arcs. Nature 220, (1994)A double-planed seismic zone in Kamchatka from
363. local and teleseismic data. Geophys. Res. Lett. 21,1675-8.
Frankel, H.(1988) From continental drift to plate tecton Gordon,RG.& Stein,S.(1992)Global tectonics and space
ics. Nature 335, 127-30. geodesy. Science 256, 333-42.
Fujita, K. & Kanamori, H.(1981) Double seismic zones Grand, S.P. (1994) Mantle shear structure beneath the
and stresses of intermediate depth earthquakes. Geophys. Americas and surrounding oceans. J. geophys. Res. 99,
J. Roy. astr. Soc. 66,131-56. 11591-621.
Fuller, M.(1987) A mechanism for reversals? Nature 326, Green, D.H. & Ringwood, A.E. (1967) An experimental
132-3. investigation of the gabbro to eclogite transformation
Fyfe, W.S. & Londsdale, P. (1981) Ocean floor hydro- and its petrological applications. Geochim. cosmochim.
thermal activity. In Emiliani, C.(ed.) The Oceanic Litho- Acta, 31, 767-833.
sphere, The Sea 7, pp. 589-638. Wiley, New York. Green, H.W. (1994) Solving the paradox of deep earth
Gaal, G. & Schulz, K.J. (eds)(1992) Precambrian metal- quakes. Sci. Am. 271(3), 50-7.
logeny related to plate tectonics. Precambrian Res. 58, Green, J.C.(1983) Geologic and geochemical evidence for
1-446. the nature and development of the Precambrian Mid-
Gao, S., Davies, P.M., Lin, H. et al. (1994) Asymmetric continent Rift ofNorth America. Tectonophysics 94,413-
upwarp ofthe asthenosphere beneath the Baikal rift zone, 38.
Sihenz.J.geophys. Res. 99, 15 319-30. Green,T.H.(1970)High pressure experimental studies on
Garfunkel, Z. (1975) Growth, shrinking and long term the mineralogical constitution of the lower crust. Phys.
evolution of plates and their tectonic implications for the Earth planet. Interiors 3,441-50.
flow pattern in the mzntXe.J.geophys. Res. 80,4225-32. Grow,J.A.(1973)Crustal and upper mantle structure ofthe
Garland, G.D.(1979)Introduction to Geophysics(2nd edn). central Aleutian arc. Bull.geol. Soc. Am. 84,2169-92.
W.B. Saunders Co., Philadelphia. Gubbins, D.(1987) Mechanism for geomagnetic polarity
Garson, M.S.& Krs, M.(1976)Geophysical and geological reversals. Nature 326,167-9.
evidence ofthe relationships ofRed Sea transverse tecton Gurnis, M.(1988) Large-scale mantle convection and the
ics to ancient fractures. Bull.geol. Soc. Am. 87, 169-81. aggregation and dispersal of supercontinents. Nature
Gass, l.G. (1980) The Troodos massif; its role in the 332,695-9.
unravelling of the ophiolite problem and its significance Hadley,D.& Kanaiiiori, H.(1977)Seismic structure ofthe
in the understanding of constructive plate margin pro Transverse Ranges, California. Bull. geol. Soc. Am. 88,
cesses. In Panayistou, A. (ed.) Ophiolites^ pp. 23-35, 1469-78.
310 REFERENCES
Halbach, P. et al (1989) Probable modern analogue of effect of attenuation on gross Earth models. J. geophys.
Kuroko-type massive sulphide deposits in the Okinawa Res. 82,1647-54.
Trough back-arc basin. Nature 338,496-9. Hasegawa, A., Umino, N. 8c Takagi, A. (1978) Double-
Hallam, A.(1972) Continental drift and the fossil record. planed seismic zone and upper-mantle structure in the
Sci. Am. 227(5), 56-66. northeastern Japan arc. Geophys. J. Roy. astr. Soc. 54,
Hallam, A. (1973a) A Revolution in the Earth Sciences. 281-96.
Clarendon Press, Oxford. Hasemi,A.H.,Ishii, H.8c Takagi, A.(1984)Fine structure
Hallam, A.(1973b) Provinciality, diversity and extinction beneath the Tohoku district, northeastern Japan arc, as
of Mesozoic marine invertebrates in relation to plate derived by the inversion ofP-wave arrival times from local
movements. In Tarling, D.H. & Runcorn, S.C. (eds) earthquakes. Tectonophysics 101,245-65.
Implications of Continental Drift to the Earth Sciences, 1, Haxby, W.F.,Turcotte, D.L. 8c Bird, J.M.(1976)Thermal
pp. 287-94. Academic Press, Lx)ndon. and mechanical evolution of the Michigan Basin. Tec
Hallam, A.(1975) Alfred Wegener and the hypothesis of tonophysics 36, 57-76.
continental drift. Sci. Am. 232(2), 88-97. Hayes, D.E. 8c Ewing, M.(1970) Pacific boundary struc
Hallam, A. (1981) Relative importance of plate move ture. In Maxwell, A.E.(ed.) The Sea. Ideas and Observa
ments, eustasy, and climate in controlling major biogeo- tions on Progress in the Study ofthe Seas 4(2), pp. 29-72.
graphical changes since the early Mesozoic.In Nelson,G. Wiley, New York.
& Rosen, D.E.(eds) Vkariance Biopfeqgraphy. A Critique, Hayward, A.B. 8c Graham,RH.(1989)Some geometrical
pp. 303-40. Colombia University Press, New York. characteristics ofinversion.In Cooper, M.A.8c Williams,
Ham, W.E.(1969) Regional Geolopfy ofthe Arbuckle Moun G.D.(eds)Thrust Tectonics. Spec.Pub. Geol.Soc.London
tains, Oklahoma. Oklahoma Geol. Survey Guidebook. 44,17-39.
Hamburger, M.W. & Isacks, B.L.(1988) Diffuse back-arc Hedenquist, J.W. 8c Lowenstern, J.B. (1994) The role of
deformation in the southwestern Pacific. Nature 322, magmas in the formation of hydrothermal ore deposits.
599-604. Nature 370,519-27.
Harding, A.J., Orcutt, J.A., Kappus, M.E. et al. (1989) Heirtzler, J.R, Le Pichon, X. 8c Baron, J.G.(1966) Mag
Structure of young oceanic crust at 13°N on the East netic anomalies over the Reykjanes Rdge. Deep Sea Res.
Pacific Rise from expanding spread profiles. J. £feophys. 13,427-43.
Res. 94,12 163-96. Heirtzler, J.R, Dickson, G.O., Herron, E.M., Pitman,
Harding,T.P.(1985)Seismic characteristics and identifica W.C. Ill 8c Le Pichon, X. (1968) Marine magnetic
tion of negative flower structures, positive flower struc anomalies, geomagnetic field reversals and motions of
tures and positive structural inversion. Bull. Am. Assoc. the ocean floor and contm^ats.J.geophys. Res. 73,2119-
Petroleum Geols. 69,582-600. 36.
Hargraves, RB. (1981) Precambrian tectonic style: a Heiskanen, W.A. 8c Vening Meinesz, F.A. (1958) The
liberal uniformitarian interpretation. In Kroner, A.(ed.) Earth and its Gravity Field. McGraw-Hill, New York.
Precambrian Plate Tectonics, pp. 21-56. Elsevier, Amster Helwig, J.(1973)Plate tectonic model for the evolution of
dam. the Central Andes: Discussion. Bull. geol. Soc. Am. 84,
Harland, W.B., Cox, A.V., Llewellyn, P.G., Picton, 1493-6.
C.A.G., Smith, A.G. & Walters, R (1982) A Geologic Herring, T.A. et al. (1986) Geodesy by radio inter-
Time Scale. Cambridge University Press, Cambridge. ferometry: evidence for contemporary plate motion. J.
Harland, W.B., Armstrong, RL., Cox, A.V., Craig, L.E., geophys. Res. 91,8341-7.
Smith, A.G.& Smith, D.G.(1990)A Geologic Time Scale Herron,E.M.(1972)Sea-floor spreading and the Cenozoic
1989. Cambridge University Press, Cambridge. History of the East-Central Pacific. Bull. geol. Soc. Am.
Harper, G.D. (1988) Episodic magma chambers and a 83,1671-92.
magmatic extension in the Josephine Ophiolite. Geology Herron, T.J., Ludwig, W., Stoffa, P., Kan, T. 8c Buhl, P.
16,831-4. (1978) Structure of the East Pacific Rise crest from
Harper,J.F.(1978)Asthenosphere flow and plate motions. multichannel seismic reflection data. J.geophys. Res. 83,
Geophys.J. Roy. astr. Soc. 55,87-110. 798-804.
Harrison, C.G.A. 8c Bonatti, E.(1981) The oceanic litho- Herron,T.J., Stoffa, P.L. 8c Buhl, P.(1980) Magma cham
sphere.In Emiliani. C.(ed.) The Oceanic Lithosphere. The ber and mantle reflections—East Pacific Rse. Geophys.
Sea 7, pp. 21-48. Wiley, New York. Res. Lett. 7,989-92.
Harrison, C.G.A. 8c Sclater, J.G. (1972) Origin of the Hess, H.H.(1962) History of ocean basins. In Petrologic
disturbed magnetic zone between the Murray and Molo- Studies: A Volume in Honor ofA.F. Buddington, pp. 599-
kai fracture zones. Earth planet. Sci. Lett. 14,419-27. 620. Geol. Soc. Am.
Hart, RS., Anderson, D.L. 8c Kanamori, M.(1977) The Hey, RN.(1977) A new class of pseudofaults and their
REFERENCES 311
bearing on plate tectonics: a propagating rift model. Holbrook, W.S. & Keleman, P.B. (1993) Large igneous
Earth planet. Sci. Lett. 37, 321-5. province on the US Atlantic margin and implications for
Hey, R.N., Dunnebier, F.K., & Morgan, W.J. (1980) magmatism during continental breakup. Nature 364,
Propagating rifts on mid-ocean ridges.J.pieophys. Res. 85, 433-6.
3647-58. Holmes, A. (1928) Radioactivity and continental drift
Hey, R.N., Kleinrock, M.C., Miller, S.P., Atwater, T.M.& (Report of Glasgow Geological Society meeting 12
Searle, R.C. (1986) Sea Beam/Deep-Tow investigation January 1928) Geol. Mag 65,236-8.
of an active propagating rift system, Galapagos 95.5°W. Hooper, P.R (1990) The timing of crustal extension and
J.£feophys. Res. 91,3369-94. the eruption of continental flood basalts. Nature 345,
Hey, R.N., Menard, H.W., Atwater, T.M. & Caress, D.W. 246-9.
