Plate tectonics

The idea of lithospheric plates (Figure 2.2) emerged with the acceptance of continental drift. If the continents have drifted, as Alfred Lothar Wegener (1915) claimed, then large chunks of crust (including continental cratons and deep ocean basins) have travelled several thousand kilometres without having suffered any appreciable lateral distortion. Two features indicate this lack of distortion. First, is the excellent 'fit' of the opposing South American and African coastlines, which have taken 200 million years to drift 4,000 km apart. Second, is the broad magnetic bands and faults of the deep-sea floor that have held their shape for tens of millions of years. This, and other evidence, suggests that the lithosphere is dynamic, that it changes.

Most geologists use the plate tectonic (or geotectonic) model to explain lithospheric change. This model is thought satisfactorily to explain geological structures, the distribution and variation of igneous and metamorphic activity, and sedimentary facies; in fact, it seems to explain all major aspects of the Earth's long-term tectonic evolution (e.g. Kearey and Vine 1990). Two aspects of the plate tectonic model engage interesting debates between conventional viewpoints and dissenting ideas. These aspects are the creation, destruction, and recycling of the oceanic lithosphere; and the repair and the assembly, disassembly, and reassembly of continental lithosphere.

0 1 2 3 4 Pressure (megabars) Hotspot DePth 0 4 8 12 16 Density (g/cm3) volcano Ridge (km) 1-1-1-1-1

0 1 2 3 4 Pressure (megabars) Hotspot DePth 0 4 8 12 16 Density (g/cm3) volcano Ridge (km) 1-1-1-1-1


Oceanic lithosphere

The accepted view of oceanic crust formation and maintenance involves a cooling and recycling system comprising the mesosphere, asthenosphere, and lithosphere lying under the oceans (Figure 2.3). The chief cooling mechanism is subduction. Volcanic eruptions along mid-ocean ridges form a new oceanic lithosphere. The newly formed material moves away from the ridges. In doing so, it cools, contracts, and thickens. Eventually, the oceanic lithosphere becomes denser than the underlying mantle and sinks, taking with it some of the sediment carried to the ocean floor from the continents. The sinking takes place along subduction zones, which are associated with earthquakes and volcanicity. Cold oceanic slabs with accompanying oceanic sediments from the denudation of continents may sink well into the mesosphere, perhaps to 670 km or more below the surface. The fate of the subducted slab is not clear. It meets with resistance in penetrating the lower mantle, but is driven on by its thermal inertia and continues to sink, though more slowly than in the upper mantle, causing accumulations of slab material (Fukao et al. 1994; Lay 1994; Maruyama 1994). It may form 'lithospheric graveyards' (Engebretson et al. 1992). Subduction feeds slab material (oceanic sediments derived from the denudation of conti-

Figure 2.1 Layers of the solid Earth. The capital letters, A-G, are seismic regions. The crust lies above the Moho. Its thickness ranges from 3 km in parts of ocean ridges to 80 km in collisional orogenic mountain belts. Continental crust is, on average, 39 km thick. The lithosphere is the outer shell of the solid Earth where the rocks are reasonably similar to those exposed at the surface. It includes the crust and the solid part of the upper mantle. It is the coldest part of the solid Earth. Cold rocks deform slowly, so the lithosphere is relatively rigid, it can support large loads, and it deforms by brittle fracture. On average, the lithosphere is about 100 km thick. Below continents, it is up to 200 km thick, and beneath the oceans, it is some 50 km thick. The differences in lithospheric thickness arise from temperature, and therefore viscosity, differences. The lithosphere under mid-ocean ridges is warm and thin; that under subduction zones is cold and thick; that under continents is cold, buoyant, and strong. The mantle and core constitute the barysphere. Processes in the barysphere influence processes in the lithosphere and thus, indirectly, cause changes in the ecosphere, particularly those occurring over millions of years. The mantle consists of upper and lower portions. The upper mantle comprises two shells. The asthenosphere (or rheosphere) lies immediately below the lithosphere. Temperatures increase with depth through the lithosphere. At around 100 km below the surface, lithospheric rocks are hot enough to melt partially, to weaken structurally, and behave rheidly (that is, like very slow-moving fluids). The asthenosphere, being relatively weak and ductile, more readily deforms than the lithosphere. Its base sits at about 400 km below the Earth's surface. In most places, the top part of the asthenosphere is a 50-100-km thick low-velocity zone. Beneath the asthenosphere is the mesosphere. The uppermost part, which extends down to about 650 km, is a transition zone into the lower mantle. Rocks become more rigid again in the mesosphere because the solidifying effects of high pressures increasingly outweigh the effects of rising temperatures. The mesosphere continues as the lower mantle. Extending down to a depth of 2,890 km, the lower mantle accounts for nearly one-half of the Earth's mass. The mantle rests upon the Earth's core, into which it merges through a fairly sharp and discontinuous transition zone known as the D" layer. The core consists of an outer shell of mobile and molten iron, some 2,260 km thick, with a mush zone at its base. It sits upon a solid inner ball that is 1,228 km in radius, close to melting point, and composed of iron, perhaps with some nickel. Source: After Huggett (1997a).

