Category: Tectonics

Despite the increased precision of radiometric dating and the steady accumulation of ages when segments of continental crust first formed, two nagging oddities refuse to go away.  There seem to have been spurts in continental growth, rather than a steady build up over time.  Odder still, some areas have more crust of a certain age range than seems feasible.  The problem is a fundamental one, because the Earth generates radiogenic heat continually, though the amount has declined as the heat-producing isotopes of uranium, thorium and potassium decay.  Earth scientists assume that most geothermal energy exits to space through the process of sea-floor spreading.  Hot, new oceanic crust is invaded by seawater, thereby losing heat through hydrothermal activity.  The dominantly felsic magmas that build continental crust originate through partial melting processes where old, cold ocean floor descends at subduction zones.  Although some heat escapes through volcanism associated with mantle plumes, most researchers reckon that it is unlikely that this loss has ever come close to the quiet cooling at mid-ocean ridges, except possibly during the Archaean.  Averaged out, subduction and continent formation ought to keep pace with sea-floor spreading, though slowly declining over time.  There are those who focus on massive mantle turnovers in the form of superplumes that build large volcanic plateaux on land and on the sea floor, suggesting that their subduction generates greater volumes of crust than usual.  The main problem is that such plateaux are unlikely to be subducted.

The evidence for periods of accelerated continental growth comes from restricted regions, albeit very large.  Examples are 1900 to 1650 Ma crust in North America, Greenland and Europe, and 800 to 550 Ma crust in NE Africa, whose volumes are equivalent to between 1 and 10 times the present global rate of crust production at volcanic arcs.  Jonathan Patchett and Clement Chase of the University of Arizona offer a solution to the conundrums (Patchett, P.J. and Chase, C.G. 2002.  Role of transform continental margins in major crustal growth episodes.  Geology, v. 30, p. 39-42).  They show that strike-slip movement at modern subduction zones gives a 16% probability of more than 400 km transport of new continental crust parallel to the margins of existing continents.  Such motions are likely to concentrate continental growth where such terranes become docked together.  Such relative plate motions stem from particular configurations of spreading axes and the margins of old continents, and can therefore vary – some periods may have been dominated by head-on subduction, others by a greater amount of oblique relative movements.  By bundling together new continental material generated in magmatic arcs, the second would give the appearance of extraordinary rates of crust formation in some areas.  If the transform faults that channelled such lateral movements became obscure – and early strike-slip motions in ancient terranes are not easy to find or to quantify – the special natures of  terrane dockyards could go unnoticed.  Patchett and Chase note that the seeming pandemonium of 800-550 Ma crustal growth in NE Africa and Arabia has a counterpart in an age gap in the record of the northern continents, and cite several other examples.

While variations in strike-slip motions of terranes helps to resolve the apparent episodicity of continental growth, there is another line of approach.  Not all modern subduction zones generate voluminous magmas, even where plate motions are head to head.  The Andes has two huge segments where active subduction is unaccompanied by volcanism, and the angle of subduction is unusually shallow.  Low-angled subduction is likely where warmer than usual oceanic lithosphere enters a subduction zone, which is what might happen to segments blanketed by young ocean-plateau lavas formed by mantle plumes.  Constant sea-floor spreading need not necessarily result in constant rates of magmagenesis at destructive plate margins.

The Institute for Frontier Research on Earth Evolution (IFREE) involves 100 Japanese researchers focusing on central aspects of tectonic evolution over the last 200 Ma.  These include Pangaea break-up, mid-Cretaceous global warming and Eocene plate reorganization.  One particularly interesting study begun using initial funding of US$12 million is the Bonin-Mariana subduction zone.  Details at http://www.jamstec.go.jp/jamstec-j/IFREE.

Since Pentti Eskola’s recognition in 1949 that many Precambrian granitic rocks form domes surrounded by cusp-like synclines of supracrustal rocks, such mantled gneiss domes have been found in most cratons.  Probably the best example characterizes the 3.5 Ga Pilbara province of the West Australian Shield.  How they formed has long been a vexed topic, the most popular views being as a result of low-density basement rising through denser cover that contains abundant volcanic rocks, or as a result of regional-scale fold interference.  Precise dating of the Pilbara granitic rocks and greenstones shows a common age range, with some older greenstones,  The age data suggest that the dome and cusp structure is a product of the co-evolution of both, probably from a primary oceanic-like crust of mafic composition (Zegers, T.E. and van Keken, P.E. 2001.  Middle Archean continent formation by crustal delamination.  Geology, v. 29, p. 1083-1086).

Archaean rocks of broadly granitic composition (dominantly tonalites, trondhjemites and granodiorites, or TTG) have geochemical features setting them apart from post-Archaean varieties.  Rather than signifying their origin by supra-subduction melting of the mantle wedge with fractional crystallization and crustal assimilation in the lower crust (the dominant crust-forming process in post-Archaean times), all Archaean TTG seem to have formed by partial melting of a garnet-rich mafic source.  One of several possibilities is that their source was eclogite.  Based on the peculiar regional structure of the Pilbara and its dominance of the whole crust, as shown by maps of gravitational potential and magnetic field strength, Zegers and van Keken revisit earlier ideas of dominantly vertical tectonics that underlay early crust formation.  They suggest that efficient cooling by hydrothermal circulation allowed thick mafic crust (similar in some respects to that formed in the Mesozoic beneath ocean plateaux) to enter the field of eclogite stability at its base to form a layer denser than ultramafic mantle.  Once sufficiently thick, this layer would begin to founder, or delaminate, to be replaced by hot mantle.  Rebound of the remaining crust would set in motion rapid crustal uplift and extension, together with decompression melting of rising mantle (to form high-magnesium basalts high in the crustal sequence)and melting induced in the remaining mafic crust (to generate TTG magmas).  Indeed, the kimberlites that puncture other Archaean cratons carry abundant eclogite xenoliths from mantle depths.  Seemingly well-documented, this tectonic model does not explain all Archaean crust formation, for other cratons, such as that of west Greenland, are more readily accounted for by seemingly familiar subduction-zone processes.