(1988) Changes in direction of seafloor spreading revis Houseman,G.A.(1983)The deep structure ofocean ridges
ited. Res. 93, 2803-11. in a convecting mantle. Earth planet. Sci. Lett. 64, 283-
Hilton,D.R.& Craig, H.(1989)A helium isotope transect 94.
along the Indonesian archipelago. Nature 342,906-8. Houseman, G. & England, P. (1993) Crustal thickening
Hirahara, K. (1981) Three-dimensional seismic structure versus lateral expansion in the India-Asia continental
beneath southwest Japan: the subducting Philippine sea co\\\s\f)n. J.geophys. Res. 98,12 233-50.
plate. Tectonophysics 79,1-44. Howell, b.G.(1989) Tectonics of Suspect Terranes: Moun
Hirn, A., Jiang, M., Sapin, M. et at. (1995) Seismic ani- tain Building and Continental Growth. Chapman & Hall,
sotropy as an indicator of mantle flow beneath the London.
Himalayas and Tibet. Nature 375,571-4. Hsu, V. (1990) Seismicity and tectonics of a continent-
Hirn, A., Lepine, J.C., Jobert, G. et at. (1984a) Crustal island arc collision zone at the island of Taiwan. J.
structure and variability of the Himalayan border of geophys. Res. 95,4725-34.
Tibet. Nature 307,23-5. Hsui, A.T. & Toksoz, M.N. (1981) Back-arc spreading:
Hirn, A., Nercessian, A., Sapin, M. et at. (1984b) Lhasa trench migration, continental pull or induced convec
block and bordering sutures—a continuation ofa 500-km tion? Tectonophysics 74,89-98.
Moho traverse through Tibet. Nature 307, 25-7. Hurley,P.M.(1968)The confirmation ofcontinental drift.
Hoefs, J. (1980) Stable Isotope Geochemistry. Springer- Sci. Am. 218(4), 52-64.
Verlag, Berlin. Hutchinson, C.S. (1983) Economic Deposits and their Tec
Hoffman,K.A.(1979)Behaviour ofthe geodynamo during tonic Setting. MacMillan Press, London.
reversal: a phenomenological model. Earth planet. Sci. lilies, J.H.(1977)Ancientand recent rifting in the Rhinegra-
Lett. 44,7-17. ben. Geologie Mijnb. 56, 329-50.
Hoffman, P. (1973) Evolution of an early Proterozoic Irving, E.(1979)Pole positions and continental drift since
continental margin: the Coronation Geosyncline and the Devonian. In McElhinny, M.M.(ed.) The Earth: its
associated aulacogens of the NW Canadian Shield. Phil. Origin^ Structure and Evolution., pp. 567-93. Academic
Trans. Roy. Soc. Lond. 273A,547-81. Press, London.
Hoffman, P.P.(1980) Wopmay orogen: a Wilson-cycle of Irving, E., North, P. & Couillard, R (1974) Oil, climate
early Proterozoic age in the northwest of the Canadian and tectonics. Can.J. Earth Sci. 11,1-17.
Shield. Geol. Assoc. Can. Spec. Paperj 20, 523-49. Isacks, B.L. & Barazangi, M.(1977) Geometry of Benioff
Hoffman, P.P. (1987) Continental transform tectonics. zones: lateral segmentation and downward bending of
Great Slave'Lake shear zone (ca 1.9 Ga), northwest the subducted lithosphere. In Talwani, M. & Pitman,
Canada. Geolo^ 15,785-8. W.C. Ill (eds) Island ArcSy Deep Sea Trenches and Back-
Hoffman, P.P. (1989) Speculations on Laurentia's first arc Basins^ pp.99-114. Am. Geophys. Union, Washing
gigayear (2.0 to 1.0 Ga). Geology 17,135-8. ton D.C.
Hoffman, P.P.(1991) Did the breakout of Laurentia turn Isacks, B. 8c Molnar, P.(1969) Mantle earthquake mecha
Gondwanaland inside out.> Science 252, 1409-12. nisms and the sinking of the lithosphere. Nature 223,
Hoffman, P., Burke, K.C.A. & Dewey, J.P.(1974) Aulaco 1121-4.
gens and their genetic relation to geosynclines, with a Isacks, B. 8c Molnar, P.(1971) Distribution of stresses in
Proterozoic example from Great Slave Lake, Canada. In the descending lithosphere from a global survey of
Dott, R.H. Jr & Shaver, R.H.(eds)Modem and Ancient focal mechanism solutions of mantle earthquakes. Rev.
Geosyncline Sedimentation. Soc. econ. Pal. Mineral. Spec. Geophys. Space Phys. 9, 103-74.
Pub. 19, 38-55. Isacks, B., Oliver, J. 8c Sykes, L.R.(1968) Seismology and
Hogrefe, A., Rubie, D.C.,Sharp, T.G. & Seifert, P.(1994) the new global tectonics.J.geophys. Res. 73,5855-99.
Metastability of estatite in deep subducting lithosphere. Isacks, B., Sykes, L.R. 8c Oliver, J. (1969) Pocal mecha
Nature 372, 351-3. nisms of deep and shallow earthquakes in the Tonga-
312 REFERENCES
Kermadec region and tectonics of island arcs. Bull. £ieol. Kanasewich, E.R (1976) Plate tectonics and planetary
Soc. Am. 80,1443-69. convection. Can.J. Earth Sci. 13, 331-40.
Ito, E. & Sato, H.(1991) Aseismicity in the lower mantle Kanimori, H.(1971) Great earthquakes at island arcs and
by superplasticity of the descending slab. Nature 351, the lithosphere. Tectonophysics 12,187-98.
140-1. Karig, D.E. (1970) Ridges and basins of the Tonga-
Jackson, E.D.(1976) Linear volcanic chains of the Pacific Kermadec island arc system.].pfeophys. Res. 75,239-54.
plate. In Sutton, G.H., Manghnari, M.H.& Moberly, R. Karig, D.E. (1971) Origin and development of marginal
(eds) The Geophysics of the Pacific Ocean Basin and its basins in the western Pacific. J.geophys. Res. 76, 2542-
Marpfin. Am. Geophys. Union. Geophys. Mono^. 19, 319- 61.
35. Karig, D.E.(1974) Evolution of arc systems in the western
Jackson, E.D., Silver, E.A. & Dalrymple, G.B. (1972) Pacific. Rev. Earth planet. Sci. 2, 51-75.
Hawaiian-Emperor chain and its relation to Cenozoic Karig, D.E. (1982) Deformation in the forearc: implica
circumpacific tectonics. Bull.^feol. Soc. Am. 83,601-18. tions for mountain belts. In Hsii, K.J. (ed.) Mountain
Jacob, K.H., Nakamura, K.& Davies, J.N.(1977)Trench- Buildinpf Processesy pp. 59-71. Academic Press, London.
volcano gap along the Alaska-Aleutian arc: facts and Karig, D.E. & Kay, R.W.(1981) Fate of sediments on the
speculations on the role of terrigenous sediments for descending plate at convergent margins. Phil. Trans. Roy.
subduction. In Talwani, M. & Pitman, W.C. Ill (eds) Soc. Lond. 301A,233-51.
Island ArcSy Deep Sea Trenches and Back-arc Basins. Karig, D.E. & Sharman, G.F. (1975) Subduction and
pp. 243-58. Maurice Ewing. Series I. Am. Geophys. accretion in trenches. Bull.pfeol. Soc. Am. 86,377-89.
Union. Karl, D.M., McMurty, G.M., MalahofF, A. & Garcia, M.O.
Jacobs, J.A. (1975) The Earth's Core. Academic Press, (1988) Loihi Seamount, Hawaii: a mid-plate volcano
London. with a distinctive hydrothermal system. Nature 335,
James, D.E.(1971) Plate tectonic model for the evolution 532-5.
of the central Andes. Bull.pieol. Soc. Am. 82, 3325-46. Karson, J.A. et al. (1987) Along-axis variations in seafloor
James, D.E. & Snoke, J.A. (1990) Seismic evidence for spreading in the MARK area. Nature 328,681-5.
continuity of the deep slab beneath central and eastern Kaula, W.M.(1975) Absolute plate motions by boundary
Vcm.J.geophys. Res. 95,4789-5001. velocity minimizations. Res. 80, 244-8.
Jarvis, G.T. & McKenzie, D.P. (1980) Sedimentary basin Kawakatsu, H.(1986) Downdip tensional earthquakes be
formation with finite extension rates. Earth planet. Sci. neath the Tonga arc: a double seismic zone? J. geophys.
Lett. 48,42-52. Res. 91,6432-40.
Jeanloz, R. (1993) The mantle in sharper focus. Nature Kay, R.W. & Kay, S.M. (1981) The nature of the lower
365, 110-11. continental crust: inferences from geophysics, surface
Johnston, C.R. & Bowin, C.O. (1981) Crustal reactions geology and crustal xenoliths. Rev. Geophys. Space Phys.
resulting from the mid-Pliocene to Recent continent- 20,735-60.
island arc collision in the Timor region. B.M.R.J. Aust. Kay, R, Hubbard, N.J. & Gast, P.W. (1970) Chemical
Geol. Geophys. 6, 223-43. characteristics and origin ofoceanic ridge volcanic rocks.
Jones, D.G., Silberling, N.J. & Hillhouse, J. (1977) J.pfeophys. Res. 75,1585-613.
Wrangellia—a displaced terrane in northwestern North Kearey, P.(1974) Gravity and seismic reflection investiga
America. Can.J. Earth Sci. 14, 2565-77. tions into the crustal structure ofthe Aves Ridge,eastern
Jones, D.L., Howell, D.G., Coney,P.J. & Monger,J.W.H. Caribbean. Geophys.]. Roy. astr. Soc. 38,435-48.
(1983) Recognition, character and analysis of tectono- Kearey,P.(1976) A regional structural model ofthe Labra
stratigraphic terranes in western North America. In dor Trough, northern Quebec, from gravity studies and
Hashimoto, M. & Uyeda, S. (eds) Accretion Tectonics in its relevance to continental collision in the Precambrian.
the circum-Pacific RepfionSy pp. 21-35. Terra Scientific Earth planet. Sci. Lett. 28, 371-8.
Pub. Co., Tokyo. Kearey,P.& Brooks, M.(1991) Introduction to Geophys
Jordan, T.E., Isacks, B.L., Allmendinger, R.W., Brewer, ical Exploration (2nd edn). Blackwell Scientific Publica
J.A.,Ramos,V.A.& Ando,C.J.(1983)Andean tectonics tions, Oxford.
related to geometry ofsubducted Nazca plate. Bull.pfeol. Kearey, P., Peter, G. & Westbrook, G.K.(1975) Geophysi
Soc. Am. 94, 341-61. cal maps ofthe eastern Caribbean./,pfeol. Soc. Lond. 131,
Jurdy, D.M. & Stefanick, M. (1983) Flow models for 311-21.
back-arc spreading. Tectonophysics 99., 191-206. Keen, C. & Tramontini, C. (1970) A seismic refraction
Kamp, P.J.J. (1980) Pacifica and New Zealand: proposed survey on the mid-Atlantic ridge. Geophys. J. Roy. astr.
eastern elements in Gondwanaland's history. Nature Soc. 20,473-91.