nents and oceanic crust, mantle lithosphere, and mantle-wedge materials) to the deep mantle, where they suffer chemical alteration, storage, and eventual recycling via mantle plumes (Tatsumi 2005).

High-resolution global mantle tomography confuses the neat model of subduction (Fukao et al. 2001). Narrow high-velocity zones under the Asian circum-Pacific arcs do not extend towards the lower mantle but shift horizontally into or under the 400-700 km transition zone, which suggest a horizontal flow in the mantle. In some cases, the leading edge of the cold slab turns upwards, implying a block to their downward descent. Scale experiments in a laboratory indicate that 'stiff' slabs tend to curl like wood shavings, while 'weak' slabs may suffer retrograde subduction, with retrograde trench migrations and the opening of back-arc basins concomitant with the backing of trenches (Faccenna 2000; see also Funiciello et al. 2003).

Several other problems beset the 'standard' view of basalt recycling in the plate tectonic system: Cliff Ollier (2003a, 2005) listed five:

1 Spreading sites are about three times longer than subduction sites. A consequence of this mismatch is that mid-ocean ridges produce some three times more basalt than subduction zones destroy, assuming that rates of plate movement stay the same from creation to destruction zones. To keep the system in a steady state, plates at subduction sites would need to converge faster than plates diverge at mid-ocean ridges.

Figure 2.2 Tectonic plates, spreading sites, and subduction sites. The lithosphere is not a single, unbroken shell of rock; it is a set of snugly tailored plates. At present there are seven large plates, all with an area over 100 million km-2. They are the African, North American, South American, Antarctic, Australian-Indian, Eurasian, and Pacific plates. Two dozen or so smaller plates have areas in the range 1-10 million km-2. They include the Nazca, Cocos, Philippine, Caribbean, A'abian, Somali, Juan de Fuca, Caroline, Bismarck, and Scotia plates, and a host of microplates or platelets. Source: Partly adapted from Oilier ( 1996).

Antarctic Plate And Australian Plate

North American plate


Caribbean plate

Arabian plate

Caroline plate ^-Bismarck plate

' SpmaliN subplate

African plate

Pacific plate

Australian-Indian plate

South American plate

Scotia plate

Antarctic plate

Plate boundary uncertain

Eurasian plate

Anatolian / plate

Subduction site Spreading site

Plate construction


Plate construction


Scotia Plate Subduction

Mafic lower crust modified slab

Mantle plume


Hotspot sources

Mafic lower crust modified slab

Mantle plume


Hotspot sources

Mafic lower crust-►

Enriched mantle type II

Sediments —►

Enriched mantle type !

Oceanic crust —►

High-jii mantle

Reservoir of chemically modified slab components {oceanic crust and sediments) and delaminated mafic lower arc crust


Reservoir of chemically modified slab components {oceanic crust and sediments) and delaminated mafic lower arc crust

Hotspot sources


Figure 2.3 The cooling and recycling system of the asthenosphere, lithosphere, and mesosphere. The oceanic lithosphere gains material from the mesosphere (via the asthenosphere) at constructive plate boundaries and hotspots and loses material to the mesosphere at destructive plate boundaries. Subduction feeds slab material (oceanic sediments derived from the denudation of continents and oceanic crust), mantle lithosphere, and mantle wedge materials to the deep mantle. These materials undergo chemical alteration and accumulate in the deep mantle until mantle plumes bear them to the surface where they form new oceanic lithosphere. Source: Adapted from Tatsumi (2005).

2 Much subduction occurs at island arcs, which in many cases are separated from continents by more spreading sites, and major plate boundaries fail to reach the continental margin. Thus, back-arc basins are the only sites available for recycling continental erosion products back to the continents.

3 The sinking slab comprises oceanic basalt with a variable load of sediments with different chemical compositions, which depend upon the continental rocks that supply the offshore sediments. After having undergone remelting, contamination, segregation of minerals, emplacement of batholiths, and the eruption of andesitic volcanoes, the basalt returns to the mid-ocean ridge as mid-ocean ridge basalt (MORB). Ollier questions the likelihood that the basalt produced at mid-ocean ridges could go through such a complex cycle and retain its uniformity.