Anyone who has watched a watchmaker at work may well have felt a tinge of panic at the sheer tininess of the screws, sprockets and gears, and awe at the near-superhuman patience and concentration involved in such micro-engineering.  Developments in geodesy based on the Global Positioning System of navigational satellites  push towards such aching precision.  The fixed stations of the International GPS Service (IGS) measure geographic position and topographic elevation to within less than a millimetre.  Corrections for known plate motions and Earth tides reveal motions that must be due to other forces.

Ultimately, the forces shaping the Earth’s surface are gravitational, and thus reduce to shifts of mass within and upon our planet.  By far the most rapid movements of matter are those involving the atmosphere and the water vapour that it carries.  Through variations in atmospheric density and the mass of water residing in soil moisture and snow cover, which arises from varying precipitation, surface load changes on an annual cycle.  Meteorological and remote sensing estimates of these loads allow geophysicists to model the elastic response of the surface to the seasons.  Why they have done this is not abundantly clear to me, but starting position is essential to astronavigation, hence similar attention to the Chandler Wobble (see Atmosphere linked to Earth’s rotation, Earth Pages, September 2000).  Anyhow, the records suggest an annual mass transfer from hemisphere to hemisphere of around 10¹³ tonnes, which is sufficient to cause elastic deformation within the scope of GPS measurements. Geomaticians from the universities of Nevada and Newcastle upon Tyne (Blewitt, G. et al. 2001.  A new global mode of Earth deformation: seasonal cycle detected.  Science, v. 294, p. 2342-2345) have been able to chart the actual motions over the period from 1996 to mid-2001.

Performing the necessary computations on the weekly data from 66 IGS stations, and fitting curves to the results, Blewitt et al. present convincingly repetitive cycles in the motions towards the intersection of the Greenwich meridian and the Equator, towards the North Pole, and perpendicular to the surface.  These tie very well to the theoretical model.  Interestingly, they were able to model the shifts of displacement globally, and in series of maps show that the positions of maximum displacement shift along a path linking the continents.  That is not surprising in itself, for the oceans respond by changes in water level, and only exposed continental lithosphere is likely to flex.  The poles sink by around 3 mm each winter, and the Equator swings towards the winter pole by 1.5 mm.  Results tally extremely well with estimates of seasonal mass shifts and theory.  The surprises include an anomaly in vertical displacement in 1996-7 preceding the 1997-8 El Niño event, probably due to changes in Pacific sea level driven by winds and anomalous monsoon precipitation.

Yet more on tectonics of the Tibetan Plateau

In the previous issue of Earth Pages was a resumé of a paper in Science that discusses the lateral tectonic motions that result from India’s collision with Eurasia (Continental tectonics of eastern Eurasia December 2001 Earth Pages)).  The compliment to that appeared in the 22 November 2001 issue of Science (Tapponnier et al. 2001.  Oblique stepwise rise and growth of the Tibet Plateau.  Science, v. 294, p. 1671-1677).  How and when the India-Eurasia collision zone achieved its pattern of huge elevated masses is partly an issue of tectonics, but they bear on any climatic effect that changed elevations might have had on climate, both regionally in the case of the South Asian monsoon circulation, and globally (one view is that Tibet’s deflection of atmospheric circulation may have been an important trigger for the onset of northern hemisphere glacial conditions).

Many geologists have considered the whole lithosphere of the region to have behaved in a ductile manner during collision, so that shortening and thickening were distributed more or less evenly.  They ascribe the uniform height (>5 000 metres) to gravitational rebound when part of the thickened lithosphere detached and fell into the mantle, around the mid-Miocene.  Erosion being unable to keep pace with uplift, the Tibetan Plateau is then thought to have become unstable and started to collapse laterally.  That is seen by many as an explanation for clear evidence of E-W extension from both numerous N-S rift systems and extensional first motions on Tibetan earthquakes.  However, this vast area is clearly subdivided into several major blocks by large strike-slip systems.  The prevailing notion is that these faults are effects of “soft” collisional tectonics.  Tapponier et al. assemble detailed evidence in relation to these faults and the blocks that they bound.  They support tectonic evolution which has been controlled by coherent blocks of lithosphere, a process which was episodic rather than continuous, and accompanied by decoupling of crust and mantle lithosphere.

The linchpin of their model is the diachronous calc-alkaline magmatism of the region during the Tertiary, which becomes younger towards the north.  As well as the principal site of northward subduction of Indian lithosphere beneath the Zangbo Suture, they propose that this magmatism was related to southward subduction that migrated northwards, and is hidden by thickened crust.  The huge strike-slip systems are, to Tapponier et al., nothing less that oblique suture zones.  The crustal blocks that they separate are, according to their model, large thrust wedges founded on major crustal detachments that accomplished most of the shortening.  The process did not involve destruction of oceanic basins, but subduction of sub-continental mantle lithosphere, when crust and mantle became detached.  Each successive subduction-accretion episode added its own increment to surface uplift, there probably having been three major steps in creating the highest average topography on Earth.

North America, particularly its west coast, is the best studied natural laboratory for active tectonics.  Nonetheless, the downturn in Earth Science funding in the USA has threatened an ambitious project aimed at consolidating knowledge of plate interactions there.  Nature (15 November 2001, p. 241) reports that the EarthScope initiative now has strong backing from the US National Academy of Sciences.

EarthScope has 4 elements: a mobile grid of seismometers; an observatory to monitor movement of plates below the NW Pacific Ocean; a programme aimed at drilling into the San Andreas Fault System; an interferometric radar satellite that will accurately measure ground movements in relation to tectonic and volcanic features.  The total cost is around $400 million, shared equally between NASA and the National Science Foundation, if the funding proposal wins acceptance.