288,659-64. Kellogg, J.N. & Bonini, W.E. (1982) Subduction of the
REFERENCES 313
Caribbean plate and basement uplifts in the overriding Larson, R.L. (1995) The mid-Cretaceous superplume epi
South American plate. Tectonics 1, 251-76. sode. Sci. Am. Feb., 66-70.
Kennett, B.L.N. (1977) Towards a more detailed seismic Larson, R.L. & Olson, P. (1991) Mantle plumes control
picture of the oceanic crust and mantle. Marine Geophys. magnetic reversal frequency. Earth planet. Sci. Lett. 107,
Res, 3,7-42. 437-47.
Kennett, B.L.N., & Orcutt, J.A. (1976) A comparison of Larson,R.L.&Pitman, W.C.Ill(1972) World-wide corre
travel time inversions for marine refraction profiles. J. lation of Mesozoic magnetic anomalies, and its implica
geophys. Res. 81,4061-70. tions. Bull.geol. Soc. Am. 83,3645-61.
Kent, D.V. & Gradstein, F.M. (1985) A Cretaceous and Leblanc,M.(1976)Proterozoic oceanic crust at Bou Azzer.
Jurassic geochronology. Soc.Am. 96,1419-27. Nature 261, 34-5.
Kent, G.M., Harding, A.J. & Orcutt, J.A.(1990)Evidence Leggett,J.K.(1987)The Southern Uplands as an accretion-
for a smaller magma chamber beneath the East Pacific ary prism: the importance ofanalogues in reconstructing
Rise at 9°30'N. Nature 344,650-3. palaeogeography./.^^o/. Soc. Lond. 144, 737-52.
Kidd, R-G.W.(1977)A model for the process offormation Leggett, J.K., McKerrow, W.S. & Casey, D.M.(1982) The
of the upper oceanic crust. Geophys.]. Roy. astr. Soc. 50, anatomy of a Lower Palaeozoic accretionary forearc: the
149-83. Southern Uplands of Scotland. In Leggett, J.K. (ed.)
Kingma, J.T. (1958) Possible origin of piercement struc Trench-Forearc Geology, Spec. Pub. Geol. Soc. Lond. 10,
tures, local unconformities, and sedimentary basins in 495-520.
the Eastern Geosyncline, New Zealand. N.Z.J. Geol. Lemiszki, P.J. & Brown, L.D. (1988) Variable crustal
Geophys. 1, 269-74. structure of strike-slip fault zones as observed on deep
Klein, E.M., Langmuir, C.H.,Zindler, A., Standigel, H.& seismic reflection profiles. Bull.geol. Soc. Am. 100,665-
Hamelin, B. (1988). Isotope evidence of a mantle con 76.
vection boundary at the Australian-Antarctic Discor Le Pichon, X.(1968) Sea-floor spreading and continental
dance. Nature 333,623-9. drift. J.geophys. Res. 73,3661-97.
Klein, G. de V. & Hsui, A.T. (1987) Origin of cratonic Le Pichon, X., Francheteau, J. & Bonnin, J. (1973) Plate
basins. Geology 15,1094-8. Tectonics. Elsevier, Amsterdam.
Kleinrock, M.C. & Hey, R.N. (1989) Detailed tectonics Le Pichon,X.,liyama,T., Boulegue,J. etal.(1987)Nankai
near the tip of the Galapagos 95.5°W propagator: how Trough and Zenisu Ridge: a deep-sea submersible survey.
the lithosphere tears and a spreading axis develops. ]. Earth planet. Sci. Lett. 83,285-99.
geophys. Res. 94,13 801-38. Lesquer, A.,de Almeida,F.F.M., Davino, A., Lachaud,J.C.
Knittle, E. & Jeanloz, R. (1991) Earth's core-mantle & Maillard,P.(1981)Signification structurale des anom
boundary: results of experiments at high pressures and alies gravimetriques de la partie sud du craton de Sao
temperatures. Science 251, 1438-43. Francisco (Bresil). Tectonophysics 76, 273-94.
Krogh, E.J. (1977) Evidence of Precambrian continent- Lister, C.R.B.(1975) Gravitational drive on oceanic plates
continent collision in western Norway. Nature 267, caused by thermal contraction. Nature 257,663-5.
17-19. Lister, C.R.B.(1980) Heat flow and hydrothermal circula
Kroner, A.(1981) Precambrian plate tectonics. In Kroner, tion. Ann. Rev. Earth planet. Sci. 8,95-117.
A.(ed.)Precambrian Plate Tectonics., pp. 57-90. Elsevier, Lister, G.S., Etheridge, M.A. & Simons, P.A. (1986) De
Amsterdam. tachment faulting and the evolution of passive continen
Kurten,B.(1969)Continental drift and evolution. Sci. Am. tal margins. Geology 14, 246-50.
220(3), 54-65. Livermore, R., Vine, F.J. & Smith, A.G. (1984) Plate
Kusznir, N.J. & Bott, M.H.P.(1976) A thermal study of motions and the geomagnetic field—II. Jurassic to Ter
the formation of oceanic crust. Geophys.]. Roy. astr. Soc. tiary. Geophys. J. Roy. astr. Soc. 79,939-61.
47,83-95. Logatchev, N.A. & Florensov, N.G. (1978) The Baikal
Kusznir N.J. & Park, R.G. (1984) Intraplate lithosphere system of rift valleys. Tectonophysics 45, 1-13.
deformation and the strength ofthe lithosphere. Geophys. Long,R.E. 8c Backhouse,R.W.(1976)The structure ofthe
J. Roy. astr. Soc. 79, 513-38. western flank of the Gregory rift. Part 11. The mantle.
Langmuir, C.H., Bender, J.B. & Batiza, R. (1986) Petro- Geophys.]. Roy. astr. Soc. 44,677-88.
logical and tectonic segmentation ofthe East Pacific Rise, Loper, D.E. (1985) A simple model of whole mantle
5°30'-14°30'N. Nature 322,422-9. convection. Res. 90,1809-36.
Larson, R.L.(1991a) Latest pulse of Earth: evidence for a Lowrie, A., Smoot, C. 8c Bartiza, R. (1986) Are oceanic
mid-Cretaceous superplume. Geology 19, 547-50. fracture zones locked and strong or weak? New evidence
Larson, R.L. (1991b) Geological consequences of super- for volcanic activity and weakness. Geology 14, 242-5.
plumes. Geology 19,963-6. Lyon-Caen, H. 8c Molnar, P. (1983) Constraints on the
314 REFERENCES
structure of the Himalaya from an analysis of gravity McKenzie, D.P. & Parker, R.L.(1967) The North Pacific:
anomalies and a flexural model of the lithosphere. J. an example oftectonics on a sphere. Nature 216,1276-
^eophys. Res. 88,8171-91. 80.
Lyzenga,G.A., Wallace, K.S., Fanselow,J.L., Raefsky, A.& McKenzie,D.P.& Sclater, J.G.(1971)The evolution ofthe
Groth, P.M.(1986) Tectonic motions in California in Indian Ocean since the late Cretaceous. Geophys. J. Roy.
ferred from very long baseline interferometry observa astr. Soc. 24,437-528.
tions, 1980-1984.7.^^(^/;yj. Res. 91,9473-87. McKenzie, D.P. & Weiss, N.(1975) Speculations on the
MacDonald, K.C.(1982) Mid-ocean ridges: fine scale tec thermal and tectonic history ofthe earth. Geophys.J. Roy.
tonic, volcanic and hydrothermal processes within a plate astr. Soc. 42,131-74.
boundary zone. Rep. Earth planet. Sci. 10,155-90. McKenzie, D.P., Watts, A., Parsons, B. & Roufosse, M.
MacDonald, K.C. & Fox. P.J. (1983) Overlapping spread (1980) Planform of mantle convection beneath the
ing centres: new accretion geometry on the East Pacific Pacific Ocean. Nature 288,442-6.
Rise. Nature 302, 55-8. McKerrow, W.S. & Cocks, L.R.M.(1986) Oceans, island
MacDonald, K.C., Sempere, J.-C., Fox, P.J. & Tyce, R. arcs and olistostromes: the use offossils in distinguishing
(1987) Tectonic evolution of ridge-axis discontinuities sutures, terranes and environments around the lapetus
by the meeting, linking, or self-decapitation of neigh Occzn.J.geol. Soc. Lond. 143,185-91.
bouring ridge segments. Geolq^y 15,993-7. McKerrow, W.S., Leggett, J.K. & Eales, M.H. (1977)
MacDonald, K.C., Fox, P.J., Perram, P.J. et at. (1988) A Imbricate thrust model of the Southern Uplands of
new view of the mid-ocean ridge from the behaviour of Scotland. Nature 267, 237-9.
ridge-axis discontinuities. Nature 335,217-25. Maekawa,H.,Shozul, M.,Ishii, T., Fryer, P.& Pearce, J.A.
MacKenzie, A.S. & McKenzie, D.P.(1983) Isomerization (1993) Blueschist metamorphism in an active subduc-
and aromatization ofhydrocarbons in sedimentary basins tion zone. Nature 364,520-3.
formed by extension. Geol. Mag. 120,417-70. Martin, H.(1981) The late Palaeozoic Gondwana glacia-
McCarthy, J., Mutter, J.C., Morton, J.L., Sleep, N.H. & tion. Geol. Rund. 70,480-98.
Thompson, G.A. (1988) Relic magma chamber struc Mascle, J. & Blarez, E. (1987) Evidence for transform
tures preserved within the Mesozoic North Atlantic margin evolution from the Ivory Coast-Ghana continen
crust.> Bull.geol. Sac. Am. 100,1423-36. tal margin. Nature 326,378-81.
McDougall,1.&Duncan,R.A.(1980)Linear volcanic chain Mason, R.(1985) Ophiolites. Geology Today 1,136-40.
recording plate motions? Tectonophysics 63,275-95. Mason, R.G. & Raff, A.D.(1961) Magnetic survey off the
McElhinny, M.W.(1973)Palaeomagnetism and Plate Tec west coast of North America, 32®N latitude to 42"N
tonics. Cambridge University Press, Cambridge. latitude. Bull.geol. Soc. Am. 72,1259-66.
McElhinny, M.W. & McWilliams, M.D. (1977) Precam- Mattauer, M.(1986) Intracontinental subduction, crust-
brian geodynamics—a palaeomagnetic view. Tectonophys mantle decollement and crustal-stacking wedge in the
ics 40, 137-59. Himalayas and other collision belts. In Coward, M.P. &
McElhinny, M.W., Taylor, S.R. & Stevenson, D.J.(1978) Ries, A.C.(eds) Collision Tectonics, pp. 37-50. Geol. Soc.