4 MORB basalt is distinctive. When produced at spreading sites, it supposedly pushes away older sea-floor. However, if that were the case, then all sea-floor should be MORB basalt, whereas MORB is different.

5 Helium (4He and 3He) - an inert gas that is uninvolved in the rock cycle or biogeo-chemical cycles - leaks from spreading sites, mid-ocean ridges, and rift valleys. The dis tinct composition of MORB and the release of helium from spreading sites tend to suggest that the basalt there is being erupted for the first time. Peter Francis (1993) had made this point earlier, arguing that, although some slab material may eventually be recycled to create new lithosphere, the basalt erupted at mid-ocean ridges shows signs of being new material that has not passed through a rock cycle before. The signs are its remarkably consistent composition, which, as mentioned above, is difficult to account for by recycling, and its emission of such gases as helium that seem to be arriving at the surface for the first time. On the other hand, MORB is not 'primitive' and formed in a single step by melting of mantle materials - its manufacture requires several stages (Francis 1993, 49).

Another source of dispute, even among believers in plate tectonics, is the cause of plate movement. It is unclear why plates should move. Several driving mechanisms are plausible, the chief of which are 'ridge push' and 'slab pull'. Basaltic lava upwelling at a mid-ocean ridge may push adjacent lithospheric plates to either side. Conversely, as elevation tends to decrease and slab thickness to increase away from construction sites, the plate may move by gravity sliding. Another possibility, currently thought to be the primary driving mechanism, is that the cold, sinking slab at subduction sites pulls the rest of the plate behind it. In this scenario, mid-ocean ridges stem from passive spreading - the oceanic lithosphere is stretched and thinned by the tectonic pull of older and denser lithosphere sinking into the mantle at a subduction site; this would explain why sea-floor tends to spread more rapidly in plates attached to long subduction zones. As well as these three mechanisms, or perhaps instead of them, mantle convection may be the number one motive force, though this now seems unlikely because many spreading sites do not sit over upwelling mantle convection cells. If the mantle-convection model were correct, mid-ocean ridges should display a consistent pattern of gravity anomalies, which they do not, and would probably not develop giant fractures (transform faults).

Although convection is perhaps not the master driver of plate motions, it does occur. Authorities disagree on the depth of the convective cell - is it confined to the asthenosphere, the upper mantle, or the entire mantle (upper and lower).? Whole mantle convection (Davies 1977, 1992, 1999) has gained support, although it now seems that whole mantle convection and a shallower circulation may both operate. The fact that MORB has a consistent composition world-wide and comes from decompression of shallow mantle material, while oceanic island basalt (OIB) at hotspots is different in composition from MORB and seems to come from deeper mantle sources, suggests that the mantle might comprise two rather distinct chemical reservoirs that do not mix; thus whole mantle convection seems unlikely. Certainly, it is difficult to account for ancient (about 1.8 billion years old on average) detectable variations within the mantle of trace element concentrations and isotopic compositions (heterogeneities) that have survived through nearly twenty complete convective cycles, which thoroughly stir and overturn the mantle in about one hundred million years. Simulation models replicate the heterogeneities by stirring and segregation, heavier materials tending to sink and move sideways, while upper fluids become depleted, which process seems to account for the more 'depleted' character of MORBs compared with OIBs (Davies 2001).

Continental lithosphere

The continental lithosphere does not take part in the mantle-convection process. It is 150 km thick and consists of buoyant low-density crust (the tectosphere) and relatively buoy ant upper mantle. It therefore floats on the underlying asthenosphere. The established view is that continents break up and reassemble, but they remain floating at the surface. They move in response to lateral mantle movements, gliding serenely over the Earth's surface. In breaking up, small fragments of continent (terranes) sometimes shear off. They drift around until they meet another continent, to which they attach themselves (rather than being subducted) or possibly shear along it. As they may come from a different continent than the one to which they attach themselves, they are exotic or suspect terranes. Much of the western seaboard of North America appears to consist of these exotic terranes. In short, the continents are affected by, and affect, the underlying mantle and adjacent plates. They are maintained against erosion (rejuvenated in a sense) by the welding of sedimentary prisms to continental margins through metamorphism, by the stacking of thrust sheets, by the sweeping up of microcontinents and island arcs at their leading edges, and by the addition of magma through intrusions and extrusions (Condie 1989, 62).