Information from:  http://www.earthscope.org

Continental tectonics of eastern Eurasia

Interferometric radar remote sensing provides high precision information on Earth motions associated with earthquakes (Radar analysis of Turkish earthquake, Earth Pages August 2001), but depends on “before and after” imaging.  Continental tectonics is not just the outcome of occasional large movements on major faults, but of strains that continually occur throughout the lithosphere.  Global positioning satellites provide means of precise location, particular when operated in differential mode, in which field-station signals are matched to those at fixed, geodetically precise base stations.  Precisions to within centimetres or better are now commonplace at low cost.  Structural geologists have been using GPS receivers for over a decade to check on the annual rates of plate motion across major structures such as the Alpine Fault of New Zealand and spreading centres such as that exposed on land in Iceland.  In the 19 October issue of Science, such geodetic analysis of tectonics leaped by an order of magnitude.

The jewel in the crown of continental tectonics is eastern Eurasia, where the active collision of the Indian sub-continent with Asia drives a huge array of very large faults that separate rigid blocks and others, such as the Tibetan Plateau, that are deforming en masse.  The spreading power of the Carlsberg and Central Indian Ridges is dissipated in motion of continental crust spanning 30° of latitude and 60° of longitude.  Chinese scientists and their collaborators from the US universities of Alaska and Colorado have measure GPS positions at 354 stations throughout China, every one or two years for the last decade.  Their analysis of the interim results (Wang, Q. et al.  2001.  Present-day crustal deformation in China constrained by global positioning system measurements.  Science, v.  294, p. 574-577) helps confirm or modify ideas about crustal motions that stemmed from seismic first-motion studies and regional field evidence.  More than a third of the tectonic power accounts for crustal shortening within the Tibetan Plateau.  While the western part of the huge system involves consistent motion towards the north-north-east, driving into Eurasia’s hinterland, the “free-edge” of eastern China  and Indo-China seems to encourage the escape tectonics first proposed by Molnar and Tapponier.  That involves a massive clockwise rotation around the East Himalayan Syntaxis, which takes up a great deal of motion.  Whereas Molnar and Tapponnier proposed the shoving of south-eastern China oceanwards by the “escape” of Tibet, Wang et al’s measurements reveal that its motion to the east is only between one third and a quarter that of the adjacent east Tibetan Plateau.  The lack of any sign that Tibetan crust is overriding that of south-east China, or that the latter is being shortened, may suggest that escape is funnelled around the East Himalayan Syntaxis into Burma and South-East Asia.

One of the powerful bits of evidence that support continental drift are the plots of magnetic pole positions determined from rocks of different ages exposed on a modern continent, relative to that of the present pole position. Comparing such plots from different continents sometimes reveals similarities in their shapes over long periods of time, so that the plots partly match when they are superimposed.  In such fits, other parts of the plots diverge considerably.  Such comparisons are best explained respectively by the former unity of the two modern continents and their movement together, and their separation to drift independently.  The plots are illusory, and are called apparent polar wander paths.  For most of the Phanerozoic Aeon such palaeomagnetic data tie in well with other evidence for the formation of composite continental masses, such as Pangaea, and plate movements since the Triassic.  That provides confirmation of the basic assumption in palaeomagnetic studies that the Earth’s magnetic poles remain close to those of its rotation, bar some circulation around the axis and magnetic reversals.  It is tempting to use the same assumption for earlier times, in the absence of  easy fitting of the margins of continental segments and the sea-floor magnetic stripes that are the key to plate tectonics since the early Mesozoic.  If magnetic poles did move well away from the poles of rotation at any time in the past, that would play havoc with continental reconstruction.  True polar wander is something that many tectonicians “Dinnae care to speak aboot”!  That is not surprising, for another reason.  Use of the term imply mean two things: a long-term shift in the Earth’s magnetic polarity relative to its axis of rotation (to me that warrants the adjective “true”, and ); a shift in the relative position of the whole crust and mantle relative to the core, whose dynamism determines the magnetic field.

A recent review (Irion, R. 2001.  Slip-sliding away.  New Scientist 18 August 2001, p. 34-37) concentrates on evidence for the second usage.  There is evidence that suggests a 20° shift of all continents over a period of 2 Ma, in the Cretaceous.  This is dwarfed by a suggestion of a 90° shift in 15 Ma that span the time of the Cambrian Explosion, so that a continent could have moved from the pole to the equator at a rate far faster then anything known from Mesozoic to Recent sea-floor spreading.  One explanation is destabilization of the Earth’s angular momentum by concentration of all crustal mass and the effect of a massive mantle plume beneath it at high latitudes.  That would distort the Earth’s shape.  A planet’s rotation is most stable when it its shape is fat around the equator.  The opposite, a prolate spheroid, is least stable, and a polar supercontinent could result in such instability, restoring steady state if the whole caboodle slipped to lower latitudes.  That is what is proposed to explain some odd palaeomagnetic pole positions newly and accurately gathered from early Cambrian rocks.  Such a notion takes on its own momentum, because of its association in time with the explosive diversification of animals with hard parts.

It is not a fundamentally new idea, for Alfred Wegener suggested that the mechanism for his hypothesis of continental drift was Pohlflucht (flight from the poles) of continental mass.

Geophysicists and geochemists are generally opposed on what happens to subducted lithosphere.  Seismic tomography of the deep mantle shows convincing evidence for slab-like cold bodies down to the core-mantle boundary, yet differences in trace-element and isotopic signatures of volcanic rocks formed at ridges from shallow mantle and ocean islands that relate to deep plumes persuades geochemists that restriction of convection within the upper mantle, at the 660 km deep discontinuity, best explains the differences.  There are other models that might account for geochemical differences, such as heterogeneities throughout a poorly stirred mantle or because material in slabs subducted to the bottom of the mantle rarely rises again, but displaces more pristine materials upwards.