Limits to the expansion of Earth, Moon, Mars and Spec. Pub. 19.
Mercury and to changes in the gravitational constant. Matthews, P.E. (1972) Possible Precambrian obduction
271, 316-21. and plate tectonics in southeastern Africa. Nature Phys.
McFadden,P.L. & Merrill, R.T.(1984)Lower mantle con Sci. 240, 37-9.
vection and geomagnetism. Res. 89,3354-62. Maxwell, A.E., von Herzen, KP., Hsu, K.J. et al. (1970)
McKenzie, D.P.(1969) Speculation on the consequences Deep sea drilling in the South Atlantic. Science 168,
and causes of plate motions. Geophys.J. Roy. astr. Soc. 18, 1047-59.
1-32. Megard, F. & Philip, H.(1976) Plio-Quaternary tectono-
McKenzie,D.P.(1978)Some remarks on the development magmatic zonation and plate tectonics in the central
ofsedimentary basins. Earth planet. Sci. Lett. 40,25-32. Andes. Earth planet. Sci. Lett. 33, 231-8.
McKenzie, D.P. (1983) The Earth's mantle. Sci. Am. Menard, H.W. (1964) Marine Geology of the Pacific.
249(3), 50-67. McGraw-Hill, New York.
McKenzie, D.P.(1990)Spinning continents. Nature 344, Menard, H.W.& Atwater, T.(1968) Changes in direction
109-10. ofsea floor spreading. Nature 219,463-7.
McKenzie, D.P. & Bickle, M.J. (1988) The volume and Menard, H.W. & Atwater, T. (1969) Origin of fracture-
composition of melt generated by extension ofthe litho zone topography. Nature 222,1037-40.
sphere./. Petrol. 29,625-79. Milanovsky,E.E.(1981)Aulacogens on ancient platforms:
McKenzie,D.P.& Morgan,W.J.(1969)Evolution oftriple problems of their origin and tectonic development. Tec
junctions. Nature 224,125-33. tonophysics 73,213-48.
REFERENCES 315
Minster, J.B. & Jordan, T.H. (1978) Present-day plate lites. Rev. Geophys. Space Phys. 20,737-60.
motions,J.£feophys. Res. 83,5331-54. Moores, E.M. (1991) South west U.S.-East Antarctic
Mitchell, A.H.G. & Garson, M.S.(1976) Mineralization at (SWEAT)connection: a hypothesis. Geolopfy 19,325-8.
plate boundaries. Sci. Efi0ineerin£f 8,129-69. Morel, P. & Irving, E.(1978) Tentative paleocontinental
Mitchell, A.H.G. & Garson, M.S.(1981)Mineral Deposits maps for the early Phanerozoic and Proterozoic./. Geol.
and Global Tectonic Settinpfs. Academic Press, London. 86,535-61.
Mitchell, A.H.G. & Reading, H.G. (1969) Continental Morgan,J.P.& Chen,Y.J.(1993)Dependence ofridge-axis
margins,geosynclines,and ocean floor spreading./. Geol. morphology on magma supply and spreading rate.
77,629-46. Nature 364, 706-8.
Mitchell, A.H.G. & Reading, H.G.(1986) Sedimentation Morgan, J.P. & Shearer, P.M.(1993) Seismic constraints
and tectonics. In Reading, H.G.(ed.) Sedimentary Envi on mantle flow and topography of the 660-km
ronments and Fades,, pp.471-519. Blackwell Scientific discontinuity: evidence for whole-mantle convection.
Publications, Oxford. Nature 36S, 506-11.
Miyashiro, A.(1972) Metamorphism and related magma- Morgan, W.J. (1968) Rises, trenches, great faults and
tism in plate tectonics. Am.J. Sci. 272,629-56. crustal blocks,J.£feophys. Res. 73,1959-82.
Miyashiro, A. (1973) Paired and unpaired metamorphic Morgan, W.J. (1971) Convection plumes in the lower
belts. Tectonophysics 17,241-54. mantle. Nature 230,42-3.
Miyashiro, A.(1986) Hot regions and the origin of mar Morgan, W.J. (1972a) Plate motions and deep mantle
ginal basins in the western Pacific. Tectonophysics 122, convection. Geol. Soc. Am. Mem. 132,7-22.
195-216. Morgan, W.J. (1972b) Deep mantle convection plumes
Molnar,P.(1984)Structure and tectonics ofthe Himalaya. and plate motions. Bull. Am. Assoc. Petroleum Geols. 56,
Ann, Rev. Earth planet. Sci. 12,489-518. 203-13.
Molnar, P. (1988) Continental tectonics in the aftermath Morgan, W.J. (1981) Hot spot tracks and the opening of
of plate tectonics. Nature 335, 131-7. the Atlantic and Indian Oceans.In Emiliani, C.(ed.) The
Molnar, P. & Atwater, T.(1973) Relative motion of hot Oceanic Lithosphere. The Sea 7, pp. 443-87. Wiley, New
spots in the mantle. Nature 246, 248-91. York.
Molnar,P.& Chen,W.-P.(1982)Seismicity and mountain Morgan,W.J.(1983)Hotspot tracks and the early rifting of
building.In Hsu,K.G.(ed.)Mountain Building Processes,, the Atlantic. Tectonophysics 94,, 123-39.
pp.41-57. Academic Press, London. Mount,V.S.& Suppe,J.(1987)State ofstress near the San
Molnar, P.& Stock,J.(1987) Relative motions of hotspots Andreas Fault: implications for wrench tectonics. Geoloigy
in the Pacific. Atlantic and Indian Oceans since late 15,1143-6.
Cretaceous time. Nature 327, 587-91. Mueller,S. & Landisman, M.(1966)Seismic studies ofthe
Molnar, P. & Tapponier, P.(1975) Cenozoic tectonics of Earth's crust in continents 1: evidence for a low velocity
Asia: effects of a continental collision. Science 189,419- zone in the upper part ofthe lithosphere. Geophys.J. Roy.
26. astr, Soc. 10, 525-38.
Molnar,P.,Freedman, D.& Shih,J.S.F.(1979)Lengths of Mutter,J.C., Barth, G.A., Buhl, P., Derrick, RS.,Orcutt,J.
intermediate and deep seismic zones and temperatures in & Harding, A.(1988a)Magma distribution across ridge-
downgoing slabs oflithosphere. Geophys.J. Roy. astr. Soc. axis discontinuities on the East Pacific Rise from multi
56,41-54. channel seismic images. Nature 336,156-8.
Montadert L. et al. (1979) Initial Reports of the Deep Sea Mutter, J.C., Buck, W.R & Zehnder, C.M.(1988b) Con-
Drilling Project Vol. 48. US Government Printing Office, vective partial melting. 1. A model for the formation of
Washington DC. thick basaltic sequences during the initiation of spread
Montagner, J.-P. (1994) Can seismology tell us anything ing. Res. 93,1031-48.
about convection in the mantle.> Rev. Geophys. 32, 115- Myers,J.S.(1975) Vertical crustal movements ofthe Andes
37. in Peru. Nature 254,672-4.
Moody, J.D. (1973) Petroleum exploration aspects of Nafe, J.E. & Drake, C.L. (1963) Physical properties of
wrench-fault tectonics.Aw.Assoc. Pet. Geol. 57,449-76. marine sediments. In Hill, M.N.(ed.) The Earth beneath
Moore, J.C.(1989) Tectonics and hydrogeology of accre- the Sea. The Sea 3, pp. 794-815. Interscience Publishers,
tionary prisms: role of decollement zone.J. struct. Geol. New York & London.
11,95-106. Nakajima, T., Maruyama, S., Uchiumi, S. et al, (1990)
Moore, J.C. & Silver, E.A. (1987) Continental mafmn Evidence for late Proterozoic subduction from 700-Myr-
tectonics: submarine accretionary prisms. Rev. Geophys. old blueschists in China. Nature 346, 263-5.
25,1305-12. Nataf, H.C., Nakanishi, 1. & Anderson, D.L. (1984) An-
Moores, E.M.(1982) Origin and emplacement of ophio- isotropy and shear-velocity heterogeneities in the upper
316 REFERENCES
mantle. Geophys. Bxs. Lett. 11,109-12. Opdyke, N.D., Glass, B., Hays, J.D. & Foster, J. (1966)
Nataf, H.C.& VanDecar, J.(1993)Seismoiogial detection Palaeomagnetic study of Antarctic deep sea cores. Science
of a mantle plume? Nature 364,115-20. 154,349-57.
Neumann, E.R. & Ramberg, LB. (1978) Paleorifts— Opdyke, N.D., Burckle, L.H. & Todd, A. (1974) The
concluding remarks.In Neumann,E.R.& Ramberg, LB. extension of the magnetic time scale in sediments of the
(eds) Tectonics and Geophysics of Continental Rifts, central Pacific Ocean. Earth planet. Sci. Lett. 22,300-6.
pp.409-24. D. Reidel, Dordrecht. Orcutt, J.A., Kennett, B.L.N. & Dorman, L.M. (1976)
Nicolas, A.(1989) Structure of Ophiolites and Dynamics of Structure of the East Pacific Rise from an ocean bottom
Oceanic Lithosphere. Kluwer Academic Publishers, seismometer survey. Geophys. J. Roy. astr. Soc. 45, 305-
Dordrecht. 20.
Nicolas, A., Boudier, F. & Ildefonse, B.(1994) Evidence Orowan, E. (1965) Convection in a non-Newtonian
from the Oman ophiolite for active mantle upwelling mantle, continental drift, and mountain building. PhU.
beneath a fast-spreading ridge. Nature 370,51-3. Trans. Roy. Soc. Lond. 258A,284-313.
Nicholson, C., Seeber, L., Williams, P. & Sykes, L.R. Oxburgh, E.R. (1972) Flake tectonics and continental
(1986a) Seismic evidence for conjugate slip and block collision. Nature 239, 202-4.
rotation vv^ithin the San Andreas fault system, southern Oxburgh,E.R.& Turcotte, D.L.(1974) Membrane tecton
California. Tectonics 5,629-48. ics and the East African Rift. Earth planet. Sci. Lett. 22,
Nicholson, C.L., Seeber, L., Williams, P. & Sykes, L.R. 133-40.
(1986b) Seismicity and fault kinematics through the Packham, G.H. & Falvey, D.A.(1971) An hypothesis for
Eastern Transverse Ranges, southern California: block the formation of marginal seas in the Western Pacific.
rotation, strike-slip faulting and low-angle thrusts. J. Tectonophysics 11, 79-110.
geophys. Res. 81,4891-908. Pakiser, L.C. (1963) Structure of the crust and upper
Niell, A.E., Ong, K.M., MacDoran, P.P. et at. (1979) mantle in the western United States.].geophys. Res. 68,
Comparison of a radiointerferometric differential base 5747-56.
line measurement with conventional geodesy. Tectono- Park, RG. (1983) Foundations of Structural Geology.
physics 52,49-58. Blackie, London and Glasgow.