Ollier (2005) has questioned the plate tectonic mechanisms for maintaining continents against erosion. The restoration of continents occurs only at active, collisional margins, namely, the western edge of the Americas, island arcs, and possibly sites associated with the closure of Tethys that form the Alpine-Himalayan Belt. The problems here are fourfold. First, most sediment eroded from continents ends up on the continental shelves of passive continental margins, which are about three times the length of active margins (Figure 2.4). These sediments cannot return to continents. Second, active continental margins have limited extent compared with passive margins and, in the Americas at least, sediments reaching them come from the relatively small drainage basins lying to the west of the continental divide. Third, spreading sites of back-arc basins back many island arcs, which trap sediment and prevent its passage to trenches (cf. p. 15). Back-arc basins show no signs of subduc-

Figure 2.4 Passive margins. Source: Adapted from Ollier (2005).

tion. The outer edge of western Pacific arcs subduct oceanic basalt, which carries trifling amounts of sediment derived from the islands of the arc itself. Fourth, the 'goodness-of-fit' of passive margins when reassembled into Gondawana or Pangaea suggests that passive margins are undeformed by subduction (as is generally believed) or by any other event moving material from offshore. Ollier makes a further point that erosion rates, at a conservative estimate, have run at around 25 B (B = Bubnoff unit [1 m per million years]). At this rate, erosion should have flattened the continents long ago, but it has not, partly owing to uplift. Ollier thinks that erosion and uplift rates cast doubt on the ability of plate tectonic mechanisms to restore continents.

In moving, continents have a tendency to drift away from mantle hot zones, some of which they may have produced: stationary continents insulate the underlying mantle, causing it to warm. This warming may eventually lead to a large continent breaking into several smaller ones. Best documented is the makeup and breakup of Pangaea, the Late Permian supercontinent. Pangaea began forming around 330 million years ago and reached its largest size in the Late Permian, 250 million years ago. At its largest, Pangaea did not contain North China or South China. Its component continents coalesced piecemeal, with some landmasses joining the Pangaean margins while others rifted off. Gondwana in southern Pangaea formed about 550 million years, and Laurussia (the combined terranes of Laurentia, Avalonia, and Baltica) in northern Pangaea formed between 418-400 million years ago. The eventual collision of Gondwana and Laurussia created Pangaea. Perhaps owing to increased mantle temperatures beneath the huge continental cap, Pangaea started to break up about 175 million years ago. Widespread magmatic activity preceded and accompanied the breakup.

The Precambrian supercontinent Rodinia (Dalziel 1991) is more speculative than is its Late Permian counterpart. In the 1970s, the observation that Grenville mountain belts found today on different continents were roughly 1,300 to 1,000 million-years-old planted in the minds of geologists the possibility of a single large landmass at that time. Most reconstructions of Rodinia try to match the mountain belts, with Laurentia forming the supercontinental core, Australia-East Antarctica along its western margin and Baltica-Amazonia along its eastern margin (Figure 2.5(a)). The classic view is that Rodinia began forming some 1,300 million years when three or four pre-existing continents started to coalesce in the Grenville Orogeny and formed a single landmass by perhaps 1,100 to 1,000 million years ago. This supercontinent then remain stable until, some 700 million years ago, it started to break up, over many millions of years, into three chief landmasses - West Gondwana, East Gondwana, and Laurasia went their own ways. Later reconstructions suggest that Rodinia disintegrated earlier, perhaps between 850 and 800 million years ago, and changed considerably during the few hundred million years of its existence (Figure 2.5(b)). The big problem with the reassembly of such ancient landmasses is that, for any given time, data on palaeolatitudes are sparse, an unfortunate situation that would be remedied by new palaeomagnetic studies running in tandem with radiometric age determinations (Torsvik 2003).

It has not passed the notice of geologists that supercontinents may repeatedly form and split up - there may be a 'supercontinent cycle' (Worsley et al. 1984; Nance et al. 1988). According to this hypothesis, the continents repeatedly coalesce to form supercontinents, and then break into smaller continents, owing to the pattern of heat conduction and loss through the crust. The entire cycle takes about 440 million years, or possibly 600 million years (Taylor and McLennan 1996). When a supercontinent is stationary, heat from the mantle should collect underneath it. As the heat accumulates, the supercontinent will dome

(a) Classic Rodinia reconstruction

(6) Alternative Rodinia reconstruction

1 Continental land mass I I Continents with paleomagnetic data -Subduction zone

^ Precambrian terranes or continents | 1300 to 1000 million-year-old mobile belts I Seafloor spreading axis ^

Was this article helpful?

+1 0


  • Christopher
    Which two caribbean countries are separated by a plate tectonic region?
    8 years ago

Post a comment