The more earthquakes that seismographs detect and locate, the better geophysicists are able to map in 3-D the zones on which they take place.  One destructive margin long known to have aberrant seismicity is the northern part of the Tonga system in the Pacific Ocean.  This is where the fastest subduction anywhere consumes lithosphere that has little time to warm up while it descends – surely a site for slabs to fall steeply into the deep mantle.  Much of the Tonga system shows the expected zone of steeply plunging Earthquakes, yet west and north-west of Fiji there are earthquakes that do not fit the regional pattern.  They are far too shallow to result from motion on the main subduction zone.  By detailed analysis of seismic data Wang-Ping Chen and Michael Brudzinski have revealed a strong possibility that a piece of old subducted slab has slid to the 660 km discontinuity since it parted company with the now rapid and steep motion at the Tonga trench (Chen, W-P. and Brudzinski, M.R. 2001.  Evidence for a large-scale remnant of subducted lithosphere beneath Fiji.  Science, v. 292, p. 2475-2478).  If such behaviour turns out to be more widespread, large volumes of old lithosphere may indeed sit at the discontinuity, satisfying many geochemists as a means to maintain very old differences in composition of the mantle.  The problem is, increasingly good resolution in seismic tomography has so far failed to detect the tell-tale high seismic velocity signature of such cold slabs.  Chen and Brudzinski suggest that they may be “invisible” to this method, because of their mineralogy – perhaps the crustal lithosphere has not equilibrated to eclogitic materials, or is given neutral buoyancy by being heavily hydrated.

Between a rock and a hard place

Plate theory stems from the notion that the lithosphere is overwhelmingly rigid and deforms only at the boundaries between plates, particularly at destructive margins.  The Earth’s seismicity is overwhelmed by earthquakes at discrete boundaries, and the mapping of seismic events along narrow lines by the world-wide network of seismographs (set up as a means of pinpointing nuclear weapon tests) formed on of the main planks in developing the theory of plate tectonics.  The plate whose evolution drove India into Asia bucks this definition.  It has long been known to host seismicity well inside its boundaries.  Oceanographic work has slowly built up a means of relating Indian Ocean seismicity to plate structure, whereas analysis of earthquake first motions from seismographs reveals that the deformation differs between various block of the ocean floor.  The plate suffers folding and thrusting, and transcurrent motions along ancient transform faults, such as the Ninety East Ridge.  The most likely explanation for the Indian plate’s aberrance is that sea-floor spreading from the ridge separating the Indian Plate from that carrying Antarctica can no longer be accommodated by subduction of the subcontinent beneath Asia, whereas it can be taken up by subduction beneath the Java-Sumatra island arc.  The Central Indian basin is being compressed, and must deform in some way, perhaps eventually to become a new subduction zone.

Source:  Deplus, C.  2001.  Indian Ocean actively deforms.  Science, v. 292, p. 1850-1851.

Considering the continual processes that stress continental lithosphere from the time of its formation, it is a puzzle to find large areas that preserve its earliest parts in an almost pristine state.  Greater heat production in the past demands that the frequency and power involved in continental jostling were greater as we go back in geological time.  Zones that show little sign of having been tectonically reworked for more than a billion years are termed cratons, and most of them have at their core continental material that formed in the Archaean, more than 2.5 Ga ago.  Later orogens do show isotopic signs that deformed and partially melted Archaean crust was involved, but no so much as might be expected.  Somehow, having a nucleus of Archaean lithosphere confers strength to cratonic areas.  Geophysics reveals that  the lithosphere beneath cratons uniquely extends to depths of 200 km, forming a “keel” or tectosphere.

Most geochemists consider that deep mantle beneath cratons is so rigid because it is unable to come close to the beginning of melting, due to it having once been the source of massive amounts of basaltic magma.  Loss of the constituent elements of basalt and volatiles, including heat-producing isotopes of U, Th and K, renders it more inert than mantle that still has the potential to generate basalt under appropriate conditions.  Basalt magmas also remove significant amounts of iron, thereby adding buoyancy to tectosphere materials.

Occasionally, much younger magmas that do form at the depths of the tectosphere bring samples of it to the surface, in the form of xenoliths.  Their petrography and geochemistry reinforce the general idea of how cratonic “keels” form, but they have been difficult to date with confidence.  The relatively new rhenium-187/osmium-187 method makes dating more assured.  Cin-Ty Lee and colleagues from Harvard University (Lee, C et al.  2001.  Preservation of ancient and fertile lithospheric mantle beneath the southwestern United States.  Nature, v. 411, p. 69-73) used the method on xenoliths from two adjacent areas, the actively extending Basin and Range Province and the Colorado Plateau.  Both contain ancient rocks, Archaean in the former and Mesoproterozoic in the second, which behaves as a stable craton.  Xenoliths from mantle deep beneath them have similar ages to those in the oldest crustal rocks, helping confirm the geochemical connection between crust formation and lithospheric mantle.  However, those from beneath the Basin and Range have potentially “fertile” compositions, whereas the Colorado samples show signs of the depletion thought to confer strength and buoyancy.  Paradoxically, a younger craton sits next to Archaean lithosphere that is demonstrably weak. 

Lee and colleagues suggest that if part of Archaean crust formation did not create a tectosphere, it is quite possible that younger orogens might contain considerably more ancient crust than currently suspected.  On the other hand, the mismatch between the near certainty that continents formed more rapidly during the first third of recorded geological history and the disproportionately small volume of known Archaean crustal rock could signify that a lot of it became resorbed into the mantle.  That doesn’t appear to have been a significant process in later times.  However, the total lack of sialic rocks older than 4 Ga, yet the evidence from detrital zircons up to 4.4 Ga in much younger sediments that some did indeed form, suggests that crustal resorption was efficient during early tectonics.  Perhaps the Archaean marked the waning of such processes, in which an increasing proportion remained locked at the surface.