Nisbet, E.G. & Fowler, C.M.R.(1978) The mid-Atlantic Park, RG.(1988) Geological Structures and Moving Plates.
Ridge at 37 and 45'N:some geophysical and petrological Blackie, London and Glasgow.
constraints. Geophys. J. Roy. astr. Soc. 54,631-60. Parker,E.N.(1969)The occasional reversal ofthe geomag
Nunns, A.G. (1983) Plate tectonic evolution of the netic field. Astrophys. J. 158,815-27.
Greenland-Scotland Ridge and surrounding areas. In Parmentier, E.M.,Turcotte, D.L. & Torrance, K.E.(1976)
Bott, M.H.P.,Sarno, V.S., Talwani, M.& Thiele, J.(eds) Studies of finite amplitude non-Newtonian thermal con
Structure and Development of the Greenland-Scotland vection with application to the Earth's mantle./.geophys.
Rid£fe, pp. 11-30. NATO Conference Series IV 8. Res. 81,1839-46.
Plenum Press, London. Parsons, B. & McKenzie, D.P.(1978) Mantle convection
Nur, A. & Ben-Avraham, Z.(1982) Oceanic plateaus, the and the thermal structure ofthe plates./,geophys. Res. 83,
fragmentation of continents and mountain building. J. 4485-96.
geophys. Res. 87,3644-61. Parsons, B. & Sclater, J.G. (1977) An analysis of the
O'Connell, R.J. & Budiansky, B.(1977) Viscoelastic pro variation of ocean floor bathymetry and heat flow with
perties of fluid-saturated cracked solids. J. pfeophys. Res. 2Lge.J.geophys. Res. 82,803-27.
82,5719-35. Pavoni, M.(1993) Pattern of mantle convection and Pan-
Olivei, J. (1982) Tracing surface features to great depths: gaea break-up, as revealed by the evolution ofthe African
a powerful means for exploring the deep crust. Tectono- p\2Xc.J.geol. Soc. Lond. 150,953-64.
physics 81,2S7-72. Pearce, J.A. (1980) Geochemical evidence for the genesis
Oliver, J. & Isacks, B. (1967) Deep earthquake zones, and eruptive setting oflavas from Tethyan ophiolites. In
anomalous structures in the upper mantle, and the litho Panayiotou, A.(ed.) Ophiolites, pp. 261-72. Geol. Surv.
sphere./. Res. 72,4259-75. Cyprus.
Olson,P. & Nam,I.S.(1986)Formation of seafloor swells Peltier, W.R (1985) Mantle convection and viscoelasticity.
by mantle iphimts.J.geophys. Res. 91,7181-91. Ann. Rev. Fluid Mech. 17,561-608.
O'Nions, R.K.(1987)Relationships between chemical and Peltier, W.R & Andrews, J.T. (1976) Glacial-isostatic
convective layering in the Earth./. Soc. Lmd. 144, adjustment—1. The forward problem. Geophys. J. Roy.
259-74. astr. Soc. 46,605-46.
Open University Course Team (1989) The Ocean Basins: Perry, F.V., Baldridge, W.S. & De Paolo, D.J. (1988)
their Structure and Evolution. Pergamon Press, Oxford. Chemical and isotopic evidence for lithospheric thinning
REFERENCES 317
beneath the Rio Grande Rift. Nature 332,432-4. seismic structure of young oceanic crust. J.geophys. Res.
Phillips, R.J. & Ivins, E.R. (1979) Geophysical observa 92,9351-62.
tions pertaining to solid state convection in the terrestrial Puzyrev, N.N., Mandelbaum, M.M.,Krylov, S.V., Mishen-
planets. Phys. Earth planet. Interiors 19, 107-48. kin, B.P., Petrik, G.V. & Krupskaya, G.V.(1978) Deep
Phillips, W.E.A., Stillman, C.J. & Murphy. T. (1976) A structure of the Baikal and other continental rift zones
Caledonian plate tectonic model. Sac. Land. 132, from seismic data. Tectonophysics 45, 15-22.
579-609. Quennell, A.M. (1958) The structural and geomorphic
Pigram, C.J. & Davies, H.L. (1987) Terranes and the evolution of the Dead Sea Rift. J. geol. Soc. Lond. 14,
accretion history of the New Guinea orogen. B.M.R.J. 1-19.
Austral.£feol. & Geophys. 10,193-211. Quennell, A.M.(1985)The Western Arabia rift system.In
Pilger, R.H. Jr (1982) The origin of hotspot traces: evi Dixon, J.E. & Robertson, A.H.F. (eds) The Geological
dence from eastern Australia./. Res. 87, 1825- Evolution of the Eastern Mediterranean. Spec. Pub. geol.
34. Soc. Lond. 17, 775-88.
Piper, J.D.A. (1982) The Precambrian palaeomagnetic Raff, A.D. & Mason, R.G.(1961) Magnetic survey off the
record: the case for a Proterozoic supercontinent. Earth west coast of North America, 40°N latitude to 52°N
planet. Sci. Lett. 59,61-89. latitude. Bull.geol. Soc. Am. 72,1267-70.
Piper, J.D.A.(1987)Palaeomapfnetism and the Continental Ramos, V.A. (1988) The tectonics of the central Andes;
Crust. Open University Press, Milton Keynes. 30° to 33°S latitude. In Clark, S.P., Burchfield, B.C.
Pitcher, W.S.(1978) The anatomy of a batholith. J. ^eol. & Suppe, J. (eds) Processes in Continental Lithospheric
Soc. Lond. 135,157-82. Deformation. Geol. Soc. Am. Spec. Paper 218, 31-54.
Pitcher, W.S.,Atherton, M.P.,Cobbing,E.J. & Beckinsale, Ranalli, G.(1987) Rheology of the Earth. Allen & Unwin.
KD.(1985). Ma£jmatism at a Plate Edpfe: the Peruvian London.
Andes. Blackie, London and Glasgow. Ranalli, G. & Murphy, D.C.(1987) Rheological stratifica
Pitman, W.C.Ill & Heirtzler, J.R.(1966)Magnetic anom tion ofthe lithosphere. Tectonophysics 132, 281-96.
alies over the Pacific-Antarctic ridge. Science 154,1164- Reading, H.G. (1980) Characteristics and recognition of
71. strike-slip fault systems. In Ballance, P.P. & Reading,
Pitman,W.C.Ill & Talwani, M.(1972)Sea-floor spreading H.G. (eds) Sedimentation in Oblique Slip Mobile Zones.
in the North Atlantic. Bull.pfeol. Soc. Am. 83,619-46. Spec. Pub. int. Ass. Sediment. 4,7-26.
Plank,T.& Langmuir,C.H.(1993)Tracing trace elements Reck,B.H.(1987)Implications of measured thermal gradi
from sediment input to volcanic output at subduction ents for water movement through the northeast Japan
zones. Nature 362, 739-43. accretionary prism.J.geophys. Res. 92,3683-90.
Platt, J.P. (1986) Dynamics of orogenic wedges and the Reeves, C.(1976) The delineation of crustal provinces in
uplift of high-pressure metamorphic rocks. Bull.pfeol. Soc. southern Africa from a compilation of gravity data. 20th
Am. 97,1037-53. Ann.Rep. Research Institute ofAfrican Geology., University
Plumstead,E.P.(1973)The enigmatic Glossopteris flora and of Leeds.
uniformitarianism. In Tarling, D.H. & Runcorn, S.K. Renkin, M.L. & Sclater, J.G.(1988) Depth and age in the
(eds) Implications of Continental Drift to the Earth Sci north Vzcific.J.geophys. Res. 93,2919-35.
ences., 1, pp.413-24. Academic Press, London. Richter, P.M. (1973) Convection and large-scale circu
Pollack, H.N. & Chapman, D.S.(1977a) The flow of heat lation of the mdmtic.J.geophys. Res. 78,8735-45.
from the Earth's interior. Sci. Am. 237(2),60-76. Richter, P.M. & Parsons, B.(1975) On the interaction of
Pollack, H.N. & Chapman, D.S.(1977b) On the regional two scales ofconvection in the mznxX^.J.geophys. Res. 80,
variation of heat flow,geotherms, and lithospheric thick 2529-41.
ness. Tectonophysics 38,279-96. Ringwood, A.E. (1974) The petrological evolution of is
Pollack, H.N., Hunter, S.J. & Johnson, J.R. (1993) Heat land arc systems./.geol. Soc. Lond. 130, 183-204.
flow from the Earth's interior: analysis of the global data Ringwood, A.E. (1975) Composition and Petrology of the
set. Rev. Geophys. 31,267-80. Earth's Mantle. McGraw-Hill, New York.
Powell, C.McA. & Conaghan, P.J. (1973) Plate tectonics Ringwood, A.E.(1977)Petrogenesis in island arc systems.
and the Himalayas. Earth planet. Sci. Lett. 20,1-12. In Talwani, M. & Pitman, W.C. Ill (eds) Island Arcs,
Powell,C.McA.& Conaghan,P.J.(1975)Tectonic models Deep Sea Trenches and Back-arc Basins, pp. 311-24. Mau
ofthe Tibetan Plateau. Geology 20,727-31. rice Ewing Series 1. Am. Geophys. Union.
Powell, C.McA., Johnson, B.D. & Veevers, J.J. (1980) A Ringwood,A.E.&Irifune,T.(1988)Nature ofthe 650 km
revised fit ofeast and west Gondwanaland. Tectonophysics seismic discontinuity: implications for mantle dynamics
63,13-29. and differentiation. Nature 331, 131-6.
Purdy, G.M. (1987) New observations of the shallow Ringwood,A.E.& Major,A.(1970)The system Mg2Si04-
318 REFERENCES
Fe2Si04 at high pressures and temperatures. Phys. Earth sediments from deformation zones in the Barbados accre-
planet. Interiors 3,89-108. tionary prism.In Moore,J.C.(ed.) Synthesis ofStructural
Ritger, S., Canson, B. & Snegge, E. (1987) Methane- Fabrics in Deep Sea Drillinpf Project Coresfrom Forearcs.
derived authigenic carbonates formed by subduction- Mem.^eol. Soc. Am. 166,105-16.
induced pore-water expulsion along the Oregon/ Schubert, G., Yuen, D.A. & Turcotte, D.L.(1975) Role of
Washington margin. Bull.pfeol. Soc. Am. 98,147-56. phase transitions in a dynamic mantle. Geophys. J. Roy.
Roecker,S.W.(1982)The velocity structure of the Pamir- astr. Soc. 42,705-35.
Hindu Kush region: possible evidence of subducted Sclater, J.G. & Francheteau, J.(1970) The implications of
cmst.J.£feophys. Res. 87,945-59. terrestrial heat flow observations on current tectonic and
Rom, J. (1994) A new forerunner for continental drift. geochemical models ofthe crust and upper mantle ofthe
Nature 367,407-8. Earth. Geophys. J. Roy. astr. Soc. 20,509-42.
Rona, P.A. (1977) Plate tectonics, energy and mineral Sclater, J.G.,Jaupart,C.& Galson,D.(1980)The heat flow
resources: basic research leading to payoff. Eos {Trans. through oceanic and continental crust and the heat loss
AG.C/.)58,629-39. of the Earth. Rep. Geophys. Space Phys. 18,269-311.
Rona,P.A.(1984)Hydrothermal mineralization at seafloor Scotese, C.R., Gahagan, L.M.& Larson, R.L.(1988) Plate
spreading centres. Earth Sci. Rev. 20,1-104. tectonic reconstructions ofthe Cretaceous and Cenozoic
Runcorn, S.K. (1962) Palaeomagnetic evidence for conti ocean basins. Tectonophysics 155, 27-48.
nental drift and its geophysical cause. In Runcorn, S.K. Scott, M.R., Scott, R.B., Nalwalk, A.J., Rona, P.A. &
(ed.) Continental Drift., pp. 1-40. Academic Press, New Butler, L.W. (1973) Hydrothermal manganese in the
York & London. median valley of the mid-Atlantic Ridge. Eos (Trans.
Rupke, N.A.(1970) Continental drift before 1900. Nature A.G.U.) 54,244.
227, 349-50. Scrutton, C.T.(1967)Absolute time data from palaeontol
Sabadini, R.& Yuen,D.A.(1989)Mantle stratification and ogy.In Runcorn,S.K.etal.(eds)International Dictionary
long-term polar wander. Nature 339, 373-5. of Geophysics I., p. 1. Pergamon Press, Oxford.
Sabadini, R.,Yuen,D.A.& Boschi,E.(1982)Polar wander Scrutton, R-A.(1979)On sheared passive continental mar
ing and the forced responses ofa rotating, multi-layered, gins. Tectonophysics 59,293-305.
viscoelastic p\2Sia..J.£feophys. Res. 81,2885-903. Searle, D.L. & Panayiotou, A.(1980) Structural implica
Sager, W.W.& Bleil, U.(1987) Latitudinal shift of Pacific tions in the evolution ofthe Troodos massif, Cyprus.In
hotspots during the late Cretaceous and early Tertiary. Panayiotou, A. (ed.) Ophiolites^ pp. 50-60. Geol. Surv.
Nature 326,488-90. Cyprus.
Salisbury, M.H. & Christensen, N.I. (1978) The seismic Searle, R.C. (1970) Evidence from gravity anomalies for
velocity structure ofa traverse through the Bay ofIslands thinning of the lithosphere beneath the rift valley in
ophiolite complex, Newfoundland, an exposure of oce Kenya. Geophys. J. Roy. astr. Soc. 21,13-31.
anic crust and upper m2int\t.J.£ieophys. Res. 83,805-17. Searle, R.C. (1983) Multiple, closely spaced transform
Sandwell, D.T. & Renkin, M.L.(1988) Compensation of faults in fast-slipping fracture zones. Geolqgy 11,607-10.
swells and plateaus in the north Pacific: no direct evidence Searle, R.C., Rusby, R.L,Engein, J. et al. (1989) Compre
for mantle con\cction.J.£ieophys. Res. 93,2775-83. hensive sonar imagery of the Easter microplate. Nature
Sawkins, F.J.(1976) Widespread continental rifting: some 341,701-5.
considerations of timing and mechanism. Geok(gy 4, Seely, D.R,(1979) The evolution of structural highs bor
427-30. dering major forearc basins.In Watkins,J.S., Montadert,
Sawkins, F.J. (1984) Metal Deposits in Relation to Plate L. & Dickerson, P.W. (eds) Geological and Geophysical
Tectonics. Springer-Verlag, Berlin. Investigations of Continental Margins., pp. 245-60.
Schaer, J.-P. & Rodgers, J. (eds)(1987) The Anatomy of A.A.P.G. Mem. 29y Am. Assoc. Petroleum Geols. Tulsa,
Mountain Ranpfes. Princeton University Press, Princeton, Oklahoma.
NJ. Sempere, J.-C., Purdy, G.M. & Schouten, H.(1990) Seg
Schilling, J.-G., Anderson, R-N. & Vogt. P. (1976) Rare mentation ofthe Mid-Atlantic Ridge between 24®N and
earth, Fe and Ti variations along the Galapagos spreading 30AO'N. Nature 344,427-9.
centre and their relationship to the Galapagos mantle §eng6r, A.M.C.(1976) Collision of irregular continental
plume. Nature 261,108-13. margins: implications for foreland deformation of
Scholz, C.H., Barazangi, M. & Sbar, M.L. (1971) Late Alpine-type orogens. Geolo£fy 4,779-82.
Cenozoic evolution of the Great Basin, western United Shackleton, R.M., Dewey, J.F. & Windley, B.F. (1988)
States, as an ensialic inter-arc basin. Bull, pfeol. Soc. Am. Tectonic evolution of the Himalayas and Tibet. Royal
82, 2979-90. Society, London.
Schoonmaker,J.(1986) Clay mineralogy and diagenesis of Shanti, M. & Roobol, M.J. (1979) A late Proterozoic
REFERENCES 319
ophiolite complex at Jabal Ess in northern Saudi Arabia. origin of porphyry copper-gold deposits in island arcs.
Nature 279,488-91. Geology 18,630-3.
Shearer, P.M. (1990) Seismic imaging of upper-mantle Solomon, S.C., Sleep, N.H. 8c Richardson, R.M.(1975)
structure with new evidence for a 520-km discontinuity. On the forces driving plate tectonics: inferences from
Nature 344,121-6. absolute plate velocities and intraplate stress. Geophys. J.
Shearer, P.-M.& Masters, G.M.(1992)Global mapping of Roy. astr. Soc. 42,769-801.
topography on the 660-km discontinuity Nature 355, Spada, G., Ricard, Y. 8c Sabadini, R.(1992) Excitation of
791-6. true polar wander by subduction. Nature 360,452-4.
Shermer, E.R., Howell, D.G. & Jones, D.L. (1984) The Spence, W.(1987) Slab pull and the seismotectonics of
origin of allocthonous terranes: Perspectives on the subducting lithosphere. Rev. Geophys. 25,55-69.
growth and shaping of continents. Ann. Rev. Earth Spohn, T. 8c Schubert, G.(1982) Convective thinning of
planet. Sci. 12,107-31. the lithosphere: a mechanism for the initiation of conti
Sillitoe, R-H.(1972a)Formation ofcertain massive sulphur nental rifting.J.geophys. Res. 87,4669-81.
deposits at sites of sea-floor spreading. Trans. Inst. Min. Spudich, P. 8c Orcutt, J.(1980) A new look at the seismic
Metall. 81,B141-8. velocity structure ofthe oceanic crust. Rev. Geophys. Space
Sillitoe, R.H.(1972b)A plate tectonic model for the origin Phys. 18,627-45.
of porphyry copper deposits. Econ. Geol. 67,184-97. Stauder, W. (1968) Mechanism of the Rat Island earth
Silver, E.A. & Reed, D.L. (1988) Backthrusting in accre- quake sequence of February, 1965, with reference to
tionary wedges. Res. 93,3116-26. island arcs and sea-floor spreading. J. geophys. Res. 73,
Silver, E.A. & Smith, R. (1983) Comparison of terrane 3847-58.
accretion in modern Southeast Asia and the Mesozoic Stauffer, M.R., Mukherjee, A.G. 8c Koo, J. (1975) The
North American Cordillera., Geology 11,198-202. Amisk Group: an Aphebian (.>) island arc deposit. Can.J.
Silver, E.A., Reed, D., McCaffrey, R. 8c Yoyodiwiryo, Y. Earth Sci. 12, 2021-35.
(1983) Back-arc thrusting in the eastern Sunda arc, Steckler, M.S.8c Watts,A.B.(1980)Subsidence ofa young
Indonesia; a consequence of arc-continental collision./. continental margin. Nature 287,425-9.
geophys. Res. 88,7429-48. Stein, M. 8c Hofmann, A.W.(1994) Mantle plumes and
Simpson,R,W.8c Cox,A.(1977)Palaeomagnetic evidence episodic crustal growth. Nature 372,63-8.
for tectonic rotation ofthe Oregon Coast Range. Geology Stein, R.S.(1987) Contemporary plate motion and crustal
5, 585-9. deformation. Rev. Geophys. 25,855-63.
Sinton, J.M. 8c Detrick, R.S. (1992) Mid-ocean ridge Stewart, J.A. (1990) Drifting Continents and Colliding
magma ch^sahtxs.J.geophys. Res. 97,197-216. Paradigms: Perspectives on the Geoscience Revolution.
Sleep, N.H. (1971) Thermal effects of the formation of Indiana University Press, Bloomington.
Atlantic continental margins by continental break-up. Stolper, E. 8c Newman,S.(1994) The role of water in the
Geophys.J. Roy astr. Soc. 24,325-50. petrogenesis of Mariana Trough magmas. Earth planet.
Sleep, N.H. (1975) Formation of oceanic crust: some Sci. Lett. 121, 293-325.
thermal constrzints.J.get^hys. Res. 80,4037-42. Storey,B.C.(1995)The role of mantle plumes in continen
Sleep, N.H. 8c Snell, N.S.(1976)Thermal contraction and tal breakup: case histories from Gondwanaland. Nature
flexure of mid-continent and Atlantic marginal basins. 377,301-8.
Geophys.J. Roy. astr. Soc. 45,125-54. Storey, B.C.(1993)The changing face oflate Precambrian
Smith, A.G. 8c Hallam, A.(1970) The fit of the southern and early Palaeozoic reconstructions. J. geol. Soc. Lond.
continents. Nature 225,139-44. 150,665-8.
Smith, A.G., Hurley,A.M.8c Briden, J.C.(1981)Phanero- Stothers, R.B. (1986) Periodicity of the Earth's magnetic
zoic Paleocontinental World Maps. Cambridge University reversals. Nature 322,444-6.
Press, Cambridge. Stothers, R.B. (1993) Hotspots and sunspots: surface
Smith, D.E., Christodoulidis, D.C., Kolenkiewicz, K, tracers of deep mantle convection in the Earth and Sun.
Dunn,P.J., Torrence, M.H. 8c Klosko,S.(1985) Global Earth planet. Sci. Lett. 116,1-8.
plate motion results from satellite laser ranging (ab Stow, D.A.V. 8c Lovell, J.P.B. (1979) Contourites: their
stract). Eos (Trans. A.G.U.)66,848. recognition in modern and ancient sediments. Earth Sci.
Smith, D.K. 8c Cann, J.R.(1993) Building the crust at the Rev. 14,251-91.