See also:  Nyblade, A.  2001.  Hard-cored continents.  Nature, v. 411, p. 39-39.

Partially melted zones beneath Tibet

Anomalously low seismic velocities, accompanied by a “muffling” of seismic energy, and high heat flow beneath the Tibetan Plateau have hinted at the possibility of active crustal melting, but such information cannot resolve whether that is the case or not.  Parts of the Plateau have been volcanically active in the near past, and that has been attributed by some workers  to the detachment and sinking into the mantle of a large chunk of sub-Tibetan lithosphere.  Freed of a substantial mass, the thick lithosphere beneath Tibet would then bob up, the rapid drop in pressure at depth inducing partial melting.  Being weak, a substantial partially melted zone would also help the Tibetan crust deform more easily.

One means of  adding support to the idea is looking for deep-crustal anomalies in electrical conductivity.  Because electric currents flow naturally in the Earth, the conventional means of resistivity survey can use them instead of an input current.  Such magnetotelluric surveys potentially give information down to depths of 100 km or more.  At these scales, zones of abnormally low conductivity are likely to be due either to pervasion of deep rock with watery fluids or with widespread partial melting.  A group of Chinese, Canadian and US geophysicisists (Wei, W. and 14 others 2001.  Detection of widespread fluids in the Tibetan crust by magnetotelluric studies.  Science, v. 292, p. 716-718) have shown that the middle to lower crust deeper than 15 to 20 km beneath most of the Tibetan Plateau is anomalous in this way.  The highest conductivity lies beneath the main Yarlung (Indus) – Tsangpo suture., and may be related to fluids released by subduction processes.  It is the anomaly beneath the Plateau itself that is most significant, for it extends for 4 degrees of latitude along the survey line.  Higher conductivity anomalies correlate closely with Plio-Pleistocene volcanically active areas, and much of the area is affected by hydrothermal fluids.  While adding detail to structure and rheological properties beneath Tibet, magnetotelluric studies still leave open the possibility that much of the electrical signature may be due to pervasive watery fluids, as well as to zones of melting.

Brazilian input to the growth of Gondwana

One of the most dramatic tectonic events known from the geological record is the break up of a supercontinent, dubbed Rodinia (from the Russian for motherland), in the Neoproterozoic.  From a unity of almost all earlier continental crust, this break up sent fragments scurrying across a plethora of new oceans.  Some of the fragments reassembled around 650 Ma ago to create what eventually became the southern part of the Carboniferous supercontinent of Pangaea; Gondwana.  The assembly of West Gondwana involved a vast network of orogenic belts in which juvenile arc materials were pinched between colliding continental fragments, as these oceans closed up.  Often called the Pan African event, because of its widespread signature in that continent, this assembly also affected eastern South America at the same time.

Fernando Alkmim, Stephen Marshak and Marco Fonseca (Alkmin, F.F.  2001.  Assembling West Gondwana in the Neoproterozoic: clues from the São Francisco craton region, Brazil.  Geology, v.  29, p. 319-322)  turn our attention from the much-described Pan African to its Braziliano counterpart in South America.  Their summary of current understanding suggests six stages in the rifting to collision, that involved major changes in palaeogeography.

Slivers of ancient crust make up part of the collages of accreted terranes found in many ancient orogens.  How they form is not well-known.   Clues might lie in modern microcontinents that still remain surrounded by oceanic lithosphere, such as Jan Mayen, the Seychelles and the East Tasman Plateau.  Geologists from the Universities of Sydney, and Aarhus and the Geological Survey of Canada believe that such fragments of continental crust form early in the evolution of passive margins, as a result of plume activity followed by asymmetric sea-floor spreading (Müller, D.M. 2001.  A recipe for microcontinent formation.  Geology, v.  29, p. 203-206).

One suspected microcontinent in the southern Indian Ocean is the Kerguelen Plateau – its shape is odd.  In the few places where it breaches surface in the Kerguelen Archipelago, there are rare occurrences of silicic plutonic rocks.  However, evidence from dredged samples seems to show that most of the Plateau formed by plume-related basaltic volcanism that began at the same time as the formation of the Rajmahal Traps in Bangladesh (about 117 Ma ago).  ODP drilling now reveals fluviatile sediment layers that contain high-grade gneisses, whose ages range back to the Proterozoic (Nicolaysen, K. and many others 2001.  Provenance of Proterozoic garnet-biotite gneiss recovered from Elan Bank, Kerguelen Plateau, southern Indian Ocean.  Geology, v.  29, p. 235-238).  The authors do not see this as directly supporting a Kerguelen microcontinent, but the formation of the plateau close to eastern India around 110 Ma ago, from where abundant Precambrian crustal debris would have been shed.  However, the presence of continental geochemical signatures in Kerguelen Plateau basalts, otherwise having plume affinities, might indicate a fragment of former Gondwanan lithosphere at the core of the Plateau, akin to the now exposed Danakil block in the nascent Red-Sea – Afar rift in NE Africa, that spalled off during the break-up of the Mesozoic supercontinent.

Africa to a large degree exerts a control over modern plate tectonics, because it barely moves at all.  The base of its lithosphere connects in several places with the solid mantle, so that asthenosphere is not universally present beneath the continent.  These roots slow down Africa’s motion.  One name applied to them is “tectosphere”, and they are partly governed by the low heat production in the lithosphere and underlying mantle, as a result of U, Th and K having been extracted from depth by processes that led to separation of continental crust.  These processes reach completion beneath the most ancient segments of continental crust, and result in them eventually becoming geologically inert; they become cratons. 

Studies based on samples brought from deep below cratons by volcanism, particularly that which formed the kimberlite plugs of Africa, suggest that their roots date back almost as far as the age of continental material above them.  But that natural sampling is haphazard, and relationships cannot be found.  Where large extraterrestrial bodies have excavated material to great depths, tectosphere material might well have reached the surface en masse by rebound following impact.  Such a deep section formed around the Vredfort Dome in the Kaapvaal  Craton of southern Africa after a major impact about 2 billion years ago.  It exposes the crust-mantle boundary. 

A programme of dating the Vredfort materials (Moser, D.E. et al.  2001.  Birth of the Kaapvaal tectosphere 3.08 billion years ago.  Science, v. 291, p. 465-468) shows that welding of crust to mantle in Archaean times, and formation of both craton and tectosphere, took place about 3.1 billion years ago, more than a hundred million years after crustal material itself coalesced.  Tectospheres seem not to begin forming at the same time as large masses of continental crust.  Instead they accrete to the base of the crust through later processes that probably involve subduction.  Other workers have suggested that the Kaapvaal tectosphere accumulated from masses of oceanic lithosphere that failed to descend completely into the mantle.  Curiously, the fragments in kimberlite pipes from which those conclusions were drawn are very dense eclogites.  Such material should descend easily into the deep mantle because their density exceeds that of peridotite.  That poses the question of why they came to stay close to the surface so long ago.  Perhaps their eclogite mineralogy stabilized long after they accreted beneath Kaapvaal, and they are “stuck” in the inert tectosphere that they form, out of gravitational equilibrium.  Should such high-density roots eventually become detached from their overlying materials, then the surface would pop up to become eroded dow to great depths.  The fact that most of the worlds cratons (the continental “shields”) preserve great volumes of material that crystallized at quite shallow depths, suggests that such “delamination” does not commonly happen beneath them.

The Mediterranean area is possibly the most tectonically complicated area there is.  It’s a plexus of microplates, all shuffling and jostling like guilty schoolboys accused of sticking gum under their desks.  That is a result of the misfit between the continental masses carried on the Eurasian and African plates, which was never resolved by the collision between the two that threw up the Alpine chain.  Complex as it is, the region is small enough, close enough to research institutes and pleasant enough to work in for there to have been a great deal of effort to understand its active plate tectonics.

The latest method to be applied is the analysis of seismic waves’ arrivals at seismometers in the manner of body scanning – seismic tomography.  Combining these new 3-D data of deep motions in the mantle with a review of surface geology, M.J.R. Wortel and W. Spakman of the Vening Mensz Research School of Geodynamics at the University of Utrecht build a remarkable picture of what seems to be going on (Wortel, M.J.R. and Spakman, W. 2000.  Subduction and slab detachment in the Mediterranean-Carpathian region.  Science, v. 290, p. 1910-1917).  One of their remarkable conclusions is a suggestion that subducted slabs are becoming detached, thereby changing the configuration of slab-pull forces in the region.  They sketch out how that might happen, by the formation of small ‘nicks’ in the short subducting slabs that focus slab-pull force along the reduced length of intact slab.  Thus focused, the pull more rapidly helps propagate the “nick” into a fully-fledged tear, which will migrate over the remaining length of the subduction zone.

Mechanically, that is interesting enough, but should it happen at a shallow depth influx of asthenosphere would generate magma on a small scale, and perhaps induce hydrothermal activity and unusual sequences of metamorphism in the overlying crust.  Isostatic responses might change depositional process at the surface too.  Wortel and Spakman suggest that there is geological evidence throughout the region for this process having operated in the past, with consequences such as these, as well as going on today.

Ultra-high pressure (UHP) metamorphic rocks from the Yankou region in China have been down a subduction zone to more than 200 km and then rebounded to the surface.  Kai Ye, Bolin Cong and Danian Ye of the Chinese Academy of Science in Beijing have worked on barometric indicators from eclogites and garnet peridotites to reach this conclusion (Ye, K. et al. 2000.  The possible subduction of continental material to depths greater than 200 km.  Nature, 407, 734-736).  It is no surprise to learn that basaltic and peridotitic materials have been down a subduction zone, because that is what oceanic lithosphere does continually, though how they return to the surface as intact slabs is problematic.

What is surprising is that such highly compressed rocks are associated with similarly UHP materials that are chemically normal materials of the continental crust.  The Yankou rocks now hold the record for deep diving.   Sialic subduction is not easy because of its reluctance to reach densities that exceed that of the mantle.  That being said, there are growing suspicions that continental materials may contribute to the composition of alkaline magmas formed deep beneath hot spots.  If sial does not reach 200 km depth, its density always lies above that of the mantle, and it must be buoyant.  Taken deeper, however, the situation reverses because of phase changes that compress silica and feldspar, so that at 300 km depth they become much denser than mantle, and must continue sinking to become potential contributors to later mantle melting.

In this case it seems as if the slab of Chinese sial was dragged from the lower crust by its attachment to enough basic and ultrabasic rocks that the whole lot broke the buoyancy barrier by their density change at high pressures.  Getting back to the surface poses the big problem, the authors proposing that they were rafted by rocks beneath them.  Somehow, a large mass of UHP basic-ultrabasic material must have become detached from sialic materials before the combined slab passed the 300 km boundary and became doomed to long-term mantle residence.  That would give them and any eclogites remaining attached to them sufficient buoyancy to bob up once again.

The rise of the huge Tibetan Plateau, with an average elevation of 5 km, presented a major barrier to atmospheric circulation, perhaps one of the largest that has ever existed.  With its latitude close to the down flow of the tropical Hadley cells, it has had an effect on the Asian monsoon in particular, strengthening its effects.  Many climatologists believe that Tibet has played a major role in global climatic change towards the end of the Cainozoic.  So, timing the uplift is critical in assessing the modelled effects in relation to detailed climate records of the Neogene.  This is by no means easy, for the late-Tertiary sediments are terrestrial in origin.