Mid-Atlantic Ridge. Nature 365,707-15. Snares, G., Molnar,P. 8c Burchfiel, B.C.(1983)Seismicity,
Smithson, S.B. 8c Brown, S.K. (1977) A model for the fault plane solutions, depth offaulting,and active tecton
lower continental crust. Earth planet. Sci. Lett. 35,134- ics of the Andes of Peru, Ecuador, and southern Colom
44. bia./. Res. 88,10 403-28.
Solomon, M. (1990) Subduction, arc reversal, and the Sun, S.S., Nesbitt, R.W. 8c Sharashin, A.Y. (1979)
320 REFERENCES
Geochemical characteristics of mid-ocean ridge basalts. throughout geologic time. PhU. Trans. Roy. Soc. Lond.
Earth planet. Sci. Lett. 44,119-38. 301A, 217-31.
Sutton,J.& Watson,J.V.(1974)Tectonic evolution ofcon Tarney, J., Dalziel, 1. & de Wit, M.(1976) Marginal basin
tinents in early Proterozoic times. Nature 247,433-5. *Rocas Verdes' complex from S. Chile: a model for
Sykes, L.R. (1966) The seismicity and deep structure of Archaean greenstone belt formation. In Windley. B.F.
island zrcs.J.^reophys. Res. 71, 2981-3006. (ed.) The Early History ofthe Earth., pp. 131-46. Wiley,
Sykes, L.R.(1967) Mechanism of earthquakes and nature London.
offaulting on the mid-oceanic ridges./.£feophys. Res. 72, Taylor, B., Crook, K. & Sinton, J. (1994) Extensional
2131-53. transform zones and oblique spreading centers./.geophys.
Sykes, L.R. (1978) Intra-plate seismicity, reactivation of Res. 99,19 707-18.
pre-existing zones of weakness, alkaline magmatism and Taylor, B., Goodliffe, A., Martinez, F. & Hey, R.(1995)
other tectonism post-dating continental fragmentation. Continental rifting and initial sea-floor spreading in the
Rev. Geophys. Space Rhys. 16,621-88. Woodlark Basin. Nature 374,534-7.
Tackley, P.J., Stevenson D.J., Glatzmaier, G.A. & Shubert, Taylor, F.B.(1910) Bearing ofthe Tertiary mountain belt
G.(1993) Effects of an endothermic phase transition at on the origin of the Earth's plan. Bull.geol. Soc. Am. 21,
670 km depth in a spherical model of convection in the 179-226.
Earth's mantle. Nature 361,699-704. Taylor, R.T. & Scott, M.M.(1985) The Continental Crust:
Takagi, H.(1986) Implications of mylonitic microstruc- its Composition and Evolution. Blackwell Scientific Publica
tures for the geotectonic evolution of the Median Tec tions, Oxford.
tonic Line, central Japan./, struct. Geol. 8,3-14. Thatcher, W.(1979)Systematic inversion of geodetic data
Talwani, M. & Watts, A.B.(1974) Gravity anomalies sea in central J.geophys. Res. 84,2283-95.
ward ofdeep-sea trenches and their tectonic implications. Thomas, M.D. & Kearey, P. (1980) Gravity anomalies,
Geophys. J. Roy. astr. Sac. 36, 57-90. block faulting and Andean-type tectonism in the eastern
Talwani, M., Le Pichon, X. & Ewing, M.(1965) Crustal Grenville Province. Nature 293,61-3.
structure of the mid-ocean ridges 2. Computed model Thompson, G. & Melson, W.G.(1972) The petrology of
from gravity and seismic reduction data./.£feophys. Res. oceanic crust across fracture zones in the Atlantic Ocean:
70, 341-52. evidence ofa new kind ofsea-floor spreading./. Geol. 80,
Tamaki, K. (1985) Two modes of back-arc spreading. 526-38.
Geolqgy 13,475-8. Thorpe, R.S.(ed.){\9%2)Andesites. Wiley, London.
Tanimoto, T. & Anderson, D.L.(1985) Lateral heteroge Thurber,C.H.& Aid,K.(1987)Three-dimensional seismic
neity and azimuthal anisotropy ofthe upper mantle: Love imaging. Rev. Earth planet. Sci. 15,115-39.
and Rayleigh waves \00-2S0s.J.£feophys. Res. 90,1842- Tissot, B.(1979)Effects on prolific petroleum source rocks
58. and major coal deposits caused by sea-level changes.
Tapley, B.D., Schutz, B.F. & Eanes, R.J. (1985) Station Nature 277,463-5.
coordinates, baselines,and Earth rotation from LAGEOS Toksoz,M.N.,Minear,J.W.& Julian,B.R.(1971)Temper
laser ranging: 1976-1984. Res. 90,9235-48. ature field and geophysical effects ofa downgoing slab./.
Tapponier, P. & Molnar, P. (1976) Slip-line field theory geophys. Res. 76,1113-38.
and large-scale continental tectonics. Nature 264, 319- Toomey, D.R., Purdy, G.M., Solomon, S.C. & Wilcock,
24. W.S.D. (1990) The three-dimensional seismic velocity
Tapponier, P., Peltzer, G., Le Dain, A.Y., Armijo, R. & structure of the East Pacific Rise near latitude 9®30'N.
Cobbing, P. (1982) Propagating extrusion tectonics in Nature 347,639-45.
Asia: new insights from simple experiments with plas- Tozer, D.C.(1972)The present thermal state ofthe terres
ticene. Geolo£fy 10,611-16. trial planets. Phys. Earth planet. Interiors 6,182-97.
Tapponier,P., Lacassin, R.,Leloup,P.H.etal.(1990)The Turcotte, D.L. & Oxburgh, E.R (1978) Intra-plate volca-
Ailao Shan/Red River metamorphic belt: Tertiary left- nism. PhU. Trans. Roy. Soc. Lon. 288A,561-79.
lateral shear between Indochina and South China. Nature Turcotte,D.L.,McAdoo,D.C.& Caldwell,J.G.(1978)An
343,431-7. elastic-perfectly plastic analysis of the bending of the
Tarling, D.H.(ed)(1981)Economic Geolqgy and Geotecton- lithosphere at a trench. Tectonophysics 47,193-205.
ics. Blackwell Scientific Publications. Oxford. Uyeda,S.& Kanamori,H.(1979)Back-arc opening and the
Tarling, D.H.(1983)Palaeomagnetism. Chapman & Hall, mode ofsubduction./.^^(^^yj. Res. 84,1049-61.
London. Vacquier, V. (1965) Transcurrent faulting in the ocean
Tarling,D.H.& Tarling, M.P.(1971)Continental Drift:A floor. Phil. Trans. Roy. Soc. Lond. 258A,77-81.
Study ofthe Earth^s Moving Surface. Bell, London. Valentine, J.W. & Moores, E.M.(1972) Global tectonics
Tarney, J. & Windley, B.F. (1981) Marginal basins and the fossil record./. Geol. 80,167-84.
REFERENCES 321
Valentine,J.W.& Moores,E,M.(1974)Plate tectonics and sphere and continental margin basins. Tectonophysics 36,
the history of the oceans. Sci. Am. 230(4), 80-9. 25-44.
Van der Hilst, R. (1995) Complex morphology of sub Watts, A.B. & Talwani, M.(1974) Gravity anomalies sea
ducted lithosphere in the mantle beneath the Tonga ward of deep-sea trenches and their tectonic significance.
Trench. Nature 374,154-7. Geophys.J. Roy. astr. Soc. 36, 57-102.
Van der Voo, K (1990)Phanerozoic paleomagnetic poles Watts, A.B., Cochran, J.R. & Selzer, G. (1975) Gravity
from Europe and North America and comparisons with anomalies and flexure of the lithosphere: a three-
continental reconstructions. Bsv. Geophys. 28,167-206. dimensional study ofthe Great Meteor seamount, north
Van Orman, J., Cochran, J.R., Weissel, J.K. & Jestin, F. east Atlantic. Res. 80,1391-8.
(1995) Distribution of shortening between the Indian Watts, A.B., Bodine, J.H. & Steckler, M.S.(1980) Obser
and Australian plates in the central Indian Ocean. Earth vations of flexure and the state of stress in the oceanic
planet. Sci. Lett. 133, 35-46. lithosphere. Res. 85,6369-76.
Vening Meinesz,F.A.(1951)A third arc in many island arc Weertman, J. (1978) Creep laws for the mantle of the
areas. Koninkl. Nederlandsch. Akad. Wetensch. Proc. Ser. Earth. PhU. Trans. Roy. Soc. Lond. 288A,9-26.
B. 54,432-42. Wegener, A. (1915) Die Entstehunpf der Kontinente und
Vermeersen, L.L.A. & Vlaar, N.J. (1993) Changes in the Ozeane. Vieweg, Braunschweig.
Earth's rotation by tectonic movements. Geophys. Res. Weissel,J.K.(1981)Magnetic lineations in marginal basins
Lett. 20,81-4. ofthe western Pacific. Phil. Trans. Roy. Soc. Lond. 300A,
Vine, F.J. (1966) Spreading of the ocean floor: new evi 223-47.
dence. Science 154,1405-15. Weissel, J.K., Anderson,R.N.& Geller, C.A.(1980)Defor
Vine,F.J.(1977)The continental drift debate. Nature 266, mation of the Indo-Australian Plate. Nature 287, 284-
19-22. 91.
Vine, F.J. & Matthews, D.H.(1963) Magnetic anomalies Weldon,R.& Humphries,E.(1986)A kinematic model of
over oceanic ridges. Nature 199,947-9. southern California. Tectonics 5,33-48.
Vink,G.E.(1982)Continental rifting and the implications Wellman,P.(1978)Gravity evidence for abrupt changes in
for plate tectonic reconstructions. J. geophys. Res. 87, mean crustal density at the junction of Australian crustal
10677-88. blocks. B.M.R.J. Aust. Geol. Geophys. 3,153-62.
Vink, G.E., Morgan, W.J. & Zhao, W.-L.(1984)Preferen Wellman, P. (1988) Development of the Australian Pro-
tial rifting ofcontinents: a source ofdisplaced terrains./. terozoic crust as inferred from gravity and magnetic
geophys. Res. 89,10 072-6. anomalies. Precambrian Research 40/41,89-100.
Vogt, P.R. (1975) Changes in geomagnetic reversal fre Wernicke, B.(1981) Low angle normal faults in the Basin
quency at times oftectonic change: evidence for coupling and Range Province-nappe tectonics in an extending
between core and upper mantle processes. Earth planet. orogen. Nature 291,645-8.