A team of Australian and Chinese geologists focussed on the sedimentary record in the Tarim Basin, north of the Kunlun mountains that form the northern flank of the Tibetan Plateau (Zheng, H.  et al., 2000.  Pliocene uplift of the northern Tibetan Plateau.  Geology, v. 28, p. 715-718).  Sediments there change from redbeds deposited in gently sloping flood plains to coarse debris laid down by flash floods at a rising mountain front; exactly the relationship that records the beginning of uplift in northern Tibet.  Dating this is no easy matter, however.  The technique that the team used is magnetostratigraphy, based on highly sensitive measurements of the polarity of 2500 samples of weakly magnetized sediments.

The change in facies spans a period when the Earth’s magnetic field was reversed – the Gilbert reversed chron – which occurred between 4.5 to 3.5 Ma ago.  The maximum age for the beginning of Tibetan uplift in the north is therefore 4.5 Ma, in the Pliocene.  This contrasts with the accepted age of Oligocene – Miocene for uplift of the Himalaya and southern Tibet, and with models that postulate climatic change that followed it.  Whereas the Tarim Basin today is arid, the sediments indicate that until the Pliocene abundant water flowed at the surface, to deposit great thicknesses of fine alluvium.

What goes on beneath constructive plate margins, and ocean ridges has, up to now, been largely a matter of conjecture, blended with the geology of ophiolite complexes obducted onto continents.  Ophiolites are perhaps not such a good model, since the low buoyancy of the basalt capped lithosphere that they represent prevented them from subduction, and stems from unusual conditions.  The bulk of oceanic lithosphere is destined for resorption into the mantle, and it forms at common or garden ridge systems.

One way of modelling magmatism at ridges is through geochemical analysis of mid-ocean ridge basalts matched with topographic and structural detail of the ridge itself, but this is a blurred approach.  It shows that part of the process must involve ponding of magma in chambers at shallow levels beneath the ridges.  The other aspect is the form taken by the mantle that must rise to undergo adiabatic partial melting.  For fast-spreading ridges, such as the East Pacific Rise, there are two such models: constraint of rising mantle in two-dimensional sheets descending from beneath the ridge itself; three-dimensional plumes of mantle from which magma migrates laterally to ridge segments.  Amplifying geochemical-structural models needs a better idea of the actual processes and the geometries that they take.  A means of getting this information is to use a technique well-honed by petroleum exploration; 3-D seismic reflection profiling.

A consortium of geophysicists from the universities of California and Cambridge used this costly method, involving 200 profiles, to look at 400 km² of the East Pacific Rise at 9°N (Kent, G.M. and 10 others, 2000.  Evidence from three-dimensional seismic reflectivity images for enhanced melt supply beneath mid-ocean-ridge discontinuities.  Nature, v. 406, p. 614-618).  Melts have about half the seismic velocity of solid rock, and so boundaries between melt and solid show up with better contrast on seismic records than do boundaries in piles of sedimentary rocks.  The surprising result is that instead of vertically extensive magma chambers, expected from either hypothesis, melt occurs in a narrow, continuous sill-like body beneath the ridge.  This connects to a plunging tongue that is probably the path taken by magma from the zone of partial melting in the mantle.  The sill itself occurs at a fixed depth below that predicted from ophiolite studies for the level at which vertical sheeted dykes form the lower part of the petrologically defined crust.  This suggests that the magma simply cannot rise en masse to inject along extensional fissures as the lower crust fails, the sheeted dyke layer acting like a seal in the flow of petroleum in sedimentary basins.  Instead, it seems more likely that magma ekes out as rising rivulets that follow the base of the dyke layer until the reach dilatations at the ridge.

Although results from this study are inconclusive as regards the two models for rising mantle, the detail that it reveals augurs well for further 3-D surveys of ocean magmatism that will complement seismic tomography of the deep mantle.

Accepted wisdom accounts for the bulk of lavas and intrusive igneous rocks that build island arcs and probably much of the continental crust by what is known as wedge melting.  As old, cold and wet ocean lithosphere descends subduction zones, metamorphic reactions in the top layer of basalts and sediments (oceanic crust) release water-rich fluids.  These depress the temperature at which melting can begin when they permeate the overriding mantle wedge.  The water-releasing reactions involve dehydration of altered ocean floor that work to create the garnet-pyroxene assemblages that characterize eclogites, and drive the top slab of the lithosphere further from conditions under which it will begin to melt.  Formation of abundant garnet and pyroxene also imparts the density jump that helps make oceanic lithosphere founder at destructive plate margins.  The less there is, the lower the angle of subduction.  Whether or not dense eclogite forms depends on the temperature at which lithosphere enters a subduction zone, and temperature depends to a large extent on how old the consumed lithosphere is.  As sea-floor spreading shoves newly created lithosphere sideways from oceanic ridge systems, it slowly cools by conduction and interaction with permeating seawater.  The faster the spreading or the smaller the plates involved, the sooner lithosphere can reach a subduction zone.  Both factors can give rise to shallow-angled subduction.

The Earth loses heat that radioactive decay generates in the mantle by sea-floor spreading.  Going back in time, there were more undecayed heat-producing isotopes, so more heat had to be lost.  In the Archaean Aeon (more than 2500 million years ago) heat production was perhaps 2 to 3 times higher, and either spreading was much faster or there were more plates.  Most geologists now accept that low-angled subduction was a common characteristic of Archaean geological processes.  That is highly significant, because such conditions drive the top slab of oceanic crust towards melting, and the melts produced are very different from the basalts and andesites produced by modern wedge melting.  They are much more silica-rich, and crystallize to form the trondhjemites, tonalites and dacites that are so common in Archaean continental crust. 

Today, plate movements are sluggish, and though slab melting has been detected it was long thought to be rare, taking place only where very young oceanic lithosphere (less than 5 million years old) entered the mantle.  Recent work by French geochemists (Gutscher, M-A. et al., 2000.  Can slab melting be caused by flat subduction?  Geology, 28, p. 535-538) showed that such occurrences relate to subduction of lithosphere as old as 45 million years.  Their model to explain such Archaean-like processes involves a transformation from normal steep subduction to a phase involving almost horizontal movement of the descending lithosphere.  The density reduction that this demands stems from the heating effect of the asthenosphere through which the plate travels.  Wedge melting is generally close to the site of subduction, marked by an oceanic trench.  Modern slab melting, however, needs a lengthy period of heating in a flat subduction zone, so the volcanoes that it produces lie much further away from the trench.  Eventually the asthenosphere itself is cooled by the advancing plate and volcanism stops because the slab begins to dehydrate and to lose the potential for partial melting.  This explains the lack of volcanoes over most of the known areas of flat subduction, as in the Andes of central Chile.

Geodynamics

Plate tectonics is not the be all and end all of how the world works.  It is merely the expression of the Earth’s overall behaviour by the thin surface rind of lithosphere.  Almost certainly, all rocky planets behave similarly, in the sense of producing energy by decay of radioactive isotopes inside, and losing this energy by transport to the surface, where it escapes by radiation.  How planets do this determines to a major degree the geological processes that go on at their surface.  Clearly, there are subtle differences among the Inner Planets, because only the Earth shows signs of active plate movements that give it both a geological and, in its case, a biological life.

Why the Earth is so odd depends on its internal processes, so geochemists and geophysicists have spent 30 years seeking ways of unravelling how the mantle behaves.  As well as a battery of geochemical methods to distinguish different kinds of mantle whose melting contributes to crust formation in different tectonic settings, the main arm in geodynamics is using earthquake waves in a manner akin to body scanning to image the deep interior.  This seismic tomography is just beginning to resolve some of the widely divergent views about deep-Earth processes.  So, a review of the state of the geodynamicists’ art in a recent issue of Science makes for compulsory reading (Tackley, P.J., 2000.  Mantle convection and plate tectonics: toward an integrated physical and chemical theory.  Science,  288, p. 2002-2007).

The geochemists’ problem, having discovered three chemically fundamental kinds of mantle that basalt magma production stems from, is to decide how they are arranged.  They have at least 6 basic models.  Before seismic tomography, each was as believable as the others.  Through reviewing 3-D images of where hot and cold materials sit in the mantle – the key to motions within it – Tackley shows how some of the geochemical models must probably bite the dust, and the directions that research will take in future.  There is still no self-consistent model for whole-mantle behaviour, but it is beginning to look like the various views of convection as simple cells, either from top to bottom of the mantle, or decoupled into lower and upper systems must give way to something much more complex.  What does seem well established is that many subducted slabs find their way right down to the core-mantle boundary.  The most primitive mantle ‘reservoirs’, from which the ocean island basalts over hot-spots stem in part, have an excess of ³He (formed only in stars and therefore locked in the Earth when it formed) over ⁴He (released by decay of radioactive uranium and thorium and so changing with time).  These reservoirs are now probably in two gigantic, hot bulges rising from the core-mantle boundary, that dominate the most tectonically active parts of the lithosphere.  Cooler mantle lies beneath more inert lithosphere.  It has a composition from which mid-ocean ridge basalts emerge, and which signifies its loss over time of the materials that now make up the continents.

The most important possibility emerging from growing knowledge of the deep Earth is that Earth scientists might have to break from James Hutton’s 200 year old notion that the present is the key to the past.  The plate-mantle system is something likely to change dramatically over time, and the Earth is currently in one form of many different kinds of possible behaviour.

In the same issue of Science is a review of how motions in the Earth’s core generate the geomagnetic field (Buffett, B.A., 2000.  Earth’s core and the geodynamo.  Science, 288, p. 2007-2012).

Periodically the Earth’s magnetic field flips, so that its direction reverses.  The signals of magnetic field reversals occur in well-dated continental lavas, and this chronology is one of the keys to understanding the more continuous magnetic signature preserved in surveys running at right angles to the oceanic ridge systems.  They presented to Earth scientists the now familiar patterns of magnetic ‘stripes’ of normal and reversed polarity running parallel to the ridges, which characterise oceanic lithosphere.  The ‘stripes’ permit the dating of the ocean floor, which increases more or less systematically in both directions away from the ridges.  That pointed unerringly to the formation of oceans by sea-floor spreading, and underpins the theory of plate tectonics.  That is a fine example of deduction from, in many respects, fortuitous information of an empirical kind, and has kept Earth scientists extremely busy since Vine and Matthews twigged its significance in the 1960s.

Why these magnetic upheavals take place has proved a tough nut to crack.  Not long after Earth scientists began  to speak of little else, theoretical geophysicists proposed that somehow the Earth contained a self-sustaining dynamo prone to inverting its magnetic effects.  The only conceivable source was the almost certainly iron-rich core, with an outer liquid shell and a solid inner core, proven by analysis of seismic waves travelling through the Earth’s central parts.  Motion within the core moves electrons, thereby simulating current flow, and from Maxwell’s law there must be a related magnetic field that would shift as the motion changed.  The liquid outer core is clearly the part that undergoes the most complex motion, partly as a consequence of rotation, and partly because of heat transfer.  Ideas on the nature of that motion have developed over the last 3 decades, importantly through analysis of the drift of the magnetic field itself.  The key feature however, is that the mantle, outer and inner core are mechanically decoupled, at least partly, by the outer core’s fluidity.  Discovering how the solid inner core moves is clearly important for more realistic models of the self-exciting dynamo.

Vidale and co-workers (25 May 2000 issue ofNature, vol. 405, p 445) show how they re-analysed 30 year old records of seismic wave arrivals from Soviet nuclear tests to ‘image’ inner-core motion from the scattering of these signals – one of very few useful outcomes of the Cold War, and hopefully one that will never be repeated!  Their results are not definitive, but suggest that the inner core rotates on a different axis from that of the Earth as a whole.