Sci. Lett. 25, 313-21. Wernicke, B. (1985) Uniform sense simple shear of the
Vogt, P.R, Ostenso, N.A. & Johnson, G.L. (1970) Mag continental lithosphere. Can.J. Earth. Sci. 22,108-25.
netic and bathymetric data bearing on sea floor spreading Wernicke, B. & Burchfiel, B.C. (1982) Modes of exten-
north ofIceland./,geophys. Res. 75,903-20. sional tectonism./. struct. Geol. 4,105-15.
von der Borch, C.(1979)Continent-island arc collision in Westbrook, G.K. (1982) The Barbados Ridge complex:
the Banda arc. Tectonophysics 54,169-93. tectonics of a mature forearc system. In Leggett, J.K.
Walcott, R.1. (1972) Gravity, flexure, and the growth of (ed.) Trench-Forearc Geology: Sedimentation and Tectonics
sedimentary basins at a continental edge. Bull, pfeol. Sac. on Modem and Ancient Active Plate Marpfins. Spec. Pub.
Aw. 83,1845-8. , £ieol. Soc. Lond. 10, 275-90.
Ward, M.A.(1963) On detecting changes in the Earth's Westbrook, G.K., Mascle, A. & Biju-Duval, B. (1984)
radius. Geophys. J. Roy. astr. Soc. 8,217-25. Geophysics and structure of the Lesser Antilles forearc.
Watanabe, T., Langseth, M.G. & Anderson, R.N.(1977) In Biju-Duval, B. & Moore,J.C.(eds)Init. Repts. DSDP
Heat flow in back-arc basins of the western Pacific. In 78A, pp. 23-38. US Government Printing Office, Wash
Talwani, M.& Pitman, W.C. Ill (eds)Island Arcs, Deep ington DC.
Sea Trenches and Back-arc Basinspp.137-61. Maurice Westbrook, G.K., Smith, M.J., Peacock, J.H. & Poulter,
Ewing Series 1, Am. Geophys. Union. M.J. (1982) Extensive underthrusting of undeformed
Watt J.P. & Shankland ,T.J.(1975) Uniformity of mande sediment beneath the accretionary complex of the Lesser
composition. Geolopfy 3,91-4. Antilles subduction zone. Nature 300,625-8.
Watters,B.R.(1976)Possible late Precambrian subduction White, G.W.(1980) Permian-Triassic continental recon
zone in south west Africa. Nature 259,471-3. struction of the Gulf of Mexico-Caribbean area. Nature
Watts, A.B. & Ryan, W.B.F.(1976) Flexure of the litho 283,823-6.
322 REFERENCES
White, N.(1989) Nature of lithospheric extension in the Windley, B.F. (1984) The Evolving Continents (2nd edn).
North Sea. Geolqgy 17,111-14. Wiley, London.
White,N.& McKenzie,D.(1988)Formation ofthe 'steer's Windley, B.F. (1985) The Himalayas. Geolqgy Today 1,
head' geometry of sedimentary basins by differential 169-73.
stretching ofthe crust and mantle. Geolqgy 16,250-3. Windley, B.F. (1993) Uniformitarianism today: plate tec
White, R.(1993) Breaking up is hard to understand. New tonics is the key to the p2L$t.J.pfeol. Soc. Lond. 150,7-19.
Scientist^ 3 April, 25-9. Woodcock, N.H. (1986) The role of strike-slip fault
White, R. & McKenzie, D.P. (1989) Magmatism at rift systems at plate boundaries. Phil. Trans. Roy. Soc. Lond.
zones: the generation of volcanic continental margins 317A,13-29.
and flood basalts./.^eophys. Res. 94, 7685-729. Woodcock, N.H. & Fischer, M. (1986) Strike-slip
White, R.S., Spence, G.D., Fowler, S.R., McKenzie, D.P., duplexes./, struct. Geol. 8,725-35.
Westbrook, G.K. & Bowen, A.N.(1987) Magmatism at Woodhouse, J.H. & Dziewonski, A.M. (1984) Mapping
rifted continental margins. Nature 330,439-44. the upper mantle: three dimensional modelling of Earth
Whiteman, A.J., Naylor, D., Pegrum, R.M. & Rees, G. structure by inversion of seismic waveforms./. pfeophys.
(1975) North Sea troughs and plate tectonics. Tectono- Res. 89, 5953-86.
physics 26, 39-54. Wyllie, P.J. (1981) Plate tectonics and magma genesis.
Whitmarsh, RB.(1975) Axial intrusion zone beneath the Geol. Rundsch. 70(1), 128-53.
median valley ofthe mid-Atlantic ridge at 37®N detected Wyllie, P.J. (1988) Magma genesis, plate tectonics and
by explosion seismology. Geophys. J. Roy. astr. Sac. 42, chemical differentiation of the Earth. Rev. Geophys. 26,
189-215. 370-404.
Whitmarsh,R.B., Avedik,F.& Saunders, M.R.(1986)The Yorath, C.J. & Chase, R.L.(1981) Tectonic history of the
seismic structure of thinned continental crust in the Queen Charlotte Islands and adjacent areas—a model.
northern Bay of Biscay. Geophys. J. Roy. astr. Soc. 86, Can.J. Earth. Sci. 18,1717-39.
589-602. Young,G.M.(1992)Late Proterozoic stratigraphy and the
Wiens, D.A., McGuire, J.J. & Shore, P.J.(1993) Evidence Canada-Australia connection. Geolopfy 20,215-18.
for transformational faulting from a deep double seismic Young, J.B., Lilwall, R.C. & Douglas. A. (1988) WM
zone in Tonga. Nature 364,790-3. Seismicity Maps Suitable for the Study of Seismic and
Wilson, D.S. (1993) Confirmation of the astronomical Geographical Regionalisation. Atomic Weapons Research
calibration of the magnetic polarity timescale from sea- Establishment Report 0 7/87. HMSO,London.
floor spreading rates. Nature 364, 788-90. Yungul, S.H. (1976) Tectonic implications of gravity
Wilson,J.T.(1963)Evidence from islands on the spreading anomalies associated with some continental rift
ofocean floors. Nature 197, 536-8. valleys—a model study. Tectonophysics 32, 321-9.
Wilson, J.T.(1965) A new class offaults and their bearing Zandt, G. & Ammon,C.J.(1995) Continental crust com
on continental drift. Nature 207, 343-7. position constrained by measurements of crustal Pois-
Wilson, J.T. (1988) Convection tectonics: some possible son's ratio. Nature 374,152-4.
effects upon the Earth's surface of flow from the deep Ziegler, P.A.(1993) Plate-moving mechanisms: their rela
mantle. Can.J. Earth Sci. 25,1199-1208. tive importance./.^^o/. Soc. Lond. 150,927-40.
Wilson, J.T. & Burke, K.(1972) Two types of mountain Zeil, W.(1979) The Andes: A Geological Review. Gebriider
building. Nature 239,448-9. Borntraeger, Berlin.
Wilson, M.(1989) Ipfneous Petropfenesis. A Global Tectonic Zhao,W.,Nelson,K.D.& Project INDEPTH team (1993)
Approach. Unwin Hyman,London. Deep seismic reflection evidence for continental under-
Windley, B.F.(1981)Precambrian rocks in the light of the thrusting beneath southern Tibet. Nature 366, 557-9.
plate-tectonic concept. In Kroner, A. (ed.) Precambrian Zoback, M.D., Zoback, M.L., Mount, V.S. et al. (1987)
Plate Tectonics., pp. 1-20. Elsevier, Amsterdam. New evidence on the state of stress of the San Andreas
Windley, B.F.(1983) Metamorphism and tectonics of the fault system. Science 238,1105-11.
Himalaya./.^^o/. Soc. Lond. 140,849-65.
Index
aborted rift zone, cross-section 286 Rocas Verdes region 170, 171, 172
absolute plate motion 80,81-3 seismicity 175-7
Acadian Orogeny 269, 270 structure 174-5, 176
accommodation zone 215 section 177
accretion 280 andesites 291
accretionary prism 293 angle of rotation 46
subduction zones 155-60 animals, evidence for continental drift 52-4
accretionary tectonics 198 anisotropy
accretionary wedge 139, 155,235 azimuthal 268
actualism 1 seismic 267
Afar rift system, French Somaliland 224 anorthosite 20, 292
Africa Antarctica
African Shield 265 earthquake epicentres 94
APW paths 282 and North America,SWEAT hypothesis 277
Ivory Coast-Ghana, sheared continental margin 135-6 anticrack faulting 141-3
separation from North America 74 Antilles see Lesser Antilles
southwest, Precambrian subduction zone 284 Aphebian 283
sec also East African Rift Appalachian Fold Belt 218, 223
Airy mechanism 21 apparent polar wandering(APW)curves 57-60, 276-7,281-2
hypothesis of isostatic compensation 35, 38, 39 appendix^ geological and stratigraphical timescale 301
Alaska arc-continent collisional mountain ranges 195-7
Cook Inlet 293 Archaean
Wrangellia 199-200 granulite-gneiss belts 273, 274
Aleutian island arc greenstone belts 272-3, 274-5
Andean-type orogeni? belt 150,160 plate tectonic model 275
earthquakes tectonics 272-5
focal mechanism solutions 147 Archimedes' Principle 36
foci 150 aseismic ridge 235
types 141, 146 Asia
free air gravity anomaly profile 139 deforming zone 194
magnetic lineation 167 distribution of tectonic styles 193
marginal seas 165 extrusion (escape) tectonics 187-94
Aleutian Trench, subduction 167 suture between India and Eurasia 180-2, 293-4
Alps asthenosphere 40-2
foreland basins 245 Atlantic Ocean
InsubricLine 128 aulacogens 221
section, compared with Himalaya-type orogens 195 cratons, correlations 50
Amisk group, Manitoba 283 FAMOUS area 103-4
ammonites, distribution 53 island arc systems 160-3
andalusite 151 North Atlantic
Andaman Sea 293 acoustic stratigraphy 22
Andean-type magmatic belt, cross-section 286 Great Meteor Seamount, free air anomaly 37
Andean-type mountain ranges 174-80 mid-ocean ridges see Mid-Atlantic Ridge
porphyry coppers 289 plate reconstruction and hotspot frame of reference 84
Andean-type orogenic belt 150,160 Tertiary igneous province 226
Andean-type subduction zones 207, 274-5, 290 Vema Fracture Zone 23
schema 207 rate of movement across plates 92
Anderson theory of faulting 13, 31 reconstruction of continents around 47
Andes seismic reflectors 23
Altiplano 174, 175, 176,177 attenuation 9,41, 140, 169
Andean plate, and island arc, continental and oceanic crust attenuation factor 140
section 175 aulacogens 216-18,222,293,294
BeniofFzone 175 Atlantic Ocean 221
formation 204-5 Australia
Nazca plate 175, 177 Banda Sea arc-continent collision 196-8
negative mass anomaly 34 earthquake epicentres 94
plate tectonic model 177-80 mammal migrations 55
Rocas Verdes complex 273, 274 marginal seas 165
323
324 INDEX
uniformitarianism 1
underplating 157 yield strength 33-4
uranium, decay, heat escape 43, 252 Young's modulus(E) 36